MRI and CT of The Cardiovascular System

MRI and CT of the
Cardiovascular System
Third Edition
LWBK1209-FM_pi-xvi.indd 1 5/18/13 1:22 PM

MRI and CT of the
Cardiovascular
System
Third Edition
Charles B. Higgins, MD
Professor of Radiology
University of California, San Francisco
San Francisco, California
Albert de Roos, MD
Department of Radiology
Leiden University Medical Center
Leiden, The Netherlands

Senior Executive Editor: Jonathan W. Pine, Jr.
Product Manager: Amy G. Dinkel
Production Project Manager: Marian Bellus
Senior Manufacturing Coordinator: Beth Welsh
Senior Marketing Manager: Kimberly Schonberger
Designer: Holly McLaughlin
Production Service: Aptara, Inc.
Third Edition © 2013 by Lippincott Williams & Wilkins, a WOLTERS KLUWER business
Second Edition © 2006 by Lippincott Williams & Wilkins, a WOLTERS KLUWER business
First Edition © 2003 by Lippincott Williams & Wilkins, a WOLTERS KLUWER business
Two Commerce Square

2001 Market Street
Philadelphia, PA 19103 USA
LWW.com
All rights reserved. This book is protected by copyright. No part of this book may be reproduced in any
form by any means, including photocopying, or utilized by any information storage and retrieval system
without written permission from the copyright owner, except for brief quotations embodied in critical
articles and reviews. Materials appearing in this book prepared by individuals as part of their official
duties as U.S. government employees are not covered by the above-mentioned copyright.
Printed in China
Library of Congress Cataloging-in-Publication Data
MRI and CT of the cardiovascular system / [edited by] Charles B. Higgins,

Albert de Roos.—Third edition.
p. ; cm.
Includes bibliographical references and index.
ISBN 978-1-4511-3731-6
I. Higgins, Charles B., editor of compilation. II. Roos, Albert de,
1953- editor of compilation.
[DNLM: 1. Cardiovascular Diseases--radionuclide imaging. 2. Diagnostic Techniques,
Cardiovascular. 3. Magnetic Resonance Imaging--methods. 4. Tomography, X-Ray
Computed--methods. WG 141.5.M2]
RC670.5.M33
616.1’07548--dc23
2013015144
Care has been taken to confirm the accuracy of the information presented and to describe generally
accepted practices. However, the authors, editors, and publisher are not responsible for errors or omis-
sions or for any consequences from application of the information in this book and make no warranty,
expressed or implied, with respect to the currency, completeness, or accuracy of the contents of the
publication. Application of the information in a particular situation remains the professional responsi-
bility of the practitioner.
The authors, editors, and publisher have exerted every effort to ensure that drug selection and dos-
age set forth in this text are in accordance with current recommendations and practice at the time of
publication. However, in view of ongoing research, changes in government regulations, and the constant
flow of information relating to drug therapy and drug reactions, the reader is urged to check the pack-
age insert for each drug for any change in indications and dosage and for added warnings and precau-
tions. This is particularly important when the recommended agent is a new or infrequently employed
drug.
Some drugs and medical devices presented in the publication have Food and Drug Administration
(FDA) clearance for limited use in restricted research settings. It is the responsibility of the health care
provider to ascertain the FDA status of each drug or device planned for use in their clinical practice.
To purchase additional copies of this book, call our customer service department at (800) 638-3030 or fax
orders to (301) 223-2320. International customers should call (301) 223-2300.
Visit Lippincott Williams & Wilkins on the Internet: at LWW.com. Lippincott Williams & Wilkins cus-
tomer service representatives are available from 8:30 am to 6 pm, EST.
10 9 8 7 6 5 4 3 2 1

To our residents and fellows, whose work has contributed
to progress in Cardiovascular MRI and CT

Right Running Head vii
Contributing Authors
Evan Appelbaum, MD Jens Bremerich, MD, PhD

Assistant Professor of Medicine Head of Cardiothoracic Imaging
Harvard Medical School Department of Radiology
Department of Medicine/Cardiovascular Division University Hospital Basel
Beth Israel Deaconess Medical Center Basel, Switzerland
PERFUSE Study Group
Boston, MA Arno Bücker, MD
Professor of Radiology and Chairman
Håkan Arheden, MD, PhD Clinic of Diagnostic and Interventional Radiology
Professor of Clinical Physiology University Clinic Saarland
Head, Department of Clinical Physiology Homburg, Germany
Institute for Clinical Sciences
Lund University Peter T. Buser, FASC, FESC
Department of Clinical Physiology Professor of Cardiology
Skåne University Hospital Vice Chairman of the Department of Cardiology
Lund, Sweden University Hospital Basel
Basel, Switzerland
Sonya V. Babu-Narayan, MBBS, BSc, MRCP, PhD
Clinical Senior Lecturer and Consultant Cardiologist Iacopo Carbone, MD
NIHR Cardiovascular Biomedical Research Unit Assistant Professor of Radiology
Royal Brompton and Harefield NHS Foundation Trust and Department of Radiological, Oncological and
Imperial College London Pathological Sciences
London, UK “Sapienza” University of Rome
Italy
Jeroen J. Bax, MD, PhD
Cardiologist Albert de Roos, MD, PhD
Cardiology Department Professor of Radiology
Leiden University Medical Center Department of Radiology
Leiden, The Netherlands Leiden University Medical Center
Hans-Christoph Becker, MD
University Professor of Radiology (in particular Marc Dewey, MD
Non-invasive Cardiac Imaging) Heisenberg Professor of Radiology
Section Chief of Computed Tomography Department of Radiology
Department of Clinical Radiology Charité - Universitätsmedizin Berlin
Großhadern Clinics Berlin, Germany
Ludwig-Maximilians-University of Munich
Munich, Germany Joost Doornbos, PhD
Associate Professor, MR Physicist
Jan Bogaert, MD, PhD Leiden University Medical Center
Professor of Radiology Department of Radiology
UZ Leuven Leiden University Medical Center
Department of Imaging & Pathology Leiden, The Netherlands
KU Leuven
Leuven, Belgium Petter Dyverfeldt, PhD
Assistant Professor
René M. Botnar, PhD University of California, San Francisco
Professor and Chair in Cardiovascular Imaging San Francisco, CA
Division of Imaging Sciences and Biomedical Engineering Linköping University
King’s College London Linköping, Sweden
London, UK
Rossella Fattori, MD
Head of Interventional Cardiology
San Salvatore Hospital
Pesaro, Italy
vii

viii Contributing Authors
Zahi A. Fayad, PhD J. Wouter Jukema, MD, PhD

Director, Translational and Molecular Imaging Institute Director Cardiologist
Sinai NIH/NHLBI Program of Excellence in Nanotechnology (PEN) Cardiology Department
Vice-Chair for Research, Department of Radiology Professor Leiden University Medical Center
Departments of Radiology and Medicine (Cardiology) Icahn School Leiden, The Netherlands
of Medicine at Mount Sinai, New York, NY
Philip Kilner, MD, PhD
Jasper Florie, MD, PhD Consultant in Cardiovascular Magnetic Resonance
Department of Radiology CMR Unit
Leiden University Medical Center Royal Brompton and Harefield NHS Trust
Leiden, The Netherlands London, UK
Mark A. Fogel, MD, FACC, FAHA, FAAP Han W. Kim, MD

Professor of Pediatrics and Radiology Assistant Professor of Medicine
The Perelman School of Medicine at The University of Pennsylvania Duke Cardiovascular Magnetic Resonance Center
Director of Cardiac Magnetic Resonance Duke University Medical Center
Department of Pediatrics, Division of Cardiology Durham, NC
The Children’s Hospital of Philadelphia
Philadelphia, Pennsylvania Raymond J. Kim, MD, MSE, FACC
Director, Duke Cardiovascular Magnetic Resonance Center
Matthias G. Friedrich MD, FESC, FACC Professor of Medicine and Radiology
Cardiologist Duke University Medical Center
Professor of Medicine, Departments of Medicine and Radiology, Durham, NC
Université de Montréal
Adjunct Professor of Medicine, Departments of Cardiac Sciences Sebastian Kozerke, PhD
and Radiology, University of Calgary Professor of Biomedical Engineering, MR Physicist
Montreal Heart Institute Institute for Biomedical Engineering
Montreal, Canada University and ETH Zurich
Zurich, Switzerland
Rolf Gebker, MD, PhD
Head of Cardiovascular Magnetic Resonance Dara L. Kraitchman, VMD, PhD, FACC
Department of Internal Medicine—Cardiology Professor, Cardiovascular Interventional Section Head
German Heart Institute Berlin Division of MR Research
Berlin, Germany Russell H. Morgan Department of Radiology and Radiological
Science
Jacob Geleijns, PhD The Johns Hopkins University School of Medicine
Physicist Baltimore, MD
Radiology Department
Leiden University Medical Center Lucia J.M. Kroft, MD, PhD
Leiden, The Netherlands Radiologist
Thomas H. Hauser, MD, MMSc, MPH Leiden University Medical Center
Assistant Professor of Medicine Leiden, The Netherlands
Harvard Medical School
Director of Nuclear Cardiology Titus Kuehne, MD, PhD
Beth Israel Deaconess Medical Center Professor
Boston, MA Unit of Cardiovascular Imaging – Congenital Heart Diseases
German Heart Institute Berlin and Charité – Medical University Berlin
Charles B. Higgins, MD Berlin, Germany
University of California, San Francisco Hildo J. Lamb, MD, MSc, PhD
San Francisco, California Professor of Radiology
Michael D. Hope, MD Leiden University Medical Center
Assistant Professor of Radiology Leiden, The Netherlands
San Francisco, CA Boudewijn P.F. Lelieveldt, PhD
Division of Image Processing
Raoul M.S. Joemai, PhD Department of Radiology
Physicist Leiden University Medical Center
Leiden University Medical Center Leiden, The Netherlands
Radiology Department
Leiden University Medical Center Maria Clara N. Lorca, MD
Leiden, The Netherlands Postdoctoral Research Scholar
San Francisco, CA

Contributing Authors ix
Venkatesh Mani, PhD Arno A.W. Roest, MD, PhD

Assistant Professor, Radiology Pediatric Cardiologist
Translational and Molecular Imaging Institute Deapartment of Pediatric Cardiology
Icahn School of Medicine at Mount Sinai, New York, NY Leiden University Medical Center
Warren J. Manning, MD
Professor of Medicine and Radiology David Rosenbaum, MD
Harvard Medical School Cardiovascular prevention unit.
Section Chief, Non-invasive Cardiac Imaging and Testing Pitié salpetriere Hospital, AP-HP, Paris, France
Department of Medicine/Cardiovascular Division and Radiology Functional imaging Lab, Pierre and Marie Curie University
Beth Israel Deaconess Medical Center Paris , France
Boston, MA
Vincenzo Russo, MD, PhD
Antoine Millon, MD Cardiovascular Radiologist
Department of vascular surgery Cardio-Thoracic-Vascular Department
University hospital of Lyon University Hospital S. Orsola
Lyon, France Bologna, Italy
Translational and Molecular Imaging Institute, Mount Sinai School
of medicine, Mount Sinai hospital Liesbeth P. Salm, MD, PhD
New York, NY Nuclear Medicine Physician
Department of Nuclear Medicine
Stefano Muzzarelli, MD Radboud University Nijmegen Medical Center
Department of Cardiology Nijmegen, The Netherlands
Fondazione Cardiocentro Ticino
Lugano, Switzerland Hajime Sakuma, MD, PhD
Department of Cardiology Professor and Chairman
University Hospital Lausanne, CHUV Department of Radiology
Lausanne, Switzerland Mie University Hospital
Mie University Graduate School/Faculty of Medicine
Gobinath Nadeshalingam, MSc Mie, Japan
Marvin Carsley CMR Centre at the Montreal Heart Institute
Department of Cardiology David Saloner, PhD
Université de Montréal Professor
Canada Department of Radiology and Biomedical Imaging
VA Medical Center
Eike Nagel, MD, PhD, FACC, FESC, MRCR University of California, San Francisco
Chair of Clinical Cardiovascular Imaging San Francisco, CA
Head of the Department of Cardiovascular Imaging
Division of Imaging Sciences and Medical Engineering Andreas Schuster, MBBS, MD, PhD
King’s College London Honorary Clinical Lecturer
The Rayne Institute Imaging Sciences and Biomedical Engineering
London, United Kingdom St Thomas’ Hospital
King’s College London
Karen G. Ordovas, MD, MAS London, United Kingdom
Assistant Professor of Radiology Specialist Registrar in Cardiology
University of California, San Francisco Department of Cardiology and Pulmonology
San Francisco, CA University Medical Centre
Georg-August-University
Harald H. Quick, PhD Göttingen, Germany
Director MR Imaging
Institute of Medical Physics Juerg Schwitter, MD, FESC
Friedrich-Alexander-University Erlangen-Nürnberg Director Cardiac MR Center
Germany Division of Cardiology
University Hospital Lausanne, CHUV
Frank E. Rademakers, MD, PhD Lausanne, Switzerland
Department of Cardiology
University Hospitals Leuven Freddy Ståhlberg, PhD
Leuven, Belgium Professor
Department of Medical Radiation Physics and Department of
Johan H.C. Reiber, PhD Diagnostic Radiology
Division of Image Processing Lund University
Department of Radiology Lund, Sweden
Leiden University Medical Center

x Contributing Authors
William Stanford, MD Martin N. Wasser, MD, PhD

Professor Emeritus Radiologist
Department of Radiology Department of Radiology
Roy J. & Lucille A. Carver University of Iowa College of Medicine Leiden University Medical Center
Iowa City, IA Leiden, The Netherlands
Andrew M. Taylor, MD, FRCP, FRCR Oliver M. Weber, PhD

Professor of Cardiovascular Imaging Clinical Scientist
UCL Institute of Cardiovascular Science Philips Healthcare
Great Ormond Street Hospital for Children Madrid, Spain
London, United Kingdom
Jos J.M. Westenberg, PhD
Lowie M.R. Van Assche, MD Physicist
Post Doctoral Research Fellow Department of Radiology
Duke Cardiovascular Magnetic Resonance Center Leiden University Medical Center
Duke University Medical Center Leiden, The Netherlands
Durham, NC
Susan B. Yeon, MD, JD
Rob J. van der Geest, PhD Assistant Professor of Medicine
Division of Image Processing Harvard Medical School
Department of Radiology Department of Medicine/Cardiovascular Division
Leiden University Medical Center Beth Israel Deaconess Medical Center
Leiden, The Netherlands Boston, MA
Rutger W. van der Meer, MD, PhD Michael Zellweger, FESC

Radiologist Head of Cardiac Imaging
Department of Radiology Department of Cardiology
Leiden University Medical Center University Hospital Basel
Leiden, The Netherlands Basel, Switzerland
Harrie C.M. van den Bosch, MD

Radiologist
Catharina Hospital
Eindhoven, The Netherlands

Right Running Head xi
Preface
The first images of the beating heart using CT were produced in Our goal in preparing this book was to provide fully up-
the late 1970s and the first images of the heart using MR, around dated information on the application of MR and CT for the
1982. Both techniques remained technologically limited for car- assessment of cardiovascular disease. Similar to the first two
diac imaging and unfamiliar to cardiovascular diagnosticians for editions, we recognized that the continued rapid advances in
two decades or more. The first edition of this book, published these modalities dictated a short interval between concep-
in 2003, addressed only cardiovascular MR. The second edition tion and publication of the third edition. Seven new chapters
was published in 2006, with very short intervals between edi- have been added to this edition.
tions because of the technological advances in cardiovascular The authors include basic scientists, cardiologists, and ra-
MR between 2003 and 2006 and the rapid emergence of ECG diologists from around the world. These experts have had a
referenced multi-detector interrogation CT for the interrogation long involvement in fostering the development of cardiovas-
of coronary arterial disease. Both techniques have become highly cular MR and CT.
effective and in some applications unique for the evaluation of
the cardiovascular system. These techniques can provide precise Charles B. Higgins, MD
depiction of cardiovascular morphology, quantification of physi- Albert de Roos, MD
ology and, to some degree, myocardial tissue characterization.
xi

Right Running Head xiii
Contents
contributing authors vii

Preface xi
PART 1 Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1 Clinical Approach to Cardiovascular Magnetic Resonance Techniques . . . . . . . 3
Hildo J. Lamb, Sebastian Kozerke, Joost Doornbos, and Albert de Roos
2 Magnetic Resonance Angiography Techniques . . . . . . . . . . . . . . . . . . . . . . . . 21

David Saloner
3 Techniques for Cardiovascular Computed Tomography . . . . . . . . . . . . . . . . . 39

Jacob Geleijns, Raoul M.S. Joemai, Albert de Roos, and Lucia J.M. Kroft
4 Cardiac Anatomy and Physiolog y: Imaging Aspects . . . . . . . . . . . . . . . . . . . . 57

Frank E. Rademakers
5 Blood Flow Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Håkan Arheden and Freddy Ståhlberg
6 Quantification in Cardiac Magnetic Resonance Imaging and

Computed Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Rob J. van der Geest, Boudewijn P.F. Lelieveldt, and Johan H.C. Reiber
7 CVMR: Imaging Planes and Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Harrie C.M. van den Bosch, Jos J.M. Westenberg, and Albert de Roos
8 Coronary CTA: Display Techniques and Planes . . . . . . . . . . . . . . . . . . . . . . 116

Karen G. Ordovas, Maria Clara N. Lorca, and Charles B. Higgins
9 Cardiovascular MR in Infants and Children . . . . . . . . . . . . . . . . . . . . . . . . . 125

Mark A. Fogel
PART 2 Acquired Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

10 Magnetic Resonance of Cardiomyopathies and Myocarditis . . . . . . . . . . . . . 137
Gobinath Nadeshalingam, Iacopo Carbone, and Matthias G. Friedrich
xiii

xiv Contents
11 Arrhythmogenic Right Ventricular Cardiomyopathy/Dysplasia . . . . . . . . . . 149

Peter T. Buser, Michael Zellweger, and Jens Bremerich
12 MRI and CT of Pericardial Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

Karen G. Ordovas and Charles B. Higgins
13 Cardiac and Paracardiac Masses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

Charles B. Higgins and Karen G. Ordovas
14 Valvular Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

Jan Bogaert and Andrew M. Taylor
PART 3 Ischemic Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

15 Myocardial Perfusion in Ischemic Heart Disease . . . . . . . . . . . . . . . . . . . . . 221
Juerg Schwitter
16 Assessment of Ischemic Heart Disease Utilizing Dobutamine . . . . . . . . . . . . 239

Rolf Gebker, Andreas Schuster, and Eike Nagel
17 Assessment of Myocardial Viability by Contrast Enhancement . . . . . . . . . . . 251

Lowie M.R. Van Assche, Han W. Kim, and Raymond J. Kim
18 Delayed Enhancement in Nonischemic Myocardial Disease . . . . . . . . . . . . . . 283

19 Magnetic Resonance Imaging Risk Assessment in Ischemic Heart Disease . . 292

Stefano Muzzarelli
20 Coronary Magnetic Resonance Angiography: Technical Approaches . . . . . . . 305

Oliver M. Weber
21 Coronary Artery Imaging—Clinical Applications . . . . . . . . . . . . . . . . . . . . . 324

Evan Appelbaum, Thomas H. Hauser, Susan B. Yeon, RenÉ M. Botnar,
and Warren J. Manning
22 Coronary Blood Flow Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342

Hajime Sakuma and Charles B. Higgins
23 Cardiovascular Magnetic Resonance and Computed Tomography

of Coronary Artery Bypass Grafts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
Liesbeth P. Salm, Jeroen J. Bax, Hildo J. Lamb, J. Wouter Jukema,
and Albert de Roos
24 CT Imaging of Coronary Artery Calcification . . . . . . . . . . . . . . . . . . . . . . . 365

William Stanford
25 Coronary Computed Tomography Angiography . . . . . . . . . . . . . . . . . . . . . 376

Marc Dewey and Lucia J.M. Kroft

Contents xv
26 CT Evaluation of Myocardial Perfusion, Function,

and Late Enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Hildo J. Lamb, Lucia J.M. Kroft, and Albert de Roos
27 Atherosclerotic Plaque Imaging: MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402

David Rosenbaum, Venkatesh Mani, Antoine Millon, and Zahi A. Fayad
28 Plaque Characterization—Computed Tomography . . . . . . . . . . . . . . . . . . . . 420

Hans-Christoph Becker
PART 4 Congenital Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . 431

29 Congenital Heart Disease: Magnetic Resonance Evaluation of
Morphology and Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
30 Magnetic Resonance Imaging of Function and Flow in Postoperative

Congenital Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
Arno A.W. Roest, Lucia J.M. Kroft, and Albert de Roos
31 Adult Congenital Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489

Philip Kilner and Sonya V. Babu-Narayan
PART 5 Vascular Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503

32 MRI and CT of Thoracic Aorta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505
Rossella Fattori and Vincenzo Russo
33 Thoracic Aorta Disease: Flow Evaluation by MR . . . . . . . . . . . . . . . . . . . . . 545

Michael D. Hope and Petter Dyverfeldt
34 Computed Tomographic Angiography of the Abdominal Arteries . . . . . . . . . 558

Rutger W. van der Meer, Lucia J.M. Kroft, Jasper Florie,
Martin N. Wasser, and Albert de Roos
35 Magnetic Resonance and Computed Tomography

Angiography of Peripheral Arteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
Martin N. Wasser and Harrie C.M. van den Bosch
PART 6 Interventional MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599

36 Catheter Tracking and Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
Harald H. Quick

xvi Contents
37 Endovascular Interventional Magnetic Resonance Imaging . . . . . . . . . . . . . 618

Arno Bücker
38 Endovascular Interventions—Congenital Heart Disease . . . . . . . . . . . . . . . . 634

Titus Kuehne
39 Endovascular Delivery of Gene and Stem Cell Therapy . . . . . . . . . . . . . . . . 648

Dara L. Kraitchman
Index 667

PART
1
Basic Principles
LWBK1209-ch01_p01-20.indd 1 16/05/13 9:45 PM

LWBK1209-ch01_p01-20.indd 2 16/05/13 9:45 PM
Hildo J. Lamb C h apt e r
1
Sebastian Kozerke
Joost Doornbos
Albert de Roos
Clinical Approach to Cardiovascular

Magnetic Resonance Techniques
■■ Introduction since these techniques are causing a revolution within the

field of CVMR. Magnetic resonance (MR) techniques for
■■ Coils
perfusion imaging, visualization of delayed enhancement,
■■ Cardiac Motion Compensation coronary artery MR angiography, and vessel wall imaging
■■ Respiratory Motion Compensation will be discussed in more detail in other chapters, and will
be only briefly touched here.
■■ Survey
■■ Plan Scan
Coils
■■ Anatomy
Clinical cardiac exams can be performed using the stan-
■■ Function
dard body coil, although image quality is suboptimal.
■■ Perfusion The main problem is the limited in-plane spatial resolu-
■■ Delayed Enhancement tion of around 3 mm2. High spatial resolution is especially
important for accurate assessment of wall motion abnor-
■■ Flow malities due to, for example, myocardial infarction. In the
■■ Coronary Mra past, reliable images were obtained using the body coil for
assessment of global and regional myocardial wall motion
■■ Vessel Wall Imaging (Fig. 1.1). When an imaging center is particularly interested
■■ CVMR Image Processing in CVMR, it is worth using a standard surface coil, such
as a single circular coil with a diameter of approximately
■■ Reduced Data Acquisition Methods 14 cm, which improves image quality and spatial resolu-
Sense tion substantially. The best alternative is a dedicated car-
k-t Blast and k-t Sense diac phased array coil constructed of multiple elements.
■■ Appendix: Calculations on K-Space This type of coil is now commercially available from most
scanner manufacturers (Fig. 1.1). The main advantage is the
further improved image quality, spatial resolution, and the
larger field of view. An additional advantage is that phased
Introduction array coils allow application of the SENSE technique (1).
SENSE represents a revolutionizing technology in CVMR.
Cardiovascular MR (CVMR) techniques are still changing Its principle is based on parallel imaging with use of all coil
rapidly and CVMR is currently in an exciting and crucial elements. Each coil has a different sensitivity profile, which
phase for its final clinical acceptance. The main technical can be exploited to unfold undersampled acquisitions, and
limitations are now overcome by development of improved to reduce the density of the acquired k-space data, thus
scanner hardware, software, and image processing tools. speeding up scan time (see below for details). Using SENSE,
In the present chapter, basic and advanced CVMR tech- currently a twofold increase in imaging speed can be
niques will be discussed in the context of clinical appli- obtained standard, in an experimental setting higher factors
cation. Based on a virtual patient examination, relevant were reached. In general, SENSE is believed to contribute
techniques will be discussed. The focus will be on tech- significantly to the clinical acceptance of CVMR, because it
niques for functional evaluation of heart disease. Special reduces MR scan time, making it comparable to ultrasound
attention is given to reduced data acquisition methods, or CT examinations.
LWBK1209-ch01_p01-20.indd 3 16/05/13 9:45 PM

4 Part 1 ■ Basic Principles
tHP
α α
TR
Figure 1.3. Schematic diagram of the basic principle of a turbo

field echo MR sequence (TFE) in relation to the timing within the
cardiac cycle. After each rf excitation pulse (α, one line in k-space is
acquired, which is repeated two times in this case (turbo factor 2),
leading to a total of two acquired k-lines per heart phase image seg-
ment per cardiac cycle (gray area). In total, four heart beats are shown,
resulting in an image of eight k-lines. Suppose an image of 120 k-lines
is required, the procedure needs to be repeated 15 times, leading to a
Figure 1.1. Practical setup of a dedicated cardiac phased array coil total of heart beats of 60. With a heart rate of 60 beats per minute,
and vector ECG. The standard body coil is integrated with the magnet the acquisition can be completed during continuous breathing within
bore. Courtesy of Philips Medical Systems, Best, The Netherlands. 60 seconds. White squares indicate heart phase image segments, tHP =
time between each heart phase (temporal resolution). TR = repetition
time between rf excitations.
Cardiac Motion Compensation
Cardiac motion compensation is performed by synchro- poral resolution. In general, a time resolution of <40 mil-
nizing the image acquisition to the electrocardiogram liseconds per cardiac phase image is needed to enable selec-
(ECG) signal. Image formation in MR is based on filling tion of the end-systolic time frame, for calculation of, for
“k-space” during data acquisition (see Appendix), this example, the end-systolic volume and ejection fraction. For
concept will not be explained further in this context, since high-resolution clinical imaging, these 20 cardiac images
many thorough and comprehensive publications are avail- per slice cannot be obtained at once. Therefore, ECG trig-
able on this topic (2–6). ECG-triggering is aimed at filling gering was developed, to synchronize the partial k-space
k-space in multiple steps, based on the timing within the filling to the cardiac cycle. Suppose we need 128 k-lines
cardiac cycle (Figs. 1.2–1.6). For example, to construct a for the first cardiac phase image, but we can only acquire
movie of a cardiac slice, approximately 20 cardiac phases 12 k-lines for that image per heart beat, then 11 heart beats
(time frames) are needed to obtain a sufficiently high tem- are needed to complete k-space filling for that image. This
would take a breath-hold of approximately 11 seconds at a
heart rate of 60 beats per minute. Of course all 20 cardiac
QRS of ECG
tHP tHP
α α EPI-Readout
Figure 1.2. Schematic diagram of the basic principle of a gradient Figure 1.4. Schematic diagram of the basic principle of an echo-
echo MR sequence (GRE, or fast field echo (FFE)) in relation to the planar MR sequence (EPI) in relation to the timing within the cardiac
timing within the cardiac cycle. After each radio frequency (rf) excita- cycle. After each rf excitation pulse (α), in this example five lines in
tion pulse (α, one line in k-space is acquired (gray area). In total, four k-space are acquired (EPI factor 5), leading to a total of five acquired
heart beats are shown, resulting in an image of four k-lines. Suppose k-lines per heart phase image segment per cardiac cycle (gray area). In
an image of 120 k-lines is required, the procedure needs to be repeated total, four heart beats are shown, resulting in an image of 20 k-lines.
30 times, leading to a total of heart beats of 120. With a heart rate of Suppose an image of 120 k-lines is required, the procedure needs to be
60 beats per minute, the acquisition can be completed during continu- repeated six times, leading to a total of heart beats of 24. With a heart
ous breathing within 2 minutes. White squares indicate heart phase rate of 60 beats per minute, this can be performed during a breath-
image segments, tHP = time between each heart phase (temporal hold of 24 seconds. White squares indicate heart phase image seg-
resolution). ments, tHP = time between each heart phase (temporal resolution).
LWBK1209-ch01_p01-20.indd 4 16/05/13 9:45 PM

Chapter 1 ■ Clinical Approach to Cardiovascular Magnetic Resonance Techniques 5
complex of the ECG, and stops around 80% of the cardiac

cycle. Consequently, the 20 cardiac phases are distributed
over this 80% of time. In this way, the last 20% of the car-
tHP
diac cycle is not imaged. This technique is suitable to image
systolic heart function. When one is interested in the last
part of the cardiac cycle to evaluate diastolic heart func-
α α tion, a different method can be applied. Retrospective ECG
EPI-Readout EPI-Readout
gating acquires image data irrespective of the ECG, while
the ECG is recorded in parallel. Once the MR acquisition
TR is finished, the computer calculates afterwards (retrospec-
tively) the appropriate cardiac phases, based on the stored
tshot ECG and k-space data. In this way, the last part of the
cardiac cycle can be imaged also. This technique has been
Figure 1.5. Schematic diagram of the basic principle of a turbo applied clinically mostly in conjunction with MR flow map-
field echo-planar MR sequence (TFEPI) in relation to the timing ping (see below), since it allows estimation of, for example,
within the cardiac cycle. After each rf excitation pulse (α, in this case diastolic filling pattern or regurgitation volume in patients
five lines in k-space are acquired (EPI factor 5), which is repeated with mitral valve insufficiency. Recently, retrospective gat-
two times per shot (turbo factor 2), leading to a total of ten acquired ing became available in combination with faster scan tech-
k-lines per heart phase image segment per cardiac cycle (gray area). In niques, such as echo-planar imaging and balanced gradient
total, four heart beats are shown, resulting in an image of 40 k-lines. echo acquisitions (see below), so the standard nowadays
Suppose an image of 120 k-lines is required, the procedure needs to
is retrospective ECG triggering for most CVMR imaging
be repeated three times, leading to a total of heart beats of 12. With
a heart rate of 60 beats per minute, this can be performed during a
purposes.
breath-hold of 12 seconds. White squares indicate heart phase image A major problem for clinical application of cardiac
segments, tHP = time between each heart phase (temporal resolution), CVMR is the practical worry of obtaining a reliable ECG
TR = repetition time, tshot = shot duration. signal from a patient inside the MR scanner. In about 2%
to 5% of clinical cases, no reliable ECG signal may be
obtained. The electrical ECG signal is distorted by the in-
teraction of the magnetic field and the pulsating blood
phases are acquired at once, so within 11 seconds a full
flow through the aortic arch (magnetohemodynamic ef-
movie of a cardiac slice can be obtained.
fect). Recently, a new approach was launched to correct
Currently two types of ECG triggering methods are
for this problem. The vector ECG (VCG) is based on the
available. The first is prospective triggering, meaning that
3-dimensional orientation of the QRS-complex and T-wave
image acquisition starts at a fixed delay after the QRS-
of the ECG and the distorting component (7). In this way,
the MR acquisition can be triggered by the QRS-complex
only and neither by mistake by the T-wave or ECG distor-
Flow sensitive Flow compensated
tions nor by signal induced by gradient switching. The VCG,
in conjunction with the dedicated cardiac synergy surface
coil and SENSE, has revolutionized clinical application of
CVMR (Fig. 1.1). Today, there is no practical limitation for
tHP CVMR, except for the conventional MR exclusion criteria,
α α such as unstable or sensitive implanted metal objects or
EPI-Readout EPI-Readout pregnancy.
TR
Respiratory Motion
tshot Compensation
Figure 1.6. Schematic diagram of the basic principle of a turbo A second source of image distortion is respiratory motion.
field echo-planar MR flow sequence (TFEPI flow) in relation to the A decade ago, MR acquisitions were so time consuming
timing within the cardiac cycle. After each rf excitation pulse (α, in that it was impossible to perform breath-hold imaging. At
this case five lines in k-space are acquired (EPI factor 5), which is that time, an inventive technique was introduced called
repeated two times (turbo factor 2), leading to a total of ten acquired respiratory ordered phase encoding (ROPE) (8) or phase
k-lines per heart phase image segment per cardiac cycle (gray area). encoding artifact reduction (PEAR), which was based on a
In total, four heart beats are shown for the flow sensitive image, as special k-line reordering technique, combined with a respi-
well as four heart beats for the flow compensated image, resulting in ratory tracking device around the abdomen. By positioning
images of 40 k-lines for each image type. Suppose the images require
acquired k-lines during breathing in the periphery of k-space
120 k-lines for a sufficiently high spatial resolution, the procedure
needs to be repeated three times, leading to a total of heart beats of
(this is the part of k-space less sensitive to motion), breath-
24. With a heart rate of 60 beats per minute, this can be performed ing artifacts were decreased. This type of artifact reduction
during a breath-hold of 24 seconds. White squares indicate heart phase improved image quality, but was not optimal yet, because
image segments, tHP = time between each heart phase (temporal resolu- the indirect relation between abdominal movement and the
tion), TR = repetition time, tshot = shot duration. actual heart motion.
LWBK1209-ch01_p01-20.indd 5 16/05/13 9:45 PM

Diaphragm Lung
Expiration Figure 1.7. The respiratory navigator

pencil beam (A, white line, B white dot) is
positioned on the right hemi-diaphragm at
the interface between lung and liver, based on
a survey image in the coronal (A) and sagit-
A
tal (B) plane. The one-dimensional naviga-
Inspiration C tor image is acquired repeatedly over time,
thereby constructing an image of diaphrag-
matic motion (C). The edge between lung and
Time Liver liver is traced automatically, yielding a breath-
ing curve similar to that represented in the
B
diagram of Figure 8.
Due to the development of faster MR imaging tech- Recent technical advances have made it possible to
niques, such as echo-planar imaging and turbo-field echo a cquire a stack of 12 slices, with around 25 cardiac phases
imaging, it became possible to acquire image data during each, in a single breath-hold of up to 20 seconds. These
a short breath-hold of around 15 seconds. The general dis- ultra-fast techniques are combinations of turbo-field echo,
advantage of breath-hold acquisitions is that the reproduc- echo-planar, or spiral k-space acquisition. The first appli-
ibility of the breath-hold level is not optimal. In a multi-slice cation is the assessment of cardiac function, because the
multiple breath-hold acquisition, this may introduce errors single breath-hold eliminates the previously mentioned
in, for example, the summed end-diastolic volume of differ- problem of slice level reproducibility in multiple breath-
ent slices which were acquired during different breath-holds. hold approaches. However, the image spatial resolution of
Without proper patient instruction, this can lead to high this technique is still quite low, allowing only evaluation
variability in the measurements, or even worse, to inaccu- of ventricular volume changes and some rough estimation
rate clinical data. So when using breath-hold acquisitions, of wall motion abnormalities. In the future, combination of
it is necessary to instruct the patient carefully to hold his the ultra-fast acquisition schemes with SENSE may allow
breath in expiration, which minimizes the above-mentioned higher resolution imaging. A promising technique for ultra-
problem of breath-hold level reproducibility. fast whole heart functional imaging in a single breath-hold is
Some high-resolution MR acquisitions require longer ac- the so called k-t BLAST technique (see below).
quisition times than possible during a breath-hold. For ex- Another promising development is real-time imaging of
ample, coronary artery MR angiography can be performed the heart. Real-time cardiac imaging was introduced al-
during a short breath-hold, but the optimal quality can only ready in the early days of MR (12). Recently data has been
be obtained by using the respiratory navigator technique
(9,10). Respiratory navigation is based on a 1-dimensional
image positioned at the interface between lung and liver. Leading Nav Trailing Nav QRS of ECG
In this way the motion of the diaphragm can be tracked.
Usually, the navigator beam is positioned on the right hemi-
diaphragm (Fig. 1.7) and is acquired before and after every
MR data acquisition block (Fig. 1.8), with an acquisition
duration of only 30 milliseconds. The acquisition is then
gated to the automatically traced breathing signal derived
from the respiratory navigator. A window around the end-
expiration position of the diaphragm is defined, determining
the positions in which MR data is accepted. For example,
a 3 mm acceptance window means that in end-expiration, Data accepted: leading Nav
still 3 mm motion is accepted to occur. Respiratory naviga- Data accepted: leading + trailing Nav
tors can also be used for slice tracking; this means that the Data rejected
slice position is adjusted respective to the breathing changes
as detected by the navigator. In this way, the acquisition is Figure 1.8. The MR acquisition is gated to a predefined acceptance
window (two solid horizontal lines) based on the traced respiratory
not gated to a certain predefined acceptance window, but
signal (curved solid line) around end-expiration. K-space data is
acquires all scanned data. The combination of these two acquired continuously, but only data are stored that fulfill the require-
techniques seems currently the best solution. Within the ac- ment that (1) before data acquisition the navigator was within the
ceptance window, of say 3 mm, the remaining respiratory acceptance window (leading navigator); or that (2) before and after
motion is compensated by using the tracking technique. This data acquisition the navigator was within the acceptance window
technique showed good results for coronary magnetic reso- (leading and trailing navigator). Other acquired image data is deleted.
nance angiography (MRA) (11). This is the principle of a real-time prospective respiratory navigator.
LWBK1209-ch01_p01-20.indd 6 16/05/13 9:45 PM

resented showing full ventricular coverage in real time

p α°
without the need for ECG-triggering or breath-holding RF
(13,14). In these preliminary studies, accurate quantification
was shown for ventricular function. In combination with fu- Gslice
ture real-time image analysis, this may revolutionize the way
we look at CVMR. Gphase
Greadout
Survey
Acq
The first step for the CVMR examination is the acquisition
of a localizer to determine general anatomy, which forms the Figure 1.9. Balanced FFE pulse sequences belong to the group of
basis for further acquisitions. The purpose of this first scan “Steady State Free Precession” techniques. Time-balanced gradients
is to image the cardiac region in three basic orientations: are applied for all gradient directions: slice selection, phase encoding,
coronal, transverse, and sagittal planes, using 15 slices of and frequency readout. In combination with the alternating phase of
10 mm each for every orientation. A decade ago, these the excitation pulse this enables acquisition of both FID and Echo
“scout” images were obtained by using a multi-slice spin- signal. The sequence produces a high image intensity for tissues with
echo technique, sometimes in combination with turbo a high T2/T1 ratio, independent of the repetition time TR. Balanced
FFE images are obtained after field shimming, since field homogeneity
spin-echo. This technique can still be used if your local MR
is very important.
scanner does not support faster imaging techniques. A faster
technique for acquiring survey images is, for example, a
turbo-field-echo or turbo-field-echo-planar MR sequence.
Recently, balanced gradient echo (bFFE, bTFE, true-FISP) Another approach is to use the survey images as input for
techniques became available, yielding images with high con- an algorithm, which plans the desired imaging planes au-
trast between blood and myocardium (Figs. 1.9 and 1.10). tomatically (15). Its principle is that the software package
Survey images can be acquired during continuous breath- finds the position of heart in the survey images automati-
ing, using the respiratory navigator, or during a breath-hold. cally, and uses this information for the planning of the other
Currently, excellent images can be obtained using balanced- acquisitions. This operator-independent automatic plan scan
FFE without respiratory motion compensation, with an method yields reliable quantification of, for example, end-
acquisition time of only 15 seconds. diastolic volume and ejection fraction, and reduces inter-
examination variability substantially. Therefore, this method is
ultimately suitable for patient follow-up to evaluate therapy ef-
Plan Scan fects of, for example, antihypertensive or lipid-lowering drugs.
However, the automatic plan scan procedure is currently only
Based on survey images, further MR scans can be planned. available offline as a research tool, and the time-efficiency of
This can be performed by hand, using the planning tools the real-time planning alternative seems to be higher.
available in the standard scanner software. Since the begin-
ning of the new millennium, planning can be performed in
real time, even in combination with balanced-TFE. With a Anatomy
real-time planning tool, manual changes in imaging plane
are immediately executed by the free-running MR acquisi- Today, the main patient category with a clinical referral
tion. In this way, the optimal imaging plane can be found for CVMR is congenital heart disease. The primary inter-
within seconds. Once the desired imaging plane is reached, est of the referring physician is usually evaluation of global
the geometry settings can be stored for later use during the cardiovascular anatomy. The purpose of these MR scans is
CVMR exam. The time efficiency increases substantially by to image the heart in three dimensions. In this way, com-
using this real-time planning tool, allowing routine applica- plex congenital cardiac deformations can be diagnosed and
tion of CVMR. followed over time to evaluate surgical correction of the
Figure 1.10. Example of balanced

FFE images acquired in the two-chamber
view (A), four-chamber view (B), and
short-axis view (C). Each image con-
sists of 25 cardiac time frames, which
were acquired during a breath-hold of
12 seconds for each view separately.
Images are shown in the end-diastolic
time frame. Note the excellent contrast
between blood and myocardium, as
compared to the echo-planar images in
A B C
the other figures of this chapter.
LWBK1209-ch01_p01-20.indd 7 16/05/13 9:45 PM

Non-selective Slice
180° selective Echo train 180° pulses
180°
90°
Figure 1.11. Schematic diagram of a black blood pulse

sequence. A nonselective 180º pulse inverts all signals.
Subsequently a slice-selective 180º pulse resets signal of the
Blood signal nulled studied slice. Blood with inverted signal flows into the slice
Inversion recovery plane. After a delay (Tinv = inversion time) the blood signal
of blood signal is nulled and data acquisition is performed with a fast spin
Tinv echo pulse train. The image is obtained in a breath-hold.
Note that Tinv is dependent on the heart rate.
anomalies. In the past, the time consuming (turbo) spin-echo Function

was the technique of choice. The major disadvantage of this
technique was the presence of major breathing artifacts, The second stepping stone in CVMR is evaluation of cardiac
hampering routine evaluation of cardiac anatomy. function. The second most frequent indication for cardiac
Nowadays, several alternatives are available. The latest MR is evaluation of myocardial wall motion abnormali-
technique is a multiple breath-hold dual inversion black ties in patients with suspected myocardial ischemia. Most
blood turbo spin-echo technique (Figs. 1.11 and 1.12), of these patients are referred by the echocardiography lab
which yields excellent image quality (16). Disadvantage of because some patients are hard to image using ultrasound.
this approach is that each slice has to be imaged during a Usually these patients are obese or have lung emphysema.
separate breath-hold, with again the problem of breath-hold The expectation is that, due to the experience cardiologists
level reproducibility. The expectation is that in the future are obtaining with MR through the aforementioned patient
this technique will be further optimized, for example, by group, there will be an increasing demand for MR analysis
combining it with SENSE, allowing single breath-hold or of cardiac function.
respiratory navigator gated acquisitions. Similar to echocardiography, the short-axis view of the
left ventricle is the working horse view in CVMR. To be able
to acquire the short-axis view, first two long-axis views need
to be acquired, which can yield diagnostic information itself,
especially concerning the apical wall motion pattern of both
ventricles and valvular function. First, the vertical long-
axis, or two-chamber view, needs to be acquired (Fig. 1.13).
Based on a transverse survey image at the level of the mi-
tral valve, the center of the slice is positioned approximately
in the middle of the mitral valve, and angulated in such a
way that the slice cuts through the left ventricular apex on
a lower transverse survey image. For selection of the end-
systolic cardiac time frame, the two-chamber view needs to
be acquired using a dynamic gradient echo technique. This
can be performed using a conventional gradient-echo tech-
nique during continuous breathing, but better results can
be obtained when using an ultra-fast breath-hold technique
such as echo-planar imaging (17). Currently, the optimal op-
tion is the use of a balanced-TFE sequence, because the con-
trast between blood and myocardium is excellent.
Next, the horizontal long-axis, or four-chamber view,
needs to be acquired (Fig. 1.14). Planning of the four-chamber
view is based on the diastolic and systolic images of the two-
chamber view. First, the center of the slice has to be posi-
tioned at one-third of the lower part of the mitral valve on the
end-systolic two-chamber image, hereafter the slice needs to
be angulated through the apex. After this, check the position
Figure 1.12. The “black blood” image was obtained using a dual of the slice in the diastolic two-chamber image to ensure the
inversion fast spin echo technique, in combination with SENSE atrium is imaged properly. The four-chamber image also
(factor 2). Echo spacing 4.3 milliseconds, FOV 350 × 350 mm, matrix needs to be acquired using a dynamic imaging technique,
192 × 146, slice thickness 8 mm. Scan time was 6 seconds. Courtesy of preferably balanced-TFE. The two- and four-chamber views
Philips Medical Systems, Best, The Netherlands. yield some important clinical information, mainly concerning
LWBK1209-ch01_p01-20.indd 8 16/05/13 9:45 PM

Figure 1.13. The vertical long-axis, or two-chamber view, is planned in the following way. Based on a
transverse survey image at the level of the mitral valve (A), the center of the slice is positioned approximately
in the middle of the mitral valve, and angulated in such a way that the slice intersects with the left ventricular
apex on a lower transverse survey image (B). An end-diastolic two-chamber image is shown here, acquired
during a 12-second breath-hold in expiration using echo-planar MR (C).
Figure 1.14. The horizontal long-axis, or four-chamber view, is planned in the following way. The center of
the slice has to be positioned at one-third of the lower part of the mitral valve on the end-systolic two-chamber
image (B), hereafter, the slice needs to be angulated through the apex (A, B). After this, check the position of the
slice in the diastolic two-chamber image (A) to ensure the atrium is imaged properly. An end-diastolic four-cham-
ber image is shown here, acquired during a 12-second breath-hold in expiration using echo-planar MR (C).
LWBK1209-ch01_p01-20.indd 9 16/05/13 9:45 PM

A B
C D
Figure 1.15. The short-axis view is planned in the following way. Position a stack of slices perpendicular
to the long-axis of the left ventricle, as defined from the mid-mitral point to the left ventricular apex, based on
the diastolic (A) and systolic (B) images of the four-chamber view, and the diastolic (C) and systolic (D) two-
chamber view. Be sure to include the entire ventricle in end-diastole, as can be checked on the end-diastolic
images (A, C). An end-diastolic short-axis image is shown here, acquired during a 12-second breath-hold in
expiration using echo-planar MR (E).
ventricular anatomy and dimensions, presence of hypertro- seems perfectly suitable for cardiac stress imaging during
phy, or valve insufficiencies, and the like. continuous increasing dobutamine dosage and for imaging
The final step for evaluation of ventricular function is the of the heart in the presence of arrhythmias.
planning of the short-axis view itself (Fig. 1.15). Position a
stack of 10–12 slices of 8 to 10 mm thickness each perpen-
dicular to the long-axis of the left ventricle, as defined from Perfusion
the mid-mitral point to the left ventricular apex, based on
the systolic and diastolic images of the four-chamber view Myocardial perfusion of the ischemic heart may be studied
and the systolic two-chamber view. Be sure to include the with MR (18). Most perfusion analysis approaches make
entire ventricle in end-diastole, which can be checked on the use of intravenously injected gadolinium-based MR contrast
end-diastolic images. The latter is important for calculation agents. This type of contrast medium temporarily reduces
of the end-diastolic volume, stroke volume, ejection frac- the T1-relaxation time and thereby relatively increases the
tion, and other derived functional parameters. MR-signal intensity of well-perfused tissues. After contrast
The short-axis scan with 10–12 slices covering the injection, ischemic myocardial regions show up as areas with
entire ventricles can be acquired using a conventional no or little signal intensity change as compared to well-per-
gradient-echo sequence during continuous breathing, which fused myocardium. In order to visualize the myocardial pas-
is very time consuming. A better alternative is a breath-hold sage of the injected contrast, fast T1-weighted MR-techniques
acquisition using echo-planar imaging, but again, the best should be used. The first reports on visualization of perfu-
technique available today is the balanced-TFE MR sequence sion defects in human patients described the acquisition of
(Fig. 1.10). Nowadays, these scans are performed using mul- a single slice level (19). Currently, perfusion studies apply
tiple breath-holds in expiration, with the associated prob- fast gradient echo pulse sequences (magnetization prepared
lems as pointed out before. In the near future, it is expected turbo Field Echo/EPI/FLASH) which enable the repetitive
that the entire short-axis stack of slices can be acquired registration of three or more anatomical levels at every heart
within a single breath-hold up to 15 seconds using balanced- beat, or alternatively six levels with a temporal resolution
TFE k-t SENSE (see below). One step further into the future of every other heart beat (20). A very promising technique
may lead to real-time imaging of the short-axis stack, al- for myocardial perfusion imaging is k-t BLAST perfusion
though this poses several problems on the offline image pro- (see below). The training data is acquired interleaved with
cessing. Real-time imaging of the short-axis view currently undersampling data, thereby dynamically correcting for
LWBK1209-ch01_p01-20.indd 10 16/05/13 9:45 PM

contrast changes during myocardial passage. It is expected bright area on images acquired with T1-weighted MR pulse
that k-t BLAST will allow full cardiac coverage CVMR per- sequences ranging from basic spin-echo methods to more
fusion imaging during first pass of contrast. sophisticated gradient echo sequences using an inversion or
Quantitative image analysis yields parameters to charac- saturation prepulse. Currently the principle “bright is dead,”
terize the bolus passage immediately after administration of indicating that bright areas on a delayed MR-image after con-
the contrast agent. Parameters obtained from the image series trast injection correspond with nonviable myocardium, is sub-
describing this first-pass are, for example, rate and level of en- ject to lively debate (24–26). The additional value of delayed
hancement, time-to-peak, and mean transit time. These char- enhancement to rest-stress wall motion imaging to determine
acteristic parameters may be obtained for each image pixel myocardial viability is questionable. Delayed enhancement
and can be graphically displayed in so-called parametric im- CVMR, in combination with first-pass perfusion and wall
ages visualizing the anatomical location of abnormalities (21). motion imaging at rest, has an important clinical value when
stress CVMR cannot be performed.
Delayed Enhancement
Flow
After injection of a low molecular weight contrast agent, its
plasma concentration will reach a maximum value and rap- One of the advantages of MR in cardiovascular diagnosis
idly decrease due to diffusion to the interstitial space and renal is the possibility of this technique to measure flow velo
washout. Contrast agent that diffused to the interstitial space city (cm/sec) and flow volume (mL/sec). MR flow measure-
will be resorbed into the capillary bed and undergo renal ment is based on the principle of “spin phase” (27). Usually
excretion. However, when the tissue is damaged, for example, MR-images only use the absolute value of the MR-signal aris-
due to infarction, the resorption rate of contrast agent will ing from the slice under investigation. However, the acquired
be diminished. At 15 to 30 minutes after contrast injection, data also contains information on a property called spin
washout will be complete in normal myocardium in contrast phase. MR data acquisition and post-processing can be set
to infarcted or edematous tissue. This phenomenon is the basis up in such a way that two images of each slice are produced:
of “delayed enhancement imaging”. Various authors (22,23) an image with gray values representing spatial localization
have described the relation between myocardial viability and of the protons, and an image with gray values representing
the size of the area displaying delayed enhancement. In MR the velocity of the protons present in each image element.
images, the presence of contrast agent can be detected as a The latter image is called the velocity map (Fig. 1.16). In a
A B
C D E
Figure 1.16. The mitral valve flow acquisition is planned in the following way. The center of the slice is
positioned in the middle of the mitral valve on the end-systolic two- (D) and four-chamber (B) images, and
angulated parallel to the mitral valve, also based on the end-diastolic two- (C) and four-chamber (A) images.
An end-diastolic image of the mitral valve is shown here (E), the upper image shows the normal, or modulus
image, the lower image shows the velocity-encoded image (velocity map).
LWBK1209-ch01_p01-20.indd 11 16/05/13 9:45 PM

600 mapping. These field gradients may be applied “in-plane” or

1 2 3 4 5 6
“through-plane,” thus encoding different components of the
500 flow direction. In most instances, the imaging slice is posi-
400
tioned to measure through-plane flow, but it is also feasible
Flux [mL/s]
to quantify flow in all three dimensions by adding additional

300 data acquisitions.
In order to obtain useful data on blood flow, it is neces-
200 sary to measure flow in a vessel or valve of interest at many
instances during the cardiac cycle. For example, the mea-
100
surement of flow velocity taken from a region of interest at
0 many time frames will result in a flow-velocity versus time
0 200 400 600 800 1000 1200 curve (see below), which gives information on flow velocity
Time [ms] changes during the cardiac cycle. In such a flow-velocity
curve, the area under the curve is computed to obtain the
Figure 1.17. Typical mitral valve flow curve obtained after trac-
ing the orifice of the valve in all cardiac time frames on the velocity
stroke volume, that is, the amount of blood passing through
encoded images. The integration of velocity data over time for all pixels the region of interest during a cardiac cycle. Other valuable
enclosed in the traced area results in a volume flow (flux) curve. The hemodynamic aspects related to cardiac contraction and
early peak filling rate (2) is a result of the pressure difference between relaxation may also be gleaned from the flow-velocity
the left atrium and left ventricle, and is a passive process. The atrial curve (28).
peak filling rate (5) is a result of left atrial contraction, and is an active Flow encoding MR scans are based on gradient-echo
process. Another functional parameter which can be derived from the pulse sequences, in combination with prospective-triggering
curve is the fastest change in flux between, for example 2 and 3, the so or retrospective ECG-gating. For evaluation of diastolic ven-
called “early deceleration” peak. Note the diastase between 3 and 4. tricular function, retrospective ECG-gating is required, be-
cause prospective ECG-triggering only images the first 80%
of the cardiac cycle, and the atrial contribution of ventricu-
velocity map pixels of static tissue will be displayed with lar filling occurs in the last 10% to 20% of the cardiac cycle.
intensity zero, whereas pixels of, for example, moving blood Currently, a standard flow measurement technique has the
will have a positive or negative value, dependent on the duration of 2–3 minutes, and the result will represent the
direction of flow. The use of additional magnetic field gra- average flow during the acquisition period. Recent devel-
dients in the imaging pulse sequence is the basis of velocity opments in scanner hardware and software have enabled
Figure 1.18. The tricuspid valve flow is based on an additional survey image; the right ventricular two-
chamber view. The right ventricular two-chamber view is planned on the end-systolic four-chamber view (B). The
center of the slice is positioned in the middle of the tricuspid valve and angulated through the apex of the right
ventricle based on the end-diastolic (A) and end-systolic images (B). An end-diastolic right ventricular two-
chamber image is shown here, acquired during a 12-second breath-hold in expiration using echo-planar MR (C).
LWBK1209-ch01_p01-20.indd 12 16/05/13 9:45 PM

A B
C D E
Figure 1.19. The tricuspid valve flow acquisition is planned in the following way. The center of the slice is
positioned on the center of the tricuspid valve on the end-systolic four-chamber view (B), and angulated paral-
lel to the tricuspid valve based on the end-diastolic four-chamber view (A). Thereafter, the slice is angulated on
the end-diastolic (C) and end-systolic (D) right ventricular two-chamber view parallel to the tricuspid valve.
An end-diastolic image of the tricuspid valve is shown here (E), the upper image shows the normal, or modulus
image, the lower image shows the velocity-encoded image.
realization of real-time flow measurement (29) in a clinical v entricular two-chamber view parallel to the tricuspid
setting (30), thus allowing phenomena with fast changes valve. A typical example of a tricuspid valve flow curve is
in flow to be investigated with MR. The k-t BLAST tech- given in Figure 1.20.
nique can also be used in combination with flow encoding. The acquisition to measure blood flow velocities and volume
Currently, there is not much experience with this technique, through the ascending aorta is planned on the original
but it is expected that k-t BLAST flow will allow ultra-fast
flow imaging.
The clinically most feasible cardiac flow acquisitions are 500
through the mitral valve, tricuspid valve, ascending aorta, 1 2 4
3 5 6
450
and pulmonary artery: 400
350
The mitral valve flow acquisition is planned on the two-
Flux [mL/s]
300
and four-chamber view in end-systole (Fig. 1.16). The 250
center of the slice is positioned in the middle of the 200
mitral valve on the end-systolic two- and four-chamber 150
images, and angulated parallel to the mitral valve. A
100
typical example of a mitral valve flow curve is given in
50
Figure 1.17.
0
The tricuspid valve flow is based on an extra survey image; 0 200 400 600 800 1000 1200
the right ventricular two-chamber view (Fig. 1.18). This Time [ms]
view is planned on the end-systolic four-chamber view.
The center of the slice is positioned in the middle of the Figure 1.20. Typical tricuspid valve flow curve obtained after trac-
ing the orifice of the valve in all cardiac time frames on the velocity
tricuspid valve and angulated through the apex of the
encoded images. The early peak filling rate (2) is a result of the pressure
right ventricle. The tricuspid valve flow acquisition is difference between the right atrium and right ventricle, and is a passive
first planned on the end-systolic four-chamber view (Fig. process. The atrial peak filling rate (5) is a result of left atrial contrac-
1.19). The center of the slice is positioned on the center of tion, and is an active process. Another possible functional parameter
the tricuspid valve, and angulated parallel to the tricus- which can be derived from the curve is the fastest change in flux
pid valve. Thereafter, the slice is angulated on the right between, for example 1 and 2, the so called “early acceleration” peak.
LWBK1209-ch01_p01-20.indd 13 16/05/13 9:45 PM

Figure 1.21. The flow acquisition

through the ascending aorta is planned
in the following way. An unangulated
slice is positioned perpendicular to the
ascending aorta on a coronal survey
image (A), usually this is at the level of
the bifurcation of the pulmonary artery.
On a sagittal image of the original sur-
vey, the angulation can be adjusted if
necessary, although this is usually not
the case (not shown). An end-diastolic
image of the ascending aorta is shown
A here (B), the upper image shows the
normal, or modulus image, the lower
image shows the velocity-encoded
B
image.
survey images in the coronal and sagittal view (Fig. 1.21). survey image (Fig. 1.23). In addition, the slice position
An unangulated slice is positioned perpendicular to the in caudocranial direction has to be positioned just before
ascending aorta on a coronal image; usually this is at the the bifurcation of the pulmonary artery on a transverse
level of the bifurcation of the pulmonary artery. On a image of the original survey. A typical example of a
sagittal image of the original survey, the angulation can flow curve through the pulmonary artery is given in
be adjusted if necessary; usually this is not the case. A Figure 1.24.
typical example of a flow curve through the ascending
aorta is given in Figure 1.22.
The acquisition to measure blood flow through the pulmo- Coronary MRA
nary artery is planned on some original survey images in
the coronal and transverse view (Fig. 1.23). First, another In this chapter, coronary MRA will not be discussed in
survey image needs to be acquired of 1–3 slices in the great detail, and the reader is referred to later chapters for
center of and parallel to the pulmonary artery based on in-depth explanation of the techniques and procedures.
an original sagittal image. Then, the pulmonary artery The only practical point for now is to realize that there are
flow acquisition can be planned perpendicular to the three major points of concern for coronary MRA. The
pulmonary artery based on the angulated survey image first is the cardiac motion itself, limiting the shot length
as was described just before, and on the original sagittal of image acquisition to below 100 milliseconds, and the
second is respiratory motion which can be corrected for
by using respiratory navigators with residual respiratory
700
1 2 34 5 6 motion of up to 3 mm. Currently, the in-plane image reso-
lution is limited to 0.7 mm because of the 100 millisec-
600 onds shot duration, in this period of time there still is some
500 cardiac motion, causing blurring of the image. In theory,
an in-plane resolution of around 0.3 mm can be reached,
Flux [mL/s]
400 but this is currently hampered by the long shot duration

300 of around 100 milliseconds. With improvement of scanner
hardware and software, it is expected that isotropic resolu-
200
tion of 250 micrometer can be reached in the near future,
100 possibly by performing coronary MRA at 3 Tesla instead of
the standard clinical field strength of 1.5 Tesla. All major
0
0 200 400 600 800 1000 1200 MR machine manufacturers have 3 Tesla MR scanners
–100
Time [ms]
available now. Initial results at 3 Tesla using SENSE are
very promising (31).
Figure 1.22. Typical flow curve through the ascending aorta The third point of attention is the way image contrast
obtained after tracing the contour of the ascending aorta in all cardiac
between the coronary vessel and surrounding tissue is
time frames on the velocity-encoded images. The peak ejection rate
(2) is a result of left ventricular contraction. The area under the curve optimized. Roughly there are two alternatives, the T2-
between 1 and 3, is the left ventricular stroke volume. Another inter- preparation technique, and secondly the use of a contrast
esting functional parameter which can be derived from the curve is agent, combined with spiral k-space filling. It is important to
the fastest change in flux between, for example 1 and 2, the so called realize that it is maybe more important to perfectly suppress
“aortic acceleration” peak. signal from surrounding tissue, than to enhance the signal
LWBK1209-ch01_p01-20.indd 14 16/05/13 9:45 PM

A B
E
C D
Figure 1.23. The flow acquisition through the pulmonary artery is planned in the following way. First,
another survey image needs to be acquired of 1–3 slices in the center of and parallel to the pulmonary artery
based on an original sagittal survey image, (A). Then, the pulmonary artery flow acquisition can be planned
perpendicular to the pulmonary artery based on the angulated survey image (B), and on the original sagittal
survey image (C). In addition, the slice position in caudocranial direction has to be positioned just before the
bifurcation of the pulmonary artery on a transverse image of the original survey (D). An end-diastolic image
of the pulmonary artery is shown here (E), the upper image shows the normal, or modulus image, the lower
image shows the velocity-encoded image.
inside the coronary arteries. Application of spiral acquisi- Vessel Wall Imaging
tions and the balanced-TFE technique showed a substan-
tial increase in image quality (32), and is expected to have In this chapter, vessel wall imaging will not be discussed in
advantage of acquisition at 3 Tesla. detail, see later chapters for an in-depth explanation of the
techniques and procedures. In general, vessel wall imaging is
an important feature of CVMR, which currently cannot be
600
1 2 3 4 5 6 accomplished noninvasively with other imaging modalities.
500 Therefore, much effort is now being undertaken to develop
MR techniques capable of, for example, imaging of the cor-
400 onary vessel wall (33). Again, switching to the field strength
of 3 Tesla may improve earlier results in this area. One
Flux [mL/s]
300
of the future prospects of vessel wall imaging is to determine
200 the composition and stability of atherosclerotic plaque. In
the future, vessel wall imaging may be combined with inter-
100 ventional CVMR and interventional MR procedures, such
as dottering and stenting under MR guidance.
0
0 200 400 600 800 1000 1200
−100 Time [ms]
CVMR Image Processing
Figure 1.24. Typical flow curve through the pulmonary artery
obtained after tracing the contour of the pulmonary artery in all car-
One of the main—if not the main—obstructions for routine
diac time frames on the velocity encoded images. The peak ejection
rate (2) is a result of right ventricular contraction. The area under the
clinical application of CVMR is the relatively underdevel-
curve between 1 and 3, is the right ventricular stroke volume. Another oped state of cardiac image analysis. Currently, the only
interesting functional parameter which can be derived from the curve option to obtain reliable measurements is to manually per-
is the fastest change in flux between, for example 2 and 3, the so called form image analysis. The drawing of, for example, endocar-
“pulmonary deceleration” peak. dial and epicardial contours is a very time-consuming part
LWBK1209-ch01_p01-20.indd 15 16/05/13 9:45 PM

of the cardiac exam. Several software packages are available scheme would only need to update highly dynamic informa-
to assist in this tedious task, such as the MASS and FLOW tion at a high rate, while less dynamic or static information
software (34,35). Recently, a new technique was developed can be acquired at a much lower rate. Among such tech-
to detect myocardial borders fully automatically, without niques, k-t BLAST and k-t SENSE (38) have received wide-
any user interaction. This approach uses the “active appear- spread attention. In k-t BLAST and k-t SENSE the image
ance model” algorithm, based on a statistical description of content is estimated based on the so-called training data
the shape myocardial contours can have in four dimensions which serve as guidance for image reconstruction of missing
(36). The method showed excellent first results, therefore, information. k-t BLAST and k-t SENSE permit typically five-
this technique may for the first time allow accurate, fully to eightfold accelerations in cardiac imaging.
automated evaluation of cardiac function. This technique In MR imaging, the information necessary to reconstruct
can also be applied in the future on perfusion and delayed an image is sampled sequentially, typically line-by-line in the
enhancement images. spatial frequency or k-space domain. The distance between
Future developments in image processing are aimed at the lines in k-space is inversely proportional to the size of
developing a single software package for evaluation of func- the field of view which encloses the object (see Appendix).
tion, perfusion, delayed enhancement, as well as coronary To reconstruct an image from the sampled data the Fourier
MRA images. In analogy to the development of a single transform is applied. Faster scanning by means of reduced
comprehensive cardiac acquisition scheme, the concept of data acquisition schemes implies less dense sampling with
one-stop-shop cardiac image processing is now within reach. a resultant reduction of the field of view. If the field of view
A step further will be the analysis of CVMR images in real becomes smaller than the actual object, aliasing occurs in
time, so that clinical evaluation based on quantification can the image which manifests as backfolding artifacts of image
be performed while the patient is still in the MR scanner, portions. Deliberate reduction of k-space sampling density
allowing online quality control of the previous acquisitions. for faster data acquisition is referred to as undersampling
In the ideal world, real-time CVMR image acquisition and hereafter. Reduced data acquisition schemes as outlined
real-time image analysis are combined, the latter being not above may be classified into methods operating on a car-
completely unrealistic on the time scale of a decennium. diac frame-by-frame basis such as SENSE and GRAPPA and
into methods taking into account the temporal dimension.
Examples for the latter kind are k-t BLAST and k-t SENSE.
Reduced Data Acquisition While SENSE and GRAPPA can operate with all available
Methods scan protocols, k-t BLAST and k-t SENSE are only possible
if time-resolved or cine data are acquired.
In cardiac imaging, data acquisition speed is crucial not only
because scans are typically run during breath-holding of
Sense
the subject but also because cardiac bulk motion limits the
data acquisition period during the cardiac cycle. The lim- Figure 1.25 illustrates the principle of parallel imaging. In
ited data acquisition window available frequently requires this exemplary illustration, twofold faster data acquisition
trade-offs in spatial and/or temporal resolutions. Increasing is achieved by sampling only every other line in k-space
imaging speed has mainly been related to stronger magnetic which results in two-times backfolding artifacts in the
field gradients. Ever faster gradient systems have, however, reconstructed image. For example, the image intensity seen
led to potential risks including peripheral nerve stimulation. at point P as indicated in Figure 1.25 is the superposition
Accordingly, further progress by mere engineering may be of the original signal at that point and the signal folded in
limited. The advent of parallel imaging has made alternative from position P′ which is at P plus the field of view divided
means for speeding up data acquisition available by exploit- by the reduction factor which is in this example equal to
ing the characteristics of antenna arrays for signal recep- two. In order to unfold the signal received at point P, addi-
tion. Among all the different parallel imaging techniques tional information is required, since two unknowns exist
presented so far, SENSE (1) and GRAPPA (37) have been with just the superposition thereof known. This additional
most successful and are nowadays widely available on com- information is derived from the difference of signal recep-
mercial MR systems. In parallel imaging, only a subset of tion of multiple receive coils placed around the object. In
the data required to reconstruct a full image is encoded by the example, at least two receive coils which “see” point P
magnetic gradient action. The missing information is repop- and the point P′ folded upon it with different intensity (and
ulated based on differences in perception of the object sig- phase) are required to unfold the folded image pixel. By
nal by multiple receiver antennae placed around the object. calculating the difference in sensitivity of coil #1 and coil
Typical speed-up factors achievable in cardiac imaging with #2 for all image points, all folded image pixels can be reas-
standard cardiac coil arrays range from two to four. Beyond signed successively.
this range, images are increasingly compromised by local
noise amplification.
K-T Blast and K-T Sense
A different approach to faster imaging of dynamic objects
is based on the observation of considerable correlation of In k-t BLAST and k-t SENSE, acquisition speed is increased
image information in space and time. This becomes appar- in a similar fashion to parallel imaging except that under-
ent, if for example, a cine image series of the heart is con- sampling is not only applied along spatial axes but also
sidered. Large regions of the image remain static or move along the time axis. Accordingly, k-t BLAST and k-t SENSE
in a coherent fashion. Accordingly, an optimized acquisition allow for higher reductions in scan time relative to methods
LWBK1209-ch01_p01-20.indd 16 16/05/13 9:45 PM

Coil sensitivity Undersampling
2 acceleration
Coil #1 Coil #2 P+P’
SENSE
reconstruction
Figure 1.25. Principle of parallel
imaging. Data acquisition is accelerated
by skipping data lines in k-space result-
ing in aliasing artifacts in the image
(for example, point P contains signal
from two points in the original). The
folded image pixel can be resolved by
P
exploiting differences in coil sensitivity
of antenna elements (e.g., coil #1 and
coil #2) placed around the object (i.e.,
point P and the point P′ folded upon it
are “seen” differently by coil #1 and coil
#2). For further explanation, see text. P’
operating on a frame-by-frame basis. On the other hand, the pattern and allow decoding of the compressed informa-
the methods are only possible for dynamic or cine imaging tion as sampled in the undersampling stage.
protocols. Although data acquisition is identical for both k-t
The undersampling pattern applied in the k-t BLAST BLAST and k-t SENSE, image reconstruction differs. In k-t
and k-t SENSE method can be judiciously designed to take BLAST, image reconstruction is done separately for each
into account the characteristics of dynamic objects typi-
cally imaged. This, in particular, refers to the fact that, for
example, chest wall, liver, and other structures surrounding
the heart are relatively immobile or move slowly compared Training Undersampling
to the beating heart. Furthermore, if one knows informa-
tion about the dynamics of one image point, consider- Low resolution 8 acceleration
able information about the spatiotemporal behavior of its
neighboring points can be inferred. This enables very quick
acquisition of an estimation of the object in a so-called
training stage. In the training stage, images are acquired at
very low spatial but at full temporal resolution. These data
may either be obtained in a separate scan or in an inter-
k-t BLAST
leaved fashion during the actual acquisition (39). During reconstruction
the undersampling stage, very sparse sampling is applied
along space and time resulting in, for example, eightfold
scan time reduction. At the same time, eightfold aliasing is
generated due to undersampling as illustrated in Figure 1.26.
To obtain a high-resolution, unaliased image series the
prior knowledge as derived from the training data is used
to unfold the image as outlined in Figure 1.26. In the par-
lance of image reconstruction, the training data determines
a filter which suppresses folding artifacts as a result of un-
dersampled data. In a more general view, k-t BLAST and
k-t SENSE may be compared to data compression as used Figure 1.26. Outline of the k-t BLAST method. A low-resolution
scan precedes the actual acquisition stage. Based on this training data,
on personal computers to reduce file sizes on a hard disk. an estimate of the expected signal intensities of the moving object is
By identifying redundant information such as the reoccur- obtained. In the actual acquisition stage, undersampling is applied,
rence of certain patterns of numbers, an index for the pat- resulting in multiple fold-over in the images. Using both training
tern is saved instead of all individual numbers. In a similar and acquisition data, unfolded high-resolution images can be recon-
sense, the training data in k-t BLAST/k-t SENSE identify structed. For further explanation see text.
LWBK1209-ch01_p01-20.indd 17 16/05/13 9:45 PM

receive coil. In case multiple coils are used, images from 3-dimensional k-t BLAST at a reduction factor of five has
the individual coils are combined after reconstruction. In successfully been applied to cardiac patients during interven-
contrast with k-t SENSE coil information is incorporated in tional procedures. An exemplary case is given in Figure 1.27.
the reconstruction process similar to conventional SENSE The increased scan efficiency allowed the acquisition of volu-
reconstruction to improve the accuracy of the reconstruc- metric data sets at high spatial (2 × 2 × 5 mm3) and temporal
tion result. (28 to 35 milliseconds) resolution in a single breath-hold
Both SENSE and k-t BLAST/k-t SENSE can facilitate car- lasting 20 to 22 seconds, with the acquisition strategy well-
diac applications by shortening acquisition times or, alterna- tolerated by the patients. It is noteworthy that k-t BLAST/k-t
tively, providing higher spatial and/or temporal resolutions. SENSE not only allow significant acceleration of cine imag-
In practice, SENSE has allowed reduction factors between ing of periodic object motion, but can also be used to ac-
two and three with current cardiac coil arrays while k-t celerate dynamic imaging of nonperiodic processes such as
BLAST/k-t SENSE have already been shown to provide re- contrast agent passage in perfusion studies. The time saved
duction factors of up to eight. This has, for instance, enabled during the CVMR exam by using k-t BLAST/k-t SENSE
to acquire time-resolved 3-dimensional volumetric data sets can be used for addition functional and flow acquisitions,
of the heart at high spatial and temporal resolutions in a or to allow inclusion of coronary MRA in the same CVMR
single breath-hold which has hitherto been impossible. Cine examination.
B
Figure 1.27. Example of cine 3-dimensional balanced SSFP single breath-hold imaging with 5× k-t BLAST
in a patient with repaired tetralogy of Fallot and pulmonary regurgitation causing right ventricular dilation.
The upper fifteen slices (A) are shown from base to apex of the heart (from left-to-right and from top-to-bot-
tom) at end-systole. Selected time frames showing both short-axis and reformatted four-chamber view planes
(B). Images are courtesy of R. Razavi, Guy’s Hospital, London.
LWBK1209-ch01_p01-20.indd 18 16/05/13 9:45 PM

The number of k-lines in the phase encoding (= Y) direc-

RESOLUTION
tion can be calculated by applying the formula:
RESOLUTION
RESOLUTION
FFT Number of k-lines = acquisition matrix × rectangular FOV

CONTRAST
× symmetric reduction × asymmetric reduction,
where FOV = field of view, symmetric reduction factor =
scan%, asymmetric reduction factor = halfscan (Figs. 1.28
RESOLUTION and 1.29). For example, with an acquisition matrix of 256,
a rectangular FOV factor of 0.6, a symmetric reduction fac-
k-space MR image tor of 0.8, and an asymmetric reduction factor of 1 (meaning
Figure 1.28. Schematic diagram of the relation between k-space no halfscan), the number of k-lines equals: 256 × 0.6 × 0.8 ×
and the MR image. The center of k-space contains relatively more 1 = 122.
information on contrast, whereas the periphery contains more infor- The image pixel resolution can be calculated according
mation on resolution. Note that each point in k-space is related to to the formula:
each point in the MR image. The image shown was acquired using
balanced FFE during a breath-hold of 15 seconds. CONTRAST = gray X-resolution (echo readout direction) = FOV/acquisition
values in larger areas, RESOLUTION = visibility of small structures, matrix
FFT = fast Fourier transform.
Y-resolution (phase encoding direction) = FOV/acquisition
matrix × symmetric reduction,
for example, with an FOV of 300 mm, an acquisition
Appendix: Calculations on K-Space matrix of 256, and a symmetric reduction factor of 0.8,
X-resolution will be: 300/256 = 1.2 mm, and the Y-resolution
Image formation in MR is based on filling “k-space” during will be: 300/(256 × 0.8) = 1.5 mm. Reconstructed to a 256
data acquisition. This concept will not be explained further × 256 matrix, meaning the Y-resolution is interpolated
in this context, since many thorough and comprehensive to the X-resolution (in this case 1.2 mm), resulting in an
publications are available on this topic (2–6). One of the apparent in-plane image resolution of 1.2 mm × 1.2 mm,
main aspects of k-space needed for practical understanding and with a slice thickness of 8 mm a final resolution of
of MR is to calculate the number of k-lines to be acquired, 1.2 × 1.2 × 8 mm3, displayed as an FOV of: 300 × 180
and determine the spatial resolution of cardiac images. (= 300 × 0.6) mm2.
Symmetric reduction (scan%): Increased line spacing (rFOV):

lower resolution preserved resolution
less noise Kx more noise
Ky
Full scan
Asymmetric reduction (halfscan): Partial echo:

preserved resolution preserved resolution
more noise more noise
Figure 1.29. MR imaging can be performed faster by acquiring fewer k-lines per MR image.
Symmetric reduction and increased line spacing are the most frequently used techniques to reduce
the number of k-lines. Symmetric reduction means that on both the extreme edges of k-space in
the Y- or phase-encoding direction, the part that determines image resolution (see Fig. 25), k-lines
are not acquired. This results in an image with fewer k-lines, but the trade off is a reduced in-plane
spatial resolution. Increased line spacing decreases the density of k-lines in k-space, whereas still
the entire k-space is traversed. In this way the image resolution remains the same, the image noise
increases, and the FOV is rectangular. Another way of reducing the number of k-lines is to acquire
only, for example, 5/8 part of k-space with the same k-line density, this asymmetric reduction results
in more image noise. Another way to speed up the MR acquisition is not to reduce the number of
k-lines, but to reduce the part of the generated echo that is sampled, this procedure increases image
noise since less image data is acquired.
LWBK1209-ch01_p01-20.indd 19 16/05/13 9:45 PM

References 21. Dromigny-Badin A, Zhu YM, Magnin I, et al. Fusion of cine magnetic resonance and
contrast-enhanced first-pass magnetic resonance data in patients with coronary artery
disease: a feasibility study. Invest Radiol. 1998;33(1):12–21.
1. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: sensitivity encoding for fast
22. Gerber KH, Higgins CB. Quantitation of size of myocardial infarctions by computerized
MRI. Magn Reson Med. 1999;42(5):952–962.
transmission tomography. Comparison with hot-spot and cold-spot radionuclide scans.
2. Boxerman JL, Mosher TJ, McVeigh ER, et al. Advanced MR imaging techniques for evalu-
Invest Radiol. 1983;18(3):238–244.
ation of the heart and great vessels. Radiographics. 1998;18(3):543–564.
23. de Roos A, Doornbos J, van der Wall EE, et al. MR imaging of acute myocardial infarction:
3. Boxt LM. Primer on cardiac magnetic resonance imaging: how to perform the examina-
value of Gd-DTPA. AJR Am J Roentgenol 1988;150(3):531–534.
tion. Top Magn Reson Imaging. 2000;11(6):331–347.
24. Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed contrast enhancement
4. Duerk JL. Principles of MR image formation and reconstruction. Magn Reson Imaging
to irreversible injury, infarct age, and contractile function. Circulation. 1999;100(19):
Clin N Am. 1999;7(4):629–659.
1992–2002.
5. Reeder SB, Faranesh AZ. Ultrafast pulse sequence techniques for cardiac magnetic reso-
25. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance
nance imaging. Top Magn Reson Imaging. 2000;11(6):312–330.
imaging to identify reversible myocardial dysfunction. N Engl J Med. 2000;343(20):
6. Sakuma H, Takeda K, Higgins CB. Fast magnetic resonance imaging of the heart. Eur J
1445–1453.
Radiol. 1999;29(2):101–113.
26. Gerber BL, Rochitte CE, Melin JA, et al. Microvascular obstruction and left ventricu-
7. Chia JM, Fischer SE, Wickline SA, et al. Performance of QRS detection for cardiac mag-
lar remodeling early after acute myocardial infarction. Circulation. 2000;101(23):
netic resonance imaging with a novel vectorcardiographic triggering method. J Magn
2734–2741.
Reson Imaging. 2000;12(5):678–688.
27. van Dijk P. Direct cardiac NMR imaging of heart wall and blood flow velocity. J Comput
8. Bailes DR, Gilderdale DJ, Bydder GM, et al. Respiratory ordered phase encoding (ROPE):
Assist Tomogr. 1984;8(3):429–436.
A method for reducing respiratory motion artefacts in MR imaging. J Comput Assist
28. Lamb HJ, Beyerbacht HP, van der LA, et al. Diastolic dysfunction in hypertensive heart
Tomogr. 1985;9(4):835–838.
disease is associated with altered myocardial metabolism. Circulation. 1999;99(17):
9. Wang Y, Rossman PJ, Grimm RC, et al. Navigator-echo-based real-time respiratory gating
2261–2267.
and triggering for reduction of respiration effects in three-dimensional coronary MR angi-
29. Guilfoyle DN, Gibbs P, Ordidge RJ, et al. Real-time flow measurements using echo-planar
ography. Radiology. 1996;198(1):55–60.
imaging. Magn Reson Med. 1991;18(1):1–8.
10. Danias PG, McConnell MV, Khasgiwala VC, et al. Prospective navigator correction of
30. Klein C, Schalla S, Schnackenburg B, et al. Magnetic resonance flow measurements in
image position for coronary MR angiography. Radiology. 1997;203(3):733–736.
real time: comparison with a standard gradient-echo technique. J Magn Reson Imaging.
11. Stuber M, Botnar RM, Danias PG, et al. Submillimeter three-dimensional coronary MR
2001;14(3):306–310.
angiography with real-time navigator correction: comparison of navigator locations.
31. Huber ME, Kozerke S, Pruessmann KP, et al. Sensitivity-encoded coronary MRA at 3T.
Radiology. 1999;212(2):579–587.
Magn Reson Med. 2004;52(2):221–227.
12. Chapman B, Turner R, Ordidge RJ, et al. Real-time movie imaging from a single cardiac
32. Maintz D, Aepfelbacher FC, Kissinger KV, et al. Coronary MR angiography: compar-
cycle by NMR. Magn Reson Med. 1987;5(3):246–254.
ison of quantitative and qualitative data from four techniques. AJR Am J Roentgenol.
13. Weber OM, Eggers H, Spiegel MA, et al. Real-time interactive magnetic resonance imaging
2004;182(2):515–521.
with multiple coils for the assessment of left ventricular function. J Magn Reson Imaging.
33. Botnar RM, Stuber M, Kissinger KV, et al. Noninvasive coronary vessel wall and plaque
1999;10(5):826–832.
imaging with magnetic resonance imaging. Circulation. 2000;102(21):2582–2587.
14. Yang PC, Kerr AB, Liu AC, et al. New real-time interactive cardiac magnetic reso-
34. van der Geest RJ, Buller VG, Jansen E, et al. Comparison between manual and semi-
nance imaging system complements echocardiography. J Am Coll Cardiol. 1998;32(7):
automated analysis of left ventricular volume parameters from short-axis MR images.
2049–2056.
J Comput Assist Tomogr. 1997;21(5):756–765.
15. Lelieveldt BPF, van der Geest RJ, Lamb HJ, et al. Automated Observer-independent Acquisition
35. van der Geest RJ, Niezen RA, van der Wall EE, et al. Automated measurement of
of Cardiac Short-Axis MR Images: A Pilot Study. Radiology. 2001;221(2):537–542.
volume flow in the ascending aorta using MR velocity maps: evaluation of inter- and
16. Simonetti OP, Finn JP, White RD, et al. “Black blood” T2-weighted inversion-recovery MR
intraobserver variability in healthy volunteers. J Comput Assist Tomogr. 1998;22(6):
imaging of the heart. Radiology. 1996;199(1):49–57.
904–911.
17. Lamb HJ, Doornbos J, van der Velde EA, et al. Echo planar MRI of the heart on a standard
36. Mitchell SC, Lelieveldt BP, van der Geest RJ, et al. Multistage hybrid active appearance
system: validation of measurements of left ventricular function and mass. J Comput Assist
model matching: segmentation of left and right ventricles in cardiac MR images. IEEE
Tomogr. 1996;20(6):942–949.
Trans Med Imaging. 2001;20(5):415–423.
18. Wilke N, Jerosch-Herold M, Stillman AE, et al. Concepts of myocardial perfusion imaging
37. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially
in magnetic resonance imaging. Magn Reson Q. 1994;10(4):249–286.
parallel acquisitions (GRAPPA). Magn Reson Med. 2002;47(6):1202–1210.
19. Cullen JH, Horsfield MA, Reek CR, et al. A myocardial perfusion reserve index in humans
38. Tsao J, Boesiger P, Pruessmann KP. k-t BLAST and k-t SENSE: Dynamic MRI with
using first-pass contrast-enhanced magnetic resonance imaging. J Am Coll Cardiol. 1999;
high frame rate exploiting spatiotemporal correlations. Magn Reson Med. 2003;50(5):
33(5):1386–1394.
1031–1042.
20. Lauerma K, Virtanen KS, Sipila LM, et al. Multislice MRI in assessment of myocardial
39. Kozerke S, Tsao J, Razavi R, et al. Accelerating cardiac cine 3D imaging using k-t BLAST.
perfusion in patients with single-vessel proximal left anterior descending coronary artery
Magn Reson Med. 2004;52(1):19–26.
disease before and after revascularization. Circulation. 1997;96(9):2859–2867.
LWBK1209-ch01_p01-20.indd 20 16/05/13 9:45 PM

Chapter
David Saloner 2
Magnetic Resonance Angiography
Techniques
■■ Flow Dynamics Acquisition Timing: Dynamic Subtraction

■■ Primary Magnetic Resonance Angiography Contrast-Enhanced Magnetic Resonance
Methods Angiography
Imaging Parameters
■■ Flow Compensation
Advantages of Contrast-Enhanced Magnetic
■■ Time-of-Flight Methods Resonance Angiography
Two-Dimensional Time-of-Flight Methods Limitations of Contrast-Enhanced Magnetic
Three-Dimensional Time-of-Flight Methods Resonance Angiography
■■ Presaturation ■■ Postprocessing and Display
■■ Parameter Choices Maximum Intensity Projection Algorithm
Limitations of Maximum Intensity Projection
■■ Repetition Time Reformatting
Flip Angle
■■ Summary
■■ Tof Mra Considerations and Strategies
Flow Independent Magnetic Resonance Magnetic resonance (MR) imaging is well suited as a modal-
Angiography ity for evaluating vascular disease. The principal strengths of
Disordered Blood Flow MR are that it can provide coverage of an extended vascular
Short Echo Sequences territory with information at all points in three-dimensional
Sequential Three-Dimensional Methods (3D) space while remaining noninvasive or, as in the case of
Spiral Scan venous-injected contrast-enhanced imaging, only minimally
Black Blood invasive. The 3D data sets obtained permit display of the ves-
sels in various formats, including contiguous slices that can
Radiofrequency Excitation Pulses
be oriented in any obliquity or using a projection format.
■■ Phase Difference Methods While the major effort in MR angiography (MRA) has been
■■ Additional Noncontrast-Enhanced Mra
directed at obtaining images of the vascular lumen, there is
substantial interest in delineating the disease process itself by
Methods imaging the vessel wall. Obtaining data on the dynamics of
3D Fast Spin-Echo Ecg-Gated Mra Methods blood flow, similar to that obtained in catheter-injected x-ray
Slab Nulling Preparation and Inflow Mra angiography, is also of increasing interest.
■■ Contrast-Enhanced Magnetic Resonance As in other MR methods, MRA is least successful when
Angiography there is substantial gross motion that occurs during the
acquisition of the data. This can arise from poor patient
T1 Shortening
compliance or because of physiologic motion, such as car-
Phase Encoding Considerations diac motion while imaging the coronaries or breathing
Acquisition Timing motion for the visceral vessels. Also, blood flow itself can
Acquisition Timing: Test Bolus degrade the quality of the depiction of the vascular lumen
Acquisition Timing: Signal Initiated when that flow is either extremely low, extremely high, or
Acquisition disordered.
21
LWBK1209-ch02_p21-38.indd 21 17/05/13 4:50 PM

Techniques for obtaining high-quality MR angiograms (4–9). MR imaging methods are capable of measuring
have evolved with improvements in MR instrumentation. As both the magnitude of the transverse magnetization and the
new performance capabilities become available, new tech- orientation of that magnetization in space (the phase).
niques become feasible, and this trend can be expected to Methods have therefore been devised that are designed
continue. In this chapter, a number of different methods for to create large differences in either the magnitude or the
obtaining MR angiograms will be discussed. Methods for phase of the magnetization between spins that are station-
displaying the data will also be presented. ary and spins that are moving (5,7). MR sequences that rely
on blood flow to transport fully magnetized blood into the
imaging volume and thereby create a substantial difference
between the magnetization of flowing and stationary spins
Flow Dynamics
are generally referred to as time-of-flight (TOF) methods,
which display the magnitude of the transverse magnetiza-
The signal intensity that is measured in MRA studies of
tion (4). Sequences that rely on the presence of contrast
flowing blood depends on blood flow velocities (both mag-
agents injected into the bloodstream to enhance vascular sig-
nitude and direction) and the variation of those quantities
nal are referred to as contrast-enhanced MRA (CE-MRA),
(e.g., from cardiac pulsatility) over the time that data are
and they create images that display the magnitude of the
acquired. The design and implementation of imaging strate-
transverse magnetization (10). Images that display the phase
gies and the choice of imaging parameters therefore require
of the magnetization are referred to as phase contrast (PC)
careful consideration of the anticipated hemodynamics in
images (11). These methods rely on the motion of spins,
the vessels, in both the normal and the diseased conditions.
with respect to the imaging gradients for vessel-to-stationary
Setting aside the problem of gross motion for the moment,
tissue contrast.
MRA methods provide accurate images of the lumen when
the image resolution is sufficiently high, to provide several
pixels across the diameter of the lumen, where flow is suf-
ficiently high that blood traveling through the imaging vol-
Flow Compensation
ume receives only a small number of radiofrequency (RF)
Accurate spatial localization, providing images with a high
excitations, and where the flow patterns remain regular and
signal-to-noise ratio (SNR), is accomplished by ensuring
reproducible throughout the cardiac cycle. Deviations from
that in the center of the readout interval, an echo is formed
these flow conditions can occur in both healthy and diseased
where the orientation of the transverse magnetization of all
vessels (1–3).
excited material is the same (rephased). In general, when
Variations in flow conditions can result from large dif-
imaging gradients are applied, spins will accumulate a phase
ferences in systolic and diastolic flows. This can be particu-
shift. Conventional imaging gradients that are designed
larly pronounced; for example, in the arteries of the lower
to generate a signal echo from stationary material do not
extremities where a pronounced interval of rapid antegrade
account for the motion of flowing spins and, at the center
flow is followed by a short, smaller retrograde component,
of the echo, moving spins accumulate a phase that depends
which is, in turn, followed by an extended diastolic inter-
on their motion between the RF excitation and the time the
val where the flow velocities are extremely low. In vivo
echo is collected. In a voxel containing blood spins moving
flow patterns are also profoundly impacted by variations
with a spread of velocities, various phases occur. The mean
in geometry. Even in healthy individuals, blood vessels are
magnetization in the voxel is the sum of the individual vec-
curved with numerous branches and bifurcations. These
tors, and because of the spread of phases, a drop in signal
geometric variations can substantially impact the velocity
strength, referred to as intravoxel phase dispersion, takes
fields resulting in features, such as flattened profiles across
place. For many MR angiographic sequences it is important
the vascular lumen at the entrance to branch vessels or
to implement gradient waveforms such that the magnetiza-
helical flow patterns in curved vessels, such as the aortic
tion of moving blood is also rephased at the center of the
arch. However, in the absence of disease, flow is generally
echo (12,13). These are referred to as motion compensation
laminar and reproducible from cycle to cycle. In regions of
gradients (14). Whereas conventional gradient echo (GRE)
atherosclerosis the vascular wall can be irregular, with a
sequences use gradient waveforms consisting of two lobes of
residual lumen that is highly asymmetric and that has rapid
gradient field strength applied with opposite polarity, flow
changes in diameter. These geometric variations can create
compensation gradients require additional lobes of applied
regions of disturbed flow, such as velocity jets emanating
gradients. In principle, flow compensation could be pursued
from a tight stenosis accompanied by the shedding of flow
to include all orders of motion (15). In practice, however, the
vortices representing turbulent flow that varies from one
added gradient lobes needed for higher order motion com-
cycle to the next.
pensation require additional time to play out, thus length-
ening the echo time. If motion compensation is used, the
most effective method for improving signal retention from
Primary Magnetic Resonance moving spins, even in disturbed flow that contains high-
Angiography Methods order motion terms, is to rephase constant velocity terms
only (16–19).
All MR angiographic techniques aim to create high contrast Most MRA sequences incorporate velocity compensa-
between spins that are moving and those that are stationary tion gradients along the directions of the slice selection and
LWBK1209-ch02_p21-38.indd 22 17/05/13 4:50 PM

Chapter 2 ■ Magnetic Resonance Angiography Techniques 23
f requency encoding gradients. The phase encoding direction

is not compensated; the short duration of the phase encod-
ing gradients minimizes the effects of motion during the ap-
plication of those gradients.
Time-of-Flight Methods
The contrast obtained in an MRA study is closely related
to the strength of the longitudinal magnetization in flow-
ing blood relative to that in stationary tissue (6,20,21). In
a GRE sequence (flip angle < 90 degrees), the strength of
the longitudinal magnetization of spins subjected to a series
of RF pulses decreases with each excitation. The longitudi-
nal magnetization of spins will continue to decrease with
increasing number of RF excitations received, a process
referred to as saturation, until it reaches a steady-state value
that is determined by the flip angle, the repetition time (TR),
and the T1 relaxation time for that tissue. Stationary mate-
rial remains in the imaging volume throughout data acquisi-
tion and, therefore, the magnetization strength of stationary
spins decreases to the steady-state value. In blood vessels,
the longitudinal magnetization of spins at any given loca-
tion depends on how many RF excitations those spins have
received between entering the imaging volume and reach-
ing that location. Fast moving blood may receive only a few
RF pulses, thus retaining substantial magnetization strength.
Slowly moving spins may receive a large number of pulses,
and for those spins, their longitudinal magnetization, such as
Figure 2.1. A 2D GRE sequence slice through the neck. Arteries
stationary spins, may also drop to the steady-state value. In
and veins are shown with high contrast relative to stationary tissue
magnitude images, spins that have received many RF pulses (TR/TE/flip angle = 35/9/30 degrees).
and are strongly saturated will appear dark, whereas spins
that retain substantial magnetization strength will appear
bright.
Shortcomings of Two-dimensional Time-of-flight
Two-Dimensional Time-of-Flight Methods

Sequential 2D TOF techniques provide strong inflow
enhancement when the slices are perpendicular to the ves-
The technique of engendering a strong contrast between sels. When vessels run in the same plane as the slice, or re-
the longitudinal magnetization of flowing and stationary enter the slice, the blood becomes saturated, and contrast is
spins and then using flow-compensated gradients to cor- progressively lost. This effect is more pronounced when the
rectly encode the transverse magnetization has been used to slices are thick. In addition, in examinations like an axial
good effect in thin slice strategies, also called sequential two- carotid artery study, a superior presaturation slab is applied,
dimensional time-of-flight (2D TOF) (22,23). In the thin which is close to and moves along with the current slice.
slice strategy, a single slice perpendicular to the blood ves- This removes venous structures. Vascular loops are some-
sel is acquired using a flow-compensated gradient sequence times found, especially in the cervical vertebral arteries of
(Fig. 2.1). By making the slice very thin, it is ensured that the older patients. If the loop is large enough, inferiorly moving
slice will be replenished with relaxed blood and that it does arterial blood may also be presaturated and yields no signal.
not become saturated, even if moderately large flip angles For this reason, the method is most effective when the vessel
are applied. Each single slice acquisition requires of the runs in a straight course.
order of 8 seconds. The sequence can be repeated multiple Slices collected perpendicular to the vessel result in pro-
times, each time shifting the position of the slice to permit jected images in which the plane of viewing, that is, along
the acquisition of a large set of consecutive slices in a rea- the length of the vessel, corresponds to the axis with poor-
sonable imaging time. Note that unlike standard multislice est resolution. Resolution will improve, if thin slices are
imaging in which the images are collected simultaneously, acquired, but this may not be practical because the total
images are collected sequentially so that blood is not satu- imaging time depends on the number of slices collected. In
rated by one slice before entering another. High signal con- addition, thinner slices have a lower SNR and require longer
trast is attained between blood vessels and the stationary echo times.
surround. This procedure provides a full 3D data set, and In many cases, particularly in the abdomen, reduced reso-
the single slices can then be reformatted by stacking them lution in the projection view can be successfully addressed by
and calculating the desired plane of projection. acquiring slices in the same plane as the desired projection.
LWBK1209-ch02_p21-38.indd 23 17/05/13 4:50 PM

In-plane saturation of the inferior vena cava (IVC) while ac-

quiring coronal slices may be reduced by the use of small flip
angles at the expense of contrast.
Three-Dimensional Time-of-Flight Methods

The major shortcoming of the thin slice strategy is the lim-
ited resolution that can currently be achieved in the slice
selection direction. In 2D techniques using flow compen-
sation gradients, slice thicknesses of 2 mm or greater are
typically used. The use of 3D techniques permits the acqui-
sition of a full 3D data set with isotropic voxels of less than
1 mm3 (4,7,24). High-resolution voxels are critical for the
visualization of small branch vessels, which might, other-
wise, be obscured by partial voluming. They also provide a
clear depiction of tortuous vessels with equal fidelity in each
spatial dimension.
Three-dimensional methods further permit the acquisi-
tion of data sets with echo times that are shorter than those
achieved with 2D methods. In acquiring data from a 3D
volume, the excitation volume in the “slice select” direction
is chosen to be a slab that is of the order of 40 mm thick.
Compared to 2D methods, an additional cycle of phase
encoding is imposed to permit the data to be reduced to a
set of partitions. The use of 3D acquisitions requires sub-
stantially greater acquisition times than do 2D methods but Figure 2.2. MIP from a 1,024 matrix time of flight acquisition at
provides advantages in increased SNR. In 3D acquisition, 3 T with a voxel size of 0.13 mm3 true resolution. The high resolution
blood flowing through the excitation volume receives more MRA is of a patient with a symptomatic right middle cerebral artery
aneurysm at the level of the M3 segment. In addition, perforating
excitation pulses than in the case for 2D imaging. To avoid
arteries and the ophthalmic arteries are clearly visible. (Courtesy of
excessive saturation effects, the flip angle must be reduced Dr. Winfried Willinek, University of Bonn, Germany.)
(<30 degrees). Like other MRI methods, TOF MRA benefits
from the increased SNR that is available at high field. This
permits the acquisition of images with reduced voxel sizes
(Fig. 2.2).
Presaturation
In many cases, evaluating an arterial structure is confusing
when it is obscured by an overlying vein and vice versa.
To eliminate the signal from these vessels, a presaturation
band with a flip angle ≥90 degrees can be placed adjacent
to the imaging slice or volume so that the blood is saturated
before it enters the slice (Fig. 2.3) (25,26). For example, to
A
remove IVC signal from a renal artery origin study, one
could place a transverse presaturation band across the
abdomen inferior to the kidneys. Despite their usefulness,
the application of presaturation pulses needs careful atten-
tion because their existence in a given acquisition is often
not directly noted in the resultant images, if they are placed
outside the imaging volume. Care must be taken to ensure
that the vessel of interest is not similarly presaturated.
Parameter Choices
B
In generating an MR angiogram, parameter choices are
often made that optimize one feature of the angiogram at Figure 2.3. An MIP on the coronal plane of a 3D TOF acquisition
the expense of others. In choosing optimum pulse sequence through the circle of Willis. (A) With no presaturation, the sagittal sinus
parameters, one needs to take into account the expected is clearly depicted (arrowheads) and overlays the arterial signal. (B) A
flow velocity, the thickness of the imaging volume that must superior axial saturation band eliminates signal from venous blood.
LWBK1209-ch02_p21-38.indd 24 17/05/13 4:50 PM

be traversed, and the acceptable level of contrast between from stationary material. However, as TR decreases, blood
flowing and stationary material. will receive more RF excitations in traversing the same dis-
tance than it would with a larger TR. Figure 2.4 illustrates
this, which is a transverse 3D acquisition through the carotid
Repetition Time bifurcation from a normal volunteer. The most suitable TR
will depend on the rate at which blood transits a slice and
The selection of the most suitable repetition time is a com- will, therefore, vary from vessel to vessel and, for a given
promise. The shorter the TR the more saturated the signal vessel, be somewhat patient dependent. One should also
A B
Figure 2.4. TOF acquisitions through the circle of Willis varying the
repetition time, TR. With increasing TR there is less saturation
of the stationary tissue and retention of more distal arterial signal.
TR/TE/flip angle were (A) 25/7/25 degrees, (B) 35/7/25 degrees, and
C
(C) 45/7/25 degrees.
LWBK1209-ch02_p21-38.indd 25 17/05/13 4:50 PM

note that an increased TR will dictate a larger total study more rapidly saturating signal from blood that receives mul-
time and hence, an increased possibility of patient motions, tiple excitations (Fig. 2.5). Again, there is an interaction
which can substantially degrade image quality. between the choice of this variable and the flow dynamics.
For steady flow, the distribution of magnetization strength
across the excited region will reflect the flip angle, and the
Flip Angle
variation will be larger with increasing flip angle.
The choice of flip angle, α, is similarly a compromise with As vessels pass through a volume of excitation, blood in
increasing α more heavily suppressing stationary signal but the more distal portions becomes increasingly saturated. This
A B
Figure 2.5. TOF acquisitions through the circle of Willis varying

the flip angle. With increasing flip angle there is increasing saturation
of the stationary tissue and reduction of more distal arterial signal.
TR/TE/flip angle were (A) 35/7/10 degrees, (B) 35/7/25 degrees, and
C
(C) 35/7/40 degrees.
LWBK1209-ch02_p21-38.indd 26 17/05/13 4:50 PM

tendency can be counteracted by designing an RF excitation reduced net transverse magnetization strength and will
pulse, a TONE pulse, with a flip angle that is relatively small appear in the MR image with reduced signal intensity. The
on the proximal side of the slab but that becomes progres- artifactual loss of signal associated with disturbed flow can
sively larger across the slab. Spins entering the slab retain mimic the appearance of vessel stenosis in vessels of normal
their signal strength for a greater distance. caliber and can lead to the overestimation of the degree and
extent of stenosis in diseased vessels.
TOF MRA Considerations Short Echo Sequences

and Strategies
In general, the effects of disturbed flow that result in intra-
Although significant progress has been made using con- voxel phase dispersion, can be reduced by shortening the
ventional TOF methods, a series of innovative strategies duration of the encoding gradients (17,19,29–32). The extent
have been pursued to further improve MRA. Specialized to which this can be pursued is limited by the rate at which
sequences have been proposed to reduce the well-known loss the magnetic field gradients can be altered (the slew rate) and
of signal noted at sites of disturbed flow. Other approaches the maximum achievable gradient strength. Further, acquir-
have been pursued to enhance the contrast between flowing data with a strong frequency encoding gradient results in
ing spins and stationary spins either by more effectively sup- a reduced SNR. Nevertheless, great progress in MRA qual-
pressing the signal from stationary spins or by retaining flow ity has been achieved with the development of higher perfor-
signal in the more distal portions of the vessels (as in the mance gradients.
TONE approach noted above). An important consideration in selecting the echo time of
TOF MRA sequences is the dependence of the background
signal on the echo time. In voxels where there are significant
Flow Independent Magnetic
contributions from protons attached to fat and protons at-
Resonance Angiography
tached to water, the signal strength depends on the phase
Conventional MR imaging methods exploit differences in relationship between these two components. The phase rela-
relaxation properties to obtain images with high contrast tionship is determined by the slight difference in precessional
between different tissue types. It is similarly possible to frequencies of protons in the two different environments. At
create images with suitable timing values that exploit the 1.5 tesla (T), this translates to the fat and water signal being
specific T1 and T2 properties of flowing blood and chemi- out of phase at odd multiples of 2.3 milliseconds and in
cal shift differences in surrounding tissue to provide good phase at even multiples of 2.3 milliseconds. As such, images
contrast between blood in the vascular lumen and the sur- acquired with an echo time of 7 milliseconds will have more
rounding tissue. This type of approach obviates the need for effective background suppression than will images acquired
contrast injection and is insensitive to many of the condi- at 5 milliseconds (33). Nevertheless, the benefits of short
tions that must be met by MRA methods that are dependent echo time MRA are still apparent in regions where there is
on specific flow properties of intraluminal blood. They do, little fat contribution, such as the brain.
however, require the careful implementation of a number of Reduced echo time sequences can also be achieved by re-
advanced methods, such as specialized water–fat suppres- laxing the requirement that the full echo be placed symmet-
sion methods, and face the challenge of separating arterial rically in the data acquisition window. Conventional MRA
from venous signal, which is more easily accomplished at sequences sacrifice a fraction of the echo and place the echo
higher field strengths. closer to the excitation pulse to reduce the echo time. The
slight loss in resolution that results by dropping the high-
order frequency components is compensated by the reduced
Disordered Blood Flow
sensitivity to flow disturbance (34,35). This strategy can be
Laminar flow becomes unstable at high flow velocities and further pursued by using partial Fourier techniques where
breaks down into turbulent flow. In regions where true highly asymmetric echoes are acquired. In this case, the loss
turbulence exists, such as immediately distal to a critical in resolution is unacceptable, and additional postprocess-
stenosis, the flow patterns are complicated because some ing steps must be taken, using the inherent symmetry of the
regions will have laminar flow, other regions will have tur- Fourier data to reconstruct the missing information.
bulent flow, and intermittent regions exist in which the flow
randomly fluctuates between these two conditions. In this
Sequential Three-Dimensional Methods
case, the different phase encoding steps needed to build up
the MRA image will be collected with different signal dis- Distal saturation of blood flow that occurs in 3D time-off
tribution. These fluctuations will result in inconsistent data light sequences can also be reduced by reducing the thick-
and image degradation (27,28). In addition, disturbed flow ness of the excitation slab and by using multiple 3D volumes
implies that between RF excitation and signal readout, pro- to cover the region of interest. This method has been termed
tons will move through complicated trajectories described multiple overlapping thin slab acquisition or MOTSA (36).
not only by constant velocity terms but high-order motion In this approach, the 3D slabs are placed orthogonal to the
terms that are not rephased by constant velocity compensa- principal direction of flow (e.g., axial slabs for studies of the
tion. At the time of readout there will be voxels that contain extracranial carotids). To avoid bands of signal loss where
protons with a range of motion histories and a correspond- the RF profiles trail off at the slab edges, substantial overlap
ing range of accumulated phases. Such voxels will have a of consecutive slabs is provided. This comes with a penalty
LWBK1209-ch02_p21-38.indd 27 17/05/13 4:50 PM

Fourier encoding, phase encoding occurs at an earlier time

than does frequency encoding. Consequently, the location
of flowing spins in the image is mismapped. One approach
to correcting this problem is to replace the phase encoding
gradient lobe with a bipolar lobe designed to encode spins
moving with constant velocity with the phase appropriate
to their location, when the center of the frequency encod-
ing axis occurs (40). Although helpful, this approach, again,
corrects for only spins moving with constant velocity, and
there remain displacement artifacts for spins moving with
higher-order motion.
A more fundamental approach to correcting displace-
ment artifact is to use pulse sequences, such as spiral scan
where spatial information along both axes is acquired simul-
taneously. Spiral scan methods build up k-space by design-
ing gradient trajectories that start in the center of k-space
and spiral out to the edge (41,42). In the case of flow, spiral
scan methods have the attraction that the two dimensions
of spatial encoding occur simultaneously in time; there-
fore, there is little or no misregistration artifact. They are
attractive also because they effectively have greatly reduced
echo times, particularly for the low spatial frequency infor-
mation in the center of k-space, providing added insensitiv-
ity to disturbed flow.
Black Blood
Conventional bright blood MRA methods are limited in
their ability to clearly define the lumen edge at locations of
disturbed flow. They are also poor at delineating the soft
tissue signal of disease processes in the vessel wall, such as
atheroma or mural thrombus. Black-blood angiography can
reduce the overestimation of stenotic disease (43–47) and
provide excellent information on the vessel wall (Fig. 2.7).
Typically, black-blood methods use spin-echo sequences
where flowing blood appears as a signal void. Spin-echo
sequences can be further modified to more fully obliterate
Figure 2.6. Flow-independent MR angiography image of popliteal signal from moving spins while retaining signal from sta-
trifurcation in a healthy volunteer obtained at 3T. TR/TE/flip angle =
tionary spins. This is accomplished by adding presaturation
8/4/50 degrees; resolution = 0.9 × 0.9 × 1 mm3. (Courtesy of Dr. Jean
Brittain, General Electric Medical Systems.)
pulses to selectively saturate spins before entering the ves-
sel segment. Further, flow compensation is removed, and
the gradient lobes can be designed to accentuate intravoxel
phase dispersion. Increased echo times also increase the like-
lihood that flowing spins will leave the selected slice between
of increased total acquisition time. Patient motion between the 90- and 180-degree pulses, thereby preventing echo for-
acquisition of consecutive slabs will be seen (as in sequen- mation for these spins.
tial 2D studies) as a sharp discontinuity of the lumen edge. Finally, the so-called double inversion methods can be
The reduced slab thickness limits the duration so that spins used to retain signal from stationary tissue but provide a
remain in the excited volume, resulting in better retention of condition where any blood that enters the slice yields no net
signal in the distal vessels (Fig. 2.6). This method provides signal.
the high resolution capabilities of 3D sequences (37). When using a black-blood sequence, caution must be ex-
ercised to ensure that regions of low signal strength are the
patent lumen and not regions of inherently low signal, such
Spiral Scan
as calcification or air that is close to the vessel. The other
An artifact that was recognized in early clinical studies is the potential problem is that slow or recirculating blood may
so-called misregistration artifact (38,39). In this case, signal not appear black.
from spins—that move obliquely with respect to the phase
and frequency encoding axes—appear at a different spatial
Radiofrequency Excitation Pulses
location in the image than the lumen of the vessel. This can be
noted as a bright vessel running diagonally and parallel to a Much of the progress in the development of MRA has come
dark lumen. This artifact occurs because in conventional 2D from improved methods of manipulating the magnetic field
LWBK1209-ch02_p21-38.indd 28 17/05/13 4:50 PM

Figure 2.7. A comparison of a single slab excitation

through the circle of Willis with a two-slab MOTSA
acquisition. The MOTSA acquisition retains signal in the
distal small arteries (arrowheads) and also shows signal in
the lower slab from venous signal (arrow) entering from
above (no presaturation). (A) Single slab TR/TE = 35/7. B
(B) Two-slab MOTSA; TR/TE/flip angle = 35/7/25 degrees.
gradients. The other key ingredient that the pulse program- Phase Difference Methods
mer can use for modifying signal characteristics is tailored
RF pulses. Although the TONE pulses referred to previously The contrast provided by longitudinal amplitude methods
are designed to more effectively retain signal from flowing described previously depends on the difference in longitudi-
blood, alternative pulses have been proposed that can selec- nal magnetization strength that exists between material that
tively reduce the signal from a stationary tissue while leaving has received few prior excitation pulses and stationary mate-
the signal from flowing spins unaffected. rial that has received numerous excitations at the time that
One such RF pulse is a fat saturation pulse that is designed the RF acquisition pulse is applied. A different approach to
to selectively excite protons attached to lipids (48,49). This angiographic contrast is to modulate the phase of the trans-
method uses the property that protons attached to lipids see verse magnetization between the time of RF excitation and
a slightly different chemical environment than do protons the time of signal acquisition (11,53).
attached to free water and therefore, have different preces- In this approach the phase of moving spins is purposely
sional frequencies. A saturating pulse (90 degrees) can there- altered by appropriate gradient waveforms. An additional
fore be delivered at the resonant frequency of fat protons, phase-modulating gradient is added to a sequence that
leaving the magnetization of flowing blood unchanged but is otherwise flow compensated. This gradient produces a
effectively eliminating the fat signal from the MR image pro- phase accumulation for a moving spin that is proportional
viding improved background suppression. to the spin’s velocity. A second image is acquired, with the
A mediating mechanism can also be used that exploits sense of the phase-modulation gradient reversed. In this case,
the different relaxation properties of protons that are at- the magnitude of the transverse magnetization is identical
tached to free water compared to protons that are bound to for both images, but the phase of the transverse magnetiza-
membranes and other macromolecules. Bound protons have tion is proportional to the flow velocity and to the duration
a very short T2 relaxation time and have a broad spectrum and magnitude of the phase-modulation gradients. Provided
of precessional frequencies. They can therefore be excited that the gradients are not too great, an image can be con-
with pulses that are far from the resonance of mobile pro- structed from the raw data of the two images by subtract-
tons and their magnetization strength is thereby saturated. ing their magnetization vectors. Stationary tissue magnitude
This, in turn, reduces the signal strength of mobile protons and phase are identical in both data sets and therefore, no
in the tissue parenchyma, which undergo a rapid exchange signal from the stationary tissue is present in the subtraction
with the bound protons. Moving spins do not participate in image.
this mechanism, and the magnetization of flowing blood is The gradient strength should be selected in proportion to
largely unaffected by these off-resonance excitation pulses the velocity of blood in the vessel under investigation. With
(Fig. 2.8). This approach is termed magnetization transfer strong gradients the method can be tailored to provide high
saturation (50–52). Both magnetization transfer saturation sensitivity to slow blood flow, such as that in veins. The SNR in
pulses and fat saturation pulses impose additional time re- a phase contrast image increases as the phase difference of the
quirements on pulse sequences and lengthen the minimum magnetization of flowing spins increases, which is what is rep-
possible repetition time. resented in the subtracted images. However, if the combination
LWBK1209-ch02_p21-38.indd 29 17/05/13 4:50 PM

A B
Figure 2.8. Comparison of MIP projections of 3D TOF images acquired through the circle of Willis with
and without magnetization transfer saturation. (A) Without magnetization transfer saturation, small vessels
are hard to see because of the low contrast with respect to stationary tissue. (B) With magnetization transfer
saturation. Increased signal saturation of the parenchyma occurs and improved delineation of the smaller
vessels (arrow).
of blood velocity and gradient strength is too high, there will be methods were developed (described below in the CE-MRA
a phase aliasing artifact that results in the areas of vasculature section) that provided large coverage in short acquisition
having signal dropout in regions of high flow velocities. times utilizing contrast agents that significantly shortened
Like TOF methods, phase contrast angiograms can be ob- the T1 relaxation time of intraluminal blood. Those meth-
tained of 2D or 3D volumes. However, in phase contrast, be- ods enjoyed great popularity for over a decade in a period
cause of the excellent suppression of stationary tissue signal, when the risk of nephrogenic systemic fibrosis was unappreci-
2D slabs can be acquired in which the slab thickness can be sev- ated (54). Once the risk of systemic fibrosis (NSF) was recog-
eral centimeters thick, while retaining projection views of the nized, there was a renewed interest in noncontrast-enhanced
vasculature with good contrast properties. Two-dimensional methods leading to the development and utilization of new
cine phase contrast studies are capable of giving informative acquisition strategies. Examples of this are now described,
displays of flow dynamics as they vary over the cardiac cycle. and then followed by a discussion of CE-MRA methods.
Since the signal intensity is proportional to the velocity
of flow, the phase contrast method can be used to determine
3D Fast Spin-Echo ECG-Gated MRA Methods
flow velocities. Flow components can lie along any of the
three orthogonal directions, and repeating the measurements An interesting approach to delineating arterial flow is to
for all three spatial directions is necessary to acquire a total exploit several of the concepts described earlier in this chap-
flow map. The total time can be reduced by acquiring one ter including the differential flow characteristics through the
reference image and three encoded images, although the total cardiac cycle and the signal loss that is inherent to many
acquisition time is still long relative to the TOF method, and spin-echo methods. This approach was originally proposed
it is susceptible to machine imperfections, such as eddy cur- as a 2D projection strategy by Wedeen et al. (55) where
rents. Nevertheless, because of the relative insensitivity of this cardiac-gated data were acquired in systole and diastole.
method to flow saturation effects and the strong suppression Luminal signal from arterial flow showed a flow void from
of signal from stationary tissue, it remains an attractive tech- dephasing during systole, whereas diastolic images acquired
nique for slow flow situations. in the interval of slow flow showed a bright intraluminal
signal. Venous signal was the same in both cardiac phases.
Subtraction of the systolic data from the diastolic data elim-
Additional Noncontrast- inated the signal from stationary tissue and venous signal
Enhanced MRA Methods and provided an MRA of the arterial lumen. More recently,
Miyazaki and Akahane (56) have extended this methodol-
Historically, the first MRA methods that found clinical util- ogy to acquire 3D fast spin-echo (FSE) data sets. Since the
ity were the methods described above that did not require FSE approach permits the sampling of multiple echoes for
the injection of any contrast agents. In the early 90s, MRA each magnetization sampling pulse, acquisition times for 3D
LWBK1209-ch02_p21-38.indd 30 17/05/13 4:50 PM

data sets remain realistic in the clinical setting. Care must tion time (typically 3 to 5 minutes), in regions where there
be taken to keep the echo train length sufficiently short is gross motion secondary to breathing, these acquisitions
so that T2-blurring artifacts do not become pronounced. are acquired using a navigator monitoring approach where
Furthermore, the method is sensitive to the underlying flow data are rejected if the position of the diaphragm is outside
conditions and imperfect selection of high-flow and low- a set window.
flow intervals can result in poor-quality MRA studies, as can
pathologic flow states distal to stenoses.
Contrast-Enhanced Magnetic
Slab Nulling Preparation and Inflow MRA
Resonance Angiography
We have noted that TOF MRA methods can attain desir-
able contrast either relying solely on the differential Signal contrast in MRA is determined predominantly by
response of magnetization in flowing blood relative to sta- blood flow patterns as opposed to conventional MRI, where
tionary tissue that is inherent in the timing and strength Tl and T2 values are critically important. Contrast agents
of the RF pulses used as part of image acquisition, or they have been used to advantage in MRA studies exploiting
can make use of additional preparation pulses (such as the the short Tl properties that these agents produce (58,59)
fat suppression pulses of magnetization saturation pulses). (Fig. 2.10). The use of contrast agents requires much of the
In the case of the fat saturation or magnetization satura- same analysis that is used in other angiographic methods,
tion pulses, that preparation is performed with each RF such as spiral computed tomography (CT) angiography,
pulse cycle. One additional approach to providing strong where the initiation and duration of the injection of the con-
contrast between arterial blood and stationary tissue is to trast bolus relative to the interval of data acquisition must
generate a profound difference in magnetization between be carefully timed.
inflowing blood and stationary tissue in a volume of inter-
est, and then rapidly sample a large fraction of k-space data
before magnetization returns to equilibrium. In its simplest
implementation, an RF inversion pulse is delivered to a vol-
ume of interest orthogonal to the artery providing inflow-
ing blood with a delay time before image data acquisition
(57). With appropriate selection of this inversion delay
time (of the order of 1 second or longer), image acquisi-
tion can then be initiated at a time point when the inverted
magnetization of stationary tissue has still not returned to
steady state and therefore appears with low signal inten-
sity. Image acquisition can be performed either with an FSE
approach or with balanced steady-state sequences which
also work well in this application. In the long inversion
delay time, blood enters the volume with full magnetiza-
tion strength providing strong contrast to the stationary
tissue (Fig. 2.9). The extent to which the distal vasculature
will be well visualized depends on the rate of blood inflow
for the specific subject, and this can vary substantially with
age and disease condition. Because of the extended acquisi-
A B
Figure 2.10. Comparison of contrast-enhanced MRA and MOTSA

TOF through the carotid bifurcation of a patient with a dilated inter-
nal carotid artery proximal to a tight stenosis. The CE-MRA image
has crisper edges because of the reduced total acquisition time and the
Figure 2.9. Coronal MIP projection of 3D balanced steady-state reduced signal saturation in blood recirculating within the dilatation.
acquisition through the renal arteries. A transverse inversion slab (A) MIP from a longitudinal slab acquired in 25 seconds with a pixel
is followed after a 1.2-second inversion delay by the 3D imaging size of 0.5 × 1 × 1.5; TR/TE/flip angle = 6/2.2/25 degrees. (B) MIP
sequence acquisition. Data acquisition is acquired with free breathing from a two-slab MOTSA acquired in 4 minutes, 30 seconds with a
and navigator-based data acquisition in a 5-mm acceptance window. pixel size of 0.7 × 0.7 × 1 mm; TR/TE/flip angle = 35/7/25 degrees.
LWBK1209-ch02_p21-38.indd 31 17/05/13 4:50 PM

T1 Shortening interest are in the center of the anatomy, this edge artifact
can be tolerated.
At the dosages applied in MRA studies, contrast agents act
The overall contrast in an MR image is largely dictated by
to reduce the T1 relaxation time of intraluminal blood. The
the signal measured at the center of k-space, whereas much
T1 relaxation time of blood is 1.2 seconds at field strengths
of the high resolution detail is determined by the periphery
of 1.5 T, the field strength of a high-field magnet. When gad-
of k-space. Because of the time course of the intravascular
olinium contrast agents are used at sufficiently high concen-
magnetization strength following contrast injection, the use
trations, the T1 relaxation time of blood is reduced to less
of linear phase encoding can result either in large portions of
than 150 milliseconds, well below the T1 of all tissue mate-
k-space being measured before the magnetization enhance-
rial. This means that contrast-enhanced blood will rapidly
ment occurs or in the center of k-space being sampled well
recover magnetization and will have high signal strength,
after the peak enhancement is over.
even for short values of the repetition time.
A wide range of alternative methods can be used for
The acquisition of MR angiograms when using contrast
sampling the k-space data points needed to create an image.
agents requires a different approach than standard MRA.
Particularly relevant to CE-MRA methods are techniques
With a bolus injection of the contrast agent, a short inter-
that permit the center of k-space to be measured when there
val exists, during which the agent will be in the arterial
is peak magnetization strength. One such method is centric
phase (60,61). Timing the MR data acquisition is important
phase encoding. For 2D this consists of first collecting the
so that it coincides with the period during which there is
echo with zero y-phase encoding strength, and then moving
peak arterial signal. After reaching a peak, the arterial signal
out from the center of k-space to the periphery, collecting
strength drops, and the venous signal starts to increase. In
echoes with alternately positive and negative values of phase
conventional MRA studies of arteries, presaturation pulses
encoding strength. This can be extended to 3D, where both
are applied superior or inferior to the volume to eliminate
z-phase encoding and y-phase encoding values can be varied
signal from the veins. In CE-MRA, the application of pre-
so that the center of k-space is collected at the beginning of
saturation pulses is not a viable strategy for eliminating ve-
the data acquisition, and the edges of k-space are collected
nous signal. Most contrast-enhanced studies rely on using
later. This has been referred to as elliptical centric phase en-
parameters providing the shortest possible data acquisition
coding view order (61).
time, and the addition of a presaturation pulse substantially
CE-MRA methods are particularly well suited for parallel
increases that time. Presaturation has limited use because
imaging methods, such as SENSE or SMASH. In these meth-
the reduced T1 values of the blood rapidly restore saturated
ods, coil arrays with coil elements of differing spatial sensi-
magnetization strength.
tivity are used to acquire data from small fields of view (with
In certain applications, to be able to acquire a 3D study
concomitant savings in acquisition time). The wrap-around
in a short time is important. This includes the extracranial
artifact that would normally result is resolved by the redun-
carotids, where there is a short interval when the first pass
dant information from the different coil elements. Provided
of the bolus provides maximal intra-arterial signal and when
that the SNR is sufficiently high, these methods can be
the venous enhancement, which occurs shortly after the ar-
used to either reduce acquisition times or increase spatial
terial phase because of the blood–brain barrier, has not yet
resolution (Fig. 2.11).
occurred. Similarly, short acquisition times are desirable for
the vessels of the abdomen, so that studies can be obtained
within a breathhold. The use of current high-performance Acquisition Timing
gradient systems permits the echo time, TE to be reduced to
As in other angiographic techniques that use a contrast
less than 2 milliseconds, and the repetition time, TR, to be
agent, timing of image acquisition relative to the passage of
5 milliseconds. This short TR value permits acquisition of
the contrast agent is important for CE-MRA. In some appli-
3D studies from 10 to 20 seconds. Conventional TOF MRA
cations, multiple injections of contrast material can be used.
methods rely on spins with full magnetization strength flow-
However, to reduce effects from venous enhancement, and
ing into the imaging volume between one excitation pulse
to fully exploit the high magnetization strength that prevails
and the next. With short repetition times, very little inflow
immediately following injection of the contrast agent, timing
enhancement takes place. The signal strength in contrast-
of data acquisition remains important. Appropriate timing
enhanced studies, therefore, depends heavily on the T1 re-
of data acquisition will depend on the specifics of the acqui-
laxation time of the material in the volume.
sition strategy (as discussed later) but can be achieved with
several different approaches.
Phase Encoding Considerations
Acquisition Timing: Test Bolus
To achieve a rapid acquisition time, choosing a small num-
ber of phase encoding lines is desirable. However, one is lim- A straightforward approach to sequence timing is the use of
ited by the need to achieve adequate resolution—the pixel a test bolus (60,62). A small amount of contrast agent (e.g.,
size varying as (field of view)/(number of phase encoding 2 mL of gadolinium) is injected and immediately followed
lines). The field of view (FOV) along the phase encoding axis with a saline flush to ensure that the contrast agent does not
must also be chosen large enough to accommodate the anat- remain in the injection tubing. A rapid imaging sequence is
omy along that axis. Failure to choose a large enough FOV applied covering the anatomy, and images are collected at
will result in misinterpretation of signal and the appearance 1-second intervals for about 50 seconds. If the image acqui-
of “wrap-around” artifact. In cases where the vessels of sition is correctly prescribed; for example, a thick slab in
LWBK1209-ch02_p21-38.indd 32 17/05/13 4:50 PM

Figure 2.11. MIP from a contrast-enhanced MRA study of the aorta and renal arteries obtained in
23 seconds with a nearly isotropic spatial resolution of 0.9 × 0.8 × 0.9 mm. Parallel imaging methods are used
with an acceleration factor of 2. Also shown are the location of two planes that are drawn transverse to the
vessel at the level of the stenosis in the right renal artery and more distal, and the luminal cross sections at
these levels. (Courtesy of Dr. Stefan Schonberg, Ludwig-Maximilians-University, Munich.)
the plane of the vessel or using saturation slabs, no tempo- and with immediate reconstruction and display of images,
rally varying in flow enhancement effects will occur, and the a method referred to as magnetic resonance fluoroscopy,
arrival of contrast agent at the target area will be apparent. or using a line scan study where signal can be sampled as
This provides a determination of the time interval between rapidly as every 20 milliseconds. After contrast injection, the
injection and peak magnetization enhancement in the region. sampling study is terminated as soon as signal enhancement
The full study is then performed using the knowledge of exceeds a preset threshold, and the CE-MRA study is begun.
the time delay between injection and peak magnetization This technique accounts for variations in transit time that
strength. The volume of gadolinium appropriate for the tar- might occur when a test bolus is used as compared to when
get vessels in question is injected, again followed by a saline a full dose is injected. Since the contrast agent has already
flush, and after a delay time, the CE-MRA sequence is initi- begun its rapid enhancement phase when the leading edge is
ated. Generally, the delay time in the CE-MRA sequence is detected, the use of an imaging sequence that is constructed
chosen such that the center of k-space is collected when the to acquire the central portion of k-space early on in the data
magnetization strength reaches a peak. Since the center of k- acquisition period is important. For this purpose the centric
space determines the overall contrast properties of the image, or elliptic phase encoding strategies can be used.
this strategy ensures maximum contrast between blood and
stationary material. With this approach, the high spatial fre-
Acquisition Timing: Dynamic Subtraction
quency data are collected when there is less magnetization
Contrast-Enhanced Magnetic Resonance
strength, although this compromise, which results in edge
Angiography
blurring, is not readily apparent in CE-MRA images.
The ideal solution to the problem of determining injection
timing would simply be to collect 3D image sets continuously,
Acquisition Timing: Signal
beginning prior to injection and ending well after the agent
Initiated Acquisition
has passed into the venous phase. If that were possible, it
An alternative approach to the test bolus technique is to use would no longer be necessary to synchronize data collection
an automated method (61,63,64). A pulse sequence is initi- with the injection. Magnetization enhancement following a
ated that samples the magnetization strength in the vessel or bolus injection occurs relatively rapidly (in 5 to 10 seconds);
in a parent vessel. The sampling sequence is chosen to be a and to capture the phase of strong magnetization enhance-
low resolution 2D study with rapid image acquisition time ment and to differentiate the arterial phase from the venous
LWBK1209-ch02_p21-38.indd 33 17/05/13 4:50 PM

phase, collection of the k-space data for the full 3D volume This increases the likelihood that the important regions of
in an interval of 5 seconds is desirable. Currently, 3D data k-space that determine the overall contrast to noise ratio are
sets require acquisition times that are generally 10 seconds. captured in the phase when there is high intravascular mag-
This limitation can be addressed by constructing image netization strength.
data sets, neighboring in time and sharing k-space data An alternative approach for acquiring time resolved
(65,66). In this method, a desired temporal resolution is CE-MR angiograms is simply to use an ultrashort repetition
specified. 3D data sets can then be created within any de- time (<2 milliseconds), to reduce the through-plane resolu-
sired temporal window, using the k-space data that are actu- tion to provide quasiprojection type angiograms, and to use
ally acquired in that window and k-space data that are cre- parallel imaging methods to further reduce acquisition times
ated by interpolation from the measurements made before (67). This combination can provide subsecond acquisitions
and after the desired time window that are closest in time to of 3D data sets (albeit with resolution that is so poor that
the specified interval. Many specific implementations of this data can be viewed in the plane of acquisition only). This
strategy are possible, and a typical approach is to sample the same strategy can then be extended to include view shar-
blocks of k-space that cover the middle of k-space more fre- ing as described previously, to provide images with high
quently than the blocks that cover the periphery of k-space. dynamic information content (Fig. 2.12).
Figure 2.12. A series of MIP images acquired of the dynamic passage of contrast shown in the sagittal plane
from a patient with an arteriovenous malformation. Images were acquired with a T1-weighted, spoiled, 3D gra-
dient-echo pulse sequence on a 3T system with an 8-channel head coil. Parallel imaging was implemented, and
scan time was further reduced using partial Fourier acquisition on all three axes. Images were obtained at a frame
rate of 3 seconds, with a voxel size of 1.04 × 0.78 × 2 mm; TR/TE/flip angle = 3.2 milliseconds/1.6 milliseconds/
15 degrees. Successive frames are shown beginning at 10.5 seconds after injection. (Courtesy of Dr. Timothy Carroll,
Northwestern University, Illinois.)
LWBK1209-ch02_p21-38.indd 34 17/05/13 4:50 PM

Imaging Parameters
The MR technologist can select several imaging parameters
that affect the quality of the CE-MR angiogram. These
include parameters, such as the flip angle, and timing param-
eters, such as TR and TE. Other parameters, such as the
phase encode ordering schemes, are determined by the pulse
sequence programmer. In conventional MRA, evaluation of
the specific impact of parameter variations on image appear-
ance is straightforward. To accomplish this in CE-MRA is
difficult because the initial contrast bolus affects the image
appearance of all subsequent injections. Parameter choices
have been made by theoretic considerations, from phantom
studies, and by limited human studies.
The choice of repetition time is dictated by the desirabil-
ity of having short acquisition times that permit breathhold
studies and capture of the phase when contrast material
reaches peak concentration in the arteries. The minimum
value of TR that can be prescribed depends on the structure
of the pulse sequence. This again is related to the need to
apply gradients of magnetic field to encode the spatial lo-
cation of the magnetization and, particularly, to form and
measure the “echo.” Increasing the strength of the applied
gradient used while measuring the echo decreases the time
needed to measure the signal. This, however, results in a de-
crease in SNR. The choice of the minimum TR is therefore a
compromise between SNR and acquisition time. Figure 2.13. MIP from a contrast-enhanced MRA study of the
In conventional TOF MRA, the choice of flip angle is a aorta and renal arteries showing excellent signal retention through the
sensitive determinant of the overall quality of the image. In aorta into the distal renal arteries and into the iliac arteries. TR/TE/flip
particular, because the T1 of blood is longer than that of most angle = 4.6/1.8/25 degrees; total acquisition time = 21 seconds.
tissue, choosing a flip angle greater than 30 degrees in a 3D
study typically results in saturation of signal in the distal terri-
This results in long acquisition times, data inefficiencies
tories of the vessels. In CE-MRA, the T1 of contrast-enhanced
because of overlap requirements, and the increased possibil-
blood is significantly shorter than the surrounding tissue, and
ity of patient motion. Provided that contrast material fills
these studies are more forgiving, with respect to the specific
the vessels, CE-MRA can be used to cover a very large vol-
flip angle chosen. Flip angles between 30 and 60 degrees are
ume with excellent contrast to noise properties (Fig. 2.13).
found to be satisfactory. The optimum flip angle depends
The third major benefit of CE-MRA is that because of the
somewhat on the repetition time chosen, and some saturation
signal strength, these sequences can be applied using a high
will occur for larger flip angles with very short TRs.
bandwidth to give very short echo times, while still retaining
an adequate SNR. All MRA sequences benefit from reduced
echo times because they restrict the extent of signal loss that
Advantages of Contrast-Enhanced
is associated with disordered flow.
Three main advantages to the use of contrast agents in MRA
Limitations of Contrast-Enhanced
are described here. First, the total study time required to col-
lect the data for a 3D study is quite short, between 10 and
20 seconds. This means that gross patient motion can be Several disadvantages to the use of CE-MRA are described
substantially reduced. Studies of the visceral arteries can be in the following. A principal disadvantage is the need to
performed in a single breathhold (68–70). With TOF meth- inject a gadolinium-chelated contrast agent. When CE-MRA
ods, studies of the extracranial carotid arteries take up to 10 methods were first introduced the risks of nephrogenic sys-
minutes to acquire, and patient motion, such as swallow- temic fibrosis in patients with impaired kidney function was
ing, snoring, and neck movement, can substantially degrade not well recognized (54). As in any other MR study requir-
image quality. Short duration CE-MRA avoids these probing Gd injection, appropriate precautions should be taken to
lems (Fig. 2.10). It also makes possible the study of rapid ensure that patients have adequate kidney function before
dynamic processes, such as flow through an arteriovenous performing a CE-MRA study. In addition, the administra-
malformation (71,72). The second major advantage is the tion of an injection requires the presence of additional per-
increased coverage that is available with CE-MRA. Since sonnel, which, together with the cost of the contrast agent,
TOF methods rely on inflow enhancement, signal strength adds to the cost of the study. As noted previously, a major
in distal vessels is diminished, and the only way to ensure concern is the presence of venous signal that increases with
uniformly high vascular signal through a large volume is to increasing time following injection. This has proved to be
use multiple overlapping subvolumes to cover the region. a major obstacle in studies of the intracranial circulation,
LWBK1209-ch02_p21-38.indd 35 17/05/13 4:50 PM

particularly for the Circle of Willis, where the venous signal stationary material typically has signal strength, albeit sub-
in the cavernous sinus obscures a delineation of the arterial stantially smaller than that of the flow signal. If one were
lumen and in locations, such as the lower extremities, where to perform a summation projection of an MR data set, the
the veins and arteries abut each other. An additional limita- flow signal would be buried in the signal from the stationary
tion of CE-MRA is the time constraints imposed by the need background.
to capture the high intensity signal in the short interval that The MIP algorithm is attractive because it is not compu-
is in the arterial phase. Even with the short repetition times tationally intensive. The algorithm is easy to prescribe and
used, 3D studies can be acquired only by compromising in helpful in providing access to the projection views. The MIP
terms of coverage and/or resolution. provides a road map to the overall vessel anatomy and can
The very strong suppression of signal from stationary ma- indicate sites of probable disease that must be more carefully
terial is advantageous for the visualization of vascular con- evaluated on the source images.
tours. Conventional MRA sequences also strongly suppress
stationary material signal but still retain considerably more
Limitations of Maximum Intensity
of that signal than do CE-MRA sequences. Stationary mate-
Projection
rial signal can add valuable information to a study of vas-
cular pathology. At regions of stenosis, conventional MRA Although the MIP algorithm is quick and easy to use, it is
sequences often show features of the atheromatous plaque known to generate significant artifacts (74,75). A common
that cannot be seen in CE-MRA. The extent of atheroma problem is that low intravascular signal can appear in the
can be assessed, and the presence of features, such as high base images with lower signal strength than stationary mate-
signal hematomas, is easily noted on TOF studies. rial (such as fat, hemorrhage, or when contrast agents have
been administered). In this case, the stationary signal will
be mapped to the projection image, and information will be
Postprocessing and Display lost. This can result in images with the appearance of clini-
cally significant disease, such as stenosis or occlusion. One
MRA studies cover a 3D volume either by acquiring mul- should keep in mind that the MIP is only a convenience and
tiple 2D slices or by an explicit 3D study of the volume that the base images contain additional information that can
of interest that produces a series of 2D partition images. be pivotal in correctly interpreting a study.
Rarely, one may find that a substantial portion of the ves- Since artifacts of the kind referred to previously are a
sel lies within a single partition. More commonly, each slice function of the volume of data that are being processed (the
will contain a small segment of the vessel. Overall interpre- probability of including high signal stationary material in-
tation of the relationship of one vessel segment to another creases with increasing volume), an important consideration
is difficult. The analysis of complex vascular anatomy, in is to limit the postprocessing to as small a volume as possible
particular, the tortuous 3D path followed by most vessels, is to fully include the vessel but still exclude unwanted station-
facilitated by postprocessing algorithms that provide a gen- ary signal. The probability of including a voxel from the sta-
eral overview of the anatomy. Several algorithms have been tionary material of higher signal strength than a voxel from
investigated that provide projection capability, and the one the flowing material increases as the thickness of stationary
most commonly used is the maximum intensity projection material included in the data set increases. The contrast in
(MIP) technique (73). the MIP can therefore be improved by restricting the MIP
to a volume that includes only the immediate vicinity of the
vessels. This can be performed iteratively, using MIP images
Maximum Intensity Projection Algorithm
on three orthogonal planes as a guide in defining the volume
The MIP algorithm exploits the property of MRA sequences to be included for postprocessing. Although fairly user in-
that signal from moving spins is high, whereas that from tensive, the algorithms are sufficiently fast that the manipu-
stationary spins is low. The MIP is a postprocessing algo- lation of these data sets can be performed in a few minutes,
rithm—it can be implemented at any time after the data are using high performance array processors. In addition, this
collected. Processing of the data can be performed without technique can be used to eliminate overlapping vessels.
additional acquisitions or increased scan time.
The algorithm operates by first specifying the desired
Reformatting
viewing plane. Once the viewing plane is specified, an imagi-
nary ray is projected perpendicular to a given pixel in that Viewing vascular contours in planes different from that of
viewing plane through the 3D data set. The algorithm then the native acquisition is often useful. A change in luminal
searches through all voxels in the acquired 3D data set that cross section; for example, is often more readily identified
lie along the specified ray and determines the signal strength from longitudinal views of the vessel than from slices trans-
of the voxel with the highest signal intensity. That inten- verse to the vessel (76). Since no overlapping background
sity value is then assigned to the image pixel in the viewing signal occurs in this case, which might obscure the vessel,
plane that is being generated. This process is repeated for this technique provides the strongest achievable contrast.
every pixel in the viewing plane, thereby creating a projec- Both 2D and 3D data sets covering a volume of interest can
tion image. This approach should be compared with the be reformatted to generate views with arbitrary orienta-
projection image created in x-ray projection. In that case, tion to the original data set. The presentation is particularly
the projection image represents a summation of all attenu- appealing for 3D data sets that have been acquired with
ation effects along the ray traversed by the x-ray. In MRA, isotropic voxels; that is, voxels that have the same length
LWBK1209-ch02_p21-38.indd 36 17/05/13 4:50 PM

along all three axes. In this case, reformatted images can 20. Gao JH, Holland SK, Gore JC. Nuclear magnetic resonance signal from flowing nuclei in
rapid imaging using gradient echoes. Med Phys. 1988;15(6):809–814.
be created with comparable resolution in any obliquity. The 21. Saloner D. Determinants of image appearance in contrast-enhanced magnetic resonance
disadvantage of this approach is its labor intensive quality. angiography. A review. Invest Radiol. 1998;33(9):488–495.
22. Gullberg GT, Wehrli FW, Shimakawa A, et al. MR vascular imaging with a fast gradi-
The appropriate planes must be interactively specified by the ent refocusing pulse sequence and reformatted images from transaxial sections. Radiology.
user and cannot be automated. Further, the resultant image 1987;165(1):241–246.
can provide a depiction of the data through one plane only. 23. Keller PJ, Drayer BP, Fram EK, et al. MR angiography with two-dimensional acquisition
and three-dimensional display. Work in progress. Radiology. 1989;173(2):527–532.
Creating an MIP image of a small number of reformatted 24. Ruggieri PM, Laub GA, Masaryk TJ, et al. Intracranial circulation: Pulse-sequence
images is sometimes advantageous, thereby retaining good considerations in three-dimensional (volume) MR angiography. Radiology. 1989;171(3):
contrast, an overview of the vessels, and a desirable viewing 785–791.
25. Edelman RR, Atkinson DJ, Silver MS, et al. FRODO pulse sequences: A new means of
plane. eliminating motion, flow, and wraparound artifacts. Radiology. 1988;166(1 pt 1):231–
Further processing can be invoked, which makes use of 236.
26. Felmlee JP, Ehman RL. Spatial presaturation: A method for suppressing flow artifacts and
depth information. Volume rendering techniques can be used improving depiction of vascular anatomy in MR imaging. Radiology 1987;164(2):559–
to provide spatial relationships between vessels and soft tis- 564.
sue structures (77,78). Surface rendering can also be used 27. Gao JH, Holland SK, Gore JC. Effects of gradient timing and spatial resolution on the
NMR signal from flowing blood. Phys Med Biol. 1992;37(7):1581–1588.
with virtual light sources to provide images with intensity 28. Oshinski JN, Ku DN, Pettigrew RI. Turbulent fluctuation velocity: The most significant
proportional to the surface gradients, providing information determinant of signal loss in stenotic vessels. Magn Reson Med. 1995;33(2):193–199.
on the texture of vessels. 29. Duerk JL, Simonetti OP, Hurst GC, et al. Experimental confirmation of phase encoding of
instantaneous derivatives of position. Magn Reson Med. 1994;32(1):77–87.
30. Ehman RL, Felmlee JP. Flow artifact reduction in MRI: A review of the roles of gradient
moment nulling and spatial presaturation. Magn Reson Med. 1990;14(2):293–307.
Summary 31. Nishimura DG, Macovski A, Jackson JI, et al. Magnetic resonance angiography by
selective inversion recovery using a compact gradient echo sequence. Magn Reson Med.
1988;8(1):96–103.
MR methods provide a wide range of approaches that can 32. Schmalbrock P, Yuan C, Chakeres DW, et al. Volume MR angiography: Methods to achieve
very short echo times. Radiology. 1990;175(3):861–865.
be used to obtain images of blood vessels, with high contrast 33. Wehrli FW, Perkins TG, Shimakawa A, et al. Chemical shift-induced amplitude modu-
and good resolution. The ability to determine the course of lations in images obtained with gradient refocusing. Magn Reson Imaging. 1987;5(2):
the vessel through 3D space and the relation of the vascu- 157–158.
34. Gatenby JC, McCauley TR, Gore JC. Mechanisms of signal loss in magnetic resonance
lature to the surrounding soft tissue is valuable. Although imaging of stenoses. Med Phys. 1993;20(4):1049–1057.
a number of exciting new capabilities are in development, 35. Evans AJ, Richardson DB, Tien R, et al. Poststenotic signal loss in MR angiography:
Effects of echo time, flow compensation, and fractional echo. AJNR Am J Neuroradiol.
and the true usefulness of others is still under evaluation, a 1993;14(3):721–729.
core of reliable and dependable methods are available for 36. Parker DL, Yuan C, Blatter DD. MR angiography by multiple thin slab 3D acquisition.
clinical use. Magn Reson Med. 1991;17(2):434–451.
37. Davis WL, Blatter DD, Harnsberger HR, et al. Intracranial MR angiography: Comparison
of single-volume three-dimensional time-of-flight and multiple overlapping thin slab acqui-
sition techniques. AJR Am J Roentgenol. 1994;163(4):915–920.
38. Nishimura DG, Jackson JI, Pauly JM. On the nature and reduction of the displacement
References artifact in flow images. Magn Reson Med. 1991;22(2):481–492.
39. von Schulthess GK, Higgins CB. Blood flow imaging with MR: Spin-phase phenomena.
1. Milnor WR. Hemodynamics. Baltimore, MD: Williams & Wilkins; 1989:xii, 419. Radiology. 1985;157(3):687–695.
2. Strandness DE, Sumner DS. Hemodynamics for Surgeons. NewYork, NY: Grune & 40. Frank LR, Buxton RB, Kerber CW. Pulsatile flow artifacts in 3D magnetic resonance imag-
Stratton; 1975:xi, 698. ing. Magn Reson Med. 1993;30(3):296–304.
3. Stroud JS, Berger SA, Saloner D. Influence of stenosis morphology on flow through severely 41. Meyer CH, Hu BS, Nishimura DG, et al. Fast spiral coronary artery imaging. Magn Reson
stenotic vessels: Implications for plaque rupture. J Biomech. 2000;33(4):443–455. Med. 1992;28(2):202–213.
4. Dumoulin CL, Cline HE, Souza SP, et al. Three-dimensional time-of-flight magnetic reso- 42. Nishimura DG, Irarrazabal P, Meyer CH. A velocity k-space analysis of flow effects in
nance angiography using spin saturation. Magn Reson Med. 1989;11(1):35–46. echo-planar and spiral imaging. Magn Reson Med. 1995;33(4):549–556.
5. Dumoulin CL, Souza SP, Walker MF, et al. Three-dimensional phase contrast angiography. 43. Finn JP, Edelman RR. Black-blood and segmented k-space magnetic resonance angiogra-
Magn Reson Med. 1989;9(1):139–149. phy. Magn Reson Imaging Clin N Am. 1993;1(2):349–357.
6. Haacke EM, Masaryk TJ. The salient features of MR angiography. Radiology. 1989; 44. Chien D, Goldmann A, Edelman RR. High-speed black blood imaging of vessel stenosis in
173(3):611–612. the presence of pulsatile flow. J Magn Reson Imaging. 1992;2(4):437–441.
7. Laub GA, Kaiser WA. MR angiography with gradient motion refocusing. J Comput Assist 45. Edelman RR, Chien D, Kim D. Fast selective black blood MR imaging. Radiology. 1991;
Tomogr. 1988;12(3):377–382. 181(3):655–660.
8. Lenz GW, Haacke EM, Masaryk TJ, et al. In-plane vascular imaging: Pulse sequence design 46. Jara H, Barish MA. Black-blood MR angiography. Techniques, and clinical applications.
and strategy. Radiology. 1988;166(3):875–882. Magn Reson Imaging Clin N Am. 1999;7(2):303–317.
9. Masaryk TJ, Tkach J, Glicklich M. Flow, radiofrequency pulse sequences, and gradient 47. Melhem ER, Jara H, Yucel EK. Black blood MR angiography using multislab three-
magnetic fields: Basic interactions and adaptations to angiographic imaging. Top Magn dimensional TI-weighted turbo spin-echo technique: Imaging of intracranial circulation.
Reson Imaging. 1991;3(3):1–11. AJR Am J Roentgenol. 1997;169(5):1418–1420.
10. Prince MR. Contrast-enhanced MR angiography: Theory and optimization. Magn Reson 48. Li D, Haacke EM, Mugler JP 3rd, et al. Three-dimensional time-of-flight MR angiog-
Imaging Clin N Am. 1998;6(2):257–267. raphy using selective inversion recovery RAGE with fat saturation and ECG-triggering:
11. Moran PR. A flow velocity zeugmatographic interlace for NMR imaging in humans. Magn Application to renal arteries. Magn Reson Med. 1994;31(4):414–422.
Reson Imaging. 1982;1(4):197–203. 49. Robison RO, Blatter DD, Parker DL, et al. Fat suppression in combination with multiple
12. Constantinesco A, Mallet JJ, Bonmartin A, et al. Spatial or flow velocity phase encoding overlapping thin-slab 3-D acquisition MR angiography: Proposed technique for improved
gradients in NMR imaging. Magn Reson Imaging. 1984;2(4):335–340. vessel visualization. AJNR Am J Neuroradiol. 1992;13(5):1429–1434.
13. Pattany PM, Phillips JJ, Chiu LC, et al. Motion artifact suppression technique (MAST) for 50. Lin W, Tkach JA, Haacke EM, et al. Intracranial MR angiography: Application of magne-
MR imaging. J Comput Assist Tomogr. 1987;11(3):369–377. tization transfer contrast and fat saturation to short gradient-echo, velocity-compensated
14. Saloner D. Flow and motion. Magn Reson Imaging Clin N Am. 1999;7(4):699–715. sequences. Radiology. 1993;186(3):753–761.
15. Xiang QS, Nalcioglu O. Differential flow imaging by NMR. Magn Reson Med. 1989; 51. Mathews VP, Ulmer JL, White ML, et al. Depiction of intracranial vessels with MRA:
12(1):14–24. Utility of magnetization transfer saturation and gadolinium. J Comput Assist Tomogr. 1999;
16. Lee JN, Riederer SJ, Pelc NJ. Flow-compensated limited flip angle MR angiography. Magn 23(4):597–602.
Reson Med. 1989;12(1):1–13. 52. Pike GB, Hu BS, Glover GH, et al. Magnetization transfer time-of-flight magnetic reso-
17. Tkach JA, Lin W, Duda JJ Jr, et al. Optimizing three-dimensional time-of-flight MR angi- nance angiography. Magn Reson Med. 1992;25(2):372–379.
ography with variable repetition time. Radiology. 1994;191(3):805–811. 53. Bryant DJ, Payne JA, Firmin DN, et al. Measurement of flow with NMR imaging using a
18. Tkach JA, Ruggieri PM, Ross JS, et al. Pulse sequence strategies for vascular contrast in gradient pulse and phase difference technique. J Comput Assist Tomogr. 1984;8(4):588–593.
time-of-flight carotid MR angiography. J Magn Reson Imaging. 1993;3(6):811–820. 54. Niendorf HP, Alhassan A, Geens VR, et al. Safety review of gadopentetate dimeglumine.
19. Urchuk SN, Plewes DB. Mechanisms of flow-induced signal loss in MR angiography. Extended clinical experience after more than five million applications. Invest Radiol.
J Magn Reson Imaging. 1992;2(4):453–462. 1994;29(suppl 2):S179–S182.
LWBK1209-ch02_p21-38.indd 37 17/05/13 4:50 PM

55. Wedeen VJ, Meuli RA, Edelman RR, et al. Projective imaging of pulsatile flow with 67. Carr JC, Laub G, Zheng J, et al. Time-resolved three-dimensional pulmonary MR angi-
magnetic resonance. Science. 1985;230(4728):946–948 ography and perfusion imaging with ultrashort repetition time. Acad Radiol. 2002;9(12):
56. Miyazaki M, Akahane M. Non-contrast enhanced MR Angiography: Established tech- 1407–1418.
niques. J Magn Reson Imaging. 2012;35(1):1–19 68. Leung DA, Hany TF, Debatin JF. Three-dimensional contrast-enhanced magnetic resonance
57. Miyazaki M, Lee VS. Nonenhanced MR angiography. Radiology. 2008;248(1):20–43 angiography of the abdominal arterial system. Cardiovasc Intervent Radiol. 1998;21(1):1–10.
58. Marchal G, Michiels J, Bosmans H, et al. Contrast-enhanced MRA of the brain. J Comput 69. Leung DA, McKinnon GC, Davis CP, et al. Breath-hold, contrast-enhanced, three-
Assist Tomogr. 1992;16(1):25–29. dimensional MR angiography. Radiology. 1996;200(2):569–571.
59. Prince MR, Yucel EK, Kaufman JA, et al. Dynamic gadolinium-enhanced three- 70. Prince MR, Narasimham DL, Stanley JC, et al. Breath-hold gadolinium-enhanced MR
dimensional abdominal MR arteriography. J Magn Reson Imaging. 1993;3(6):877–881. angiography of the abdominal aorta and its major branches. Radiology. 1995;197(3):
60. Kim JK, Farb RI, Wright GA. Test bolus examination in the carotid artery at dynamic 785–792.
gadolinium-enhanced MR angiography. Radiology. 1998;206(1):283–289. 71. Du J, Carroll TJ, Brodsky E, et al. Contrast-enhanced peripheral magnetic resonance
61. Wilman AH, Riederer SJ, King BF, et al. Fluoroscopically triggered contrast-enhanced angiography using time-resolved vastly undersampled isotropic projection reconstruction.
three-dimensional MR angiography with elliptical centric view order: Application to the J Magn Reson Imaging. 2004;20(5):894–900.
renal arteries. Radiology. 1997;205(1):137–146. 72. Ozsarlak O, Van Goethem JW, Maes M, et al. MR angiography of the intracranial vessels:
62. Earls JP, Rofsky NM, DeCorato DR, et al. Hepatic arterial-phase dynamic gadolinium- Technical aspects and clinical applications. Neuroradiology. 2004;46(12):955–972.
enhanced MR imaging: Optimization with a test examination and a power injector. 73. Laub G. Displays for MR angiography. Magn Reson Med. 1990;14(2):222–229.
Radiology. 1997;202(1):268–273. 74. Tsuruda J, Saloner D, Norman D. Artifacts associated with MR neuroangiography. AJNR
63. Foo TK, Saranathan M, Prince MR, et al. Automated detection of bolus arrival and initia- Am J Neuroradiol. 1992;13(5):1411–1422.
tion of data acquisition in fast, three-dimensional, gadolinium-enhanced MR angiography. 75. Anderson CM, Saloner D, Tsuruda JS, et al. Artifacts in maximum- intensity-projection
Radiology. 1997;203(1):275–280. display of MR angiograms. AJR Am J Roentgenol. 1990;154(3):623–629.
64. Ho VB, Choyke PL, Foo TK, et al. Automated bolus chase peripheral MR angiography: 76. Anderson CM, Lee RE, Levin DL, et al. Measurement of internal carotid artery stenosis
Initial practical experiences and future directions of this work-in-progress. J Magn Reson from source MR angiograms. Radiology. 1994;193(1):219–226.
Imaging. 1999;10(3):376–388. 77. Hu X, Alperin N, Levin DN, et al. Visualization of MR angiographic data with segmenta-
65. Korosec FR, Frayne R, Grist TM, et al. Time-resolved contrast-enhanced 3D MR angiog- tion and volume-rendering techniques. J Magn Reson Imaging. 1991;1(5):539–546.
raphy. Magn Reson Med. 1996;36(3):345–351. 78. Shapiro LB, Tien RD, Golding SJ, et al. Preliminary results of a modified surface rendering
66. Mistretta CA, Grist TM, Korosec FR, et al. 3D time-resolved contrast-enhanced MR DSA: technique in the display of magnetic resonance angiography images. Magn Reson Imaging.
Advantages and tradeoffs. Magn Reson Med. 1998;40(4):571–581. 1994;12(3):461–468.
LWBK1209-ch02_p21-38.indd 38 17/05/13 4:50 PM

Jacob Geleijns Chapter
3
Raoul M.S. Joemai
Albert de Roos
Lucia J.M. Kroft
Techniques for Cardiovascular

Computed Tomography
■■ Technical Principles of Multislice Computed Fast multislice computed tomography (MSCT) offers
Tomography three-dimensional (3D) opportunities for simultaneously im-
aging the coronary arteries (1,2) and myocardium (3) at any
■■ Phase Selection and Temporal
time point in the cardiac cycle. As a noninvasive examina-
Resolution tion, cardiac CT has a low complication rate and allows for
■■ Multislice Computed Tomography Imaging generating consistent good image quality at low radiation
Requirements doses to patients (4).
Required Spatial Resolution The speed and relative simplicity of the examination con-
tribute to patient comfort. Cardiac CT imaging offers good
Required Temporal Resolution
spatial and good low-contrast resolution. Limitations of
Low-Contrast Resolution CT imaging include radiation exposure, not providing real-
Scan Time time guidance for coronary interventions, administration
■■ Patient Preparation of potentially allergenic, and nefrotoxic intravenous con-
trast material and image artifacts. Other diagnostic imaging
■■ Special Applications
techniques for cardiac pathology include two-dimensional
Assessment of Ventricular Function (2D) projection imaging of the coronary arteries and by-
Coronary Artery Calcification Score pass grafts with invasive coronary angiography and imaging
Myocardial Imaging of the myocardium with echocardiography, cardiovascular
Assessment of Pulmonary Veins MRI, SPECT, or PET.
Pediatric Computed Tomography of the Heart Visualization of cardiac morphology by CT was already
■■ Cardiac Computed Tomography reported in 1976. The infarcted area of the myocardium,
both in unenhanced and in enhanced acquisitions, was at
Artifacts that time identified in ex vivo CT examinations in dogs after
High-Contrast Artifacts deliberate coronary occlusion (5–7). These early experi-
Cardiac Motion ments provided a first indication for the potential of cardiac
Respiratory Motion CT and they illustrated the excellent low-contrast resolution
Arrhythmias of axial CT imaging compared to x-ray projection radiogra-
Technical Errors phy. Being ex vivo experiments, these first cardiac CT stud-
■■ Patient Dose in Multislice Computed ies did not address the main challenge of cardiac imaging;
that is, to reduce motion artifacts to an acceptable level.
Tomography Motion artifacts cause blurring of CT images. Cardiac CT
■■ Future Perspectives is particularly susceptible to respiratory and cardiovascular
motion. In 1976, Harell et al. (8) performed experiments to
Computed tomography (CT) is a relatively new versatile demonstrate the feasibility of in vivo cardiac CT by imaging
clinical tool for assessing acquired and congenital cardiovas- the heart of a patient using a CT scanner with 6-second rota-
cular diseases. A CT scan is used not only for the assessment tion time. They described different approaches to minimize
of ischemic heart disease, but also for congenital heart dis- the effect of cardiac motion and presented “stop-action im-
ease, large vessel disease, pulmonary embolism, and pulmo- ages” (Fig. 3.1). The methodologies that were developed in
nary vein imaging. CT imaging can contribute to screening, these early years of cardiac CT are still the conceptual basis
diagnosis, treatment optimization, and follow-up of patients for current techniques in cardiac CT and include reducing
with (suspected) cardiovascular diseases. the duration of data acquisition and synchronization with
39
LWBK1209-ch03_p39-56.indd 39 17/05/13 4:50 PM

A B
C D
E F
Figure 3.1. CT cardiovascular “stop action” images at five phases of the cardiac cycle obtained in 1976,
using 6-second rotation time and 12-second acquisition time.A retrospective segmented sinogram space recon-
struction was performed. (From: Harell GS, Guthaner DF, Breiman RS, et al. Stop-action cardiac computed
tomography. Radiology. 1977;123(2):515–517, with permission.)
LWBK1209-ch03_p39-56.indd 40 17/05/13 4:50 PM

Chapter 3 ■ Techniques for Cardiovascular Computed Tomography 41
the ECG. Cardiac CT, nowadays, is a technique capable of

imaging the small, fast-moving coronary arteries, includ-
ing their calcifications, stenoses, and stents; imaging bypass
grafts; assessment of ventricular function; and visualization
of myocardial perfusion. This chapter focuses on cardiac CT
with scanners employing a fast-rotating x-ray tube and de-
tector array assembly. It does not address other implemen-
tations of 3D x-ray cardiac imaging with x-rays, such as
cardiac electron beam CT (9), which nowadays has become
extinct, and the new application of 3D rotational coronary
angiography with C-arm systems by cardiologists in the
catheterization laboratories (10,11).
Technical Principles of
Multislice Computed
Tomography
It took more than 25 years of innovation in mechanical and
electronic engineering after Harell’s description of “stop-
action cardiac CT” to develop CT scanners that fulfill the
minimal requirements for the introduction of cardiac CT
imaging of the entire heart and coronary arteries in routine
clinical practice. Excellent in-plane spatial resolution and
low-contrast resolution became available in the 1980s with
axial single slice CT scanners. Coverage of the entire heart,
from apex to the base, within one breath-hold became fea-
sible 10 years later with the introduction of helical CT in the Figure 3.2. Contours of the x-ray beams for three types of CT
early 1990s; with these scanners, studies were still performed scanners. From left to right: A 64-slice CT scanner (coverage 32 mm),
in rather thick (≥5 mm) slices (12). The combination of thin a volumetric CT scanner (coverage 160 mm), and a dual source CT
slices (≤2 mm) and sufficient volume coverage for imaging scanner (coverage 38 mm).
the entire heart within one breath-hold was realized with the
introduction of MSCT scanners operating at rotation times arteries in the range of +200 HU, and calcified plaque and
well below 1 second (13). CT technology developed further, bone well above several hundreds of HU (Fig. 3.3). The HU
and currently cardiac CT imaging can be performed within for voxels that contain more than one tissue represents the
one heartbeat by using axial wide cone beam volumetric CT average attenuation within the voxel; this averaging is re-
or fast, helical dual-source CT (Fig. 3.2). ferred to as the partial volume effect. The partial volume
X-ray imaging is based on the registration of the trans-
mission of x-rays through the body. In radiography and fluo- 1200
roscopy, the x-ray transmission is measured in one fixed di- 2200
rection yielding a 2D projection of organs and tissues within 2000
the exposed part of the human body. In CT, x-ray transmis- 1800 Cortical 1000
sion is registered under a large number of projection angles 1600 bone
enabling the reconstruction of axial images and thus avoid- 1400
ing superposition of organs and tissues. 1200 Calcified
Reconstructed axial CT images show pixels, typically 800
plaque
1000
in a 5122 matrix. In cardiac CT, the length and width of
800
these pixels measure about 0.4 mm, depending on the field Cancellous
of view. The height of the associated voxels is in the order 600
bone 600
of magnitude of 0.5 to 3 mm, depending on the clinical ap- 400
plication, with thin slices (<1 mm) in coronary CT angiog- 200
Soft tissue
raphy, and 3-mm thick slices in coronary artery calcium 0 400
fat
scoring. The numerical value associated to the pixels repre- −200
sents the average linear x-ray attenuation coefficient of the −400
tissues within the associated voxel relative to the attenua- Enhanced
−600 200
tion of water, and these values are referred to as Hounsfield Lung vessel lumen
−800
units (HU). The pixel value for air is by definition –1000 HU
−1000 Air
and for water, 0 HU. Human tissues and the lumen of en- Soft plaque
hanced vessels can be characterized by a range of Hounsfield 0
units; for example, fat slightly below 0 HU, soft tissue and Figure 3.3. Range of Hounsfield units for several tissues and for
noncalcified plaque slightly above 0 HU, enhanced coronary plaque and enhanced lumen of the coronary arteries.
LWBK1209-ch03_p39-56.indd 41 17/05/13 4:50 PM

The small, generally 0.5 to 0.625 mm thick, scanned vol-

ume corresponding to one registered transmission profile at
any moment is referred to as a slice; it is an acquisition param-
eter that is selected by the operator of the scanner and deter-
mines the minimal voxel height within the reconstructed slice.
The latest MSCT scanners register from 64 to 320 trans-
mission profiles simultaneously. Slice thickness and the num-
ber of simultaneously acquired transmission profiles together
are referred to as the acquisition configuration; for example,
an acquisition configuration of 64 × 0.625 mm refers to re-
cording simultaneously 64 transmission profiles each with
a slice thickness of 0.625 mm. In this example, the nominal
beam width is 64 × 0.625 = 40 mm, and coverage of the
entire heart is achieved either in successive axial acquisitions
or one helical acquisition. An acquisition configuration of
up to 320 × 0.5 = 160 mm can be selected for a wide cone
beam volumetric CT scanner; this provides enough coverage
for scanning the entire heart in one axial scan.
Rotation time is the duration of a 360-degree rotation of
the x-ray tube. The higher the rotation speed, the faster data
acquisition and the shorter the rotation time. In helical CT, in
order to achieve coverage of the entire volume of interest, the
rotation of the x-ray tube is coupled with a continuous transla-
tion of the patient through the gantry of the CT scanner. The
helical pitch factor is the ratio of patient displacement per
360-degree rotation of the x-ray tube divided by the total
nominal thickness of all simultaneously acquired transmis-
Figure 3.4. Cardiac CT involves simultaneous registration of trans-
mission profiles and an ECG. Each horizontal line in sonogram space
sion profiles, or in other words, it is the table movement per
represents one transmission profile, at a certain acquisition angle and is gantry rotation relative to the beam width.
associated with the registered ECG signal. MSCT yields multiple sinogram The acquisition window is duration of the data acquisition
space, that is, each acquisition channel generates its own sinogram space. at a specific level within the scanned volume. In axial scans,
the acquisition window is the actual duration of the axial
scan, the axial scan must be at least long enough to cover a
effect affects image quality especially in case of large voxels, 180-degree acquisition, but may be longer in order to cover
such as within thick reconstructed slices and within voxels more cardiac phases or heartbeats. The acquisition window in
at the border of two tissues with a large difference in HU, or helical cardiac CT can be estimated by dividing rotation time
at the border of coronary artery lumen and calcified plaque. by the pitch factor. In helical cardiac CT with retrospective
The detectors of a CT scanner record x-ray transmission ECG gating, the acquisition window should be long enough
as a transmission profile. Each transmission profile acquired to catch at least the entire cardiac cycle at any level along the
with the rotating x-ray tube and detector array is associated entire volume of interest. A sufficiently long acquisition win-
with a specific projection angle. Transmission profiles mea- dow can be achieved by a small pitch factor. To facilitate syn-
sured at consecutive angles yield the lines from which the chronization, cardiac CT scanners are equipped with a device
so-called sinogram space is composed (Fig. 3.4). The hori- for recording an electrocardiogram (ECG). Data acquisition
zontal axis in sinogram space is associated to the position in cardiac MSCT thus consists of simultaneous registration of
along the detector array and represents the transmission transmission profiles together with the associated ECG signal.
measured by all detector elements along one detector array.
The vertical axis of sinogram space represents the consecu-
tive projection angles. The sinogram space is visualized by Phase Selection and Temporal
gray levels that correspond to the recorded x-ray transmis- Resolution
sion for each detector element and for each projection angle.
The technique of filtered back projection provides the recon- The main challenge of cardiac CT is to minimize blurring
structed images, but other reconstruction algorithms have due to motion. This is achieved by optimizing phase selec-
become available, like iterative reconstruction techniques. tion; that is, reconstruction of the scan at the cardiac phase
The basic acquisition parameters in helical and axial car- with least motion and good temporal resolution. For phase
diac MSCT are slice thickness, number of active detector selection usually ECG is used. Phase selection is complicated
arrays, x-ray tube rotation time, and in helical CT, also the since the cardiac rest phase depends on the heart rate. At
pitch factor. These parameters must be optimized to achieve lower heart rates, least cardiac motion occurs predominantly
contradictory goals such as minimal motion artifacts, cover- at the middiastolic phase, but at higher heart rates minimal
age of the entire volume of interest, coverage of the cardiac motion may occur at end systole.
phases of interest, good spatial resolution, and as less radia- Phase selection can be performed during the acquisition
tion dose as possible. (prospective) and the reconstruction (retrospective). These
LWBK1209-ch03_p39-56.indd 42 17/05/13 4:51 PM

two techniques for phase selection are generally referred to as modulating the tube current in real time between higher and
prospective ECG triggering of the acquisition and retrospec- lower levels, or by periodically switching the tube current on
tive ECG gating of the reconstruction. Often a combination and off. Prospective ECG triggering is a common practice
of prospective ECG triggering of the acquisition and retro- in cardiac CT and can be used for both helical and axial
spective ECG gating of the reconstruction is applied. If no acquisitions, and for wide cone beam and dual source CT
prospective phase selection is performed then the acquisition scanners.
window must cover the entire RR interval, and in this case, Cardiac phase points define a reconstruction window
retrospective-gated reconstructions can be made any cardiac within the cardiac cycle. Techniques for synchronization
phase point but at the cost of elevated patient exposure. often assume a linear relationship between the ECG and the
During prospective ECG-triggered acquisitions, ECG is actual contraction and relaxation of the heart chambers,
used to trigger the start of the acquisition. ECG can also be which might be a good assumption if heart rate does not
used to limit the acquisition window to only that part of the vary too much, but artifacts might occur even in case of mild
RR interval that covers the cardiac rest phase; this is done in arrhythmias.
order to reduce patient radiation dose. ECG can also be used Figure 3.5 shows the acquisition technique and result-
to optimize the tube current; this can be achieved by either ing image quality of a low-dose prospective ECG-triggered
B C
Figure 3.5. A prospective ECG-triggered helical scan in a 64 × 0.5 mm acquisition configuration in a 74-year-
old female patient after bypass surgery. The ECG (A) shows the identified R-peaks (red), the short intervals of data
acquisition (or acquisition windows, in gray), and marked in yellow the reconstruction windows. This acquisition
mode allows for low-dose full thorax imaging (B). A range of 260 mm was acquired in 11 consecutive heartbeats
as to visualize the full bypass trajectory after left and right internal mammary artery use (C). Patient had a low heart
rate of 50 to 51 beats/minute and was exposed to x-rays only at diastole (effective dose 5.7 mSv, DLP 405 mGy cm,
13.9 seconds scan time, acquisition window at 75% RR, and a helical pitch factor of 0.23).
LWBK1209-ch03_p39-56.indd 43 17/05/13 4:51 PM

elical acquisition. Note that tube current was only on dur-

h 300
Pitch factor = 0.4
ing the gray-marked intervals (acquisition windows). The Pitch factor = 0.3
figure also shows that the ECG-triggered acquisition was 250 Pitch factor = 0.2
performed with an acquisition window that was slightly lon-
Reconstruction window, ms
Pitch factor = 0.1
ger than the reconstruction window. This allows for some 200
retrospective optimization of the phase selection.
In general noncardiovascular applications of CT, the 150
reconstruction window (also referred to as temporal reso-
lution) of any reconstructed image simply equals the full 100
360-degree rotation time. Rotation times of modern scan-
ners are about one-third of a second. A reconstruction 50
window of this order of magnitude is sufficiently short for
most applications of CT, but not for cardiac CT. In CT,
0
opposing projections yield identical transmission profiles 40 60 80 100 120
meaning that 360-degree sinogram space contains redun- Beats/minute, min−1
dant information. The minimal requirements for the CT Figure 3.7. Theoretic reconstruction window as a function of heart
reconstruction are achieved by excluding all redundant rate and pitch factor. The graphs are calculated, assuming optimal seg-
transmission profiles from sinogram space. This yields a mentation of sinogram space and thus optimal adaptation of rotation
180-degree sinogram space corresponding to a reconstruc- time to the heart rate. A short reconstruction window reduces motion
tion window of approximately half the rotation time. For artifacts, it requires a low pitch factor, especially at low heart rates.
typical rotation times of respectively 0.35, 0.5, and 0.6
seconds, the reconstruction window associated with a re-
thus explains the feasibility of cardiac CT by Harell et al. (8)
construction from a 180-degree sinogram space would be
even at a CT scanner with a rotation time of 6 seconds. To
approximately 175, 250, and 300 milliseconds. A quarter
achieve the optimal curves in Figure 3.7, it was assumed that
of a rotation is sufficient for covering 180-degree sinogram
rotation time was adapted accurately to the heart rate. In
space with a dual source CT scanner; thus for these scan-
general practice, the theoretically achievable lowest recon-
ners, the reconstruction window is about one-fourth of the
struction window cannot always be realized, since optimal
rotation time (14).
adaptation of rotation time to the actual heart rate cannot
To achieve an extra short reconstruction window, tech-
be achieved due to the fixed values of rotation time and vari-
niques have been developed to record transmission profiles
ations in heart rate during the CT scan.
during more than one heart cycle throughout the entire vol-
Volume rendering of two scans of one and the same
ume of interest. The reconstructed slices can then be derived
patient (Fig. 3.8) illustrates the difference in image quality
from a so-called 180-degree segmented sinogram space,
where small complementary subsegments of sinogram space
are derived from successive heartbeats together yielding the
180-degree segmented sinogram space. This technique was
already described in 1977 by Harell et al. (8) (Fig. 3.6), and
is based on a relatively long acquisition window. The recon-
struction window (temporal resolution) that can be achieved RCA
theoretically with this technique in retrospective ECG-gated LAD
reconstructions is shown in Figure 3.7. The figure applies
to a reconstruction algorithm that allows for including all
available subsegments in sinogram space. Note that the re-
construction window is independent of rotation time and A
Figure 3.6. Three of the acquired 13½ cardiac cycles. The RR

interval for each cycle was subdivided into seven cardiac segments B
representing different cardiac phases. In reconstructing axial images
at different cardiac phases, angular projections (or raw data) from all Figure 3.8. Volume rendering of two scans of the same patient
cardiac cycles, corresponding with the cardiac phase of interest, were illustrates the difference in image quality obtained with a retrospec-
used. (From: Harell GS, Guthaner DF, Breiman RS, et al. Stop-action tively gated 180-degree segmented sinogram space reconstruction
cardiac computed tomography. Radiology. 1977;123(2):515–517, with (ECG synchronized) (A) versus a regular 360-degree sonogram space
permission). reconstruction (not synchronized) (B).
LWBK1209-ch03_p39-56.indd 44 17/05/13 4:51 PM

obtained with a cardiac CT scan (ECG synchronized) versus Multislice Computed Tomography
a regular CT scan (nonsynchronized). The volume render- Imaging Requirements
ing of the synchronized images shows the heart in the ret-
rospectively selected diastolic phase. All different phases of
Required Spatial Resolution
the cardiac cycle express in the nonsynchronized images as
smooth deformations of the heart. The long reconstruction Spatial resolution, or high-contrast resolution, is the ability to
window leads to poor temporal resolution and blurring observe the contours of small objects within the scanned vol-
of the coronary arteries in the nonsynchronized images ume. Small objects can be resolved properly only when they
whereas the short reconstruction window of the ECG syn- provide a rather large difference in signal (HU’s) compared to
chronized and retrospectively gated reconstruction enables the direct environment. Spatial resolution plays an important
good temporal resolution and visualization of the coronary role in the visualization of contrast-enhanced distal segments
arteries. of coronary arteries, calcifications, or stents (Fig. 3.9). In CT,
Figure 3.9. Stent reconstructed with medium-soft and sharp kernel. Visualization of the stent in the left
anterior descending coronary artery. The medium-soft kernel reconstruction is generally used for coronary artery
evaluation (A), but suffers, at the level of the stent, from blooming artifacts that hamper evaluation of the stent,
especially within the stent lumen. Additional sharp kernel reconstruction allows better delineation of the stent
itself and usually also stent lumen (B). Note the proximal stent-edge intima hyperplasia/plaque causing stenosis.
LWBK1209-ch03_p39-56.indd 45 17/05/13 4:51 PM

umen Diameter of Normal

L that the reconstruction window should be lower than 100
milliseconds for coronary angiography in middiastole at
Table 3.1 Coronary Artery Segments
62 ± 10 beats/minute (17) and that a 100-millisecond recon-
and Venous Bypass Grafts
struction window is relatively optimal for most patients at
Coronary Arteries First Segment Last Segment heart rates up to 90 beats/minute (16). These considerations
assume imaging at the cardiac phase point that is associ-
LM 4.3 (3.5–5.1) mm — ated with least motion; for example, reconstruction window
LAD 3.5 (2.9–4.2) mm 0.8 (0.5–1.2) mm starting between 60% and 80% of the interval between
LCX 3.2 (2.4–3.9) mm 1.3 (0.6–1.8) mm two consecutive R waves. In practice, the requirements on
RCA 3.7 (2.7–4.6) mm 1.8 (1.2–2.5) mm the reconstruction window appear to be less demanding:
Bypass grafts 6 (4–8) mm A reconstruction window for coronary CT angiography
of 250 milliseconds seems to be sufficiently short at heart
LM, left main; LAD, left anterior descending; LCX, left circumflex;
RCA, right coronary artery.
rates below 70 beats/minute (18). More strict criteria for the
From: Dodge JT, Jr., Brown BG, Bolson EL, et al. Lumen diameter reconstruction window apply if the heart should be assessed
of normal human coronary arteries. Influence of age, sex, anatomic at more than one cardiac phase point, including those that
variation, and left ventricular hypertrophy or dilation. Circulation. are associated with rapid movement of the heart wall; for
1992;86:232–246. example, for studying the dynamics of the myocardium.
Alfidi et al. (19) concluded already in 1976 that for imag-
ing the dynamics of the heart with CT, such as the detection
spatial resolution is limited not only by the acquisition geom- of akinetic and dyskinetic myocardium, it should become
etry; that is, the dimensions of the detector elements, small- possible to achieve a reconstruction window smaller than 50
est available slice thickness, focus size, and focus to detector milliseconds. However, accurate evaluations of cardiac func-
distance, but also by the reconstruction filter and the recon- tion by CT have been reported under less strict conditions,
structed slice thickness. at reconstruction window exceeding 50 milliseconds (20).
The reconstruction of small isotropic voxels facilitates
advanced postprocessing and 3D image viewing, and is par-
Low-Contrast Resolution
ticularly advantageous for review of the coronary arteries.
The actual diameters of the lumen of normal coronary ar- Low-contrast resolution is the ability to detect structures
tery segments range from 5 mm in the proximal segments to that express only a small difference in signal (HU’s) com-
less than 1 mm in the distal segments. Venous bypass grafts pared to their direct environment. For the detection of low
typically range from 4 to 6 mm (Table 3.1). This means that contrasts in CT images, the lesions must be fairly large in
a voxel size of 1 mm3 should be sufficient for imaging of the size. Native tissue contrasts are, in general, not sufficient to
coronary arteries, except for distal segments that would re- differentiate between structures, such as the vessel wall, its
quire a spatial resolution of at least 0.5 mm3. A spatial reso- unenhanced lumen, and the myocardium, or to distinguish
lution of 2 mm3 might be sufficient for imaging the lumen pathology within the myocardium. Contrast enhancement is
of venous bypass grafts, but for imaging of structures within thus mandatory. However, even with proper enhancement
the coronary arteries, such as atherosclerotic plaque and using iodinated contrast media, the contrast between nor-
stents, excellent spatial resolution even better than 0.5 mm3 mal, ischemic, and infarcted myocardium remains rather
might be required. The requirements for spatial resolution low (see Fig. 3.10). Image noise is the main limitation for CT
are of course less demanding for imaging larger structures when imaging structures that exhibit rather low contrast.
such as the myocardium, large thrombi, pulmonary veins, or Image noise may be decreased either by raising tube current
heart chambers. (mA) at the cost of patient exposure, or by increasing the
reconstructed slice thickness at the cost of spatial resolution.
Required Temporal Resolution
In addition, low-contrast resolution depends on tube volt-
age, beam filtration, and reconstruction filter.
The reconstruction window determines the ability to resolve
fast-moving objects in the displayed CT image. The recon-
Scan Time
struction window is also referred to as the temporal resolu-
tion and plays an essential role in the visualization of a beat- Scan time is the time interval between the start and the end of
ing heart and its coronary arteries. A short reconstruction an acquisition of the entire volume of interest. To avoid breath-
window avoids degradation of both low-contrast resolution ing artifacts and to limit the amount of contrast material, scan
and spatial resolution in the image, generally referred to as time should remain at least below 30 seconds but preferably
blurring of the image, due movement of the cardiac wall below 20 seconds. The extent of the target volume, as well as
and coronary arteries. Actual in-plane velocities of human acquisition parameters such as rotation time, pitch factor, slice
coronary arteries have been measured at different moments thickness, and number of simultaneously acquired slices define
during the cardiac cycle by means of imaging techniques scan time. Short scan times, well below 10 seconds, can be
that allow for a very short reconstruction window; that is, realized with multichannel CT scanners that acquire 64 slices
gradient-echo MRI (15) and prospectively triggered electron simultaneously. Very short scan times can be realized with
beam CT (16). With these techniques, a reconstruction win- wide cone beam CT scanners (21,22) and ultrafast dual source
dow as short as 15 milliseconds (MRI) and 50 milliseconds CT scanners (23); these scanners are capable to scan the entire
(EBT) was achieved. In these publications, it is concluded heart within a fraction of one single heartbeat.
LWBK1209-ch03_p39-56.indd 46 17/05/13 4:51 PM

A B C
Figure 3.10. Imaging of a perfusion defect requires good low-contrast resolution. ECG-synchronized CT
in a 38-year-old male patient after left ventricular anterior wall infarction. Axial plane (A), parasagittal (B),
and left ventricular short axis (C) reformats. A subendocardial perfusion defect is shown in the left ventricular
anterior wall (white arrows). Also, a small subendocardial perfusion defect can be observed at the inferior wall
(black arrow, B, C).
Patient Preparation time and according to scan volume. Standard cranio-

caudal scanning is used for coronary artery scans. If the
Contraindications for cardiovascular MSCT include severe superior mediastinum is included for visualization of the
arrhythmias and allergy for iodinated intravenous contrast aorta or mammary artery bypass grafts, a caudo-cranial
media. Good fixation of the ECG wires is essential. The direction is preferred.
ECG signal should be clear and specifically yield a clear
demarcation between the R-peaks and the rest of the signal. Special Applications
It should be avoided that lines may drop due to motion of
the table or movements of the patient during the scan since MSCT provides special opportunities for cardiovascular CT
they may cause signal distortion on the ECG. Beta-blockers in addition to imaging of the coronary arteries, coronary
may be used to reduce the heart rate to preferably below bypass grafts, and the myocardium. These options include
65 beats/minute. The resulting performance is more pre- assessment of left ventricular (LV), right ventricular (RV),
dictable and of more consistent quality when using medi- and left atrium (LA) function; coronary calcification score;
cation. The less variation in heart rate occurs during the myocardial imaging; imaging of anatomy of pulmonary
scan, the better the result. Also, when total scanning time veins in patients with atrial fibrillation; and pediatric CT of
is short; for example, below 15 seconds, the quality of the the heart. Each of these applications can be characterized
scan improves because of the reduction of total amount of by its specific techniques for acquisition and reconstruction.
heartbeats in the scan and less variation in the heart rate.
Furthermore, the patient should be positioned supine with
Assessment of Ventricular Function
the heart at the center of rotation and the arms outside the
region of interest. It is important in patient preparation to A contrast-enhanced MSCT examination of the entire heart
provide a good explanation to the patient about the scan that allows for reconstructions at any cardiac phase is suit-
procedure. A patient should respond adequately on breath- able for assessment of ventricular function (20). This can
ing instructions and practicing the breath-hold before be achieved; for example, by using retrospective ECG-gated
scanning is advised. The patient should be informed about reconstructions, or by an ECG-triggered acquisition with a
the breath-holding time, and should be instructed that the wide cone beam CT scanner during an entire RR interval.
breath-hold should be at the same depth. Hyperventilation Assessment of ventricular function can thus be achieved
and administration of oxygen may be used to reduce the simultaneously with the MSCT coronary angiography acqui-
heart rate particularly at scan times of approximately sition. Cardiac function is generally derived from recon-
20 seconds or longer. structions at 10 to 20 cardiac phase points. Not only left
In cardiac CT, it can be difficult to exactly set the scan- ventricular function, but also right ventricular function can
ning range with the 2D-scan projection radiograph (SPR) be measured by dynamic CT. Contrast-enhanced CT is now
that is acquired prior to the actual scan. Therefore, a low- routinely used as a first-line imaging tool in patients with
dose axial orientation scan may be useful. Areas difficult suspected pulmonary embolism. Right ventricular enlarge-
to recognize in the SPR are the apex of the heart, the origin ment on chest CT has been shown to be a predictor of early
of the coronary arteries, and the origin of bypass grafts. If death in patients with acute pulmonary embolism (24,25).
exact positions cannot be found on the SPR and an orien- Even the dimensions of the right ventricle on nondynamic
tation scan is not performed, longer scanning ranges are routine CT images may be predictive for mortality.
set to avoid the risk of not scanning the total target vol- Global ventricular function is generally measured as the
ume at the cost of elevated radiation exposure. Scanning end-systolic volume (ESV) and the end-diastolic volume (EDV).
directions are to be set according to total breath-holding Subsequently, stroke volume (SV) and ejection fraction (EF) can
LWBK1209-ch03_p39-56.indd 47 17/05/13 4:51 PM

ECG dose modulation

Tube current (mA)
High
Low
0
0 200 400 600 800 1000 1200
RR interval (ms)
Figure 3.11. The coverage of a
volumetric CT scanner is sufficient
Cardiac function for imaging the entire heart during
250 one single heartbeat thus allowing
Left ventricular for simultaneous single-beat CT
coronary angiography and cardiac
200 Right ventricular function. (A) shows above the ECG
Left atrium signal and the tube current modula-
Volume (mL)
tion. Tube current is high during the

150 diastolic phase for high-quality CT
coronary angiography and low dur-
100 ing systole. Graphs below represent
left ventricular (LV), right ventricu-
lar (RV), and left atrial (LA) func-
50 tion. (B) shows from left-to-right,
two-chamber view, four-chamber
view, short axis, and volume-
0
rendered segmentations. The colors
0 200 400 600 800 1000 1200 in the graph and the volume
RR interval (ms) rendering represent LV (blue),
B
RV (red), and LA (green).
LWBK1209-ch03_p39-56.indd 48 17/05/13 4:51 PM

easily be derived from ESV and EDV. Software may be used adenosine. Limitations of myocardial imaging with CT are
for automatic ventricular cavity contour detection and for the that CT provides a relatively poor signal-to-noise ratio (com-
calculation of global ventricular function. Figure 3.11 shows pared to, for example, MRI (33)), that myocardial imaging
software for the assessment of LV, RV, and LA functions. is sensitive for the occurrence of artifacts, and the required
Regional LV wall motion can be assessed by visual scor- extra radiation dose.
ing of cinematic loops of well-described myocardial seg-
ments (26). A wall motion score of 1 to 4, ranging from
normal, hypokinetic, akinetic to dyskinetic myocardium can Assessment of Pulmonary Veins
be assigned to each segment (26).
Atrial arrhythmias often originate in the pulmonary veins
Integrated CT assessment of the coronary arteries and
and can be treated with percutaneous radiofrequency cath-
regional myocardial function allows more complete evalu-
eter ablation. With this technique, the arrhythmic foci are
ation of the functional consequences of a coronary artery
electrically disconnected from the left atrium by means
stenosis. The usefulness of this combined approach has been
of catheters placed in the left atrium (34). Preprocedural
reported in patients with hypertension and diabetes mellitus
MSCT examination is helpful to depict the anatomy of the
(27,28).
pulmonary veins and left atrium and particularly to demon-
strate additional pulmonary veins (e.g., middle lobe vein),
which is important for planning the interventional proce-
Coronary Artery Calcification Score
dure. Variations in pulmonary venous anatomy are quite
Coronary artery calcification is a manifestation of athero- common and comprise variation in the number of veins as
sclerosis in the vessel wall, which becomes more exten- well as the occurrence of common ostia and early branching
sive in advanced lesions. Coronary artery calcification is (35). Three-dimensional, surface-rendering reconstructions
strongly correlated to coronary events. CT provides a non- provide a quick overview of the pulmonary venous anatomy,
invasive method for detecting and quantifying coronary but cross-sectional reconstruction in coronal, sagittal, and
artery calcification (29). This may contribute to improved transverse orientations are necessary for full appreciation of
sensitivity in the stratification of high-risk, asymptom- the morphology of the pulmonary veins (36). MSCT allows
atic individuals. Coronary calcification is best detected visualization of pulmonary vein anatomy, providing the car-
and measured in a plain (unenhanced) MSCT acquisi- diologist with a roadmap prior to ablation in patients with
tion. Prospective ECG-triggered acquisitions are generally atrial fibrillation (36). Postprocedural MSCT also offers
used to avoid motion artifacts. Translation of the patient an opportunity for follow-up of the pulmonary vein after
through the gantry of the CT scanner is achieved between ablation (37).
successive acquisitions. These ECG-triggered acquisitions MSCT pulmonary venography requires a contrast-
are most often axial acquisitions, but can also be helical enhanced acquisition. Imaging requirements for the rather
acquisition. There is, however, a lack of standardization of large pulmonary veins with, in general, diameters well above
the MSCT techniques with regard to image acquisition as 10 mm are not as strict as in coronary CT angiography, and
well as to the methodologies for quantitative coronary cal- pulmonary venography does not require synchronization of
cification scoring. The development of standardized and transmission profiles with the ECG since cardiac phase se-
reproducible algorithms for measurements is a technical lection is not essential in the evaluation of the pulmonary
prerequisite for the use of coronary calcification scoring veins. This means that a regular MSCT acquisition, with
to become a useful clinical tool. In addition, the coronary a high pitch factor and resulting rather low patient dose,
calcification score will have to be established as an inde- can be performed. However, ECG synchronization is rec-
pendent predictor of existing risk factors for cardiovascu- ommended and for assessment of complications related to
lar disease (30). the ablation procedure, such as pulmonary venous stenoses,
ECG-synchronized MSCT may be preferred for better visu-
alization of details (Fig. 3.12).
Myocardial Imaging
Using MSCT, myocardial perfusion defects can be observed
Pediatric Computed Tomography
in the early phase of the contrast bolus phase (early defect).
of the Heart
In the late phase, residual defects and late enhancement
can be observed. MSCT can provide information about With a wide cone beam CT scanner, the entire heart can
myocardial pathology. MSCT allows for distinguishing be imaged within a fraction of one heartbeat (38,39). This
between viable and nonviable myocardium with a technique is particularly advantageous in pediatric CT since both the
similar to delayed enhancement imaging in MRI. MSCT also dose to the patient and the acquisition time are very low dur-
allows for myocardial perfusion imaging in rest and stress ing such acquisitions. Cardiac CT can be applied safely for
for assessment of stenoses that limit the blood flow mainly imaging congenital heart disease in neonates and small chil-
under stress and less at rest (31,32). Such applications of dren (39). The subsecond acquisition time with wide cone
myocardial imaging generally start with a regular coronary beam scanning may obviate the need for sedation in these
CT angiography scan, followed by one additional scan of children. Figure 3.13 shows an ECG-triggered CT of a child
the entire heart that provides information about the delayed and a 3-year-old male patient with congenital heart disease.
enhancement or stress perfusion of the myocardium. Stress Contrast-enhanced wide cone beam CT was performed in a
imaging can be performed by using a vasodilator such as 280 × 0.5 mm acquisition configuration.
LWBK1209-ch03_p39-56.indd 49 17/05/13 4:51 PM

A B
Figure 3.12. Single heartbeat, low-dose prospectively ECG-triggered pulmonary vein imaging with a 320-
slice volumetric CT scanner in a 49-year-old male patient with paroxysmal AF, before radio-frequency ablation
therapy of the pulmonary veins. The heart rate was 59 beats/minute, prospective ECG triggering at 75% of
the cardiac cycle was used (A). Left atrium (LA), left atrial auricle (LAA). Pulmonary veins: Left common (LC)
ostium and right superior (RS) and right inferior (RI) pulmonary veins. In this low-dose prospectively triggered
CT, the coronary arteries are displayed sharply as well (B). Coronary artery stenosis can be confidentially
excluded. Radiation dose was 0.9 mSv.
Cardiac Computed Tomography tions, or a high concentration of iodinated contrast material.

Artifacts The high-contrast artifacts are caused by the partial volume
effect, beam hardening, and a relatively low detector signal lead-
CT is sensitive for the occurrence of image artifacts. Well- ing to a nonlinear response of the detector and its electronics. In
known CT artifacts occur due to the partial volume effect, the CT images, this results in blooming and streaks (Fig. 3.14).
beam hardening, and motion. Typical artifacts in general Blooming artifacts caused by calcifications, stents, or clips may
MSCT are helical artifacts (40), “windmill” and “cone beam” seriously affect the evaluation of the lumen of coronary arteries
artifacts (41). Cardiac CT is particularly susceptible to arti- or bypasses. Streaks caused by high concentrations of iodin-
facts, including both general CT artifacts as well as artifacts ated contrast material or clips extend throughout a relatively
that are introduced by the dedicated reconstruction algo- large part of the image and may obscure structures of inter-
rithms that are used in cardiac CT. est. The effect of high-contrast artifacts, especially blooming,
was most prominent at four-slice scanners and decreased with
16- and 64-slice CT scanners that allow for scanning the heart
High-Contrast Artifacts
at a smaller slice thickness. Reconstruction techniques based on
High-contrast artifacts are caused by high attenuating objects, iterative algorithms that have become available may be more
either metal objects such as stents and surgical clips, calcifica- robust with regard to avoiding high-contrast artifacts.
Figure 3.13. ECG-triggered

CT of a child, 3-year-old male
patient with congenital heart dis-
ease. CT image of thin maximum-
intensity projection in axial orien-
tation showing the origins of the
right coronary artery (RCA) and
left main coronary artery (LM).
Contrast-enhanced volumetric CT
was performed at 45% of the RR
cycle (end-systolic) at high heart
rate of 98 beats/minute. The child
weighted 10 kg, effective dose was
2 mSv, DLP was 70 mGy cm, 100
kV, and volumetric CT with a 280
× 0.5 mm acquisition configuration.
LWBK1209-ch03_p39-56.indd 50 17/05/13 4:51 PM

A B
C D
Figure 3.14. Surgical vascular clip artifacts. A 67-year-old male patient after coronary bypass operation.
Coronary calcifications (A, small arrow) and metal clip artifacts in the course of the left internal mammary
artery (LIMA) (A–D, arrow) after left anterior descending coronary artery (LAD) bypass. Because of these arti-
facts, the LIMA and native coronary artery at the sites of the metal clip artifacts cannot be evaluated properly.
Cardiac Motion A countermeasure for cardiac motion artifacts is to use

heart-rate–reducing medication (beta-blockers) in case of a
Cardiac motion leads to blurring of the coronary arteries too high heart rate.
and the myocardium, and may affect visualization of rele-
vant structures (Fig. 3.15). Motion-related artifacts are most
Respiratory Motion
severe for the right coronary artery (RCA) since it moves at
the highest velocity and least for the left anterior descend- Respiratory motion leads to deformations in the reconstructed
ing (LAD) artery (16). Breath-holding also influences the images that can be easily recognized in multiplanar reformat
heart rate because during the breath-hold, heart rate initially (MPR) or 3D reconstructions. Four-slice CT scanners were par-
decreases then gradually increases (42). Cardiac motion particularly susceptible to breathing artifacts since the small number
ticularly complicates the evaluation of small distal segments of simultaneously acquired slices results in a rather long scan time
of coronary arteries and, in particular, the distal anastomo- and consequently from 30 to 40 seconds long breath-hold. With
ses of bypass grafts to the coronary arteries. the 64-slice CT scanners, the breath-hold dropped to well below
LWBK1209-ch03_p39-56.indd 51 17/05/13 4:51 PM

since a wide cone beam scanner allows for arrhythmia rejec-

tion. Arrhythmia rejection is applied during the acquisition,
and the CT scanner automatically stops the exposure in case
a too short RR interval is recorded, until the heartbeat is long
enough for proper image acquisition at the desired cardiac
phase point. Figure 3.17 shows that after two rejected RR
intervals, the third heartbeat was accepted, and the recon-
struction provided good diagnostic image quality.
Technical Errors
Figure 3.15. Although temporal resolution of MSCT has improved,
blurring of the vessel wall still occurs.Compare, for example, MSCT Cardiac CT is also prone to technical errors. Incomplete
coronary angiography (curved MPR) with conventional invasive coro- coverage of the volume of interest might occur in case of
nary angiography. The latter technique is superior to cardiovascular inaccurate planning of the scan range. Erroneous timing of
MSCT, with regard to temporal resolution. the start of the scan relative to contrast administration might
cause poor contrast enhancement. Synchronization artifacts
may occur in case of a spurious ECG signal, leading to mis-
10 seconds, and a wide cone beam CT scanner or dual-source registrations of the R-peaks, or may occur in case of serious
CT scanner can image the entire heart within a fraction of a arrhythmias (Fig. 3.18). Selection of a pitch factor that is
second. Figure 3.16 shows a breathing artifact, as a smooth too high, or a rotation time that is too long in relation to the
deformation of the heart wall. Countermeasures for respira- heart rate of the patient, will result in missing one or more
tory motion artifacts include proper patient preparation, and slices within the reconstructed volume of interest.
instruction and hyperventilation just before the acquisition.
Arrhythmias
Patient Dose in Multislice
Computed Tomography
Arrhythmias that are common in patients with cardiovascular
disease often cause severe artifacts in the cardiac CT scans. Radiation protection of patients is based on the principles of
Artifacts from mild arrhythmias may be reduced by ECG justification and optimization. Justification implies that the
editing. With ECG editing, the phase selection for the most benefit for the patient; for example, exclusion of pathology,
appropriate reconstruction window is performed for each sep- diagnosis of disease, or follow-up of treatment, outweighs
arate RR interval after that the scan was performed. Artifacts the risk of radiation exposure. Generic justification requires
resulting from severe arrhythmias cannot be corrected after- that CT imaging will usually improve the diagnosis or treat-
ward with ECG editing. Wide cone beam scanning is a robust ment of a well-described clinical problem and patient pop-
acquisition technique for patients with irregular heart rate, ulation. Individual justification means that the cardiac CT
scan is expected to do more good than harm to the individ-
ual patient. Once justified, incorporating measures that tend
to lower radiation doses should optimize the CT examina-
tion. Patient dose assessment is thus required for balancing
harm and benefit of the CT examination and to assess the
effect of measures for optimization of cardiac CT. Note that
neither dose constraints nor dose limits apply to patients.
Nowadays, most CT scanners provide the user with an indi-
cation of the radiation exposure.
There is general consensus concerning the dose descrip-
tors to be used in CT, these are the Computed Tomography
Dose Index (CTDI) (43) and the Dose Length Product (DLP)
(44). The CTDI is generally expressed as the volume CTDI
(CTDIvol) (45).
The CTDIvol is a measure of the local absorbed dose in
an axial plane. The CTDI (mGy) depends primarily on tech-
nical acquisition parameters such as tube current, rotation
time, tube voltage, beam filtration, pitch factor, and geomet-
ric efficiency. The CTDI for scans of the body is derived from
measurements in a 32-cm diameter cylindrical PMMA CT
dosimetry phantom representing the adult body. The CTDI
is particularly suitable for comparison of different acquisition
protocols irrespective of the type of MSCT scanner. In addi-
Figure 3.16. Motion artifact: In helical CT the smooth deformation, a quantity is required that expresses patient exposure
tion of the cardiac wall, as a result of breathing during the scan, can from the complete MSCT examination. Such a quantity
be recognized easily in the 3D reconstruction (arrow). should take into account the extent of the exposed range
LWBK1209-ch03_p39-56.indd 52 17/05/13 4:51 PM

Figure 3.17. Wide cone-beam CT of the heart in a patient with irregular heart rate. Two unexpectedly short
RR intervals occurred after starting the scan (respectively 417 and 751 ms), the scanner rejected these heartbeats
and could successfully acquire the cardiac CT scan during the third heartbeat (RR interval 1,145 ms). The red
dots mark the R-peaks, tube current was on for the gray areas, the blue line marks the rejected heartbeats,
and the yellow line marks the reconstruction window. The acquisition configuration was 280 × 0.5 mm. Good
diagnostic image quality was achieved.
and the exposures during all sequences of the examination. modern CT scanners by multiplying DLP with the normal-
A dosimetric quantity that fulfils these conditions is the dose- ized effective dose coefficient of 0.03 mSv/mGy cm (49); this
length product (DLP, mGy cm). DLP is the CTDIvol multiplied value is valid for a typical cardiac CT acquisition at a tube
by the length of the exposed range for each sequence. voltage of 120 kVp. Table 3.2 provides typical values for the
There is sometimes a need to assess the effective dose radiation exposure for patients, dose limits for workers, and
(46) for cardiac CT to allow comparison with other types natural radiation levels.
of radiologic examination such as invasive coronary angi- Effective dose from cardiac CT was relatively high,
ography or for risk assessment. The effective dose for a par- mainly due to lack of prospective ECG triggering and thus
ticular cardiac CT scanning protocol can be estimated from the need to catch all cardiac phases at any level of the scan
the CTDIvol and the acquisition protocol by using dedicated (50). Concern about radiation exposure stimulates the de-
software (47,48). velopment of methods for dose reduction in cardiac CT such
Alternatively, broad estimates of effective dose (E) may be as ECG-triggered modulation of the tube current which re-
derived from the DLP that is available for the operator at all duces or eliminates x-ray output during the systolic phases
Figure 3.18. Technical error, syn-

chronization artifact in helical CT:
Wrongly identified R-waves, for exam-
ple, as a result of a bad ECG signal,
result in synchronization errors during
the reconstruction leading to abrupt
deformations that can be recognized
easily in a 3D reconstruction.
LWBK1209-ch03_p39-56.indd 53 17/05/13 4:51 PM

ypical Effective Doses (mSv) for Chest Examinations of

T
Table 3.2 Patients (mSv per Examination), Regulatory Dose Limits
(mSv per Year) and Natural Background (mSv per Year)
Cardiac MSCT Angiographya

Helical retrospective ECG-gated reconstruction 15 (10–20) mSv
Helical with ECG-triggered tube current modulation 12 (6–12) mSv
Axial prospective ECG-gated acquisition 4 (2–6) mSv
Other x-ray Examinations
Chest radiographs (PA and LAT) 0.2 (0.1–0.3) mSv
Selective coronary angiography (CAG) 5 (2–10) mSv
MSCT coronary calcification scoreb 2 (1–3) mSv
MSCT angiography, pulmonary veinsc 1–2 mSv
PET
Heart 18F-fluorodeoxyglucose (185 MBq) 3.5 mSv
SPECT
Heart 99mTc-sestamibi (400 MBq, at rest) 4 mSv
Heart 99mTc-sestamibi (400 MBq, at stress) 3 mSv
Regulatory Dose Limits (46)
General public 1 mSv/yr
Radiation worker 20 mSv/yr
Natural Background (57)
The Netherlands (average) 2 mSv/yr
The World (average) 2.5 mSv/yr
Kerala and Madras (India) (average)d 15 mSv/yr
Certain locations in Kerala (India)d (58) 70 mSv/yr
a
Cardiac MSCT angiography, contrast enhanced study of the heart enabling assessment and diagnosis of
the myocardium, coronary arteries, bypass grafts, and cardiac function. Retrospective-gated, segmented
180-degree sinogram space reconstruction.
b
MSCT calcium scoring, native study of the heart, enabling assessment of the calcium load of the coronary
arteries for risk stratification. Prospective triggering of the acquisition, 180-degree nonsegmented sinogram
space reconstruction.
c
Whole chest MSCT, contrast enhanced study of the entire chest, enabling assessment and diagnosis of the
lungs (tumors), aorta and pulmonary arteries (pulmonary embolism). Nontriggered acquisition, nongated
reconstruction, 360-degree nonsegmented sinogram space reconstruction.
d
There is no evidence of increased cancers or other health problems arising from these high natural radiation
levels.
that are expected to be of less interest for the evaluation of Future Perspectives
the coronary arteries. The clinical implementation of pro-
spective ECG-triggered acquisitions had the biggest impact Cardiac CT has developed into a reliable, operator indepen-
on reduction of patient dose in cardiac CT. The acquisition dent, and minimally invasive clinical technique. Further devel-
window in retrospective ECG-gated scans covers the entire opments in the field of MSCT technology and reconstruction
RR interval; whereas in prospective ECG-triggered scans, algorithms are expected to contribute to a further improvement
the acquisition window is shorter, and the associated radia- of the quality of the cardiovascular MSCT scans. Compared to
tion exposure is therefore lower. Patient effective dose from other 3D imaging modalities, MSCT offers excellent spatial
coronary CT angiography may remain well below 5 mSv resolution but still at a rather long reconstruction window,
with the prospective ECG-triggered acquisitions. typically exceeding 100 milliseconds (Table 3.3). Best in-plane
omparison of Techniques for Coronary Angiography and

C
Table 3.3
Imaging of the Myocardium
Technique In-plane Thickness Temporal Window

2
MSCT, 3D volume acquisition 0.4 mm 0.5 mm 50–200 ms
EBT, 2D tomographic acquisition 0.8 mm2 3 mm 100 ms
MRI, 3D volume acquisition 1.2 mm2 3 mm 30–50 ms
Selective CAG, 2D projection 0.15 mm2 n.a. 10–20 ms
LWBK1209-ch03_p39-56.indd 54 17/05/13 4:51 PM

A B
Figure 3.19. Iterative algorithms improve image quality of CT scans. (A) is a filtered back projection and
(B) is reconstructed with iterative noise reduction algorithms working both in raw data and image domain. The
coronal views show that iterative algorithms effectively reduce noise and streak artifacts.
resolution and temporal resolution can still be achieved with permits identification of coronary plaques and that CT den-
2D invasive coronary angiography. sity values measured within plaques reflect plaque composi-
With regard to CT scanner engineering, a further reduction (55). However, it should be noted that the performance
tion of the rotation time is expected to contribute to im- of cardiovascular MSCT plaque imaging is seriously limited
provement of temporal resolution. Detector technology may by the current technical performance of MSCT, especially
improve further, perhaps toward the technique of photon with regard to its spatial resolution and even more its tem-
counting for which the characteristics are being explored in poral resolution resulting in considerable blurring of small
preclinical experiments (51). plaques.
Improvement of the analytical CT-filtered back projection
reconstruction algorithms, and even alternatives for filtered
back projection might provide specific benefits in cardiac
References
applications, for example, the incorporation of iterative al-
gorithms in the filtered back projection, and (statistical) it- 1. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary artery stenoses by
contrast-enhanced, retrospectively electrocardiographically-gated, multislice spiral com-
erative reconstructions (52). Iterative processes may be used puted tomography. Circulation. 2001;103:2535–2538.
not only for preprocessing raw data prior to a filtered back 2. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography with multi-slice com-
projection, but may also be used for postprocessing of im- puted tomography. Lancet. 2001;357:599–603.
3. Koyama Y, Mochizuki T, Higaki J. Computed tomography assessment of myocardial per-
ages that are derived from a filtered back projection. Model- fusion, viability, and function. J Magn Reson Imaging. 2004;19:800–815.
based, fully iterative reconstructions offer an alternative for 4. de Feyter PJ, Nieman K. Noninvasive multi-slice computed tomography coronary angiog-
the filtered back projection; they may provide advantages raphy: An emerging clinical modality. J Am Coll Cardiol. 2004;44:1238–1240.
5. Adams DF, Hessel SJ, Judy PF, et al. Computed tomography of the normal and infarcted
in terms of flexibility, resolution, contrast, and image noise. myocardium. Am J Roentgenol. 1976;126:786–791.
Potential benefits of using iterative methods include the re- 6. Higgins CB, Siemers PT, Schmidt W, et al. Evaluation of myocardial ischemic damage of
various ages by computerized transmission tomography. Time-dependent effects of contrast
moval of streak artifacts and blooming, and better noise material. Circulation. 1979;60:284–291.
properties in low-dose studies (Fig. 3.19) (53,54). 7. Ter Pogossian MM, Weiss ES, Coleman RE, et al. Computed tomography of the heart.
The technique of rotational coronary angiography with Am J Roentgenol. 1976;127:79–90.
8. Harell GS, Guthaner DF, Breiman RS, et al. Stop-action cardiac computed tomography.
C-arm–mounted flat detectors has potential to provide addi- Radiology. 1977;123:515–517.
tional 3D functionality within the catheterization laboratory 9. McCollough CH, Morin RL. The technical design and performance of ultrafast computed
during x-ray–guided interventions (10). However, it is not tomography. Radiol Clin North Am. 1994;32:521–536.
10. Raman SV, Morford R, Neff M, et al. Rotational X-ray coronary angiography. Catheter
expected that rotational coronary angiography will provide Cardiovasc Interv. 2004;63:201–207.
the same diagnostic quality of coronary MSCT angiography 11. Lauritsch G, Boese J, Wigstrom L, et al. Towards cardiac C-arm computed tomography.
IEEE Trans Med Imaging. 2006;25:922–934.
due to the limitation of the flat detector with regard to its 12. Kalender WA, Vock P, Polacin A, et al. Spiral-CT: A new technique for volumetric scans.
dynamic range and response time. I. Basic principles and methodology. Rontgenpraxis. 1990;43:323–330.
Coronary plaque imaging may be important to improve 13. Klingenbeck-Regn K, Schaller S, Flohr T, et al. Subsecond multi-slice computed tomogra-
phy: Basics and applications. Eur J Radiol. 1999;31:110–124.
risk stratification and to monitor progression of coronary 14. McCollough CH, Primak AN, Saba O, et al. Dose performance of a 64-channel dual-
atherosclerosis. It is reported that contrast-enhanced MSCT source CT scanner. Radiology. 2007;243:775–784.
LWBK1209-ch03_p39-56.indd 55 17/05/13 4:52 PM

15. Hofman MB, Wickline SA, Lorenz CH. Quantification of in-plane motion of the coronary 36. Jongbloed MR, Dirksen MS, Bax JJ, et al. Atrial fibrillation: Multi-detector row CT of
arteries during the cardiac cycle: Implications for acquisition window duration for MR pulmonary vein anatomy prior to radiofrequency catheter ablation—initial experience.
flow quantification. J Magn Reson Imaging. 1998;8:568–576. Radiology. 2005;234:702–709.
16. Lu B, Mao SS, Zhuang N, et al. Coronary artery motion during the cardiac cycle and 37. Maksimovic R, Cademartiri F, Scholten M, et al. Sixteen-row multislice computed tomog-
optimal ECG triggering for coronary artery imaging. Invest Radiol. 2001;36:250–256. raphy in the assessment of pulmonary veins prior to ablative treatment: Validation vs.
17. Hoffmann MH, Shi H, Manzke R, et al. Noninvasive coronary angiography with 16-detec- conventional pulmonary venography and study of reproducibility. Eur Radiol. 2004;14:
tor row CT: Effect of heart rate. Radiology. 2005;234:86–97. 369–374.
18. Mahesh M, Cody DD. Physics of cardiac imaging with multiple-row detector CT. 38. Kroft LJ, Roelofs JJ, Geleijns J. Scan time and patient dose for thoracic imaging in neo-
Radiographics. 2007;27:1495–1509. nates and small children using axial volumetric 320-detector row CT compared to helical
19. Alfidi RJ, MacIntyre WJ, Haaga JR. The effects of biological motion on CT resolution. Am 64-, 32-, and 16- detector row CT acquisitions. Pediatr Radiol. 2010;40:294–300.
J Roentgenol. 1976;127:11–15. 39. Tsai IC, Lee T, Chen MC, et al. Visualization of neonatal coronary arteries on multidetector
20. Dirksen MS, Bax JJ, de Roos A, et al. Usefulness of dynamic multislice computed tomogra- row CT: ECG-gated versus non-ECG-gated technique. Pediatr Radiol. 2007;37:818–825.
phy of left ventricular function in unstable angina pectoris and comparison with echocar- 40. Wilting JE, Timmer J. Artefacts in spiral-CT images and their relation to pitch and subject
diography. Am J Cardiol. 2002;90:1157–1160. morphology. Eur Radiol. 1999;9:316–322.
21. Dewey M, Zimmermann E, Deissenrieder F, et al. Noninvasive coronary angiography by 41. Manzke R, Grass M, Hawkes D. Artifact analysis and reconstruction improvement in heli-
320-row computed tomography with lower radiation exposure and maintained diagnostic cal cardiac cone beam CT. IEEE Trans Med Imaging. 2004;23:1150–1164.
accuracy: Comparison of results with cardiac catheterization in a head-to-head pilot inves- 42. Nieman K, Rensing BJ, Van Geuns RJ, et al. Non-invasive coronary angiography
tigation. Circulation. 2009;120:867–875. with multislice spiral computed tomography: Impact of heart rate. Heart. 2002;88:
22. Rybicki FJ, Otero HJ, Steigner ML, et al. Initial evaluation of coronary images from 470–474.
320-detector row computed tomography. Int J Cardiovasc Imaging. 2008;24:535–546. 43. Shope TB, Gagne RM, Johnson GC. A method for describing the doses delivered by trans-
23. Maffei E, Martini C, Rossi A et al. Diagnostic accuracy of second-generation dual-source mission x-ray computed tomography. Med Phys. 1981;8:488–495.
computed tomography coronary angiography with iterative reconstructions: A real-world 44. Bongartz G, Golding SJ, Jurik A et al. Quality criteria for computed tomography. EUR
experience. Radiol Med. 2011;117:725–738. 16262 2000; The European Commission’s Study Group on Development of Quality
24. van der Meer RW, Pattynama PM, van Strijen MJ, et al. Right ventricular dysfunction and Criteria for Computed Tomography.
pulmonary obstruction index at helical CT: Prediction of clinical outcome during 3-month 45. IEC. 60601-2-44 Ed 2 Amendment 1: Medical electrical equipment—Part 2–44: Particular
follow-up in patients with acute pulmonary embolism. Radiology. 2005;235:798–803. requirements for the safety of x-ray equipment for computed tomography. International
25. Schoepf UJ, Kucher N, Kipfmueller F et al. Right ventricular enlargement on chest com- standard of IEC 2003.
puted tomography: A predictor of early death in acute pulmonary embolism. Circulation. 46. ICRP. The 1990 recommendations of the International Commission on Radiological
2004;110:3276–3280. Protection. ICRP Publication 60, ISSN 0146-6453 1991;21.
26. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation 47. Shrimpton PC, Edyvean S. CT scanner dosimetry. Br J Radiol. 1998;71:1–3.
and nomenclature for tomographic imaging of the heart: A statement for healthcare profes- 48. Schmidt B, Kalender WA. A fast voxel-based Monte Carlo method for scanner- and patient-
sionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the specific dose calculations in computed tomography. Physica Medica. 2002;18:43–53.
American Heart Association. Circulation. 2002;105:539–542. 49. Geleijns J, Joemai RM, Dewey M, et al. Radiation exposure to patients in a multicenter
27. Schuijf JD, Bax JJ, Jukema JW et al. Noninvasive evaluation of the coronary arteries with coronary angiography trial (CORE 64). AJR Am J Roentgenol. 2011;196:1126–1132.
multislice computed tomography in hypertensive patients. Hypertension. 2004;45:227–232. 50. Dewey M, Schnapauff D, Laule M, et al. Multislice CT coronary angiography: Evaluation
28. Schuijf JD, Bax JJ, Jukema JW, et al. Noninvasive angiography and assessment of left ven- of an automatic vessel detection tool. Rofo. 2004;176:478–483.
tricular function using multislice computed tomography in patients with type 2 diabetes. 51. Shikhaliev PM. Photon counting spectral CT: Improved material decomposition with
Diabetes Care. 2004;27:2905–2910. K-edge-filtered x-rays. Phys Med Biol. 2012;57:1595–1615.
29. Girshman J, Wolff SD. Techniques for quantifying coronary artery calcification. Semin 52. Nelson RC, Feuerlein S, Boll DT. New iterative reconstruction techniques for cardiovascu-
Ultrasound CT MR. 2003;24:33–38. lar computed tomography: How do they work, and what are the advantages and disadvan-
30. Thompson GR, Partridge J. Coronary calcification score: The coronary-risk impact factor. tages? J Cardiovasc Comput Tomogr. 2011;5:286–292.
Lancet. 2004;363:557–559. 53. Zbijewski W, Beekman FJ. Characterization and suppression of edge and aliasing artefacts
31. Blankstein R, Shturman LD, Rogers IS, et al. Adenosine-induced stress myocardial perin iterative x-ray CT reconstruction. Phys Med Biol. 2004;49:145–157.
fusion imaging using dual-source cardiac computed tomography. J Am Coll Cardiol. 54. Renker M, Nance JW, Jr., Schoepf UJ, et al. Evaluation of heavily calcified vessels with
2009;54:1072–1084. coronary CT angiography: Comparison of iterative and filtered back projection image
32. George RT, Arbab-Zadeh A, Miller JM, et al. Adenosine stress 64- and 256-row detector reconstruction. Radiology. 2011;260:390–399.
computed tomography angiography and perfusion imaging: A pilot study evaluating the 55. Leber AW, Knez A, Becker A, et al. Accuracy of multidetector spiral computed tomog-
transmural extent of perfusion abnormalities to predict atherosclerosis causing myocardial raphy in identifying and differentiating the composition of coronary atherosclerotic
ischemia. Circ Cardiovasc Imaging. 2009;2:174–182. plaques: A comparative study with intracoronary ultrasound. J Am Coll Cardiol. 2004;43:
33. Gerber BL, Belge B, Legros GJ, et al. Characterization of acute and chronic myocardial 1241–1247.
infarcts by multidetector computed tomography: Comparison with contrast-enhanced 56. Dodge JT, Jr., Brown BG, Bolson EL, et al. Lumen diameter of normal human coronary
magnetic resonance. Circulation. 2006;113:823–833. arteries. Influence of age, sex, anatomic variation, and left ventricular hypertrophy or dila-
34. Pappone C, Rosanio S, Oreto G, et al. Circumferential radiofrequency ablation of pul- tion. Circulation. 1992;86:232–246.
monary vein ostia: A new anatomic approach for curing atrial fibrillation. Circulation. 57. United Nations Scientific Committee on the effects of atomic radiation. Sources and effects
2000;102:2619–2628. of ionizing radiation. Volume 1: Sources. 2000.
35. Ghaye B, Szapiro D, Dacher JN, et al. Percutaneous ablation for atrial fibrillation: The role 58. Nair MK, Nambi KS, Amma NS, et al. Population study in the high natural background
of cross-sectional imaging. Radiographics. 2003;23 Spec No:S19–S33:discussion S48–S50. radiation area in Kerala, India. Radiat Res. 1999;152:S145–S148.
LWBK1209-ch03_p39-56.indd 56 17/05/13 4:52 PM

Chapter
Frank E. Rademakers 4
Cardiac Anatomy and Physiolog y:
Imaging Aspects
■■ Functional Anatomy ■■ Valvular Function

General Anatomy Stenosis
Defining Regions and Surfaces Insufficiency
Circulation ■■ Future Perspective
Atria
Valves The technical capabilities of cardiovascular magnetic
Ventricles resonance (CVMR) and computer tomography (CT) and,
Myofiber Structure consequently, the clinical indications in which they can con-
Collagen Matrix tribute to better patient management, continue to expand.
Pericardium For some indications, they have become the gold standard
Interventricular Dependence whereas for others they are the preferred alternative when
Endocardium echo Doppler fails. However, to be able to translate these
superior imaging capabilities into improved patient care
Conduction System
requires a good understanding of cardiovascular physiology
■■ Contraction and Relaxation and pathophysiology.
Introduction The purpose of this chapter is to reintroduce some of the
Myocytes general concepts of cardiovascular physiology that can be
Cardiac Cycle useful in the interpretation of CVMR and CT still images,
Heart Sounds perfusion and flow curves, and cine loops.
From the early days of Galen (about 200 ad) and through
■■ Loading and Contractility as the times of Servetus (1511), Vesalius (1514), Harvey (1578),
Determinants of Systolic and Frank (1895), Starling (1918), and many others, the anat-
Diastolic Function omy and function of the heart have fascinated scientists all
Intrinsic Deformation over the world (Fig. 4.1). Even today, not all of the anatomic
Loading and functional aspects of the heart are fully understood. The
Velocity of Contraction fiber organization in bundles, lamina, and sheets remains a
Contractility topic of discussion, and the famous Frank–Starling law can-
not be fully explained by present knowledge. Experimental
■■ Evaluation of Cardiac Function techniques using molecular technology are rapidly adding
Dimensions to our understanding of some of the basic mechanisms,
Deformation but a noninvasive method like CVMR/CT—with the abil-
Flow ity to provide a comprehensive evaluation of morphology,
■■ Peripheral Circulation regional and global function, flow, and perfusion—is needed
to shed new light on the function of the human heart in vivo.
■■ Coronary Blood Flow and Myocardial Cardiovascular disease remains the most important killer
Perfusion in the Western world, so a technique that provides compre-
■■ Ventriculoarterial Coupling hensive information noninvasively and repetitively in the
aging population is valuable.
57
LWBK1209-ch04_p57-72.indd 57 16/05/13 9:43 PM

Figure 4.1. Andreas Vesalius (1514–1564)

Tabulae Anatomicae: schematic of the circulation.
Because of the need for a combination of advanced inferiorly toward the left hip. It has four chambers, two infe-
technical skills and a good knowledge of anatomy, physi- riorly located ventricles and two superiorly located atria. The
ology, pathophysiology, and patient management, a close heart is divided longitudinally by the fibrous interatrial and
cooperation between all disciplines involved will be required the muscular interventricular septum. The right ventricle is
for the successful application of these technologies. the most anterior part of the heart, and the pulmonary trunk
is the most anterior vessel. The left ventricle (LV) defines the
inferoposterior part of the heart and the apex.
Functional Anatomy
Several grooves (usually filled with epicardial fat) are
visible on the surface of the heart, delineating the different
General Anatomy
cavities and carrying the blood vessels supplying the myo-
The heart, as a muscle and pump, consists of two serial sys- cardium. The atrioventricular groove, or coronary sulcus,
tems joined in one organ enveloped by the pericardium. It encircles the junction between the atria and ventricles and
weighs between 250 and 350 g and is located in the medias- contains the right coronary artery, the circumflex coronary
tinum. It extends obliquely from the second rib to the fifth artery, and the coronary sinus. The anterior interventricular
intercostal space and spans about 13 cm. groove lies at the anterior junction of right and left ventricles
The heart rests on the superior surface of the diaphragm and contains the anterior descending coronary artery and
and is flanked and partially covered by the lungs. Its broad the great cardiac vein. The posterior interventricular sulcus
base is directed toward the right shoulder, and the apex points marks the same junction on the inferoposterior surface
LWBK1209-ch04_p57-72.indd 58 16/05/13 9:43 PM

Chapter 4 ■ Cardiac Anatomy and Physiolog y: Imaging Aspects 59
of the heart and contains the posterior descending artery, closed, the atria have a reservoir function and, during
which can originate from the circumflex or right coronary fast-filling and diastasis, they function as conduits. Only during
artery, as well as the middle cardiac vein. The small cardiac atrial contraction do they have an active, boosting function,
vein runs along the right inferior margin of the heart and optimizing ventricular filling. Because no valves are present
empties in the coronary sinus, just before the latter opens at the entrance of the veins into the atria, some retrograde
into the right atrium (RA). blood flow occurs at the time of atrial contraction.
Only the inflow regions of the atria are smooth; the au-
ricula and the anterior wall are muscular and have muscle
Defining Regions and Surfaces
bundles called the pectinate muscles. Several other structures
Definitions of the several surfaces and regions of the heart can be recognized, mainly in the right atrium (RA). The ori-
differ depending on the imaging technique, so that a great gin of the inferior vena cava (IVC) is delineated by the eu-
deal of confusion can arise when results from echocardio- stachian valve, which is intended to direct the blood from
graphic, nuclear, angiographic, CT, or CVMR studies are the IVC toward the interatrial septum, where it crosses the
compared. Part of the confusion derives from the reference foramen ovale in the prenatal circulation. After birth (and
point used and part from the changing position of the heart the subsequent normal closure of the foramen ovale, which
in the thoracic cavity that is associated mainly with aging, remains as a depression on the interatrial septum), this flow
but also with disease. At a younger age, the heart is more deviation persists, often with some acceleration and turbu-
vertically positioned, with a smaller “diaphragmatic” part lence. Higher in the atrium, but continuous with the eusta-
and a larger “posterior” part. chian valve, the Chiari network spans the atrium with small
With aging and in pulmonary disease, the heart assumes chords that can be mistaken for thrombi or vegetations. On
a more horizontal position, and the apex is more laterally the posterior surface of the atrium, the crista terminalis de-
located; as a result, the contact area with the diaphragm is lineates the smooth inlet portion of the right atrium from
larger, and the “posterior” face of the heart is more “in- the highly trabecular auriculum and can sometimes be very
ferior.” Clearly, the confusion in nomenclature stems from prominent and mistaken for an abnormal structure.
definitions, which are based on external references, body The left atrium (LA) receives the four pulmonary veins
anatomy, or cardiac anatomy. A good example is the com- and has an appendage with a smaller orifice than the one in
parison of echo with angiography. Echo adapts the acoustic the RA. The LA appendage may contain thrombi, as in the
window to obtain a standard view of the heart and from case of atrial fibrillation, and can be the origin of thrombotic
there defines the different regions; angiography uses a stan- cerebral or peripheral disease.
dard external positioning of the X-ray tubes and defines re- CVMR/CT can accurately identify the veins emptying in
gions from the projection boundaries thus obtained. the atria and abnormal pulmonary right venous return to
CVMR and CT provide an unlimited choice of image the RA or to the superior vena cava. Stenosis of the pulmo-
and/or reconstruction planes; one can, therefore, choose nary veins (after ablation for atrial fibrillation) is common
standardized 2-dimensional (2D) slices of the heart (e.g., two- and can be quantified, both morphologically and by flow
chamber, four-chamber, short axis slices) and use intrinsic measurements. Flow in the pulmonary veins, used to analyze
cardiac anatomy (insertion of right ventricle, papillary mus- diastolic function, can be easily obtained, but optimal tem-
cles) as the reference for naming the different regions. When poral resolution is required. Abnormal structures or masses
standard transversal, sagittal, or frontal views are used, the seen or suspected on echocardiography can be identified as
same problems encountered with angiography arise, and normal anatomic variants or true thrombi or tumors.
one has to be aware of the impact of differing positions of
the heart in the thorax on naming the surfaces and regions. Valves
On the other hand, because of its large field of view,
CVMR/CT is extremely well suited to depict the relationship During diastole, blood flows from the atria into the ven-
of the heart to the surrounding structures and to define masses tricles through the mitral and tricuspid valves: The mitral
(i.e., tumors) extending from the surrounding area into the valve, resembling a bishop’s miter, has a larger anterior and
heart or vice versa. smaller posterior leaflet; each leaflet is a flexible, thin sheath
of connective tissue that is firmly attached to the mitral valve
annulus. This annulus has the shape of a horseshoe, is flex-
Circulation
ible, does not lie in one plane, and changes shape during the
The right side of the heart accepts desaturated blood from cardiac cycle. On the right side, the tricuspid valve has three
the body through the inferior and superior vena cava and leaflets (anterior, posterior, and septal) that differ in size and
from the heart itself through the coronary sinus and pumps shape; therefore, this valve is difficult to evaluate morpho-
it through the pulmonary circulation. The left side of the logically because it can never be captured in one plane. The
heart receives oxygenated blood from the four pulmonary mitral and tricuspid valves are suspended in the ventricle by
veins and pumps it into the aorta, from where it is distrib- the chordae tendineae. These are thin, tendinous structures,
uted through the systemic circulation to the entire body. connecting the free edges but also some neighboring parts of
the valves to the papillary muscles, which are elongated pro-
trusions of the muscular wall of the ventricles. Contraction
Atria
of the papillary muscles during systole prevents the atrio-
Blood is continuously received in the thin-walled, muscular ventricular valves from prolapsing into the atria as pressure
atria. During systole, when the atrioventricular valves are rises in the ventricles.
LWBK1209-ch04_p57-72.indd 59 16/05/13 9:43 PM

During systole, blood is expelled from the ventricles must be overcome is lower; therefore, ejection starts earlier
through the aortic and pulmonary semilunar valves into the (the pulmonary valve opens before the aortic valve) and is
aorta and pulmonary artery, respectively. Both the aortic mainly accomplished by segmental shortening (mostly in the
valve and the pulmonary valve are tricuspid, consisting of long axis) of the crescent-shaped ventricle rather than by
three pocket-like cusps that are freely suspended in the valve wall thickening. Because the resistance to ejection is lower,
ring. the contraction continues for a longer time, and the pulmo-
With regular CVMR, it is difficult to image the thin, nary valve closes after the aortic valve.
fast-moving valve structures, but CT and newer MR tech- CVMR/CT can image both ventricles in detail, includ-
niques, also involving valve tracking, provide an accurate ing the marked intraventricular trabeculation. Because
visualization of valve leaflets and openings and permit the the entire ventricle can be imaged or reconstructed three-
direct quantification of a stenotic valve area. However, dimensionally from a stack of short-axis slices, ventricular
structures with fast, irregular, or erratic movements, such mass and volume can be derived from the epicardial and
as endocarditis lesions, can be missed by the average endocardial contours. For mass calculations, the trabecula-
CVMR/CT images. tion and papillary muscles are included in the myocardial
volume; for functional measurements (wall thickening), they
are excluded. Volume and mass measurements are accurate
Ventricles
and reproducible and can be used for the follow-up of in-
The major difference between the left and right sides of the dividual patients. When both end-diastolic and end-systolic
heart is the difference in load level: The pulmonary circula- images are acquired, wall thickening, stroke volume, cardiac
tion is a short, low-pressure system that normally works at a output, and ejection fraction can be calculated. Some dif-
peak of 25 mm Hg whereas the left side of the heart operates ficulties arise at the ventricular apex and base as a result of
at a level of up to 125 mm Hg and drives a longer circuit. the tapering of the wall and the erroneous inclusion of some
This difference has major structural consequences; the wall atrial volume secondary to long-axis shortening during ejec-
of the right side of the heart is thinner than that of the left tion; this can be avoided by combining short- with long-axis
side (2 to 3 mm vs. 9 mm). The purpose of this difference is images and full 3-dimensional reconstruction.
to normalize wall tension and the tension on the myocardial
fibers in the wall (Fig. 4.2).
Myofiber Structure
Fiber tension is difficult to measure or calculate, but it is
roughly directly proportional to the pressure in the cavity As alluded to earlier, the myofiber structure in the wall of the
and the diameter of the cavity and inversely proportional ventricles is intricate. Although some layers can be recog-
to the thickness of the wall. Other factors, such as the shape nized in the right ventricle, they are more prominent in the
of the ventricle and the curvature, also play a role: The more left ventricle. The mid layer is mostly circumferential (cir-
curved, the lower the tension; the flatter the surface, the cular in short-axis views) whereas the more epicardial and
higher the tension. endocardial layers are obliquely oriented, but in an opposite
A major consequence of the structural difference between sense: When the ventricle is viewed with the base at the top,
the ventricles is the way in which they eject blood. The left the epicardial fibers run from base left to apex right, the
ventricle must overcome a larger afterload and the pressure endocardial fibers in just the opposite direction and, at the
in the cavity must increase to aortic pressure before ejec- cavity surface, several bundles run completely along the long
tion can start; the former is accomplished mainly by the axis of the ventricle (Fig. 4.3).
contraction of fibers running circumferentially (in the mid The consequence of this changing fiber orientation is that
portion of the wall) and the latter by contraction of more fibers at the epicardium and endocardium are at nearly right
obliquely oriented fibers, which leads to thickening, longitu- angles to each other and will show an interaction or teth-
dinal shortening and torsion, and inward motion of the en- ering; also, when the oblique fibers contract, the ventricle
docardium and ejection. On the right side, the pressure that exhibits a twisting motion or torsion which contributes to
the efficiency of ejection. A physical effect of LV torsion is
elevation and motion of the apex toward the chest wall,
which can be felt as the apical impulse or apex beat; finally,
long-axis shortening during systole, produced by oblique
and longitudinal fibers, causes the base to move toward the
apex, which is nearly stationary because of the fixation of
the pericardium to the diaphragm (this fixation is partially
lost after the pericardium has been opened, i.e., after cardiac
surgery). As a result, when a short-axis slice of the left ven-
tricle is imaged near the base, different parts of myocardium
are visualized at end-diastole versus end-systole so that er-
roneous calculations of segmental shortening and thickening
of the wall are possible.
Although MR diffusion imaging can identify local fiber
orientation, this is difficult in the in vivo situation because
of the concomitant overall motion of the heart. On normal
Figure 4.2. Transverse section through the left and right ventricles. CVMR/CT images, the ventricular wall is seen as a solid,
LWBK1209-ch04_p57-72.indd 60 16/05/13 9:43 PM

Pericardium
The pericardium is a thin, fibrous structure that surrounds the
entire heart, except at the entry and exit of the great vessels. It
consists of two layers, one adhering to the outer surface of the
heart, and another in contact with lungs and other surrounding
tissues. The former, the visceral serous layer, is an integral part
of the heart wall. The latter consists of a fibrous part and the
parietal serous layer. The fibrous part is made of tough, dense
connective tissue and anchors the heart to surrounding struc-
tures, that is, the diaphragm and the great vessels. Although the
pericardium is stretchable, it resists large, sudden increases in
cardiac volume; when the volume increase is slow, the pericar-
dium will grow to accommodate the enlarged content.
The parietal serous layer lines the internal surface of
the fibrous pericardium and is continuous with the visceral
layer as it folds over where the pericardium is attached to
the great vessels. The two serous layers are separated by a
small amount of lubricating fluid, which accommodates the
twisting and shortening movements of the heart during con-
traction and relaxation. Epicardial and pericardial fat are
present in differing amounts, depending on an individual’s
constitution. It is found mostly in the atrioventricular and
interventricular grooves, containing the epicardial vessels.
Increased amounts of epicardial fat are often mistaken for
a pericardial effusion on echocardiography, which is readily
recognized on CVMR/CT. Pericardial thickening (>4 mm)
can cause constrictive pericarditis, a clinical entity difficult
to diagnose. CVMR/CT can easily measure this pericardial
thickness; this needs to be done at various locations because
thickening and constriction can be confined to one ventricle;
on the other, a normal pericardial thickness does not exclude
constriction because a thin pericardium can also be exces-
sively stiff. Pericardial calcifications are readily seen on CT.
Figure 4.3. Dissection of myofiber layers (Torrent-Guasp). AIS, Interventricular Dependence

anterior insertion septum; LV, left ventricle; PIS, posterior insertion Because of the presence of the constraining pericardium and
septum; RV, right ventricle.
the fact that they share the interventricular and interatrial sep-
tum, the right and left sides of the heart show a great deal
of interdependence not only in pathologic circumstances
omogeneous structure. To visualize the different com-
h but also in cases of (sudden) ventricular enlargement and in
ponents of myocardial deformation which ensue from the healthy persons. The lower pressure and thinner wall of the
variable fiber orientation (i.e., circumferential, radial, and right side of the heart make it more vulnerable to compressive
longitudinal deformation as well as shearing or twisting), a forces, but the interdependence between the ventricles works
marking or tagging technique must be used. Speckle tracking both ways. When filling increases on the right side, as dur-
techniques, as developed for echocardiography, can be used ing inspiration, less filling occurs on the left side; in addition,
to automatically obtain (regional) deformation. pooling of blood in the pulmonary circulation decreases fill-
ing pressures on the left side of the heart. The reverse occurs
during expiration, further augmented by the increase in intra-
Collagen Matrix
thoracic pressure, increasing LV afterload. Interdependence
The myocardium also has a “scaffold” structure, the fibrous is increased when ventricular volumes are increased (dilated
skeleton, consisting of a network of collagen and elastin cardiomyopathies) or when the pericardium is more resistant
fibers. This fibrous skeleton supports the myofibers and to stretch (stiffer pericardium of constrictive pericarditis or
transmits force throughout the myocardium. Some regions increased intrapericardial pressure of pericardial effusion and
are thicker, that is, at the valvular rings and where the myo- tamponade). Pressures in the cavity should always be referred
cardium attaches to the great vessels. Because the fibrous to the pressure in the pericardium which, in normal circum-
skeleton is not electrically excitable, the ring plane, joining stances, is about the same as pleural pressure (0 to 2 mm Hg)
the atrioventricular valves, provides a barrier to electrical and becomes negative during inspiration. In pathologic con-
conduction, which can be passed only at the atrioventricular ditions, however, intracavitary pressure can be significantly
node. increased by elevated intrapericardial pressure or effective
LWBK1209-ch04_p57-72.indd 61 16/05/13 9:43 PM

pericardial resistance. Because this extra pressure is “felt” cava. From the sinoatrial node, the impulses spread through
inside the heart only and not in the supplying veins, there is the gap junctions between myocytes in both atria, activating
a resistance to filling; also, during some phases of the cardiac atrial muscle and initiating atrial contraction. The impulse
cycle when the pressure drops in the atria and in the (right) then reaches the AV node, in the inferior part of the inter-
ventricle, pericardial pressure can become higher than intra- atrial septum, where it slows for optimal timing and coordi-
cavitary pressure, so that collapse of the cavity ensues. nation between atrial and ventricular contraction (atrioven-
Another example of the major impact of the pericardium tricular time delay is about 150 milliseconds). From there,
on interventricular dependence is the change in regional the impulse runs through the His bundle and is conducted
motion that occurs after cardiac surgery in which the peri- to the left and right ventricles over the bundle branches and,
cardium was opened. The motion of the interventricular finally, the Purkinje fibers. Because the course of the latter is
septum can be changed for months afterward, although subendocardial, the impulse must spread through the thick,
thickening and intrinsic contractility are unhampered. This left ventricular wall; it does so preferentially along the fiber
is a consequence of the changed interaction between the left bundles, which have an inward trajectory or imbrication
and right ventricles after removal of the constraints of an angle and ultimately reach the subepicardial myocardial lay-
otherwise normal pericardium, whereby the left ventricle ers. The time between the onset of activation at the sino-
shows a more eccentric contraction, and the interventricular atrial node and the activation of the last ventricular myocyte
septum is nearly immobile in an external reference system. takes about 220 milliseconds in the normal heart. The cells
If one were to reposition the images to fix the center of the that are activated last have the shortest action potential and
left ventricle on the screen, contraction would look more are, therefore, inactivated first. The activation front, there-
normal, but the right ventricle would tend to move outward. fore, runs from endocardium to epicardium and the inac-
The abnormal motion of the pericardium with respira- tivation front from epicardium to endocardium, so that a
tion, which is a discerning sign in the differential diagnosis T wave has the same orientation as the QRS complex on the
between restrictive cardiomyopathy and constrictive peri- surface electrocardiogram (ECG). Consequently, contrac-
carditis, can be well observed and quantified with CVMR. tion starts at the endocardium and moves to the epicardium,
The newer, fast imaging techniques allow the identification but the epicardium relaxes first. In addition to this local in-
of the typical morphologic and flow (velocity encoding) homogeneity of activation and inactivation, the spread of
changes with the different phases of the respiration. The the impulse over the ventricle—reaching the septum first,
specific causes of pericardial constriction, such as pericardial traveling toward the apex and free wall, and arriving at the
thickening, accumulation of fluid, blood, tumor material, or basal parts latest—causes an inhomogeneity in contraction
thrombi in the pericardium, can be identified. and relaxation, with respect to both timing and extent. This
comes on top of the regional inhomogeneity of contraction
Endocardium because of differences in local loading, secondary to varia-
tions in wall thickness and curvature.
The endocardium covers the entire inner surface of the heart
A fairly high degree of temporal and spatial resolutions is
and valves and has a large surface area because of the exten-
required to fully appreciate this inhomogeneity, but it should
sive trabeculation in both ventricles. Only the subaortic septal
not be mistaken for abnormalities of conduction (bundle
region of the LV is completely smooth. The function of the
branch block) or contraction (caused by ischemia, hypertro-
endocardium is still under investigation, but alterations mainly
phy, or loading).
to the relaxation phase have been shown. The endocardium
Knowledge of the microscopic and macroscopic anatomy
also covers the valves and is continuous with the endothelial
of the heart is important to understand the mechanics of
lining of the blood vessels as they enter or leave the heart.
the heart. A technique like CVMR/CT can contribute to
Thickening of the endocardium, as in endomyocardial
a better characterization of cardiac anatomy by providing
fibrosis, can be well visualized.
in vivo images with a high level of spatial resolution and
nearly “pathologic” characterization. Imaging of the cavity
Conduction System
and wall provides detailed information on volumes, mass,
The conduction system of the heart consists of specialized, and global and regional deformation; it also provides quan-
spontaneously active cells and a conduction system. The titative information on shape and curvature, which enable
trigger cells are concentrated in specific areas (the sinoatrial a better evaluation of local loading conditions. This knowl-
node and the atrioventricular node), but impulses can origi- edge of loading is needed to derive intrinsic contractility or
nate all along the conduction system; the intrinsic frequency true systolic function from measurements of deformation.
of these cells, however, decreases from the sinoatrial node
to the atrioventricular (AV) node to the His bundle and fur-
ther in the subendocardial Purkinje system. In normal cir- Contraction and Relaxation
cumstances, the sinoatrial node is the driving pacemaker of
the heart at an average pace of 60 to 70 beats per minute. Introduction
This cardiac frequency is the result of a balance between
parasympathetic and orthosympathetic tone which, at rest, Like the skeletal muscle, the cardiac muscle is also striated,
is shifted toward parasympathetic (the intrinsic rate of the and the sliding of the myofilaments generates force and short-
sinoatrial node is about 100 beats per minute). ening. In contrast to the long, multinucleate fibers of skeletal
The sinoatrial node is located in the posterior wall of the muscle, the cardiac cells are short, wide, branched, and inter-
right atrium just inferior to the entrance of the superior vena connected with one or two nuclei. A loose, connective tissue
LWBK1209-ch04_p57-72.indd 62 16/05/13 9:43 PM

Actin Myosin
(thin filaments) (thick filaments)
Z line
Actin filament
ADP
Figure 4.4. Sarcomeres and M line Myosin head
their actin and myosin compo- region
nents.
or endomysium surrounding the cells connects them to one and right ventricles, can be unwrapped in a single continu-
another and to the fibrous skeleton of the heart. Whereas ous bundle, which is single-coiled for the right ventricle and
skeletal muscle fibers are structurally and functionally inde- double-coiled for the left ventricle.
pendent, cardiac cells are interlocked by intercalated disks,
containing desmosomes and gap junctions; the cells are
Myocytes
firmly attached to one another, and electrical impulses can be
easily transmitted from one cell to adjacent cells. Once stimu- Large mitochondria are abundantly present in the myocar-
lated, the heart contracts as a unit, although in sequence, as dial cells because these cells must operate aerobically nearly
the electrical impulse is spread through the conduction sys- exclusively. On the other hand, cardiac myocytes can switch
tem and along the myocytes. The entire myocardium, thus, readily from burning carbohydrates to fats and even lactic
behaves as a large functional syncytium, structured in bun- acid; lack of oxygen, not lack of nutrients, is therefore the
dles and sheets and wrapped around the ventricles in a figure- main cause of problems.
eight configuration; a bundle originating at the mitral valve The myofilaments are typically organized in sarcomeres.
ring can be followed at the epicardium as it runs obliquely They consist of thin actin filaments and thick myosin fila-
over the heart and penetrates the wall to turn on itself at the ments. The head region of the myosin filament can attach
apical dimple, and it returns as an endocardial bundle, run- to the actin and generate shortening of the sarcomere
ning again obliquely but at about 90° to the epicardial part (Fig. 4.4). Calcium is needed for this contraction.
of its trajectory, and finally inserts again at the valve ring. Calcium homeostasis is governed by sarcolemmal and sar-
Some investigators believe that the entire heart, including left coplasmic reticular transport (Fig. 4.5). Because of the long,
Na-Ca
Catecholamines exchanger Na-K pump Cardiac
1 Ca++ Na+ glycosides
Ca channel Ca pump
Ca ++
β R
Sarcolemma ATP ATP
Adenylyl
ATP cyclase cAMP Ca ++ Ca ++ 3 Na+ K+
Cardiac glycosides
cAMP-PK Ca ++
Ca pump inhibit Na-K pump which
T-tubule
results in intracellular Na+

Phosphorylates Phospholamban ATP
accumulation
SR Ca ++ +
inhibition
Troponin l activation
Binding of
Ca++ by Troponin C
Ca++-Troponin
complex
Myofilaments
Figure 4.5. Schematic of calcium homeostasis.
LWBK1209-ch04_p57-72.indd 63 16/05/13 9:43 PM

absolute refractory period, which is nearly as long as the ventricle and displacing blood in the cavity, mainly from the
contraction itself, cardiac myocytes cannot be tetanized. apical region toward the outflow tract in preparation for the
subsequent ejection. As soon as the pressure in the cavity
exceeds the one in the connecting artery, the semilunar valve
Cardiac Cycle
opens and blood is accelerated and ejected. Although the
As a result of the cyclic increase and decrease in intracellular ventricles start to relax after about 100 milliseconds (i.e.,
calcium, the myocytes and the myocardial syncytium exhibit after one-third to one-half of the ejection phase has elapsed),
a cycle of contraction and relaxation. The cardiac cycle can flow is maintained because of inertia and compliance of the
be described in terms of changes in blood volume and pres- aorta (Windkessel effect). Relaxation of the ventricle is ac-
sure and is divided into phases related to the opening and companied by a decrease of ventricular pressure to levels
closing of the valves (Fig. 4.6). below aortic pressure. Pressure in the ventricle drops further,
With the onset of ventricular contraction, pressure in- and flow reverses briefly; consequently, the semilunar valve
creases in the cavity, and the atrioventricular valves are closes and the isovolumic relaxation phase starts. Again,
forced to close. Because the impulse through the Purkinje this phase is characterized not only by a decrease in pressure
system is transmitted to the papillary muscles before the wall but also by specific mechanical events: Untwisting of the left
itself, these papillary muscles contract first and prevent the ventricle occurs for the largest part during this phase, to-
atrioventricular valves from bulging under the rising ventric- gether with some longitudinal lengthening. Consequently,
ular pressure. During the subsequent isovolumic contraction the mitral ring starts to move upward, engulfing blood of
phase, in which both the atrioventricular and ventriculo- the left atrium. The mitral valve, which was flattened to-
arterial, semilunar valves are closed, pressure in the ventricu- ward the atrium during the high-pressure phase, returns to
lar cavities rises to equal to that in the aorta and pulmonary a more pointed configuration, and blood is shifted from the
artery. Parts of the ventricles sequentially contract and inflow region toward the apex. These phenomena prepare
shorten during this phase, thereby changing the shape of the the ventricle for an efficient subsequent filling. As soon as
the ventricle, through its active relaxation and the release
of restoring forces, lowers its pressure below the level of
atrial pressure, the atrioventricular valves open, and fast, ac-
tive filling follows. The “active” character of early filling is
120 Incisura proven by the continuous drop of ventricular pressure, even
Closes while the ventricle starts to fill, which is incompatible with
Pressure (mmHg)
Aortic
Aortic pressure a passive filling phenomenon. After relaxation is completed,
80 valve filling continues, but ventricular pressure now concomi-
Opens
tantly rises. Flow is dependent on inertia during this passive
diastasis period. Upon electrical activation of the atria, they
contract and further optimize ventricular filling by adding a
Atrial
a c
pressure
final blood volume.
v y
0 x In normal resting conditions at a heart rate of 60 beats
Ventricular
Closes Opens
pressure per minute, the cardiac cycle takes 1,000 milliseconds.
130
AV Isovolumic contraction takes about 30 milliseconds; ejec-
Ventricular tion, 260 milliseconds; isovolumic relaxation, 60 millisec-
Volume (mL)
valve
volume onds; and filling, the remaining 650 milliseconds. When
the heart rate and contractility increase, mainly the filling
period is shortened; therefore, in normal circumstances, the
R diastolic or filling function is more stressed during dynamic
60
exercise than the systolic function.
CVMR/CT is especially appropriate to evaluate the
P T global and regional changes in shape and morphology of the
mV ECG ventricles during systole and early diastole. When combined
O
S with tagging or other deformation analysis tools, quantify-
1st
2nd
3rd ing these deformations and studying the underlying mecha-
Heart
sounds
nisms of ejection and filling is possible. Because long-axis
shortening is an important component of the efficient ejec-
Diastole Systole Diastole Clinical Definition tion of blood, obtaining this parameter should be a routine
Diastole Systole Diastole measurement in the evaluation of systolic function.
Wiggers Definition
Diast Atr CtrIC Ejection IR FF Diast Mechanical Definition
Diastole Systole Dia Heart Sounds
Physiologic Definition Closure of the atrioventricular and semilunar valves is audi-
Contr Relaxation
ble through the chest wall on auscultation. The first heart
0 0.25 0.55 0.85s
sound is caused by the nearly simultaneous closure of the
Figure 4.6. The cardiac cycle, divided in phases, according to atrioventricular valves. The second heart sound represents
different definitions. closure of the semilunar valves, with aortic valve closure
LWBK1209-ch04_p57-72.indd 64 16/05/13 9:43 PM

preceding pulmonary valve closure. With respiration and or references are used that are part of the cardiac structure
because of ventricular interdependence, an audible difference itself—these whole body motions are excluded. Similarly,
occurs in the separation between the aortic and pulmonary when deformation of the myocardial wall must be quan-
components of the second heart sound: During inspiration, tified, a local cardiac coordinate system must be adopted:
increased filling of the right side of the heart, prolonged Usually, a perpendicular coordinate system is used, in which
ejection, and delayed pulmonary closure result in a wider one axis points outward, perpendicular to the surface of the
splitting of the second heart sound. The third heart sound is heart (radial axis R), one is aligned to the short axis or cir-
caused by sudden deceleration of the inflowing blood during cumference (circumferential axis C), and the last one is per-
early filling, either by a very fast inflow (young, healthy indi- pendicular to the other two axes and is aligned to the long
viduals) or by a decreased compliance of the left ventricle axis of the ventricle (longitudinal axis L). At each site of the
(pathologic S3). The fourth heart sound occurs during atrial left ventricle, such a coordinate system would have other
contraction on the same basis as in cases of a decreased com- orientations when viewed from the outside; for example, it
pliance. Finally, an opening snap refers to the abbreviated would point a little downward near the apex to account for
opening of a stenotic mitral (or tricuspid) valve. the tapering of the wall in this region.
Heart sounds have been used extensively to identify the Radial deformation will then quantify everywhere true
different periods of the cardiac cycle. As an alternative, 3-dimensional (3D) thickening and can be compared from
CVMR/CT can register valve openings and closures to ob- region to region. If thickening is calculated from a simple
tain the same information, provided that the temporal reso- short axis, this would give falsely high values near the apex
lution is high enough (minimally 25 milliseconds). as a result of the oblique cut through the wall. Although
in most instances, simple calculations of deformation will
suffice for a given clinical indication, the limitations of not
Loading and Contractility using true, oblique cuts and real 3D calculations of deforma-
as Determinants of Systolic tion must be remembered. Only by using a marker system
and Diastolic Function (i.e., myocardial tagging, wall tracking) can all deforma-
tion components (normal and shear strains) be quantified in
All imaging techniques attempt to optimally characterize such a local cardiac coordinate system. One of these shear
motion and deformation of the heart. To move from motion strains is the torsion motion of the ventricle, which has an
to function or performance, several factors must be taken important function in equalizing fiber load and optimizing
into account. LV performance, transforming 18% of fiber shortening into
an ejection fraction of 70%. This can be accomplished only
with an amplification mechanism in which the different lay-
Intrinsic Deformation ers of the LV myocardium work together and influence each
Whole body motion of the myocardium (i.e., swinging or other to enhance thickening in the endocardial part of the
rotation) in the thoracic cavity must be “subtracted” to wall (endocardial thickening can exceed 70% whereas epi-
obtain intrinsic motion or deformation. In other words, cardial thickening is only 20%).
myocardial deformation must be measured in a cardiac All imaging techniques these days are capable of 3D im-
coordinate system (Fig. 4.7). When motion and deforma- aging of the heart in clinical routine. With a combination
tion are referenced to the heart itself—that is, when markers of short- and long-axis information, a reliable and accurate
evaluation of global and regional parameters is feasible. Full
3D visualization is not usually required in clinical routine, but
the parameters can be extracted, if needed. Further evolution
of automated contouring, segmentation, and reconstruction
could bring this into the clinical realm and provide further
useful information for the evaluation and follow-up of indi-
vidual patients in several cardiac conditions, that is, ischemic
heart disease, valvular, and dilated cardiomyopathies.
Loading
Lo Whichever technique is used, still only deformation is quan-
tified, and whatever the level of sophistication, deformation
is only half of the contractility equation. The other part is
load. The higher the load, the lower the deformation will
Longitudinal
C
be for a given degree of contractility. Conversely, a small
deformation (low ejection fraction or thickening) can be
R caused by decreased contractility or high loading conditions
Cir
or both. In most cases, we are interested in intrinsic con-
cu
mf tractility to make decisions on therapy and revascularization
ere
nti Radial procedures. Thus, it is important to be able to at least judge
al
and qualitatively evaluate the loading conditions, because
Figure 4.7. Definition of cardiac coordinate system. quantifying them has proved very difficult.
LWBK1209-ch04_p57-72.indd 65 16/05/13 9:43 PM

Contractile– Types of contraction ventricle, beside color M-mode propagation and myocardial
elastic
Isotnic Ismetric velocity imaging (Fig. 4.9). In healthy, young individuals, LV
model Afterloaded contraction
contraction contraction compliance is high and relaxation is fast, so that a large,
1 2 1 2 1a 2b 2a 2b
Contractile–
early filling volume and velocity and a smaller volume and
element velocity on atrial contraction are seen. With aging, filling
Elastic pressures may drop and/or relaxation may slow down, so
element AL that isovolumic relaxation is prolonged and early filling vol-
PL PL- ume and velocity are lower; in compensation, atrial filling
Preload = PL increases, which is possible because the compliance of the
Afterload = AL 2b
ventricle is still normal. With hypertrophy caused by systolic
T T T overload (hypertension, aortic valve disease), a similar pat-
2 2a
Length = L 2 1 tern develops but mainly as a result of prolonged ejection
1 1 ab
Tension = T
L L L
with a subsequent slowed relaxation. When the changes in
the myocardium lead to a decreased compliance (fibrosis),
Length tension diagrams
filling pressures increase, thereby shortening the isovolumic
Figure 4.8. Definition of preloaded and afterloaded contractions. relaxation time and increasing early filling velocity; because
resistance to filling is higher at larger volumes (at atrial con-
traction after early filling), the volume/velocity on atrial con-
Load is usually divided into preload and afterload, although traction drops, and the pattern resembles the normal values.
the muscle feels only one load at each instant (Fig. 4.8). Load This is called “pseudo-normalization.” In the early stages
has several components, dependent on the size and shape of the of disease, this pattern can be reversed to a “slowed relax-
ventricle. Although muscle load is what we are really interested ation, aging” pattern by decreasing the filling pressure, such
in, tension at the myofiber level is difficult to obtain. We, there- as during a Valsalva maneuver. In the final stage of restric-
fore, try to infer tension in the wall from a simplification of tive disease, such an intervention will not change the filling
Laplace’s law: Tension in the wall increases with a larger cavity pattern any more, and this is designated as “irreversible,”
size and a thinner wall for a given pressure. and treatment will have less effect.
Load during ejection is mainly dependent on the level Such a restrictive syndrome can be the final stage of vari-
of blood pressure. Aortic stenotic disease is another obvious cardiac abnormalities, that is, ischemic heart disease,
ous cause of increased systolic load. Another is dilatation dilated cardiomyopathy, and hypertrophic heart disease (hy-
without adequate compensatory hypertrophy or wall thin- pertension). It also exists in primary restrictive cardiomy-
ning, as in nontransmural myocardial infarction. Changes opathies that often involve some infiltrative abnormalities
in the shape of the ventricle can also increase wall tension: (amyloidosis, metabolic abnormalities). The same pattern,
A rounder ventricle, in comparison with a more ellipsoidal finally, can be found in constrictive pericarditis, where the
normal ventricle, will have a higher load at the same pres- stiff pericardium does not allow any filling during the last
sure. In physiologic circumstances (intermittent volume or phase of diastole (atrial contraction).
pressure load), an increase of the cavity size (endurance ath- Differentiating constrictive pericarditis from restrictive
letes) is compensated for by an adequate hypertrophy, which cardiomyopathy can be difficult; CVMR can help by dem-
consists of myofiber hypertrophy. In contrast to this patho- onstrating a thickened pericardium and the abnormal, early
logic enlargement exists, where hypertrophy is inadequate diastolic motion of the interventricular septum in constric-
and/or consists of both myofiber hypertrophy (to a smaller tion. The influence of respiration on the filling parameters
extent) and hypertrophy of the collagen matrix with fibrosis. (an important criterion in echo Doppler), can be studied
Fewer muscle fibers must carry an increased load that leads using real-time velocity measurements.
to a negative, vicious circle with further dilatation and, ulti-
mately, cardiac failure.
Velocity of Contraction
CVMR/CT is uniquely appropriate to measure the dif-
ferent components of loading: Cavity size, wall thickness, A third component of ventricular performance, beside defor-
shape, and wall curvature. Although no absolute wall ten- mation and loading, is the velocity of deformation. This
sion can be calculated, this approach, thus far, provides the parameter is well known from experimental studies in papil-
best approximation that can be obtained. lary muscle preparations, but it has been the development of
During filling, the load is determined by the pressure dif- myocardial velocity imaging with echo Doppler that has made
ference between atrium and ventricle, but resistance to flow it possible to study velocity of deformation or strain rate in
by the mitral valve (including the subvalvular apparatus, mi- vivo. Although strain rate remains a load-dependent param-
tral ring, and so forth) and resistance to filling by the left eter, strain rate is less load dependent than deformation itself
ventricle also determine the ultimate filling dynamics. The and, therefore, could provide a more robust way to interpret
resistance to filling by the LV depends on the rate of myocar- and follow myocardial performance in an individual patient.
dial relaxation and on the stiffness of the myocardium. The CVMR has the intrinsic ability to measure velocity in each
less compliant the myocardium, the more difficult it is to fill voxel. When this is applied to the myocardium, strain rate can
the ventricle, the higher the filling pressures, and the lower be calculated by subtraction of velocities between two parts of
the filling volume. the myocardium. The advantage of CVMR over echo Doppler
Filling dynamics can be addressed by measuring the du- is that all three components of velocity can be measured
ration of isovolumic relaxation, the mitral inflow pattern, (whereas echo Doppler presently obtains only velocity along
the pulmonary vein flow, and the dimensions of atrium and the insonating beam, while speckle tracking does so in one or
LWBK1209-ch04_p57-72.indd 66 16/05/13 9:43 PM

Stage I Stage II Stage III Stage IV

Normal Diastolic Fixed
Impaired Pseudonormal Reversible
Function Restrictive
Relaxation Restrictive
0.75<E/A>1.5 0.75<E/A<1.5 E/A>1.5 E/A>1.5

E/A ≤0.75 DT>140 ms DT<140 ms
DT>140 ms DT<140 ms
Velocity.m/s 2.0
Mitral Inflow E
A
0
Adur
Time, ms Time, ms Time, ms Time, ms Time, ms
∆E/A<0.5 ∆E/A<0.5 ∆E/A ≥0.5 ∆E/A ≥ 0.5 ∆E/A<0.5

2.0
Velocity.m/s
Mitral Inflow at
Peak Valsalva E
Maneuver A
0
S≥D S>D S<D or S<D or S<D or

ARdur<Adur ARdur<Adur ARdur>Adur+30ms ARdur>Adur+30ms ARdur>Adur+30ms
2.0
Velocity.m/s
Pulmonary
S D
Venous Flow
ARdur
0
Vp > 45 Vp < 45 Vp < 45 Vp < 45

5
Distance,cm
Vp
Color M-Mode
Propogation Velocity
0 Time, ms
Time, ms Time, ms Time, ms Time, ms
Velocity.m/s
Doppler Tissue E/Ea<10 E/Ea<10 E/Ea ≥10 E/Ea ≥10 E/Ea ≥10
0
Imaging of Mitral
Annular Motion 0.15
LV Relaxation Normal Impaired Impaired Impaired Impaired

LV Compliance Normal Normal to
Atrial Pressure Normal Normal
Figure 4.9. Parameters of diastolic function.
two dimensions, but with less accuracy). A disadvantage is the tension relation of cardiac myocytes is made possible in
limited temporal resolution; it has been shown that during pe- that, in normal conditions, the length of cardiac myocytes
riods of fast motion, a temporal resolution of over 150 frames is below their optimal length. Any increase in stretch, there-
per second is required to fully resolve the strain rate. fore, increases the strength of contraction. An example of a
physiologic factor that causes an increase in venous return
and end-diastolic volume is dynamic exercise. Conversely, a
Contractility
high heart rate or bleeding with decreased filling can cause
Contractility is governed by intrinsic and extrinsic factors. a decrease in intrinsic contractility. A second determinant of
Intrinsic regulation is described by the Frank–Starling law of intrinsic contractility is heart rate per se: Contractile force of
length-dependent activation: Within certain limits, the more normal myocardium increases with stimulation frequency,
the muscle fibers are stretched, the more tension they can the staircase phenomenon. However, a failing myocardium
develop (Fig. 4.10). exhibits a negative staircase; that is, contractility decreases
At the organ level, this can be translated as follows: with increases in heart rate.
When the ventricle is filled more, the subsequent ejection is Extrinsic contractility is any change in contractile force
more forceful, stroke volume is increased, and end-systolic independent of muscle stretch. This can be caused by an in-
volume returns to previous values. This positive length– crease in sympathetic stimulation, through nerve stimulation
LWBK1209-ch04_p57-72.indd 67 16/05/13 9:43 PM

z z z < 1.6 µ > z z z ventricular and atrial dimensions. End-diastolic volume is

the major determinant of intrinsic contractility, but on the
1.9 µ 2.2 µ 2.8 µ other hand, the Frank–Starling relationship has its limits,
120
and dilatation without adequate compensatory hypertro-
phy increases load and decreases performance. Enlargement
100
of the left ventricle is the final common pathway in many
% of maximum active tension
Skeletal
disorders; it leads, through increases in load, to decreases
80 Active in cardiac output and activation of neurohumoral systems
that drive a negative, vicious cycle. In a first phase, cardiac
60 Cardiac
output at rest can be normal, even with a decreased ejec-
tion fraction. However, an appropriate increase in cardiac
40
output cannot be provided during exercise, and through fur-
ther elimination of overloaded cardiomyocytes (apoptosis),
Passive a downward path is inevitable, unless preventive measures
20 are taken.
cardiac
Increases in atrial dimensions point to valvular abnor-
0 malities, increases in filling pressures, or both.
1.8 2.0 2.2 2.4 2.6
Sarcomere length - µ
Deformation
Figure 4.10. Intrinsic contractility depends on the stretch of the
sarcomeres. Next, motion and deformation of the ventricles, with
attention to circumferential and longitudinal shortening
and thickening, have to be addressed. This can be done by
or circulating catecholamines. Thyroxin, glucagon, digitalis, cine imaging. Transverse images can provide information
and some other drugs also increase extrinsic contractility but are often more difficult to interpret than true short-
(positive inotropy) whereas acidosis, hyperkalemia, hypo- and long-axis orientations or reconstructions. Sometimes
calcemia, and drugs like calcium antagonists and beta block- spatial modulation of magnetization (SPAMM) tagging or
ers have a negative inotropic effect. tissue tracking can be helpful to clarify whether a certain
Parallel to the changes in systolic or ejection function, region of myocardium is being pulled and moved by the
positive and negative lusitropic or filling effects can be de- surrounding tissue or is undergoing true contraction and
scribed. In most instances, inotropic and lusitropic effects go deformation.
hand-in-hand: When contraction force is increased, relax- When the entire ventricle is covered in cine mode, the
ation rate and filling velocities are also augmented. end-diastolic and end-systolic volumes, stroke volume, and
ejection fraction can be obtained (Fig. 4.11). In the future,
true 3-dimensional information will make this technique even
Evaluation of Cardiac Function
Because the function of the heart is to deliver oxygen and
nutrients to the cells and remove carbon dioxide and waste,
the performance of the circulation can be described in terms
of maintaining an appropriate cardiac output and blood
pressure while keeping filling pressures at acceptable lev- Aorta
els. Although these are referred to as systolic and diastolic
functions and are viewed as separate aspects of cardiac func-
tion, they cannot really be separated. An abnormal systolic
function causes a decreased stroke volume and, subsequently,
an increased end-diastolic volume and pressure; conversely, Left
a decreased filling volume, for whatever reason, immediately atrium
causes a diminished stroke volume. Symptoms of systolic or
forward failure (decreased cardiac output, low blood pres-
sure, fatigue, exercise intolerance) go hand-in-hand with
symptoms of diastolic or backward failure (pulmonary con-
gestion, dyspnea, peripheral edema). In a patient at a certain
moment of clinical evolution, one or the other may be more
prominent, but in essence they are inseparable. A
Dimensions Y
The regulation and parameters of systolic and diastolic func-
tion, therefore, must be dealt with as a whole. To understand Figure 4.11. Multislice coverage of the ventricle allows for accurate
and describe cardiac function, it is essential to first observe measurement of volumes. A = area;Y = slice thickness
LWBK1209-ch04_p57-72.indd 68 16/05/13 9:43 PM

more reliable, avoiding partial volume effects at the apical the resistance part: The greatest resistance to flow and the
slices and the impact of through-plane motion at the base point of reflection (returning waves) is situated in the small
(with atrial volume in the most basal slice in systole). arterial vessels or arterioles.
On the other side of the capillary bed, the veins are ca-
pacitance vessels: They contain 63% of the blood volume
Flow
in the circulation (arteries, 13%; arterioles and capillaries,
Subsequently, flow can be obtained at the inflow and out- 7%; heart, 7%; pulmonary circulation, 10%). Depending on
flow regions. These measurements will provide a control gravity, muscle exercise, and integrity of the venous valves,
value of stroke volume and information on the presence or the capacitance of the veins can profoundly change and influ-
absence of intracardiac shunts (comparing aortic with pul- ence venous return and cardiac output.
monary flow volumes), and they will give insight regarding Although the blood travels through the lumen of vessels
diastolic performance by interpretation of the mitral and at relatively low speed (in the aorta 1 m per second, to less
pulmonary vein flows. than 0.05 cm per second in the capillaries), the pulse wave
Depending on the indication for the study, specifically di- travels through the arterial wall at high speed (5 to 10 m/sec
rected information can then be acquired. per second) and is reflected in the small arterioles to return
during the same cardiac cycle, so that it is superimposed
on the forward traveling pulse. The pressure measured at
Peripheral Circulation one point in the vasculature is, therefore, composed of both
the forward and returning waves. In young healthy indi-
Blood is conducted through the systemic and pulmonary cir- viduals, the pulse wave velocity is slower, and the return-
culation via arteries, capillaries, and veins (Fig. 4.12). ing wave in the aorta is superimposed after aortic valve
The exchange of oxygen, carbon dioxide, nutrients, and closure, augmenting early diastolic pressure and increas-
waste takes place in the capillaries, and the system tries to ing the perfusion pressure of the coronaries (Fig. 4.13B).
keep pressure (25 to 30 mm Hg) in this part of the circu- With aging and, prematurely, with hypertension, the arte-
lation as stable as possible while adapting flow to the re- rial wall becomes stiffer, and the wave travels faster and is
quirements of the tissue. Flow depends on the existence of superimposed before aortic valve closure, during ejection
a pressure gradient, generated by the cyclic cardiac mus- (Fig. 4.13A). It, therefore, contributes to the ejection load,
cular pump. Pulsatile flow in the aorta must be converted, and the perfusion pressure of the coronaries is less optimal
therefore, to a continuous, laminar flow in the capillaries. (lack of early diastolic augmentation). This situation is less
The Windkessel effect in the aorta is an important adap- favorable for the myocardium: More energy must be spent
tive process, whereby potential energy from the LV ejec- to overcome the increased load, and less efficient perfusion
tion is stored in the aortic wall and converted to kinetic is available.
energy (flow) during the subsequent diastole. Through the CVMR can quantitatively measure flow and flow profiles
ever increasing total surface area of the branching arterial in the large arteries. Although this type of study has not been
tree, the flow becomes more continuous (although the size pursued very intensively, knowledge of the aortic flow profile
of each vessel after branching is smaller than the original could help to decide on specific therapy for a patient with ar-
one, the total area after branching is larger than the one terial hypertension to decrease compensatory left ventricular
before). hypertrophy; such hypertrophy is, in origin, a required com-
The decreasing size of each single vessel and the presence pensation for the increased afterload but becomes a vicious
of precapillary sphincters make this side of the circulation circle of neurohumoral stimulation and an independent risk
Endothelial cell
Capillary
(just tunica intima)
Tunica adventitia
Tunica media
Venous
Tunica intima
valve
Lumen
From heart To heart
Artery Vein
Figure 4.12. Structure of arteries and veins. (thick tunica media) (thin walls, relatively thick tunica adventitia)
LWBK1209-ch04_p57-72.indd 69 16/05/13 9:43 PM

A of interventions (lifestyle, medication) on the early ath-

erosclerotic process, a repetitive, noninvasive technique is
needed; CT has the potential to fill this gap. Subsequently,
M.S.P.
the atherosclerotic lesion continues to grow and encroaches
on the lumen, decreasing the diameter. Depending on the
thickness and stability of the cap covering the lesion, it can
rupture or abrade, causing a local thrombus and arterial
P.P. occlusion with a myocardial infarction as the consequence.
M.P
Exercise or stress-induced complaints of angina, on the
dP other hand, are caused most often by progressive narrowing
dt M.D.P. of the lumen (possibly in steps by nonocclusive thrombi on
ruptured plaques). A stenosis of more than 75% is usually
needed to compromise flow.
Myocardial perfusion is phasic and occurs mainly
(80%) during diastole in the left ventricle because intramyo-
cardial pressure is increased during systolic contraction. In
normal conditions, flow and perfusion are governed by the
B
driving pressure across the coronary bed (diastolic aortic
M.S.P. pressure minus right atrial pressure) and the resistance to
flow in the small arterioles. Increases in diastolic volume
P.P. M.P and pressure (dilated cardiomyopathies) can, therefore, de-
M.D.P.
dP crease perfusion pressure in the absence of any significant
dt stenosis of the epicardial coronaries. Changes in the arte-
riolar resistance (hypertrophic disease) can have a similar
effect.
Figure 4.13. Returning wave superimposing in systole (A) and Coronary flow and perfusion are maintained constant,
diastole (B). MSP, mean systolic pressure; MDP, mean diastolic pres-
sure; MP, mean pressure; PP, pulse pressure; dP/dt, rate of pressure
irrespective of the coronary perfusion pressure for a given
increase. range of pressure (200 to 50 mm Hg). When pressure
drops below this critical value, a steep decrease in flow is
observed (coronary autoregulation) (Fig. 4.14). Conversely,
during exercise, flow can be doubled or tripled by the in-
factor for cardiovascular events. LV mass, as such, can also creased mean perfusion pressure, the increased pulse pres-
be measured accurately, so that it is possible to assess the sure, and maximal vasodilatation (see right vertical arrow in
efficacy of therapy in an individual patient. Fig. 4.14). In the presence of a significant narrowing of
an epicardial vessel, a pressure drop occurs across this le-
sion, and perfusion pressure in the poststenotic segment
Coronary Blood Flow and is decreased. At baseline, this is compensated by arterio-
Myocardial Perfusion lar vasodilatation (autoregulation), but this leaves less va-
sodilatation available during increased demand and, thus,
The coronary circulation provides the myocardium with the
critically important oxygen supply needed for the nearly
exclusively aerobic metabolism of the myocytes. Because
oxygen extraction is already maximal in basal conditions, VD reserve
increased blood flow is the only way to meet increased
demands in the coronary circulation. Although stenoses in
Coronary flow (steady state) (ml/min)
be
the epicardial and large intramyocardial conductance ves-

tu
gid
sels are obvious causes for perfusion abnormalities, deficien-

Ri
cies in the smaller resistance arterioles and the capillaries

(microcirculation) can also have a major impact on oxygen Autoregulation
delivery to the myocardium. Maximum
Coronary atherosclerosis is a very slow process that dilatation
starts early in life, probably soon after childhood in the Failure of
Western world. In the first phase, fatty streaks are formed compensatory
that develop into fatty lesions, expanding outward, increas- constriction
ing the total vessel diameter without impinging on lumi-
nal size. Such vessel remodeling is, therefore, invisible for
“projection” techniques (i.e., coronary angiography, which
visualizes only the coronary lumen). Intravascular ultra- 0 50 100 150 200 250
sound, OCT, or CT/CVMR is required to image the wall Critical closing Coronary pressure (mmHg)
and observe the major changes that occur there early in pressure
the development of the disease. To investigate the effects Figure 4.14. Autoregulation of coronary blood flow.
LWBK1209-ch04_p57-72.indd 70 16/05/13 9:43 PM

c onstitutes a decrease in coronary vasodilatory reserve and Ventriculoarterial Coupling

can cause ischemia during exercise (see left vertical arrow in
Fig. 4.14). When the lesion is severe (>85%), the segment The blood that is ejected by the left ventricle must be
falls below the auto-regulatory range, and a flow deficit en- “accepted” by the aorta. Similarly, as the resistance to flow
sues at rest or during minimal exercise. Collateral flow can in the aorta is the afterload for the ventricle, the stroke
sometimes compensate for decreased antegrade flow, but is volume is the input function for the vasculature. If a sys-
usually unable to provide adequate perfusion during higher tem is to be efficient, input and output must be optimally
demand (exercise). matched, that is, ventriculoarterial coupling must be fine-
The endocardial layers of the myocardium are subjected tuned, and an equilibrium situation must be reached. If
to a greater stress, consume more oxygen, and are less well- not, performance of the system will be less than optimal,
perfused during systole, which could explain their greater and more energy will be spent to obtain the same resulting
vulnerability to ischemia. Similarly, after total coronary oc- combination of flow and pressure, that is, cardiac output
clusion, infarcts expand in a wave front from the endocar- or stroke volume and blood pressure. Although these con-
dium to the epicardium, with a greater expanse seen in the cepts are not widely used clinically, they can clarify ven-
subendocardium. tricular dysfunction in some cases and also help to direct
CVMR/CT can visualize regional perfusion using first- therapy (Fig. 4.15).
pass techniques after contrast injection and provide qualita- It is easiest to consider equilibrium as the point where
tive and semiquantitative analyses. The high spatial resolu- two performance lines (of ventricle and aorta) cross on a
tion permits study of subendocardial versus subepicardial diagram of pressure and LV stroke volume. For the ventricle,
perfusion (this is impossible with SPECT imaging for the the relation is inverse. The higher the pressure (afterload),
moment) and the ratio between both, which is a more sensi- the lower the stroke volume; for the arterial system, the
tive parameter for ischemic problems. Like most other tech- relation is direct: the more blood enters the aorta, the larger
niques (including nuclear), CVMR does not really measure the rise in pressure. Depending on ventricular performance
perfusion but rather contrast content of the tissue, and spe- and arterial compliance, the lines can be more or less steep,
cial algorithms are needed to obtain quantitative flow in mil- but equilibrium is where they cross. At that point, the trans-
liliters per gram of tissue. mission of energy is most efficient.
170 170
EES AO PES – SV
LV Pressure mmHg
LV Pressure mmHg
LV ESPVR relationship
Ea
0 150 0 150
LV volume mL LV volume mL
170
LV Pressure mmHg
LV ESPVR
AO PES – SV
Figure 4.15. Ventriculoarterial relationship
coupling. ESPVR, end-systolic pres-
sure–volume relation; Ea, arterial 0 150
LV volume mL
elastance; PES, end-systolic pressure;
EES, end-systolic elastance. 100 LV stroke volume mL 0
LWBK1209-ch04_p57-72.indd 71 16/05/13 9:43 PM

Valvular Function easures antegrade and retrograde flow over the valve,
m
from which the regurgitant fraction can be readily calcu-
The function of the valves is to direct the blood flow through lated. To find the regurgitant orifice, the maximal gradient
the circulation. They can be found in the heart and veins, velocity must also be measured, and this entails the same
and dysfunction is a major cause of symptoms and disease. difficulties as for a stenotic valve. Overall, CVMR is unde-
In the heart, the atrioventricular and the semilunar valves rused for quantification of regurgitant lesions, and even
are quite different in structure, but they both must resist a without valve tracking it performs at least as well, if not
high-pressure gradient in closed position, while allowing better, than most echo techniques. Furthermore, the effect
opening and high-volume flow in low-gradient conditions. of valvular insufficiency on chamber dimensions (atrial
Problems can be caused by an increased resistance to flow and ventricular dilatation in cases of AV valve insuf-
(stenosis) or a leakage (insufficiency). Quantification of ficiency) can be reliably and reproducibly documented,
these disorders is quite difficult, but very important when making CVMR the technique of choice for individual fol-
decisions must be made regarding prognosis, therapy, and low-up of these patients.
timing of intervention.
Stenosis
Future Perspective
An evaluation of the severity of stenosis can be based on the Although the insights into physiology and pathophysiol-
stenotic area, gradient across the valve, or resistance to flow. ogy do not develop as fast as CVMR and CT themselves,
A stenotic area can be measured directly with an imaging the data obtained with these techniques have enhanced our
technique that is capable of choosing an image plane, per- understanding in numerous areas, and this positive interac-
pendicular to the valve, and tracking valvular through-plane tion will continue in the future.
motion. CVMR/CT has such capabilities and is the most A major aspect where pathophysiology could impact on the
reliable technique to perform such direct measurements. The clinical use of CVMR/CT is the paradigm of ischemic heart
stenotic area can also be derived from gradient measurements disease. At the moment, cardiology focuses on the presence of
and the continuity equation. In echo Doppler, this technique luminal narrowing of the coronary tree, and treatment is aimed
is commonly used. The velocity at one site is measured at relief of the stenosis, often irrespective of the impact on flow
together with the area at that same site, as well as the veloc- or flow reserve. It might be more appropriate to first evaluate
ity at the stenotic site. Because the velocity integral times the the presence of impaired flow/flow reserve and to go to non-
area equals flow volume, and the flow at two sites of a con- invasive coronary imaging with CT only in cases where such
tinuous circuit must be the same, the one unknown—which an abnormality is present. This would require studies looking
is the stenotic area—can be computed. With echo Doppler, at prognosis of asymptomatic patients with intermediate coro-
the peak stenotic velocity is measured with continuous-wave nary lesions but without abnormal flow reserve (as performed
Doppler that captures the highest velocities along the trajec- for nuclear techniques). The noninvasive, non-radiation char-
tory, wherever they occur. This point, also called the vena acter of CVMR could position the technique as the optimal
contracta, is usually situated a little beyond the physically tool for evaluation of patients with ischemic heart disease. On
smallest opening because the flow lines of a stenotic lesion the other hand, the superior noninvasive visualization of the
continue to converge after passing this smallest area. With coronary anatomy with CT as well as the size and content
CVMR, velocity can be measured in-plane and through- of lesions and plaques will improve our insight into the evo-
plane, but choosing the location where the highest velocity lution of atherosclerotic disease and the possibilities and re-
can be obtained is not always easy. Using the void created quirements to interfere with this evolution either by preventive
by dephasing on cine imaging effectively detects the presence measures or by invasive or surgical treatments.
of a stenotic lesion but cannot quantify it.
Insufficiency
Suggested Reading
Berne RM, Levy MN. Cardiovascular Physiology. 8th ed. St. Louis, Missouri: Mosby; 2001.
If quantifying stenosis is difficult, this case is even more Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. 2nd ed. Boston:
true for insufficiency. Many techniques have been used in Kluwer Academic Publishers; 2001.
Boron WF, Boelpaep EL. Medical Physiology. Revised edition. Elsevier Health Sciences;
echo Doppler to quantify insufficiency, but all have sig- 2010.
nificant disadvantages. Regurgitant orifice, fraction, and Bonow R, Mann DL, Zipes DP, et al. Braunwald’s Heart Diseases. 9th ed. Philadelphia: Elsevier
volume, in addition to semiquantitative grading on color Saunders; 2011.
McManus BM, ed. Atlas of Cardiovascular Pathology. 2nd ed. Philadelphia: Springer; 2008.
mapping, are all being used. CVMR velocity mapping, Page E, Fozzard HA, Solaro RJ, eds. The heart. In: Handbook of Physiology. New York:
especially when combined with valve tracking, effectively Oxford University Press; 2002.
LWBK1209-ch04_p57-72.indd 72 16/05/13 9:43 PM

Chapter
Håkan Arheden
Freddy Ståhlberg 5
Blood Flow Measurements
■■ Background motion on the NMR signal was described, and the possible use
Historical Overview of Flow Measurements of NMR for motion detection was investigated. Among the pio-
with Magnetic Resonance Imaging neers in this field were Suryan (3), who demonstrated a signal-
enhancing inflow effect from water flowing through an NMR
■■ Flow Measurement Techniques probe; Bowman and Kudravcev (4), who proposed a method for
Modulus-Based Techniques the construction of an NMR blood flowmeter where this phe-
Phase-Sensitive Techniques nomenon was used; Singer (5), who proposed a motion-induced
■■ Validation signal enhancement effect for measurements of blood flow;
and Hahn (6), who developed a method for the detection of
Validation In Vitro
the motion of seawater using phase dispersion effects. In 1960,
Validation In Vivo Grover and Singer (7) proposed a spin-echo–based method for
■■ Congenital Heart Disease the measurement of in vivo blood velocity distributions, while
Coarctation of the Aorta NMR flowmeters for the measurement of flow in the extremities
Pulmonary Vessels using magnetic tagging of protons were developed by Battocletti
Quantification of Shunts et al. (8). Wash-out effects, creating an apparent T2 shorter than
Valvular Obstruction and Regurgitation that for stationary spins were also used early for velocity mea-
surements (9).
Conduits
After the revolutionary introduction of NMR as an imag-
Tetralogy of Fallot ing modality (10), a large number of potential methods for
■■ Acquired Heart Disease the in vivo visualization and quantification of blood flow
Cardiac Failure using magnetic resonance imaging (MRI) emerged. Flow
Valvular Heart Disease effects were theoretically, as well as experimentally exam-
Coronary Disease ined, and flow measurement methods with varying degrees
of complexity were presented, ranging from straightforward
■■ Basic Physiology methods, using standard sequences for the visualization of
■■ Summary flow-induced signal voids, to quantitative methods requir-
ing specially designed pulse sequences. Among the first
■■ Acknowledgments
to explore the field were Herfkens et al. (11), who stated
that the information in a conventional MR image was flow
Background dependent; Grant and Back (12), who showed wash-out
effects in tubes; and Crooks et al. (13), who published a signal-
Blood flow is a central physiologic parameter that can be mea- versus-velocity curve obtained with a spin-echo sequence.
sured, using magnetic resonance (MR) velocity techniques The possibility of using MRI for studies of obstructions in
with high accuracy and precision, noninvasively, without vessels was pointed out by, among others, Kaufman et al.
ionizing radiation, in any part of the body and at any angle. (14) while Singer and Crooks (15), as well as Wehrli et al.
Under normal physiologic conditions, blood flows with mini- (16), proposed quantitative methods for velocity mea-
mal resistance in conducting vessels. Disturbances to normal surements in vessels, based on wash-in/wash-out effects.
physiology, such as decreased pumping energy in heart failure, Magnetic tagging methods, often known as bolus tracking
regurgitation in valve disease, increased resistance in stenosis, or time-of-flight methods, were earlier proposed for use in
or rerouting of blood in shunting, can thus be quantified. MRI by Shimizu et al. (17). With respect to flow-induced
phase effects on motion, the attenuation in modulus images
as a result of phase spread or loss of phase coherence, within
Historical Overview of Flow Measurements
the voxel, was investigated by Waluch and Bradley (18), and
with Magnetic Resonance Imaging
the so-called even-echo rephasing effect was pointed out.
Shortly after the first essential discoveries regarding the nuclear Development along another line more directly took advan-
magnetic resonance (NMR) phenomenon (1,2), the effect of tage of the phase-altering properties of motion. Here, the
73
LWBK1209-ch05_p73-92.indd 73 17/05/13 4:55 PM

theoretic framework was made by Moran (19), who probe taken regarding confounding phase effects introduced by,
posed a method for the creation of velocity images by the for example, nonlinear gradients, eddy currents, and con-
addition of flow-sensitive or flow-encoding gradients to a comitant gradients.
standard pulse sequence. Several quantification methods On the basis of the type of image representation that is
using this basic idea were rapidly suggested (20,21) and led used, MRI flow measurement techniques are often divided
to clinical application. into modulus-based and phase-sensitive techniques, respec-
As a result of the work outlined above, quantitative flow tively.
measurement sequences and evaluation tools have become
available on standard MRI scanners, providing the pos-
Modulus-Based Techniques
sibility not only to measure velocity and flow, but also to
evaluate important physiologic parameters, such as wall Slice-selective pulse sequences using two or more spatially
shear stress (WSS) and compliance. Furthermore, increased selective RF pulses prior to the sampling of the echo are sensi-
knowledge concerning flow effects in MR imaging has led to tive to the transport of spins into and out of the excited slice.
improved image quality in standard imaging because flow- If the flow has components perpendicular to the imaging
induced image artifacts have been effectively reduced by the slice, the signal will decrease owing to the outflow of spins
use of techniques such as flow refocusing, respiratory gating, during the time between RF pulses (wash-out). On the other
retrospective/prospective ECG triggering, and presaturation. hand, for sequences that require repetition, the inflow of non-
saturated spins during the TR will increase the magnetiza-
tion compared with the static, saturated spins and thus give
Flow Measurement Techniques an increase in signal (wash-in). In a conventional spin-echo
sequence, these competing mechanisms make the modulus
In general, MR signal acquisition and image reconstruc- signal behavior versus velocity biphasic and, possibly, difficult
tion mathematically requires so-called complex or vectorial to interpret, although, with a rough knowledge of the veloci-
treatment. After Fourier transformation in two dimensions ties involved, the experimental parameters could be adjusted
of the spatially encoded signal, the reconstructed volume so that one mechanism is dominant. If, however, only one RF
element (voxel) magnetization Mxy is obtained as a vector pulse is used prior to the sampling of the echo, as in basic gra-
in the transverse or (x, y) plane. Mxy can be decomposed dient-echo sequences, the signal decrease due to wash-out will
into real and imaginary parts according to Figure 5.1. MR not occur, and wash-in effects may give prominent increase
images are normally displayed in absolute (modulus) mode, of vessel signal. For repeated sequences including multiple
that is, the picture element (pixel) intensity is proportional RF pulses—such as inversion recovery, fast spin-echo, and
to the length of the vector. This vector length is dependent fast inversion recovery, as well as in multislice imaging—
on the object parameters (e.g., proton density, relaxation, a complicated relation between velocity and signal is to be
flow, diffusion) as well as on acquisition parameters (echo expected.
time [TE], repetition time [TR], inversion time [TI], and so Wash-out and wash-in effects are predominantly used for
forth) and may hence be used for flow estimation. qualitative evaluation of flow phenomena, and several clini-
Also possible is to reconstruct images where the inten- cally useful techniques have been proposed for this purpose.
sity is proportional to the real or imaginary part of the vec- In cardiac applications, “black blood imaging” based on
tor or simply to the phase angle φ between Mxy and the real combined effects of wash-out phenomena and spin dephas-
axis. The latter type of image is referred to as a phase image ing is used to visualize vessel walls in triggered multislice
and can, in principle, be reconstructed after acquisition of all spin-echo imaging, whereas the inflow effect in gradient-
types of pulse sequences. The phase angle is ideally inde- echo imaging forms the contrast basis in time-of-flight MR
pendent of relaxation parameters and is zero in the absence angiography, as well as in rapid cine imaging of the heart.
of macroscopic motion. Presence of such motion will give It has been shown that quantifying flow from transport
rise to a flow-dependent phase angle, but caution should effects reflected in modulus images is possible, if suitable
models describing the signal-versus-velocity behavior are ap-
plied (16,22–24). However, signal-versus-velocity relations
Abs/Mod obtained using modulus images are generally nonlinear and
influenced by relaxation times, which are difficult to deter-
Mxy mine accurately in vivo (25), and presently such methods are
less frequently used than the phase-sensitive techniques de-
Im scribed in more detail below. An alternative modulus-based
Φ
method uses the so-called bolus tracking (17). Here, a bolus
Re is first labeled or tagged with an RF pulse at a specific posi-
tion and then observed downstream using a second RF pulse.
Bolus tracking, which can be performed as a through-plane
as well as an in-plane flow measurement method, is rapid
Figure 5.1. Illustration of different types of information in an MR and does not use relaxation time information, although
image: The modulus or absolute value is the length of the signal vector, methodologic drawbacks are, for example, that a mini-
which creates a phase angle φ with the real axis. Components (Re) and mum velocity is required, that the obtained information is
(Im) are the projections of the modulus vector on the real and imaginary generally not two-dimensional (2D) and that the method is
axis, respectively. best suited for nonpulsating flow. In this context, it should
LWBK1209-ch05_p73-92.indd 74 17/05/13 4:55 PM

Chapter 5 ■ Blood Flow Measurements 75
be mentioned that several myocardial tagging techniques ydrogen nuclei of water is 42.6 MHz/T. Since accumulated
h
have emerged during the last years, as recently reviewed by phase (φ) is proportional to the time integral of frequency,
Ibrahim (26). Based on early work by Zerhouni et al. (27) a proton will accumulate a phase shift proportional to the
and Axel and Dougherty (28), techniques such as HARP time integral of the magnitude of the magnetic field in each
(29), DENSE (30), and SENC (31) have shown the possi- position.
bility for qualitative as well as quantitative evaluation of In the presence of a linear magnetic field gradient along
cardiac tissue motion, including determination of derived a direction x, the magnetic field at a position x can be de-
parameters such as strain. scribed as B(x) = B0 + Gx(t) × x, where B0 is the main mag-
netic field (T) and Gx(t) is the time-dependent gradient (T/m)
along the direction x (m). Hence, the additional phase shift
Phase-Sensitive Techniques introduced by the gradient can be calculated as
Background φ ∼ ∫ Gx(t) × x(t) dt φ ∼ ∫ Gx(t) × x(t) dt (1)
As mentioned earlier, reconstruction of images is possible, Consequently, application of a gradient will result in an
where the intensity is proportional to the phase angle that the additional phase shift according to the location along this
Mxy vector describes with the real axis. In MRI, the phase angle gradient. Now consider application of two consecutive gra-
is used in the spatial-encoding procedure. After the Fourier dients with same duration and magnitude but with opposite
decoding, ideally the phase information has been translated sign (Fig. 5.2).
into position information, and no phase information remains. When the first gradient is applied, all stationary protons
However, objects that move in a varying magnetic field (e.g., will accumulate a phase shift determined by their location x
a pulsed gradient field) change their precession frequency and according to Eq. 1. Immediately after the first gradient, the
therefore obtain an offset phase angle not removed by the second gradient is applied and, again according to Eq. 1, all
decoding procedure (19). The exact phase behavior for differ- the stationary protons lose their accumulated phase shift and
ent types of gradients and motion patterns can be calculated, obtain a net phase shift of zero. This procedure is known as
and a very simple linear relationship between constant veloc- refocusing and is used in MRI to prevent gradient-induced
ity (so-called first-order motion) and phase angle is predicted signal dephasing along the slice-encoding and the frequency-
by theory as well as confirmed in experiments. This relation- encoding gradient directions. However, a proton moving
ship forms the basis for most of the clinically used flow mea- along the magnetic field gradients during their execution
surement techniques in MRI. In its 2D form, the method is will experience unequal positive and negative magnetic gra-
known by many names, for example, velocity mapping, phase dients and, consequently, accumulate a net phase shift φ. In
mapping, or phase contrast MR (PC-MRI). the case of constant velocity v along x, the position of each
proton can be written x(t) = v × t, and an additional velocity-
dependent phase shift φv is obtained from Eq. 2:
Theory
φv ∼ v × ∫ Gx(t) × t dt φv ∼ v × ∫ Gx(t) × t dt (2)
The Larmor equation states that the precession frequency
f of a proton is proportional to the external magnetic field where the phase shift is proportional to the velocity along
B, where the proportionality constant for protons in the the magnetic gradient. Consequently, the phase image can
Gx(t)
t
t1 t2 t3
B B
x x
B0 B0
Figure 5.2. Application of two consecutive magnetic

field gradients with same duration and magnitude but with t1 t2 t3
opposite sign. When the first gradient is applied, all station-
Stationary spin
ary protons will accumulate a phase shift determined by their
location in the magnetic gradient field. Immediately after the
first gradient, the second magnetic gradient is applied, and all
of the stationary protons lose their accumulated phase shift
and obtain a net phase shift of zero. A proton moving along
Moving
the direction of the magnetic field gradients during their exe- spin
cution will experience unequal positive and negative magnetic
gradients and, consequently, accumulate a net phase shift φ.
LWBK1209-ch05_p73-92.indd 75 17/05/13 4:55 PM

Figure 5.3. Prospectively, ECG-triggered velocity map-

ping with interleaving of two differently velocity-encoded
sequences 1 and 2. After collection of a sufficient number
of k-space lines for each sequence, pairs of raw data images
are obtained for a number of time frames (cardiac phases)
in the cardiac cycle. After Fourier transform, the resulting
phase image pairs are subtracted to give the final phase
map. Note that if synchronization is performed on every
heart beat, k-space sampling is stopped prior to the end of
the cardiac cycle because the R–R interval can be expected
to vary in vivo, and hence end-diastolic phase maps cannot
be obtained.
be regarded as a velocity map, where the phase shift in each encoding or the phase-encoding direction is used in a similar
voxel is proportional to velocity with a proportionality con- way for velocity sensitization. Subsequent complex voxel-
stant depending on the time course of the applied gradient wise subtraction of the two phase images (34,35) will result
in the investigated direction. Theoretically, the phase shift in a net phase image where the phase shift is ideally deter-
in voxels with stationary tissue should be zero, but addi- mined only by motion, as shown in Figure 5.3.
tional phase shifts may be caused by many factors other An important prerequisite for PC-MRI of pulsatile mo-
than motion—for example, main magnetic field inhomo- tion is synchronization of the sequence execution to the
geneities, eddy currents, concomitant gradient effects, and time course of the flow pattern regulated by the cardiac
local magnetic field gradients induced by magnetic suscepti- cycle. Two strategies are commonly used for this pur-
bility variations—but also by specific acquisition strategies, pose: Prospective ECG triggering and retrospective gating
such as asymmetric echo sampling. These problems may be (36,37). In ECG-triggered PC-MRI, sampling of each phase-
partly overcome by creating an additional phase image using encoding line is triggered by the R-wave in the ECG (Fig.
a different set of gradient amplitudes which, in turn, gives a 5.3). The total number of image frames that can be ob-
different velocity sensitivity or often no velocity sensitivity at tained within an R–R interval depends upon the minimum
all but with similar nonmotion-related phase shift behavior. sequence TR and the heart rate of the patient. Typical values
Subtraction of these two phase images results in a velocity for adults are 30 to 40 frames, but the number of frames
map, ideally showing φ = 0 for stationary voxels. will, as a rule, be lower in children as a result of a higher
heart rate in children than in adults. ECG triggering involves
drawbacks related to an almost certain change of the pa-
Basic Sequences and Strategies for PC-MRI
tient’s heart rate during the examination. Hence, it is diffi-
Magnetic field gradients are present in all types of pulse cult to obtain an image exactly at end-diastole during which
sequences, and standard sequences usually exhibit some atrial contraction occurs, unless the experiment is performed
flow-dependent phase behavior unless this is specifically over two cardiac cycles. Furthermore, additional longitudi-
prevented by using motion-compensated gradient patterns. nal relaxation in the delay time previous to each trigger pulse
However, if flow is to be carefully quantified, additional gra- may lead to increased signal intensity and ghost artifacts in
dients are applied to maximize the sensitivity for a particular the first frames. These drawbacks can be overcome by using
velocity range since the velocity-to-noise ratio (VNR) in a retrospective gating, where the MR signal is acquired con-
phase map scales with the selected motion sensitivity as well tinuously, asynchronously with the ECG. By postprocessing,
as with the signal-to-noise ratio (SNR) in the corresponding the MR information is sorted and interpolated to fixed times
modulus image (32). in the cardiac cycle before reconstruction of the image frames.
Phase-sensitive flow MRI (henceforth denoted PC-MRI) Although potential disadvantages in retrospective gating in-
is in its simplest form performed in single-slice mode, using clude signal manipulation, such as filtering and interpolation
two interleaved gradient-echo sequences with different flow have been reported (38), Ley et al. (39) in a comparative study
sensitivity in one spatial direction (21,33), either through between ECG-gating methods concluded that retrospective
the imaging plane or in the imaging plane (2D PC-MRI). ECG-gated free-breathing measurements allow for the most
If flow is determined in the through-plane direction, either precise assessment of the bronchosystemic blood flow.
the slice-selective gradient can be redesigned to create suf- If the studied vessel also moves periodically in space be-
ficient velocity sensitivity or a bipolar gradient can be added cause of respiratory motion, the basic strategy described
in the slice-selective direction, although at the cost of pro- above may not be sufficient. A method to overcome mo-
longed TE. For in-plane measurements, either the frequency- tion artifacts induced by respiration in PC-MRI is to use the
LWBK1209-ch05_p73-92.indd 76 17/05/13 4:55 PM

s o-called segmented k-space technique, where the entire data

acquisition can be made within a breath-hold by the sam-
pling of several phase-encoding lines within a limited time
window during each heart cycle (40,41). Although early
studies have pointed out that the long acquisition windows
introduced by the segmentation technique may cause unac-
ceptable blurring in vessels moving rapidly with the cardiac
rhythm (42,43), the use of techniques such as view-sharing
(44) or shared velocity encoding (VENC) (45) can improve
the temporal resolution, the latter technique also enabling
real-time imaging.
Furthermore, since breath-holding puts demands on the
patient’s physical ability and has been shown to be inad-
equate for different subpopulations (46), several methods
for combined synchronization to cardiac and respiratory
motion during free breathing have been proposed. One ex-
ample is the use of so-called navigators, a technique where
a small excitation pulse is positioned, for example, over the
diaphragm for monitoring of the patient’s breathing so that Figure 5.4. Calculation of a flow-versus-time curve from time-
separated phase maps. Volume flow F is obtained using the relation F =
signal sampling is made only during end-expiration (47– 49).
v × A for each phase map, where A is the area of the region of interest
The resulting velocity-induced phase shift in the sub- (ROI) encompassing the vessel and v is the average velocity measured
tracted images can (without additional phase unwrapping) in the ROI.
only be unambiguously determined in the phase angle inter-
val (−180 degrees, +180 degrees). The velocity correspond-
ing to a phase angle of 180 degrees is known as the VENC. PC-MRI can also be extended to all three spatial di-
The VENC can be calculated for a specific velocity-sensitive rections, although, for time-resolved measurements, at
sequence pair, using Eq. 2. Using the VENC concept, veloc- the cost of prolonged acquisition time or reduced tempo-
ity is obtained from measured phase angle by: ral resolution, because in such a case at least four images
(one reference image and three velocity-sensitized images
v = VENC × φv/180°; φv = [-180°,180°]v = VENC × Φv/180°; in each geometrical direction) has to be obtained to cre-
Φv = [-180°,180°] (3) ate subtracted net phase images for each time point in the
cardiac cycle. The dimensional extension has been referred
Basic parameters that can be determined from each to as, for example, time-resolved three-dimensional (3D),
time point in a velocity map with an adequately designed four-dimensional (4D), or even seven-dimensional (7D)
through-plane experiment are: flow imaging, the latter referring to three VENC dimen-
The (average) linear velocity in each voxel (cm/s): This sions, three spatial dimensions, and time resolution as one
basic parameter is obtained with direct use of Eq. 3. Mean dimension. Initial suggestions for this extension emerged
velocity within a region of interest (ROI) is obtained by av- already in the 1990s (53,54), and thereafter proposals for
eraging linear velocities in each voxel within the ROI. methodologic development of the technique, here denoted
The cross-sectional flow area (cm2): An accurate way to 4D PC-MRI, as well as suggestions for potential clinical
obtain this parameter in stenotic vessels is to use the half- applications have been numerous (55), as also described in
value of the maximum velocity (vmax/2) as threshold for pix- the following.
els included in the orifice, giving a precision of over 90% for Several methods for reduction of the acquisition time in
realistic orifice areas (50); see Section “Selection of Region PC-MRI have been suggested. For example, strategies using
of Interest.” reduction to one spatial dimension (56–58) give possibilities
The volume flow (cm3/s): Found by multiplying the entire for either dynamic recording of flow in 1D projections or
flow area with the mean velocity within it (Fig. 5.4). full so-called Fourier VENC within a reasonable acquisition
Parameters such as net flow and stroke volume during time, and spiral readout techniques allow for real-time 2D
an average heart beat, can be subsequently derived from the PC-MRI (59). Echo-planar imaging methods for flow quan-
flow-versus-time curve in Figure 5.4, provided that the sam- tification using the PC-MRI strategy have been proposed for
pling is made over the whole cardiac cycle. 2D PC-MRI as well as for 4D PC-MRI in vessels and struc-
tures with a reasonably large area (60–62).
Technical Extensions of PC-MRI Reduction of acquisition time in PC-MRI can also be
achieved by combination with scan-time reduction strate-
Although most frequently performed with spoiled gradient- gies, such as k-t-BLAST and SENSE (63). Initial evaluations
echo sequences, the PC-MRI has also been combined with show that PC-MRI in combination with parallel imaging
the balanced steady-state free precession (SSFP) pulse reconstruction can be accurately performed at reasonable
sequence (51,52). Since the SNR in blood is higher in SSFP reduction factors (64–66), while k-t-BLAST holds promise
acquisitions compared to gradient-echo acquisitions, the to reduce scan times in SSFP PC-MRI and Fourier VENC
combination of SSFP and PC-MRI is favorable in terms of (67,68). Both techniques have been evaluated for 2D as well
reduced phase noise. as for 4D PC-MRI (69,70).
LWBK1209-ch05_p73-92.indd 77 17/05/13 4:55 PM

Furthermore, in particular for 4D PC-MRI, where the de- underestimation. For larger vessels at normal resolution and
mand for time reduction is of major concern, radial under- SNR level, selection of an ROI slightly larger than AROI will
sampling techniques such as PC HYPR (highly constrained give accurate values of F (78).
back projection) and PC VIPR (vastly undersampled imag- A different type of problem becomes important, if the
ing with projections) have recently been suggested (55,71). number of pixels per vessel diameter (Nd) is low, a situation
occurring for imaging of smaller vessels and cavities or even
for intermediate vessel sizes at low spatial resolution. In this
Potential Error Sources
case, partial-volume effects at the edges of the vessel may
By now, the PC-MRI technique, and especially in its 2D form become significant (79). The partial-volume effects originate
with VENC in the through-plane direction, has been thor- from within a voxel where stationary and flowing tissue is
oughly validated. For an overview of validation results, please mixed, and the modulus signal is a vector summation be-
see later Section “Validation.” Nevertheless, several sources tween signals from these two compartments. The net phase
of error might hamper the accuracy of a PC-MRI measure- angle will correctly relate to average voxel velocity, only
ment, and such error sources are described in the next section. if magnitude signal components reflect the concentration
relation between the two compartments. Normally, flow-
aliasing and slice misalignment. As stated pre- ing liquid has higher magnitude than stationary tissue in a
viously, the net phase shift can only be unambiguously through-plane investigation as a result of inflow effects, and
determined in the phase angle interval (−180 degrees, +180 the net phase angle may therefore overestimate velocity if
degrees) and hence, velocities resulting in phase angles out- an ROI is chosen so that all edge pixels are included. Errors
side this interval will result in wrapped or aliased phase in- may reach over 10% if Nd ≤ 5 (79). Several methods for re-
formation. To avoid aliasing, a certain a priori knowledge duction of the partial-volume effect have been suggested, for
of expected velocities in the investigated vessel is useful. If example, selection of an ROI slightly smaller than the entire
average flow (F) and vessel area (A) can be estimated, aver- visible flow area (53,77), empiric corrections based on phan-
age velocity (v) is easily calculated by v = F/A, and assum- tom measurements (80), thresholding in combination with
ing laminar flow in the vessel, the highest expected velocity seed-growing (81), magnitude-based corrections (82), cor-
is vmax = 2 × v. By setting VENC slightly higher than vmax, rections based on complex difference (CD) techniques (35),
aliasing can be avoided. Methods to correct for aliasing (so- automatic active contour models (83), and methods based
called unwrapping) have been proposed (72,73), but these on point-spread function shapes (84).
techniques are rarely included in standard scanner software. Furthermore, regardless of vessel size, great care must
In this context it should be noted that if VENC is increased be taken that AROI does not encompass areas with signal at
to avoid aliasing, the velocity sensitivity is reduced and or close to the noise level (e.g., air spaces), because in such
thereby also VNR. Accordingly, strategies for VENC adap- areas the phase value will be random unless the phase image
tion during execution of the PC-MRI sequence have been is post-processed to give zero phase. Hence, the proper
proposed (74,75). placement of the ROI may require inspection of modulus, as
Misalignment between the direction of flow and the di- well as phase images (85,86).
rection of the motion-encoding magnetic gradients may
also lead to erroneous MR velocity measurements. The background correction. As described earlier, several
phenomena can be avoided in through-plane PC-MRI by factors may contribute to the net phase information in an
careful adjustment of the imaging plane so that it is aligned MR phase image. The subtraction routine in PC-MRI out-
perpendicular to the flow direction. However, if the angle φ lined previously is intended to cancel all phase effects not
of misalignment is known, then the true velocity value vtrue related to encoding of motion, but residual effects are un-
can theoretically be calculated from the measured value vmea avoidable and may cause significant errors, for example,
using vmea = vtrue × cos φ. In the absence of partial-volume er- when flow over a whole cardiac cycle is calculated from a
rors, a misalignment of as much as 20 degrees will produce large number of time-resolved phase maps (87). Several rea-
only a 6% error, whereas a more realistic misalignment of sons for nonzero phase background can be found.
5 degrees causes an error <1% (76). Volume flow measure- First, the use of alternating magnetic gradient fields
ments are ideally unaffected by misalignment because the induces eddy currents in the conducting parts of the MR
measured cross-sectional area increases proportionally to system (88). These eddy currents generate transient mag-
the decrease in vmea; however, in practice partial-volume er- netic fields, which create phase offsets that vary linearly in
rors may result in overestimation of flow in the presence of space. Since different gradient patterns are used in the two
misalignment (77). sequences in a basic PC-MRI experiment, different phase ef-
fects can be expected, and a residual phase effect will appear
selection of region of interest. A general, and perin the subtracted image.
haps slightly overseen, problem hampering accurate volume Second, when magnetic field gradients are applied for
flow measurements, specifically in through-plane applica- image formation and VENC, concomitant magnetic fields
tions, is the selection of the position, size, and shape of the are created in accordance with Maxwell’s equations. These
ROI that encompass the vessel. Ideally, average flow can be additional magnetic fields may create significant errors in ve-
calculated by the product between average velocity and ROI locity measurements (89) because, again, different gradient
area, that is, F = v × AROI, provided that the ROI area is patterns are used in the two basic sequences, giving a net
chosen equal to or larger than the cross-sectional vessel area, effect from concomitant gradients in the subtracted phase
while selection of an area smaller than AROI will lead to flow image. It should be noted that effects from concomitant
LWBK1209-ch05_p73-92.indd 78 17/05/13 4:55 PM

gradients scale quadratically with gradient amplitude and In PC-MRI of laminar flow the phase–velocity relation is,
distance from the magnet isocenter but inversely with mag- however, generally preserved. When the Reynolds number
netic field strength. The largest effects can be anticipated in (which is related to the ratio between velocity and viscosity
measurements performed at positions significantly displaced of the liquid) is increased, the flow profile changes and at
from the isocenter with a combination of strong magnetic very high Reynolds numbers, a flat velocity profile, although
field gradients and low main magnetic field strength (90). randomly fluctuating in time at every position, is obtained
Third, nonlinearity in the applied magnetic field gradi- (turbulent flow). However, even with this time-dependent
ents can be caused by the finite size of the coils in the gra- random motion, there is a well-defined average velocity pro-
dient system. In the images, gradient nonlinearities result file in the vessel, and the instantaneous velocity at a point is
in image distortion as a result of differences between the given by this average velocity plus the fluctuating velocity
actual and expected magnetic field strengths. In the case at that instant. Transition from laminar to turbulent flow
of PC-MRI, nonlinear gradients will introduce differences in straight, smooth-walled vessels is not associated with a
between expected and actual VENC and, similar to the decreased modulus signal in gradient-echo imaging (98), and
Maxwell effects, these errors become more prominent the the phase–velocity relation in PC-MRI is not affected, owing
more distant the investigated voxels are from the isocenter to that the so-called turbulent intensity (the ratio between
of the magnet. fluctuating velocity and average velocity) in a turbulent flow
Consequently, several methods for background phase field is relatively low in such cases (99).
corrections have been proposed. Removal of linear phase Another situation may occur in a constriction, where
effects can be done by modeling linear functions based on so-called separated or complex flow is at hand. It has been
the measured background in the stationary regions in the shown that if a constriction is present, turbulence intensity
image (32,91). Correction is then achieved by subtracting values may drastically increase (99). Although several other
the position-dependent modeled background phase from explanation models, such as influence of higher-order mo-
measured phases in the vessel. A simplified version of linear tion, have been proposed (100), fluctuations of velocity in
correction is to select two ROIs on opposite sides of and time as well as in space might be the most important reason
at equal distances from the vessel, and subtract the average for reduction in modulus signal and breakdown of phase-
background phase value from the vessel data (92). However, versus-velocity linearity after a constriction. Systematic in-
selection of adequate background ROIs may be difficult, es- vestigations using flow phantoms have revealed that several
pecially in the thoracic region. Other more advanced meth- factors in relation to the complex flow patterns influence
ods include combined noise reduction and semiautomatized the phase-versus-velocity linearity. Thus, for a given volume
background correction (93), and correction methods for flow, the breakdown in linearity becomes more pronounced
errors caused by nonlinear gradient fields (94,95). as the cross-sectional area of the constriction is reduced.
In summary, although technical developments such as Furthermore, for a certain constriction the phase–velocity
the use of self-shielded gradient coils and the use of built- relation is more likely to become nonlinear as the volume
in analytical corrections for concomitant gradients may reflow increased. Even though the phase–velocity relation is
duce the phase offset problem substantially, in our opinion regained as the imaging plane for MR velocity mapping is
other development lines such as the use of increased gradi- moved downstream the constriction, disturbances may ap-
ent strengths, shorter magnets and gradient coils, and larger pear several centimeters distal to the tightest constrictions.
magnet bore sizes may act in the opposite direction. The Regardless of the correct explanation model for such find-
effects of background phase should therefore certainly not ings, results from several groups show that a reduction of
be neglected in precise velocity and volume flow measure- TE, or, more formally, VENC gradient duration, will re-
ments using PC-MRI. In fact, background errors in modern duce signal loss and regain unambiguous phase information
MRI scanners appear to be highly variable in multicenter (100–103). In this context, it should be mentioned that radi-
comparisons (96). cally reduced TEs can be achieved with so-called UTE (ultra-
short TE) sequences, which have recently been proposed for
phase dispersion. In PC-MRI, each voxel may contain flow quantification in complex flow fields (104).
a spectrum of velocities as well as higher-order motion com-
ponents, such as acceleration and jerk. This inhomogene- displacement. In MR imaging of flow, time differences
ity of motion components will give rise to a distribution of between phase and frequency encoding may lead to a dis-
phase shifts within the voxel (phase dispersion). The net placement artifact that is manifested as a distortion of the
flow-induced phase shift within a voxel is obtained from a vessel lumen (105) and which can be prominent for rapid
vectorial summation of all of the elementary phase contribu- flow moving obliquely relative to these gradients. Alternative
tions, and, when this vectorial summation is made, several names for this phenomenon are misregistration and oblique
situations can occur that will affect the phase-versus-velocity flow artifact.
linearity. For example, errors in velocity estimated from av- Furthermore, when the moving spins are subjected to ac-
erage voxel phase may occur, if modulus signal differs within celeration and other higher orders of motion, the obtained
the voxel because of different degrees of inflow enhancement phase shift no longer depends solely on velocity, and Eq. 2
(97). Furthermore, phase dispersion inevitably leads to loss does not hold (106). The induced phase shift in the presence
of magnitude signal and increased uncertainty in the cor- of higher orders of motion has been identified as a poten-
responding phase information. In extreme cases phase in- tially confounding factor in velocity measurements (107),
formation cannot be unambiguously determined, and the but it can be further analyzed theoretically by so-called
linearity between velocity and measured phase is lost. Taylor expansion of the expression for the position x(t) of
LWBK1209-ch05_p73-92.indd 79 17/05/13 4:55 PM

a spin as a function of time. According to such analysis, the the signal from PC-MRI can be used to measure the standard
total phase shift in a bipolar experiment can be regarded as a deviation of the velocity distribution within a voxel, in turn
velocity measurement free from the influence of higher-order giving possibilities to measure parameters related to turbu-
motion but spatially misplaced in the reconstructed velocity lence, such as the turbulent kinetic energy (TKE) (117).
image owing to differences in time between true VENC, oc- A proposed marker for localization of areas where for-
curring at the so-called moment center time or gravity center mation of atherosclerosis will appear is the WSS (118). The
and spatial encoding (106,108,109). calculation of the WSS τ is given by τ = μ × dv/dr, where
Displacement errors can therefore be expected when the quotient dv/dr corresponds to the velocity gradient at
measuring physiologic flows in the heart and great vessels the vessel wall and μ is the viscosity, which for the blood
with the phase-contrast technique. One way of reducing the is 0.004 N s/m2. Several studies have shown the possibility
oblique flow artifact is to reduce the time difference between of calculating the WSS in vessels using MRI velocity data
phase and frequency encoding and, similarly, the displace- (119,120), recently including 4D PC-MRI measurements
ment effects of higher-order motion components can be re- also encompassing the oscillatory shear index (OSI) which
duced by reduction of the time difference between the true describes the existence and magnitude of WSS changes over
VENC time and the time for spatial encoding (110). the cardiac cycle (121).
Another parameter of clinical interest which can be mea-
calculation of derived parameters. Apart from sured with 2D as well as 4D PC-MRI is the velocity of propa-
blood flow measurements, there are many examples where gation of a pulse wave along a vessel, the pulse wave velocity
the acquired velocity data can be used for extended visu- (PWV) which relates to aortic stiffness (122). Furthermore,
alization of flow patterns (e.g., based on virtual particle the elasticity of the greater blood vessels can be assessed by
behavior) as well as for calculation of different physiologic calculation of the vascular compliance. Compliance provides
parameters derived from the velocity data. From PC-MRI an estimate on the vessel conditions because low compliance
data, visualization of flow can hence be made, for exam- is associated with a number of different cardiovascular dis-
ple, using velocity vector mapping, streamline visualization, eases whereas large values correlate to overall fitness and
pathline visualization, and volume tracking (111–115); the youth. It has been shown feasible to calculate vascular com-
three latter techniques developed for 4D PC-MRI (Fig. 5.5). pliance using time-resolved velocity data (123).
Furthermore, several quantitative parameters of physi- The contraction and relaxation of the pumping heart
ologic value can be derived as described below. imply varying strains in different regions of the myocardial
The transstenotic pressure gradient (in case of stenotic ves- muscle. Quantification of strain and strain rate in the myo-
sels), which gives information about the severity of constric- cardium has been suggested as methods to identify ischemic
tions in the vessels, can be estimated by inserting vmax (m/s) areas in the heart, and velocity data in the myocardium can
in the stenotic jet in the modified Bernoulli equation ΔP = be used to calculate strain rate (124–126). In the case of
4 = vmax2 (116). However, the simplified Bernoulli equation is velocity-mapping–based strain and strain-rate calculations,
based on several assumptions, for example, that the velocity no tracking of fictive markers has to be performed because
of blood is zero before the constriction. The measurement ac- the measured velocities provide enough information for the
curacy of this parameter can also be compromised by partial- calculation of this parameter.
volume effects and malpositioning of the imaging slice (50). In this context, it should be noted that with the evolution
In this context it should be mentioned that the magnitude of of PC-MRI, possibilities to compare measured PC-MRI data
Figure 5.5. Comparison of three methods for visualization of blood flow in the human heart
during rapid filling of the left ventricle. Blood flow was measured using 4D phase contrast mag-
netic resonance (4D PC-MR). Left: Streamlines are lines that are instantaneously tangent to the
flow. Middle: Pathlines, released near the mitral annulus during rapid filling. Pathlines show the
path of virtual particles released in the flow. Note that streamlines and pathlines differ in pulsatile
flows such as the heart and great vessels. Right: Volume Tracking (114) of the blood flowing into
the left ventricle. The blue surface separates the blood flowing from the left atrium into the left ven-
tricle from the blood that was already in the ventricle. LV, left ventricle; LA, left atrium; Ao, aorta;
RV, right ventricle. Color scale: velocity.
LWBK1209-ch05_p73-92.indd 80 17/05/13 4:55 PM

with results from computed fluid dynamic (CFD) modeling may become less accurate because of complex flow and
in the cardiovascular system (127), and also to use PC-MRI limited spatial resolution in MR. In vitro studies have inves-
data as input for improved CFD analysis on a patient-to- tigated the performance of PC-MRI to measure jet veloci-
patient basis has emerged, as shown for example by Wood ties (101) and derive pressure gradients using the modified
et al. (128) and Torii et al. (129). Bernoulli equation with promising results (50).
Validation in Vivo
Validation
Volume flow, flow velocity, and flow profiles have been exten-
In our opinion, flow velocity mapping sequences used for sively validated in both animals and humans by comparison
clinical or scientific purposes should be validated in-house to planimetric measurements of ventricular stroke volume,
before implementation. The reasons for this are twofold: oximetric measurements, transthoracic and intravascular
First, the investigator should be familiar with the error ultrasound, ultrasonic flow probes, and radionuclides.
sources described in the previous section and their effects in
the investigator’s own equipment environment and second,
Planimetric Measurements
hardware and software are continuously updated, and new
flow velocity sequences are developed and delivered auto- Forward stroke volumes derived from MR velocity maps in
matically with new software updates. the great vessels have been validated by comparison to pla-
Validation of the specific flow sequence can be performed nimetric measurements of left and right ventricular stroke
both in vitro using a flow phantom and in vivo where flow volumes, with excellent agreement. Standard errors are
conditions can be assumed to be completely normal, as in typically in the lower range of 3 to 8 mL (135,145–150).
healthy volunteers or in patients where flow velocities can Tricuspid volume flow has shown good agreement to right
be compared to Doppler ultrasound. ventricular stroke volumes measured planimetrically (151).
Lately, perhaps due to introduction of newer versions of One study demonstrated high agreement to calculate the
hardware and software, some investigators have had reasons change in total heart size throughout the cardiac cycle
to believe that flow measurements may have become less ro- between planimetric measurements of cine short-axis images
bust (130,131). A multicenter, multivendor study using stan- on the one hand and measurements of all venous inflows
dard clinical flow sequences in standard measurement planes and arterial outflows from the heart on the other (152).
demonstrated significant phase offset errors when applied to
a static gelatin phantom (132). These types of errors are in-
Doppler Velocity Measurements
dividual to each individual scanner and can be accounted
and corrected for in each scanner underscoring the need for Flow velocities from PC-MRI and Doppler ultrasound have
in-house validation of flow sequences. Vendors and the sci- been shown to be in good agreement (21,133–137). PC-MRI
entific community are well aware of these challenges, work can also be used for assessment of the severity of stenosis by
together to provide a solution, and should be able to provide using the modified Bernoulli equation. Several studies have
advice on these matters. shown a tendency for underestimation of stenotic jet veloci-
ties by PC-MRI in comparison to Doppler ultrasound. The
general findings, however, show that PC-MRI may be used
Validation in Vitro
to estimate severity of aortic valve stenosis (101,153,154),
Volume flow and velocity have been validated in vitro using aortic coarctation (101,155–157), mitral stenosis (158,159),
flow phantoms. In 1984, Bryant et al. (133) described a con- and pulmonary venous obstruction (160).
ventional spin-echo sequence with balanced gradient pulses
on either side of the pi radiofrequency pulse and showed
Flow Versus Flow
that the flow rate measured by MR agreed with the volume
flow rate through a constant flow phantom. Several studies Velocity mapping of venous return by the superior and infe-
have confirmed high accuracy and precision of MR volume rior caval veins has been compared to velocity mapping
flow quantification in vitro, using phantoms with constant of cardiac output (CO) in the aorta with good agreement
(92,134–137) or pulsatile flow (77,138–142). between the measurements (161).
Volume flow quantification in small vessels entails special A commonly used way to validate flow measurements
difficulties owing to an increased number of edge pixels over by MR that can easily be practiced in-house is by com-
total pixels because edge pixels introduce partial-volume paring pulmonary to aortic volume flow ratio (QP:QS)
errors (77,81). In vitro studies, however, where most condi- (66,92,138,139,162). These measurements should be nearly
tions can be controlled, have confirmed that it is theoreti- identical if they have been performed at the level of the
cally possible to also reach high accuracy and precision of valves. Flow measurement in the aorta, however, is usually
MR volume flow measurements in small vessels with pulsa- performed a few centimeters distal to the coronary ostia.
tile flow (143,144) even in the presence of in-plane, through- Aortic flow should therefore be about 3% to 5% less than
plane, and in- and through-plane motion after motion flow in the pulmonary trunk because of coronary runoff.
correction (Fig. 5.6) (77). This difference, however, is in the order of the error of the
Estimation of flow velocities using PC-MRI compared measurements.
to Doppler ultrasound is generally in good agreement Four-dimensional flow measurements have been vali-
(101,137,142,144). Assessment of jet velocities, however, dated against 2D flow measurements at 1.5 T (163,164).
LWBK1209-ch05_p73-92.indd 81 17/05/13 4:55 PM

Non-corrected Background corrected
40 A A corr
Mean flow velocity (cm/s)
Stationary
30
20
10
0
-10
-20
40 B In-plane B corr
30
20
10
0
-10
-20
C C corr
40 Through-plane
30
20
10
0
-10
-20
D D corr
40 In + through-plane
30
20
Figure 5.6. Pulsatile flow shown for small
(6-mm diameter) moving tubes using a phantom
10 with simulated cardiac motion. Left: Before back-
ground correction when phantom is stationary
0
(A), moving in-plane (B), through-plane (C), and
-10 in- and through-plane (D). Right: Same as left
panels but after background correction. (Reprinted
-20 from Arheden H, Saeed M, Tornqvist E, et al.
CONV Accuracy of segmented MR velocity mapping to
0 200 400 600 800 1000 1200 14001600 SEGM5 measure small vessel pulsatile flow in a phantom
SEGM7
Time (ms) simulating cardiac motion. J Magn Reson Imaging.
Phantom motion
2001;13:722–728, with permission.)
4D flow accelerated with SENSE has good accuracy at both Radionuclide Angiography
1.5 T and 3 T whereas 4D flow accelerated with k-t-BLAST
underestimates flow velocities and thereby yields too high a Comparison of QP:QS ratio from PC-MRI to radionuclide
bias for intracardiac quantitative in vivo use at the present angiography shows good agreement up to QP:QS values of
time (70). For intracardiac 4D flow visualization, however, 2.5 (92).
1.5 T and 3 T as well as SENSE or k-t-BLAST can be used
with similar quality (70). Small Vessels
Small vessels are of special interest in cardiovascular applica-
Oximetry
tions because measurement of coronary artery flow reserve
Comparison of pulmonary to aortic flow ratio (QP:QS) by PC-MRI would provide a powerful tool in assessment of
by PC-MRI and oximetry has shown good agreement coronary artery disease. Such measurements, however, intro-
(147,162,165–168). duce specific difficulties because of substantial vessel motion,
LWBK1209-ch05_p73-92.indd 82 17/05/13 4:55 PM

partial-volume errors caused by proportionally more edge over time. To minimize the effect of such variations on blood
pixels compared to big vessels. Phantom studies of pulsa- flow measurements, acquisition of blood flow measurements
tile flow in small vessels moving in- and through-plane have should preferentially not be obtained directly after the sub-
demonstrated that the technique as such is surprisingly accu- ject has been positioned in the MR scanner, but rather after
rate and robust as long as the size and position of the ROI the person has accommodated to the situation. Curiousness,
is correct (77) (Fig. 5.6). Experimental animal studies report nervousness, and wakefulness tend to decline during the
good agreement between quantitative blood flow measure- scan, thereby decreasing the sympathetic outflow from the
ments, using PC-MRI and ultrasonic flow probes placed cardiovascular center. This lowers heart rate, venous return,
directly around the investigated vessel in femoral (81,169) CO, and blood pressure. An uprising urge to urinate after a
and coronary arteries (170,171). Validation has also been prolonged scan session may act in the opposite direction by
done in humans by comparison to intracoronary ultrasound. activating the sympathetic system.
With few exceptions, most studies found that peak blood Short-term variations in CO, for example, may introduce
velocities are underestimated by PC-MRI in comparison to errors in shunt quantification (QP:QS) but can usually be
intravascular ultrasound but that flow reserve demonstrates compensated for by interleaved measurements and averag-
good agreement (144,172–176). Blood flow velocities in ing, for example, by measuring QP then QS and QP again
coronary arteries have also been compared to transthoracic and average the QP measurements (92).
Doppler ultrasound (177) and myocardial perfusion single
photon emission tomography (178) with good results.
Global left ventricular perfusion can be calculated as cor- Congenital Heart Disease
onary sinus blood flow rate per gram left ventricular mass
(179). Comparison with positron emission tomography Congenital heart disease encompasses malformations of
shows good correlation in healthy volunteers (180–182) in the heart and great vessels that affect blood flow dynamics.
patients that have undergone orthotopic heart transplan- PC-MRI can provide important hemodynamic information,
tation (180) and in patients with coronary artery disease in addition to morphologic mapping before and after cor-
(183). rective surgery and during follow-up (184–186). Assessment
of the hemodynamic situation using PC-MRI is of special
importance in patients with congenital heart disease who
Validation During Patient Examination
otherwise risk to be exposed to repeated catheterizations
(Internal Control Measurements)
and ionizing radiation.
Validation, or internal control measurements, of flow in
patients during the examination can be achieved by measur-
Coarctation of the Aorta
ing flow before and after branching of vessels, for example,
in the pulmonary trunk on the one hand and in each of the In coarctation of the aorta, forward flow is hindered because
two pulmonary arteries on the other (136). Flow measure- of inborn obstruction of the aorta at the junction of the aor-
ments in the pulmonary arteries should add up to the same tic arch and descending aorta (Fig. 5.7). Assessing the state
volume flow as in the pulmonary trunk. of this disease, besides mapping of anatomy, is best achieved
In patients with patent ductus arteriosus, the sum of the by measuring resistance in the coarctation and quantification
flow in the ductus and in the proximal pulmonary trunk of collateral flow. Measurement of the resistance is accom-
should add up to the same flow as in the two pulmonary plished by measuring maximal flow velocity in the lesion and
arteries together. If flow in the ductus cannot be measured calculating the pressure gradient using the modified Bernoulli
directly, then flow in the two pulmonary arteries should add equation (101,137,155–157). Assessment of collateral flow
up to the same flow as in the ascending aorta, and the dif- is achieved by measuring volume flow immediately after
ference between flow in the ascending aorta and in the pul- the coarctation and at the level of the diaphragm (Fig. 5.7).
monary trunk is an indirect quantification of shunt flow in Blood flow decreases by about 7% from the proximal part
the ductus. to the distal part of the descending aorta in normal, healthy
In patients with atrial septal defects, the difference of individuals (187), while it increases in significant coarctation
flow in the pulmonary trunk and the aorta should equal the (Fig. 5.7). The increase in volume flow at the level of the dia-
flow over the septal defect. Measurement of blood flow over phragm compared to the slice location immediately distal to
atrial septal defects, however, are difficult and may show the coarctation gives a quantitative measure of the degree of
considerable scatter in comparison to the difference in flow collateralization (187–189). A study of experimental coarcta-
between the pulmonary trunk and aorta (162). tion has confirmed that PC-MRI is an accurate, noninvasive
Aortic regurgitation can be accurately quantified by flow method of measuring collateral blood flow in coarctation and
velocity mapping (see later). A straightforward control mea- that collaterals develop within a few weeks (190).
surement is usually derived from planimetric calculation of
left ventricular stroke volume from short-axis cine images
Pulmonary Vessels
covering the left ventricle (see later and also Chapter 14).
PC-MRI can measure blood flow individually in the right
and left pulmonary arteries (136) as well as in the four lung
Physiologic Drift
veins (152,160). Measurements of blood flow in all of these
Physiologic parameters, such as blood pressure, heart rate, vessels, leading to and from the lungs, enable quantifica-
stroke volume, CO, and vascular resistance, normally vary tion of pulmonary blood volume variation throughout the
LWBK1209-ch05_p73-92.indd 83 17/05/13 4:55 PM

Figure 5.7. MR velocity mapping
Flow (mL/s)
in a patient with severe coarctation.
Left: Blood flow is measured in aorta
ascendens, immediately distal to the
coarctation and at the level of the
diaphragm. Note the extensive collat-
eralization from the intercostal arter-
ies. Right: Flow in aorta ascendens
show a normal flow pattern, whereas
flow distal to the coarctation is very
low and increases at the level of the
diaphragm, clearly demonstrating
the physiologic significance of the
coarctation.
c ardiac cycle, and PC-MRI of CO in combination with a by measuring the velocity of the tricuspid jet in PC-MRI as
first-pass study of Gd-contrast enables estimation of pulmo- is done routinely in echocardiography (130,201). Another
nary blood density in health and disease (191–194). approach found that the average blood velocity in the pul-
It has been shown that arterial switch repair after trans- monary trunk throughout the cardiac cycle correlates to the
position may lead to temporary stenosis of the primary pul- pulmonary artery pressure (202). It has also been suggested
monary arterial branches. The hemodynamic effects of these that the time of existence of vortices in the main pulmonary
can be evaluated by PC-MRI but go undetected with other artery, as assessed with 3D PC-MRI, can identify patients
imaging modalities (195). with manifest pulmonary artery hypertension (203).
Blood flow distribution to the right and left lungs and the Attempts have also been made to estimate pulmonary
contribution to venous flow from the superior and inferior vascular resistance by PC-MRI (204,205).
caval veins can be measured after total cavopulmonary con-
nection (TCPC) (196,197).
Quantification of Shunts
PC-MRI has demonstrated 100% specificity and 100%
sensitivity in detecting pulmonary venous obstruction in 7 Quantification of shunt size may be performed to evaluate
pediatric patients with pulmonary obstruction and 27 con- the need for surgery and to follow up the result postop-
trols, using Doppler echocardiography and catheterization eratively. Quantification of intracardiac shunt by PC-MRI
as the gold standard (160). is accomplished by measuring pulmonary to systemic flow
Several attempts have been made to estimate pulmo- (QP:QS) in the pulmonary trunk and in the aorta (Fig. 5.8).
nary artery pressure (198–200). It is, however, presently A specific feature of PC-MRI is that it, as opposed to other
not possible to reliably estimate pulmonary artery pressure modalities, also enables quantification of right-to-left shunt.
800
Aorta
600 Pulmonary artery
400
Flow (mL/s)
200 Figure 5.8. Measurement of pulmo-

nary to systemic flow (QP:QS). Left:
Healthy volunteer with a QP:QS ratio of
0 1.03, which is normal. Right: A patient
0 200 400 600 800 1000 0 400 with patent foramen ovale has increased
200 600 800 1000
volume flow in the pulmonary trunk
compared to the aorta, which demon-
–200 strates left-to-right intracardiac shunting
Time (ms) with a QP:QS ratio of 1.8.
LWBK1209-ch05_p73-92.indd 84 17/05/13 4:55 PM

Shunted volume is the difference between the pulmonary Diastolic ventricular function can be assessed by PC-MRI
and the aortic flow. of blood flow patterns over the mitral (111,212–218) and
Extracardiac shunt size may prove more difficult to mea- the tricuspid valve (151,207,219).
sure unless the complete nature of the shunting is known. In Measurements of 4D intraventricular blood flow with
the case of anomalous veins, shunt size can be measured in focus on ventricular hemodynamics (115,220–222) or ki-
the same way as intracardiac left-to-right shunt (see above) netic energy (223) are promising but elaborate methods that
or by direct measurement of volume flow in the anomalous may provide information on both systolic and diastolic func-
veins that may serve as an internal control. tion in health and disease.
Measurement of QP:QS ratio by PC-MRI has been An experimental study using PC-MRI has demonstrated
extensively validated by comparison to results from flow that the pulmonary blood volume variation decreased after
phantoms (92,139) to planimetry of the left ventricle, in experimental acute cardiac failure whereas the pulmonary
end-systole and end-diastole (147,150,165), to catheter- blood volume surprisingly remained unchanged (193).
based oximetry (147,162,165–168) to radionuclide angi-
ography (92), with convincing accuracy. Preliminary results
Cardiac Output
using parallel imaging (65) and real-time MR imaging (66)
to shorten acquisition time have been presented. CO is the volume flow delivered from the left ventricle to
The usefulness of MR in assessment of shunt size the aorta or from the right ventricle to the main pulmonary
(QP:QS) is considered “Class I = provides clinically relevant artery. CO is often corrected for by body surface area to
information and is frequently useful; may be used as first- give cardiac index (CI). CI at resting conditions varies from
line imaging technique; usually supported by substantial about 3.5 L/min/m2 for teenagers to about 2.5 L/min/m2 in
literature” (185). the eighth decade of life (224,225). CO has been measured
by PC-MRI and compared to measurements by the Fick prin-
ciple and thermodilution after cardiac catheterization, with
Valvular Obstruction and Regurgitation
high agreement (226). CO has been measured by PC-MRI in
See “Valvular Heart Disease” in next section under “Acquired a cohort of healthy persons, elite athletes, and patients with
Heart Disease.” congestive heart failure (227) (Fig. 5.9). The results show
that CO in athletes is similar to normal subjects whereas it is
decreased in patients with various degrees of cardiac failure.
Conduits
Measurements of CO as such using PC-MRI are not very
Extracardiac ventriculopulmonary conduits have a tendency common for clinical purposes. Such measurements, how-
to develop both obstruction and regurgitation over time and ever, may provide important insights into unknown effects
are better examined by echocardiography (85,101,206). of oxygen inhalation on CO under emergency conditions
(228). A peculiar application is to elucidate the diving reflex
in juvenile elephant seals (229) which demonstrates the abil-
Tetralogy of Fallot
ity of PC-MRI to elucidate mechanisms of normal physiol-
PC-MRI can provide insights into hemodynamics and dia- ogy not only in humans but also in the animal kingdom.
stolic ventricular function after repair of Tetralogy of Fallot Low CO may be caused by myocardial infarction, severe
(TOF) (148,207–209). Early diastolic dysfunction with valvular heart disease, myocarditis, cardiac tamponade, and
delayed onset of tricuspid valve flow compared to aortic certain cardiac metabolic derangements or by peripheral
flow is common after TOF repair, and indicative of reduced factors leading to decreased venous return, as in decreased
RV ejection fraction (210).
Acquired Heart Disease

Acquired heart disease encompasses derangements to nor-
Cl (L/min/m2)
mal anatomy and physiology that cause loss of pumping

efficiency as a result of myocardial ischemia and failure, or
loss of flow energy as a result of stenosis or valvular regurgi-
tation. Situations in acquired heart disease, where PC-MRI
provides important quantitative physiologic information
are therefore cardiac failure, valvular disease, and coronary
artery disease.
Cardiac Failure
Figure 5.9. Cardiac index in patients with cardiac failure, healthy
Ventricular Function subjects, and elite athletes. There is no significant difference between
healthy subjects and athletes suggesting similar basal metabolism
Systolic ventricular function, at least under stress conditions, whereas patients with cardiac failure have lower cardiac index. CI,
may be assessed by PC-MRI of ejected blood into the aorta. cardiac index. Adapted from: Carlsson M, Andersson R, Bloch KM
Peak flow acceleration has been shown to be a sensitive et al. Cardiac output and cardiac index measured with cardiovascular
indicator of the effect of ischemia on global ventricular func- magnetic resonance in healthy subjects, elite athletes and patients with
tion (211). congestive heart failure. J Cardiovasc Magn Reson. 2012;14:51.
LWBK1209-ch05_p73-92.indd 85 17/05/13 4:55 PM

blood volume or obstruction of large veins (224,225). When When both valves of a ventricle are regurgitant, each re-
CI decreases below a certain level, physiologic compensatory gurgitant volume can be measured as, for example, in aor-
mechanisms come into play that serve a purpose in the short tic and mitral regurgitation (241). The aortic regurgitation
perspective, increasing preload and afterload to restore is measured directly as backward flow in the valve plane
blood pressure, but worsen the situation in the long run by assessed by velocity mapping. The mitral regurgitation is
loading the failing heart. Measurement of CI may therefore calculated indirectly, as previously mentioned, minus the
provide information on the degree of cardiac failure (227). regurgitant volume of the aortic valve.
Four-dimensional flow measurements have shown excel-
lent agreement to 2D flow measurements for both mitral
Valvular Heart Disease
and tricuspid flow even in the presence of regurgitation
Regurgitation (163,164). It was recently demonstrated that 4D PC-MRI
with an acquisition time of 4 minutes using echo planar
Valvular regurgitation leads to increased preload and thus a
imaging with retrospective valve-tracking off-line enabled
loss of pumping energy. The regurgitant volume can be quan-
accurate quantification of net flow through all four heart
tified with high accuracy by MR in three principally different
valves (164).
ways: (a) Directly by velocity mapping in the valvular plane
The usefulness of MR in valvular regurgitation is consid-
(Fig. 5.10); (b) as the difference between the left ventricular
ered “Class I = provides clinically relevant information and
stroke volume assessed by planimetry and ejected volume
is frequently useful; may be used as first-line imaging tech-
assessed by velocity-mapping (see Chapter 14); or (c) by 4D
nique; usually supported by substantial literature” (185).
flow measurements.
Direct measurements of valvular regurgitation can be
achieved by measuring forward and backward flow at the Stenosis
level of the valve in aortic regurgitation (91,230,231), mi-
Severity of valvular stenosis is derived essentially in the same
tral regurgitation (91,232), and pulmonary regurgitation
way as for Doppler ultrasound, using the modified Bernoulli
in congenital heart disease (148,233,234). This can be per-
equation to estimate the pressure gradient across a stenotic
formed for all valves but is most reliable for the aortic and
lesion by measuring peak velocity in the lesion: ΔP = 4v2,
pulmonary valves. A complicating factor is through-plane
where ΔP is the peak pressure gradient (mm Hg) across the
motion of the valves, during the measurement, which may
lesion and v is the peak blood velocity (m/s). Shortcomings
introduce errors. These errors can be minimized by plac-
of derived pressure gradients are discussed previously in
ing the measurement plane as close as possible to the valve
Section “Calculation of Derived Parameters.”
plane (235,236). A promising feature is “moving slice” PC-
Assessment of the severity of stenosis can be achieved with
MRI (237,238). With this technique, the measurement plane
good accuracy by PC-MRI compared to Doppler ultrasound
follows the motion of the valve and thus minimizes errors
in mitral (153,158) and aortic stenosis (153). Methods to
caused by through-plane motion of valves.
assess the severity of aortic stenosis by calculating valve ef-
Indirect measurement of valvular regurgitation is
ficient orifice area (EOA) using PC-MRI have shown good
achieved by subtracting the net ejected volume (forward
correlation to transthoracic echocardiography and phantom
stroke volume) into the aorta or the pulmonary trunk from
measurements (103,243–245) and may serve as alternative
the planimetrically derived stroke volume of the ventricle
methods in cases when transthoracic echocardiography is
(total stroke volume) (239–242). This method works better
limited by insufficient acoustic window or calcification.
for the mitral and tricuspid valves in comparison to direct
measurement of regurgitant volume by velocity mapping.
Prosthetic Valves
Metallic prosthetic valves cause signal loss in the immediate
1200 vicinity of the metal parts. Despite this, PC-MRI has been
shown to provide valuable information on velocity fields
1000 around prosthetic aortic valves (62,246–250). In vitro stud-
800 ies have shown that 4D flow mapping may provide valu-
able information on both metallic and bioprosthetic valves
600 (251,252).
Flow (mL/s)
400
200 Coronary Disease

0 Coronary Flow Measurements
0 500 1000
–200 Coronary flow measurements using PC-MRI are sparsely
used in clinical routine. The technique, however, is subject
–400
to development and great interest for research applications.
Time (ms) Several in vitro and in vivo studies have validated PC-MRI
Figure 5.10. Patient with aortic regurgitation amounting to 54%. for coronary flow measurements (see Section, “Validation
Note that the forward flow rate during systole is high and that retro- in Vivo”). The following section describes measurements in
grade flow continues throughout diastole. humans.
LWBK1209-ch05_p73-92.indd 86 17/05/13 4:55 PM

A B C
Figure 5.11. Components of left ventricular flow can be divided into “direct flow, retained
inflow, delayed ejection, and residual volume” and identified by their behavior in early diastole
(A); late diastole (B); and early systole (C). Direct flow is shown in red to yellow and retained
flow in blue. (From Bolger AF, Heiberg E, Karlsson M, et al. Transit of blood flow through the
human left ventricle mapped by cardiovascular magnetic resonance. J Cardiovasc Magn Reson.
2007;9:741–747, with permission).
Global left ventricular perfusion and perfusion reserve can the left atrium (272), which have substantially increased the
be estimated by measuring coronary sinus blood flow, which understanding of fundamental cardiovascular hemodynamics.
is an approximation of the amount of blood that has perfused Measurements of all inflows and outflows of the heart have
the left ventricular myocardium, divided by left ventricular been used to derive total heart volume change during a heart
mass. Perfusion reserve is the amount of increased blood flow beat (152) and likewise, measurements of all inflows and out-
in absolute terms (mL/min) over resting perfusion. Perfusion flows of the lungs have been used to measure pulmonary intra-
reserve index is the ratio between maximal stress-induced per- vascular blood volume change (192), also during a heart beat.
fusion divided and resting blood flow. Perfusion reserve index Four-dimensional flow measurements provide new data
typically increases by a factor of 3 to 5 at pharmacologic and insight into normal cardiac pumping allowing separation
stress using dipyridamole or adenosine infusion. of blood paths and compartmentalization of blood during a
Measurements of global left ventricular perfusion have heart beat (220) (Fig. 5.11), calculation of kinetic energy of
been performed in healthy volunteers to estimate resting perventricular blood during diastole (221,222), and visualiza-
fusion (253), as well as perfusion reserve, which has been tion and calculation of total kinetic energy in both ventricles
shown to be in the range of 2.8 to 4.3 (180,182,254,255). The over the full cardiac cycle (223) (Fig. 5.12).
perfusion reserve is decreased in patients with hypertrophic
obstructive cardiomyopathy (254), orthotopic heart trans-
plant (180), ischemic heart disease (183), and chronic heart
failure (255). Administration of oxygen decreases systemic
and coronary oxygen delivery in healthy subjects as measured
by blood partial pressure of oxygen and coronary sinus flow
which may have important clinical implications in emergency
care and in care of patients with cardiac failure (228).
Measurement of blood flow velocity or volume flow is
more difficult in the coronary arteries compared to the coro-
nary sinus because of smaller dimensions and a higher degree
of motion during the cardiac cycle. Flow velocity or flow vol-
ume and flow reserve have been investigated in the left an-
terior descending (LAD) and right coronary artery (RCA) of
healthy subjects, and flow velocities typically increased three
to five times using pharmacologic stress (40,179,256,257)—
the higher values usually obtained by adenosine stress and
the lower by using dipyridamole. The same type of mea-
surements has been undertaken in the LAD artery in patient
populations and has shown high agreement with intravas-
cular ultrasound for measurement of flow and flow reserve KE [mJ/mL]
(172,173,175,258) and may be used to identify patients
with significant coronary artery stenosis (176,258,259), re- Figure 5.12. Visualization of the location and magnitude of
stenosis after stent implantation (260,261), or to evaluate kinetic energy (KE) in the heart at three time points during the cardiac
the physiologic status of venous grafts (262–264). cycle during systole, early diastole (E-wave), and atrial contraction
(A-wave). Kinetic energy is located near the outflow tracts during
systole and near the mitral and tricuspid valves during diastole. (From
Basic Physiology Carlsson M, Heiberg E, Toger J, et al. Quantification of left and right
ventricular kinetic energy using four-dimensional intracardiac mag-
Properties of flowing blood have been studied in the aorta netic resonance imaging flow measurements. Am J Physiol Heart Circ
(146,265–269), in the left ventricle (113,270,271), and in Physiol. 2012;301(4):H893–H900, with permission).
LWBK1209-ch05_p73-92.indd 87 17/05/13 4:55 PM

A promising feature of PC-MRI is derivation of intra- 18. Waluch V, Bradley WG. NMR even echo rephasing in slow laminar flow. J Comput Assist
Tomogr. 1984;8:594–598.
vascular and intracavitary pressure gradients from accelera- 19. Moran PR. A flow velocity zeugmatographic interlace for NMR imaging in humans.
tion and deceleration of blood, using an algorithm based on Magn Reson Imaging. 1982;1:197–203.
20. van Dijk P. Direct cardiac NMR imaging of heart wall and blood flow velocity. J Comput
the Navier–Stokes equations for incompressible Newtonian Assist Tomogr. 1984;8:429–436.
fluid (273–279). 21. Nayler GL, Firmin DN, Longmore DB. Blood flow imaging by cine magnetic resonance.
Important to mention in this context is that PC-MRI can J Comput Assist Tomogr. 1986;10:715–722.
22. Stahlberg F, Nordell B, Ericsson A, et al. Quantitative study of flow dependence in NMR
also be used to calculate strain rate in myocardium (124,280– images at low flow velocities. J Comput Assist Tomogr. 1986;10:1006–1015.
282) and to track myocardial motion (126,283,284). 23. Gao JH, Holland SK, Gore JC. Nuclear magnetic resonance signal from flowing nuclei in
rapid imaging using gradient echoes. Med Phys. 1988;15:809–814.
24. Gullberg GT, Simons MA, Wehrli FW. A mathematical model for signal from spins flow-
ing during the application of spin echo pulse sequences. Magn Reson Imaging. 1988;6:
Summary 437–461.
25. Henriksen O, Stahlberg F, Thomsen C, et al. In vivo evaluation of femoral blood flow
measured with magnetic resonance. Acta Radiol. 1989;30:153–157.
Blood flow measurements using PC-MRI have been thor- 26. Ibrahim el-SH. Myocardial tagging by cardiovascular magnetic resonance: evolution of
oughly validated in vitro and in vivo and have several impor- techniques—pulse sequences, analysis algorithms, and applications. J Cardiovasc Magn
Reson. 2011;13:36.
tant applications in congenital and acquired heart disease for 27. Zerhouni EA, Parish DM, Rogers WJ, et al. Human heart: tagging with MR imaging–a
diagnosis, estimation of severity, and follow-up after treat- method for noninvasive assessment of myocardial motion. Radiology. 1988;169:59–63.
ment. Quantification of intracardiac shunt and regurgitation 28. Axel L, Dougherty L. MR imaging of motion with spatial modulation of magnetization.
Radiology. 1989;171:841–845.
in valvular disease are considered first-line examinations. 29. Osman NF, Kerwin WS, McVeigh ER, et al. Cardiac motion tracking using CINE har-
Other important applications are measurement of collateral monic phase (HARP) magnetic resonance imaging. Magn Reson Med. 1999;42:1048–
1060.
blood flow in coarctation of the aorta and assessment of ste- 30. Aletras AH, Ding S, Balaban RS, et al. DENSE: displacement encoding with stimulated
nosis in coarctation and valve disease whereas applications echoes in cardiac functional MRI. J Magn Reson. 1999;137:247–252.
to coronary disease are mainly scientific. 4D blood flow mea- 31. Osman NF, Sampath S, Atalar E, et al. Imaging longitudinal cardiac strain on short-axis
images using strain-encoded MRI. Magn Reson Med. 2001;46:324–334.
surements are rapidly developing and show great promise for 32. Pelc NJ, Sommer FG, Li KC, et al. Quantitative magnetic resonance flow imaging. Magn
both clinical and scientific applications. Reson Q. 1994;10:125–147.
33. Stahlberg F, Mogelvang J, Thomsen C, et al. A method for MR quantification of flow
velocities in blood and CSF using interleaved gradient-echo pulse sequences. Magn Reson
Imaging. 1989;7:655–667.
Acknowledgments 34. Bernstein MA, Ikezaki Y. Comparison of phase-difference and complex-difference process-
ing in phase-contrast MR angiography. J Magn Reson Imaging. 1991;1:725–729.
35. Polzin JA, Alley MT, Korosec FR, et al. A complex-difference phase-contrast technique for
The authors would like to thank the following colleagues measurement of volume flow rates. J Magn Reson Imaging. 1995;5:129–137.
for valuable contributions to the contents of this chap- 36. van Dijk P. ECG-triggered NMR imaging of the heart. Diagn Imaging Clin Med. 1984;
53:29–37.
ter: Johannes Töger, Lund University, Sweden and Anders 37. Lentz GW, Haacke EM, White RD. Retrospective cardiac gating: A review of technical
Nilsson, Lund University, Sweden. aspects and future directions. Magn Reson Imaging. 1989;7:445–455.
38. Sondergaard L, Stahlberg F, Thomsen C, et al. Comparison between retrospective gat-
ing and ECG triggering in magnetic resonance velocity mapping. Magn Reson Imaging.
1993;11:533–537.
References 39. Ley S, Ley-Zaporozhan J, Kreitner KF, et al. MR flow measurements for assessment of
the pulmonary, systemic and bronchosystemic circulation: impact of different ECG gating
1. Bloch F, Hansen WW, Packard M. Nuclear induction. Phys Rev. 1946;69:127. methods and breathing schema. Eur J Radiol. 2007;61:124–129.
2. Purcell EM, Torrey HC, Pound RV. Resonance absorption by nuclear magnetic moments 40. Edelman RR, Manning WJ, Gervino E, et al. Flow velocity quantification in human coro-
in a solid. Phys Rev. 1946;69:37–38. nary arteries with fast, breath-hold MR angiography. J Magn Reson Imaging. 1993;3:
3. Suryan G. Nuclear resonance in flowing liquids. Proc Indian Acad Sci. 1951;A33:107– 699–703.
111. 41. Chatzimavroudis GP, Zhang H, Halliburton SS, et al. Clinical blood flow quantifica-
4. Bowman RL, Kudravcev V. Blood flowmeter utilizing nuclear magnetic resonance. IRE tion with segmented k-space magnetic resonance phase velocity mapping. J Magn Reson
Trans Med Electr. 1959:267. Imaging. 2003;17:65–71.
5. Singer JR. Blood flow rates by Nuclear Magnetic Resonance measurements. Science. 42. Hofman MB, van Rossum AC, Sprenger M, et al. Assessment of flow in the right human
1959;130:1652–1653. coronary artery by magnetic resonance phase contrast velocity measurement: effects of
6. Hahn EL. Detection of sea-water motion by nuclear precession. J Geophys Res. 1960; cardiac and respiratory motion. Magn Reson Med. 1996;35:521–531.
42:776. 43. Hofman MB, Wickline SA, Lorenz CH. Quantification of in-plane motion of the coronary
7. Grover T, Singer JR. NMR spin-echo flow measurements. J Appl Physics. 1971;42:938. arteries during the cardiac cycle: implications for acquisition window duration for MR
8. Battocletti JH, Halbach RE, Salles-Cunha SX, et al. The NMR blood flowmeter—theory flow quantification. J Magn Reson Imaging. 1998;8:568–576.
and history. Med Phys. 1981;8:435–443. 44. Foo TK, Bernstein MA, Aisen AM, et al. Improved ejection fraction and flow velocity
9. Hemminga MA, de Jager PA, Sonneveld A. The study of flow by pulsed nuclear magnetic estimates with use of view sharing and uniform repetition time excitation with fast cardiac
resonance. I. Measurement of flow rates in the presence of a stationary phase using a dif- techniques. Radiology. 1995;195:471–478.
ference method. J Magn Reson. 1977;27:359–370. 45. Lin HY, Bender JA, Ding Y, et al. Shared velocity encoding: a method to improve the
10. Lauterbur PC. Image formation by induced local interactions. Examples employing temporal resolution of phase-contrast velocity measurements. Magn Reson Med. 2012;68:
nuclear magnetic resonance. Nature. 1973;242:190–191. 703–710.
11. Herfkens R, Davis P, Crooks L, et al. Nuclear magnetic resonance imaging of the abnor- 46. Jahnke C, Paetsch I, Achenbach S, et al. Coronary MR imaging: breath-hold capability
mal live rat and correlations with tissue characteristics. Radiology. 1981;141:211–218. and patterns, coronary artery rest periods, and beta-blocker use. Radiology. 2006;239:71–
12. Grant JP, Back C. NMR rheotomography: feasibility and clinical potential. Med Phys. 78.
1982;9:188–193. 47. Oshinski JN, Hofland L, Mukundan S Jr, et al. Two-dimensional coronary MR angiogra-
13. Crooks L, Sheldon P, Kaufman L, et al. Quantification of obstructions in vessels by nuclear phy without breath holding. Radiology. 1996;201:737–743.
magnetic resonance (NMR). IEEE Trans Nucl Sci. 1982;29(3):1181. 48. Baltes C, Kozerke S, Atkinson D, et al. Retrospective respiratory motion correction for
14. Kaufman L, Crooks LE, Sheldon PE, et al. Evaluation of NMR imaging for detection and navigated cine velocity mapping. J Cardiovasc Magn Reson. 2004;6:785–792.
quantification of obstructions in vessels. Invest Radiol. 1982;17:554–560. 49. Johnson K, Sharma P, Oshinski J. Coronary artery flow measurement using navigator echo
15. Singer JR, Crooks LE. Nuclear magnetic resonance blood flow measurements in the gated phase contrast magnetic resonance velocity mapping at 3.0 T. J Biomech. 2008;
human brain. Science. 1983;221:654–656. 41:595–602.
16. Wehrli FW, Shimakawa A, MacFall JR, et al. MR imaging of venous and arterial flow 50. Sondergaard L, Stahlberg F, Thomsen C, et al. Accuracy and precision of MR velocity
by a selective saturation-recovery spin echo (SSRSE) method. J Comput Assist Tomogr. mapping in measurement of stenotic cross-sectional area, flow rate, and pressure gradient.
1985;9:537–545. J Magn Reson Imaging. 1993;3:433–437.
17. Shimizu K, Matsuda T, Sakurai T, et al. Visualization of moving fluid: quantitative analysis 51. Overall WR, Nishimura DG, Hu BS. Fast phase-contrast velocity measurement in the
of blood flow velocity using MR imaging. Radiology. 1986;159:195–199. steady state. Magn Reson Med. 2002;48:890–898.
LWBK1209-ch05_p73-92.indd 88 17/05/13 4:55 PM

52. Markl M, Alley MT, Pelc NJ. Balanced phase-contrast steady-state free precession 87. Buonocore MH, Bogren H. Factors influencing the accuracy and precision of velocity-
(PC-SSFP): a novel technique for velocity encoding by gradient inversion. Magn Reson encoded phase imaging. Magn Reson Med. 1992;26:141–154.
Med. 2003;49:945–952. 88. Jehenson P, Westphal M, Schuff N. Analytical method for compensation of eddy-current
53. Pelc NJ, Bernstein MA, Shimakawa A, et al. Encoding strategies for three-direction phase- effects induced by pulsed magnetic field gradients in NMR systems. J Magn Reson. 1990;
contrast MR imaging of flow. J Magn Reson Imaging. 1991;1:405–413. 90:264–278.
54. Wigstrom L, Sjoqvist L, Wranne B. Temporally resolved 3D phase-contrast imaging. Magn 89. Bernstein MA, Zhou XJ, Polzin JA, et al. Concomitant gradient terms in phase contrast
Reson Med. 1996;36:800–803. MR: analysis and correction. Magn Reson Med. 1998;39:300–308.
55. Frydrychowicz A, Francois CJ, Turski PA. Four-dimensional phase contrast magnetic reso- 90. Norris DG, Hutchison JM. Concomitant magnetic field gradients and their effects on
nance angiography: potential clinical applications. Eur J Radiol. 2011;80:24–35. imaging at low magnetic field strengths. Magn Reson Imaging. 1990;8:33–37.
56. Feinberg DA, Mark AS. Human brain motion and cerebrospinal fluid circulation demon- 91. Sondergaard L, Thomsen C, Stahlberg F, et al. Mitral and aortic valvular flow: quantifica-
strated with MR velocity imaging. Radiology. 1987;163:793–799. tion with MR phase mapping. J Magn Reson Imaging. 1992;2:295–302.
57. Kahn T, Muller E, Lewin JS, et al. MR measurement of spinal CSF flow with the RACE 92. Arheden H, Holmqvist C, Thilen U, et al. Left-to-right cardiac shunts: comparison of
technique. J Comput Assist Tomogr. 1992;16:54–61. measurements obtained with MR velocity mapping and with radionuclide angiography.
58. Maier SE, Hardy CJ, Jolesz FA. Brain and cerebrospinal fluid motion: real-time quantifica- Radiology. 1999;211:453–458.
tion with M-mode MR imaging. Radiology. 1994;193:477–483. 93. Walker PG, Cranney GB, Scheidegger MB, et al. Semiautomated method for noise reduc-
59. Gatehouse PD, Firmin DN, Collins S, et al. Real time blood flow imaging by spiral scan tion and background phase error correction in MR phase velocity data. J Magn Reson
phase velocity mapping. Magn Reson Med. 1994;31:504–512. Imaging. 1993;3:521–530.
60. Guilfoyle DN, Gibbs P, Ordidge RJ, et al. Real-time flow measurements using echo-planar 94. Markl M, Bammer R, Alley MT, et al. Generalized reconstruction of phase contrast
imaging. Magn Reson Med. 1991;18:1–8. MRI: analysis and correction of the effect of gradient field distortions. Magn Reson Med.
61. Debatin JF, Leung DA, Wildermuth S, et al. Flow quantitation with echo-planar phase- 2003;50:791–801.
contrast velocity mapping: in vitro and in vivo evaluation. J Magn Reson Imaging. 1995; 95. Giese D, Haeberlin M, Barmet C, et al. Analysis and correction of background velocity
5:656–662. offsets in phase-contrast flow measurements using magnetic field monitoring. Magn Reson
62. Kozerke S, Hasenkam JM, Pedersen EM, et al. Visualization of flow patterns distal to Med. 2012;67:1294–1302.
aortic valve prostheses in humans using a fast approach for cine 3D velocity mapping. J 96. Rolf MP, Hofman MB, Gatehouse PD, et al. Sequence optimization to reduce velocity
Magn Reson Imaging. 2001;13:690–698. offsets in cardiovascular magnetic resonance volume flow quantification–a multi-vendor
63. Tsao J, Boesiger P, Pruessmann KP. k-t BLAST and k-t SENSE: dynamic MRI with high study. J Cardiovasc Magn Reson. 2011;13:18.
frame rate exploiting spatiotemporal correlations. Magn Reson Med. 2003;50:1031– 97. Polzin JA, Korosec FR, Wedding KL, et al. Effects of through-plane myocardial motion on
1042. phase-difference and complex-difference measurements of absolute coronary artery flow.
64. Thunberg P, Karlsson M, Wigstrom L. Accuracy and reproducibility in phase contrast J Magn Reson Imaging. 1996;6:113–123.
imaging using SENSE. Magn Reson Med. 2003;50:1061–1068. 98. Evans AJ, Blinder RA, Herfkens RJ, et al. Effects of turbulence on signal intensity in gradi-
65. Beerbaum P, Korperich H, Gieseke J, et al. Rapid left-to-right shunt quantification in chil- ent echo images. Invest Radiol. 1988;23:512–518.
dren by phase-contrast magnetic resonance imaging combined with sensitivity encoding 99. Oshinski JN, Ku DN, Pettigrew RI. Turbulent fluctuation velocity: the most significant
(SENSE). Circulation. 2003;108:1355–1361. determinant of signal loss in stenotic vessels. Magn Reson Med. 1995;33:193–199.
66. Korperich H, Gieseke J, Barth P, et al. Flow volume and shunt quantification in pediatric 100. Stahlberg F, Sondergaard L, Thomsen C, et al. Quantification of complex flow using MR
congenital heart disease by real-time magnetic resonance velocity mapping: a validation phase imaging—a study of parameters influencing the phase/velocity relation. Magn Reson
study. Circulation. 2004;109:1987–1993. Imaging. 1992;10:13–23.
67. Baltes C, Kozerke S, Tsao JT, et al. Eight-fold acceleration of PC-SSFP velocity mapping 101. Kilner PJ, Firmin DN, Rees RS, et al. Valve and great vessel stenosis: assessment with MR
using k-t-BLAST. Kyoto, Japan: Poc ISMRM; 2004:1858. jet velocity mapping. Radiology. 1991;178:229–235.
68. Hansen MS, Baltes C, Tsao J, et al. Accelerated dynamic Fourier velocity encoding by 102. Gatenby JC, McCauley TR, Gore JC. Mechanisms of signal loss in magnetic resonance
exploiting velocity-spatio-temporal correlations. MAGMA. 2004;17:86–94. imaging of stenoses. Med Phys. 1993;20:1049–1057.
69. Stadlbauer A, van der Riet W, Globits S, et al. Accelerated phase-contrast MR imaging: 103. O’Brien KR, Gabriel RS, Greiser A, et al. Aortic valve stenotic area calculation from
comparison of k-t BLAST with SENSE and Doppler ultrasound for velocity and flow mea- phase contrast cardiovascular magnetic resonance: the importance of short echo time.
surements in the aorta. J Magn Reson Imaging. 2009;29:817–824. J Cardiovasc Magn Reson. 2009;11:49.
70. Carlsson M, Toger J, Kanski M, et al. Quantification and visualization of cardiovascular 104. O’Brien KR, Myerson SG, Cowan BR, et al. Phase contrast ultrashort TE: A more reli-
4D velocity mapping accelerated with parallel imaging or k-t BLAST: head to head com- able technique for measurement of high-velocity turbulent stenotic jets. Magn Reson Med.
parison and validation at 1.5 T and 3 T. J Cardiovasc Magn Reson. 2011;13:55. 2009;62:626–636.
71. Gu T, Korosec FR, Block WF, et al. PC VIPR: a high-speed 3D phase-contrast method for flow 105. von Schulthess GK, Higgins CB. Blood flow imaging with MR: spin-phase phenomena.
quantification and high-resolution angiography. AJNR Am J Neuroradiol. 2005;26:743–749. Radiology. 1985;157:687–695.
72. Axel L, Morton D. Correction of phase wrapping in magnetic resonance imaging. Med 106. Frayne R, Rutt BK. Understanding acceleration-induced displacement artifacts in phase-
Phys. 1989;16:284–287. contrast MR velocity measurements. J Magn Reson Imaging. 1995;5:207–215.
73. Xiang QS. Temporal phase unwrapping for CINE velocity imaging. J Magn Reson 107. Oshinski JN, Ku DN, Bohning DE, et al. Effects of acceleration on the accuracy of MR
Imaging. 1995;5:529–534. phase velocity measurements. J Magn Reson Imaging. 1992;2:665–670.
74. Ringgaard S, Oyre SA, Pedersen EM. Arterial MR imaging phase-contrast flow measure- 108. Simonetti OP, Wendt RE 3rd, Duerk JL. Significance of the point of expansion in interpreta-
ment: improvements with varying velocity sensitivity during cardiac cycle. Radiology. 2004; tion of gradient moments and motion sensitivity. J Magn Reson Imaging. 1991;1:569–577.
232:289–294. 109. Kouwenhoven M, Hofman MB, Sprenger M. Motion induced phase shifts in MR: accel-
75. Nett EJ, Johnson KM, Frydrychowicz A, et al. Four-dimensional phase contrast MRI with eration effects in quantitative flow measurements—a reconsideration. Magn Reson Med.
accelerated dual velocity encoding. J Magn Reson Imaging. 2012;35:1462–1471. 1995;33:766–777.
76. Firmin DN, Nayler GL, Kilner PJ, et al. The application of phase shifts in NMR for flow 110. Thunberg P, Wigstrom L, Wranne B, et al. Correction for acceleration-induced displace-
measurement. Magn Reson Med. 1990;14:230–241. ment artifacts in phase contrast imaging. Magn Reson Med. 2000;43:734–738.
77. Arheden H, Saeed M, Tornqvist E, et al. Accuracy of segmented MR velocity mapping to 111. Mohiaddin RH. Flow patterns in the dilated ischemic left ventricle studied by MR imaging
measure small vessel pulsatile flow in a phantom simulating cardiac motion. J Magn Reson with velocity vector mapping. J Magn Reson Imaging. 1995;5:493–498.
Imaging. 2001;13:722–728. 112. Buonocore MH. Algorithms for improving calculated streamlines in 3-D phase contrast
78. Thomsen C, Ståhlberg F, Henriksen O. Quantification of portal venous blood flow during angiography. Magn Reson Med. 1994;31:22–30.
fasting and after a standardized meal. Eur Radiol. 1993;3:242–247. 113. Wigstrom L, Ebbers T, Fyrenius A, et al. Particle trace visualization of intracardiac flow
79. Wolf RL, Ehman RL, Riederer SJ, et al. Analysis of systematic and random error in MR using time-resolved 3D phase contrast MRI. Magn Reson Med. 1999;41:793–799.
volumetric flow measurements. Magn Reson Med. 1993;30:82–91. 114. Toger J, Carlsson M, Soderlind G, et al. Volume Tracking: A new method for quantita-
80. Nilsson C, Stahlberg F, Thomsen C, et al. Circadian variation in human cerebrospinal fluid tive assessment and visualization of intracardiac blood flow from three-dimensional, time-
production measured by magnetic resonance imaging. Am J Physiol. 1992;262:R20–R24. resolved, three-component magnetic resonance velocity mapping. BMC Med Imaging.
81. Hofman MB, Visser FC, van Rossum AC, et al. In vivo validation of magnetic resonance 2011;11:10.
blood volume flow measurements with limited spatial resolution in small vessels. Magn 115. Markl M, Kilner PJ, Ebbers T. Comprehensive 4D velocity mapping of the heart and great
Reson Med. 1995;33:778–784. vessels by cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2011;13:7.
82. Tang C, Blatter DD, Parker DL. Correction of partial-volume effects in phase-contrast 116. Holen J, Aaslid R, Landmark K, et al. Determination of pressure gradient in mitral stenosis
flow measurements. J Magn Reson Imaging. 1995;5:175–180. with a non-invasive ultrasound Doppler technique. Acta Med Scand. 1976;199:455–460.
83. Kozerke S, Botnar R, Oyre S, et al. Automatic vessel segmentation using active contours in 117. Dyverfeldt P, Gardhagen R, Sigfridsson A, et al. On MRI turbulence quantification. Magn
cine phase contrast flow measurements. J Magn Reson Imaging. 1999;10:41–51. Reson Imaging. 2009;27:913–922.
84. Lagerstrand KM, Lehmann H, Starck G, et al. Method to correct for the effects of lim- 118. Oshinski JN, Ku DN, Mukundan S Jr, et al. Determination of wall shear stress in the aorta
ited spatial resolution in phase-contrast flow MRI measurements. Magn Reson Med. with the use of MR phase velocity mapping. J Magn Reson Imaging. 1995;5:640–647.
2002;48:883–889. 119. Oyre S, Ringgaard S, Kozerke S, et al. Quantitation of circumferential subpixel vessel wall
85. Holmqvist C, Oskarsson G, Stahlberg F, et al. Functional evaluation of extracardiac ven- position and wall shear stress by multiple sectored three-dimensional paraboloid modeling
triculopulmonary conduits and of the right ventricle with magnetic resonance imaging and of velocity encoded cine MR. Magn Reson Med. 1998;40:645–655.
velocity mapping. Am J Cardiol. 1999;83:926–932. 120. Wu SP, Ringgaard S, Pedersen EM. Three-dimensional phase contrast velocity map-
86. Bloch KM, Carlsson M, Arheden H, et al. Quantifying coronary sinus flow and global LV ping acquisition improves wall shear stress estimation in vivo. Magn Reson Imaging.
perfusion at 3 T. BMC Med Imaging. 2009;9:9. 2004;22:345–351.
LWBK1209-ch05_p73-92.indd 89 17/05/13 4:55 PM

121. Frydrychowicz A, Stalder AF, Russe MF, et al. Three-dimensional analysis of segmen- 152. Carlsson M, Cain P, Holmqvist C, et al. Total heart volume variation throughout the
tal wall shear stress in the aorta by flow-sensitive four-dimensional MRI. J Magn Reson cardiac cycle in humans. Am J Physiol Heart Circ Physiol. 2004;287:H243–H250.
Imaging. 2009;30:77–84. 153. Kilner PJ, Manzara CC, Mohiaddin RH, et al. Magnetic resonance jet velocity mapping in
122. Markl M, Wallis W, Brendecke S, et al. Estimation of global aortic pulse wave velocity by mitral and aortic valve stenosis. Circulation. 1993;87:1239–1248.
flow-sensitive 4D MRI. Magn Reson Med. 2010;63:1575–1582. 154. Caruthers SD, Lin SJ, Brown P, et al. Practical value of cardiac magnetic resonance imag-
123. Urchuk SN, Plewes DB. A velocity correlation method for measuring vascular compliance ing for clinical quantification of aortic valve stenosis: comparison with echocardiography.
using MR imaging. J Magn Reson Imaging. 1995;5:628–634. Circulation. 2003;108:2236–2243.
124. Wedeen VJ. Magnetic resonance imaging of myocardial kinematics. Technique to detect, 155. Mohiaddin RH, Kilner PJ, Rees S, et al. Magnetic resonance volume flow and jet velocity
localize, and quantify the strain rates of the active human myocardium. Magn Reson Med. mapping in aortic coarctation. J Am Coll Cardiol. 1993;22:1515–1521.
1992;27:52–67. 156. Engvall J, Sjoqvist L, Nylander E, et al. Biplane transoesophageal echocardiography, trans-
125. Reese TG, Feinberg DA, Dou J, et al. Phase contrast MRI of myocardial 3D strain by thoracic Doppler, and magnetic resonance imaging in the assessment of coarctation of the
encoding contiguous slices in a single shot. Magn Reson Med. 2002;47:665–676. aorta. Eur Heart J. 1995;16:1399–1409.
126. Bergvall E, Cain P, Sparr G, et al. Very fast and highly automated method for myocardial 157. Rupprecht T, Nitz W, Wagner M, et al. Determination of the pressure gradient in children
motion analysis with phase contrast MRI. J Cardiovasc Magn Reson. 2004;6:394A. with coarctation of the aorta by low-field magnetic resonance imaging. Pediatr Cardiol.
127. Lee BK. Computational fluid dynamics in cardiovascular disease. Korean Circ J. 2002;23:127–131.
2011;41:423–430. 158. Heidenreich PA, Steffens J, Fujita N, et al. Evaluation of mitral stenosis with velocity-
128. Wood NB, Weston SJ, Kilner PJ, et al. Combined MR imaging and CFD simulation of flow encoded cine-magnetic resonance imaging. Am J Cardiol. 1995;75:365–369.
in the human descending aorta. J Magn Reson Imaging. 2001;13:699–713. 159. Lin SJ, Brown PA, Watkins MP, et al. Quantification of stenotic mitral valve area with
129. Torii R, Keegan J, Wood NB, et al. The effect of dynamic vessel motion on haemodynamic magnetic resonance imaging and comparison with Doppler ultrasound. J Am Coll Cardiol.
parameters in the right coronary artery: a combined MR and CFD study. Br J Radiol. 2004;44:133–137.
2009;82 (Spec No 1):S24–S32. 160. Videlefsky N, Parks WJ, Oshinski J, et al. Magnetic resonance phase-shift velocity
130. Kilner PJ, Gatehouse PD, Firmin DN. Flow measurement by magnetic resonance: a unique mapping in pediatric patients with pulmonary venous obstruction. J Am Coll Cardiol.
asset worth optimising. J Cardiovasc Magn Reson. 2007;9:723–728. 2001;38:262–267.
131. Chernobelsky A, Shubayev O, Comeau CR, et al. Baseline correction of phase contrast 161. Van Rossum AC, Sprenger M, Visser FC, et al. An in vivo validation of quantitative blood
images improves quantification of blood flow in the great vessels. J Cardiovasc Magn flow imaging in arteries and veins using magnetic resonance phase-shift techniques. Eur
Reson 2007;9:681–685. Heart J. 1991;12:117–126.
132. Gatehouse PD, Rolf MP, Graves MJ, et al. Flow measurement by cardiovascular magnetic 162. Beerbaum P, Korperich H, Barth P, et al. Noninvasive quantification of left-to-right shunt
resonance: a multi-centre multi-vendor study of background phase offset errors that can in pediatric patients: phase-contrast cine magnetic resonance imaging compared with inva-
compromise the accuracy of derived regurgitant or shunt flow measurements. J Cardiovasc sive oximetry. Circulation. 2001;103:2476–2482.
Magn Reson. 2010;12:5. 163. Westenberg JJ, Roes SD, Ajmone Marsan N, et al. Mitral valve and tricuspid valve blood
133. Bryant DJ, Payne JA, Firmin DN, et al. Measurement of flow with NMR imaging using flow: accurate quantification with 3D velocity-encoded MR imaging with retrospective
a gradient pulse and phase difference technique. J Comput Assist Tomogr. 1984;8: valve tracking. Radiology. 2008;249:792–800.
588–593. 164. Roes SD, Hammer S, van der Geest RJ, et al. Flow assessment through four heart valves
134. Meier D, Maier S, Bosiger P. Quantitative flow measurements on phantoms and on blood simultaneously using 3-dimensional 3-directional velocity-encoded magnetic resonance
vessels with MR. Magn Reson Med. 1988;8:25–34. imaging with retrospective valve tracking in healthy volunteers and patients with valvular
135. Kondo C, Caputo GR, Semelka R, et al. Right and left ventricular stroke volume measure- regurgitation. Invest Radiol. 2009;44:669–675.
ments with velocity-encoded cine MR imaging: in vitro and in vivo validation. AJR Am 165. Brenner LD, Caputo GR, Mostbeck G, et al. Quantification of left to right atrial shunts
J Roentgenol. 1991;157:9–16. with velocity-encoded cine nuclear magnetic resonance imaging. J Am Coll Cardiol.
136. Caputo GR, Kondo C, Masui T, et al. Right and left lung perfusion: in vitro and in vivo 1992;20:1246–1250.
validation with oblique-angle, velocity-encoded cine MR imaging. Radiology. 1991;180: 166. Hundley WG, Li HF, Lange RA, et al. Assessment of left-to-right intracardiac shunting
693–698. by velocity-encoded, phase-difference magnetic resonance imaging. A comparison with
137. Henk CB, Grampp S, Koller J, et al. Elimination of errors caused by first-order aliasing in oximetric and indicator dilution techniques. Circulation. 1995;91:2955–2960.
velocity encoded cine-MR measurements of postoperative jets after aortic coarctation: in 167. Petersen SE, Voigtlander T, Kreitner KF, et al. Quantification of shunt volumes in con-
vitro and in vivo validation. Eur Radiol. 2002;12:1523–1531. genital heart diseases using a breath-hold MR phase contrast technique–comparison with
138. Evans AJ, Iwai F, Grist TA, et al. Magnetic resonance imaging of blood flow with a phase oximetry. Int J Cardiovasc Imaging. 2002;18:53–60.
subtraction technique. In vitro and in vivo validation. Invest Radiol. 1993;28:109–115. 168. Powell AJ, Tsai-Goodman B, Prakash A, et al. Comparison between phase-velocity cine
139. Powell AJ, Maier SE, Chung T, et al. Phase-velocity cine magnetic resonance imaging magnetic resonance imaging and invasive oximetry for quantification of atrial shunts.
measurement of pulsatile blood flow in children and young adults: in vitro and in vivo Am J Cardiol. 2003;91:1523–1525, A9.
validation. Pediatr Cardiol. 2000;21:104–110. 169. Clarke GD, Eckels R, Chaney C, et al. Measurement of absolute epicardial coronary
140. Chatzimavroudis GP, Oshinski JN, Franch RH, et al. Evaluation of the precision of mag- artery flow and flow reserve with breath-hold cine phase-contrast magnetic resonance
netic resonance phase velocity mapping for blood flow measurements. J Cardiovasc Magn imaging. Circulation. 1995;91:2627–2634.
Reson. 2001;3:11–19. 170. Sakuma H, Saeed M, Takeda K, et al. Quantification of coronary artery volume flow rate
141. Greil G, Geva T, Maier SE, et al. Effect of acquisition parameters on the accuracy of veloc- using fast velocity-encoded cine MR imaging. AJR Am J Roentgenol. 1997;168:1363–1367.
ity encoded cine magnetic resonance imaging blood flow measurements. J Magn Reson 171. Lund GK, Wendland MF, Shimakawa A, et al. Coronary sinus flow measurement by
Imaging. 2002;15:47–54. means of velocity-encoded cine MR imaging: validation by using flow probes in dogs.
142. Matre K, Ersland L, Larsen TH, et al. In vitro agreement between magnetic resonance Radiology. 2000;217:487–493.
imaging and intraluminal Doppler ultrasound for high flow velocity measurements. Scand 172. Hundley WG, Lange RA, Clarke GD, et al. Assessment of coronary arterial flow and flow
Cardiovasc J. 2002;36:180–186. reserve in humans with magnetic resonance imaging. Circulation. 1996;93:1502–1508.
143. Keegan J, Firmin D, Gatehouse P, et al. The application of breath hold phase velocity map- 173. Kessler W, Moshage W, Galland A, et al. Assessment of coronary blood flow in humans
ping techniques to the measurement of coronary artery blood flow velocity: phantom data using phase difference MR imaging. Comparison with intracoronary Doppler flow mea-
and initial in vivo results. Magn Reson Med. 1994;31:526–536. surement. Int J Card Imaging. 1998;14:179–186; discussion 187–189.
144. Langerak SE, Kunz P, Vliegen HW, et al. MR flow mapping in coronary artery bypass 174. Nagel E, Bornstedt A, Hug J, et al. Noninvasive determination of coronary blood flow
grafts: a validation study with Doppler flow measurements. Radiology. 2002;222:127– velocity with magnetic resonance imaging: comparison of breath-hold and navigator tech-
135. niques with intravascular ultrasound. Magn Reson Med. 1999;41:544–549.
145. Firmin DN, Nayler GL, Klipstein RH, et al. In vivo validation of MR velocity imaging. 175. Shibata M, Sakuma H, Isaka N, et al. Assessment of coronary flow reserve with fast cine
J Comput Assist Tomogr. 1987;11:751–756. phase contrast magnetic resonance imaging: comparison with measurement by Doppler
146. Bogren HG, Klipstein RH, Firmin DN, et al. Quantitation of antegrade and retrograde guide wire. J Magn Reson Imaging. 1999;10:563–568.
blood flow in the human aorta by magnetic resonance velocity mapping. Am Heart 176. Bedaux WL, Hofman MB, de Cock CC, et al. Magnetic resonance imaging versus Doppler
J. 1989;117:1214–1222. guide wire in the assessment of coronary flow reserve in patients with coronary artery
147. Sieverding L, Jung WI, Klose U, et al. Noninvasive blood flow measurement and quantifi- disease. Coron Artery Dis. 2002;13:365–372.
cation of shunt volume by cine magnetic resonance in congenital heart disease. Preliminary 177. Koskenvuo JW, Saraste M, Niemi P, et al. Correlation of transthoracic Doppler echocar-
results. Pediatr Radiol. 1992;22:48–54. diography and magnetic resonance imaging in measuring left anterior descending artery
148. Rebergen SA, Chin JG, Ottenkamp J, et al. Pulmonary regurgitation in the late post- flow velocity and time-course of dipyridamole-induced coronary flow increase. Scand J
operative follow-up of tetralogy of Fallot. Volumetric quantitation by nuclear magnetic Clin Lab Invest. 2003;63:65–72.
resonance velocity mapping. Circulation. 1993;88:2257–2266. 178. Salm LP, Bax JJ, Vliegen HW, et al. Functional significance of stenoses in coronary
149. Rebergen SA, Ottenkamp J, Doornbos J, et al. Postoperative pulmonary flow dynamics artery bypass grafts. Evaluation by single-photon emission computed tomography perfu-
after Fontan surgery: assessment with nuclear magnetic resonance velocity mapping. J Am sion imaging, cardiovascular magnetic resonance, and angiography. J Am Coll Cardiol.
Coll Cardiol. 1993;21:123–131. 2004;44:1877–1882.
150. Rebergen SA, van der Wall EE, Helbing WA, et al. Quantification of pulmonary and sys- 179. Sakuma H, Blake LM, Amidon TM, et al. Coronary flow reserve: noninvasive mea-
temic blood flow by magnetic resonance velocity mapping in the assessment of atrial-level surement in humans with breath-hold velocity-encoded cine MR imaging. Radiology.
shunts. Int J Card Imaging. 1996;12:143–152. 1996;198:745–750.
151. Rebergen SA, Helbing WA, van der Wall EE, et al. MR velocity mapping of tricuspid 180. Schwitter J, DeMarco T, Kneifel S, et al. Magnetic resonance-based assessment of global
flow in healthy children and in patients who have undergone Mustard or Senning repair. coronary flow and flow reserve and its relation to left ventricular functional parameters: a
Radiology. 1995;194:505–512. comparison with positron emission tomography. Circulation. 2000;101:2696–2702.
LWBK1209-ch05_p73-92.indd 90 17/05/13 4:55 PM

181. Sakuma H, Koskenvuo JW, Niemi P, et al. Assessment of coronary flow reserve using fast 211. Pennell DJ, Firmin DN, Burger P, et al. Assessment of magnetic resonance velocity map-
velocity-encoded cine MR imaging: validation study using positron emission tomography. ping of global ventricular function during dobutamine infusion in coronary artery disease.
AJR Am J Roentgenol. 2000;175:1029–1033. Br Heart J. 1995;74:163–170.
182. Koskenvuo JW, Sakuma H, Niemi P, et al. Global myocardial blood flow and global 212. Hartiala JJ, Mostbeck GH, Foster E, et al. Velocity-encoded cine MRI in the evalua-
flow reserve measurements by MRI and PET are comparable. J Magn Reson Imaging. tion of left ventricular diastolic function: measurement of mitral valve and pulmonary
2001;13:361–366. vein flow velocities and flow volume across the mitral valve. Am Heart J. 1993;125:
183. Koskenvuo JW, Hartiala JJ, Knuuti J, et al. Assessing coronary sinus blood flow in patients 1054–1066.
with coronary artery disease: a comparison of phase-contrast MR imaging with positron 213. Engels G, Muller E, Reynen K, et al. Evaluation of left ventricular inflow and volume by
emission tomography. AJR Am J Roentgenol. 2001;177:1161–1166. MR. Magn Reson Imaging. 1993;11:957–964.
184. Varaprasathan GA, Araoz PA, Higgins CB, et al. Quantification of flow dynamics in con- 214. Fujimoto S, Mohiaddin RH, Parker KH, et al. Magnetic resonance velocity mapping of
genital heart disease: applications of velocity-encoded cine MR imaging. Radiographics. normal human transmitral velocity profiles. Heart Vessels. 1995;10:236–240.
2002;22:895–905; discussion 905–906. 215. Karwatowski SP, Brecker SJ, Yang GZ, et al. Mitral valve flow measured with cine MR
185. Pennell DJ, Sechtem UP, Higgins CB, et al. Clinical indications for cardiovascular magnetic velocity mapping in patients with ischemic heart disease: comparison with Doppler echo-
resonance (CMR): Consensus Panel report. Eur Heart J. 2004;25:1940–1965. cardiography. J Magn Reson Imaging. 1995;5:89–92.
186. Baumgartner H, Bonhoeffer P, De Groot NM, et al. ESC Guidelines for the management 216. Milet SF, Mayberry JL, Ivarsen HR, et al. A semi-automated method to quantify left ven-
of grown-up congenital heart disease (new version 2010). Eur Heart J. 2010;31: tricular diastolic inflow propagation by magnetic resonance phase velocity mapping. J
2915–2957. Magn Reson Imaging. 1999;9:544–551.
187. Steffens JC, Bourne MW, Sakuma H, et al. Quantification of collateral blood flow in coarc- 217. Houlind K, Schroeder AP, Egeblad H, et al. Age-dependent changes in spatial and tem-
tation of the aorta by velocity encoded cine magnetic resonance imaging. Circulation. poral blood velocity distribution of early left ventricular filling. Magn Reson Imaging.
1994;90:937–943. 1999;17:859–868.
188. Holmqvist C, Stahlberg F, Hanseus K, et al. Collateral flow in coarctation of the aorta 218. Houlind K, Schroeder AP, Stodkilde-Jorgensen H, et al. Intraventricular dispersion and
with magnetic resonance velocity mapping: correlation to morphological imaging of col- temporal delay of early left ventricular filling after acute myocardial infarction. Assessment
lateral vessels. J Magn Reson Imaging. 2002;15:39–46. by magnetic resonance velocity mapping. Magn Reson Imaging. 2002;20:249–260.
189. Araoz PA, Reddy GP, Tarnoff H, et al. MR findings of collateral circulation are more 219. Bollache E, Redheuil A, Clement-Guinaudeau S, et al. Automated left ventricular diastolic
accurate measures of hemodynamic significance than arm-leg blood pressure gradient after function evaluation from phase-contrast cardiovascular magnetic resonance and compari-
repair of coarctation of the aorta. J Magn Reson Imaging. 2003;17:177–183. son with Doppler echocardiography. J Cardiovasc Magn Reson. 2010;12:63.
190. Chernoff DM, Derugin N, Rajasinghe HA, et al. Measurement of collateral blood flow 220. Bolger AF, Heiberg E, Karlsson M, et al. Transit of blood flow through the human left
in a porcine model of aortic coarctation by velocity-encoded cine MRI. J Magn Reson ventricle mapped by cardiovascular magnetic resonance. J Cardiovasc Magn Reson.
Imaging. 1997;7:557–563. 2007;9:741–747.
191. Mischi M, van den Bosch HC, den Boer JA, et al. Intra-thoracic blood volume measure- 221. Eriksson J, Dyverfeldt P, Engvall J, et al. Quantification of presystolic blood flow orga-
ment by contrast magnetic resonance imaging. Magn Reson Med. 2009;61:344–353. nization and energetics in the human left ventricle. Am J Physiol Heart Circ Physiol.
192. Ugander M, Jense E, Arheden H. Pulmonary intravascular blood volume changes through 2011;300:H2135–H2141.
the cardiac cycle in healthy volunteers studied by cardiovascular magnetic resonance mea- 222. Fredriksson AG, Zajac J, Eriksson J, et al. 4-D blood flow in the human right ventricle.
surements of arterial and venous flow. J Cardiovasc Magn Reson. 2009;11:42. Am J Physiol Heart Circ Physiol. 2011;301:H2344–H2350.
193. Ugander M, Kanski M, Engblom H, et al. Pulmonary blood volume variation decreases 223. Carlsson M, Heiberg E, Toger J, et al. Quantification of left and right ventricular kinetic
after myocardial infarction in pigs: a quantitative and noninvasive MR imaging measure energy using four-dimensional intracardiac magnetic resonance imaging flow measure-
of heart failure. Radiology. 2010;256:415–423. ments. Am J Physiol Heart Circ Physiol. 2012;302:H893–H900.
194. Kanski M, Arheden H, Wuttge DM, et al. Pulmonary blood density quantified by CMR 224. Guyton AC. Textbook of Medical Physiology. 8th ed. Philadelphia, PA: Saunders; 1991.
is reduced in newly diagnosed systemic sclerosis, consistent with pulmonary arteriolar 225. Carlsson M, Andersson R, Bloch KM, et al. Cardiac output and cardiac index measured
proliferation. J Cardiovasc Magn Reson. 2011;13(Suppl 1):P385. with magnetic resonance imaging in healthy subjects, elite athletes and patients with con-
195. Gutberlet M, Boeckel T, Hosten N, et al. Arterial switch procedure for D-transposition of gestive heart failure. J Cardiovasc Magn Reson. 2012;14:51.
the great arteries: quantitative midterm evaluation of hemodynamic changes with cine MR 226. Hundley WG, Li HF, Hillis LD, et al. Quantitation of cardiac output with velocity-
imaging and phase-shift velocity mapping-initial experience. Radiology. 2000;214:467– encoded, phase-difference magnetic resonance imaging. Am J Cardiol. 1995;75:1250–
475. 1255.
196. Houlind K, Stenbog EV, Sorensen KE, et al. Pulmonary and caval flow dynamics after total 227. Carlsson M, Andersson R, Bloch KM, et al. Reference values for cardiac index measured
cavopulmonary connection. Heart. 1999;81:67–72. with magnetic resonance imaging in healthy subjects and comparison with patients with
197. Pedersen EM, Stenbog EV, Frund T, et al. Flow during exercise in the total cavopulmonary congestive heart failure and athletes. J Cardiovasc Magn Reson. 2012;In press.
connection measured by magnetic resonance velocity mapping. Heart. 2002;87:554–558. 228. Bodetoft S, Carlsson M, Arheden H, et al. Effects of oxygen inhalation on cardiac output,
198. Bogren HG, Klipstein RH, Mohiaddin RH, et al. Pulmonary artery distensibility and coronary blood flow and oxygen delivery in healthy individuals, assessed with MRI. Eur J
blood flow patterns: a magnetic resonance study of normal subjects and of patients with Emerg Med. 2010.
pulmonary arterial hypertension. Am Heart J. 1989;118:990–999. 229. Thornton SJ, Hochachka PW, Crocker DE, et al. Stroke volume and cardiac output in
199. Laffon E, Laurent F, Bernard V, et al. Noninvasive assessment of pulmonary arterial hyper- juvenile elephant seals during forced dives. J Exp Biol. 2005;208:3637–3643.
tension by MR phase-mapping method. J Appl Physiol. 2001;90:2197–2202. 230. Sondergaard L, Lindvig K, Hildebrandt P, et al. Quantification of aortic regurgitation by
200. Laffon E, Vallet C, Bernard V, et al. A computed method for noninvasive MRI assessment magnetic resonance velocity mapping. Am Heart J. 1993;125:1081–1090.
of pulmonary arterial hypertension. J Appl Physiol. 2004;96:463–468. 231. Honda N, Machida K, Hashimoto M, et al. Aortic regurgitation: quantitation with MR
201. Bradlow WM, Gibbs JS, Mohiaddin RH. Cardiovascular magnetic resonance in pulmo- imaging velocity mapping. Radiology. 1993;186:189–194.
nary hypertension. J Cardiovasc Magn Reson. 2012;14:6. 232. Fujita N, Chazouilleres AF, Hartiala JJ, et al. Quantification of mitral regurgitation
202. Sanz J, Kuschnir P, Rius T, et al. Pulmonary arterial hypertension: noninvasive detection by velocity-encoded cine nuclear magnetic resonance imaging. J Am Coll Cardiol.
with phase-contrast MR imaging. Radiology. 2007;243:70–79. 1994;23:951–958.
203. Reiter G, Reiter U, Kovacs G, et al. Magnetic resonance-derived 3-dimensional blood flow 233. Roest AA, Helbing WA, Kunz P, et al. Exercise MR imaging in the assessment of pulmo-
patterns in the main pulmonary artery as a marker of pulmonary hypertension and a mea- nary regurgitation and biventricular function in patients after tetralogy of fallot repair.
sure of elevated mean pulmonary arterial pressure. Circ Cardiovasc Imaging. 2008;1:23– Radiology. 2002;223:204–211.
30. 234. Li W, Davlouros PA, Kilner PJ, et al. Doppler-echocardiographic assessment of pulmonary
204. Bell A, Beerbaum P, Greil G, et al. Noninvasive assessment of pulmonary artery flow and regurgitation in adults with repaired tetralogy of Fallot: comparison with cardiovascular
resistance by cardiac magnetic resonance in congenital heart diseases with unrestricted magnetic resonance imaging. Am Heart J. 2004;147:165–172.
left-to-right shunt. JACC Cardiovasc Imaging. 2009;2:1285–1291. 235. Chatzimavroudis GP, Walker PG, Oshinski JN, et al. The importance of slice location on
205. Garcia-Alvarez A, Fernandez-Friera L, Mirelis JG, et al. Non-invasive estimation of pul- the accuracy of aortic regurgitation measurements with magnetic resonance phase velocity
monary vascular resistance with cardiac magnetic resonance. Eur Heart J. 2011;32:2438– mapping. Ann Biomed Eng 1997;25:644–652.
2445. 236. Chatzimavroudis GP, Oshinski JN, Franch RH, et al. Quantification of the aortic regurgi-
206. Martinez JE, Mohiaddin RH, Kilner PJ, et al. Obstruction in extracardiac ventriculopul- tant volume with magnetic resonance phase velocity mapping: a clinical investigation of
monary conduits: value of nuclear magnetic resonance imaging with velocity mapping and the importance of imaging slice location. J Heart Valve Dis. 1998;7:94–101.
Doppler echocardiography. J Am Coll Cardiol. 1992;20:338–344. 237. Kozerke S, Scheidegger MB, Pedersen EM, et al. Heart motion adapted cine phase-con-
207. Helbing WA, Niezen RA, Le Cessie S, et al. Right ventricular diastolic function in children trast flow measurements through the aortic valve. Magn Reson Med. 1999;42:970–978.
with pulmonary regurgitation after repair of tetralogy of Fallot: volumetric evaluation by 238. Kozerke S, Schwitter J, Pedersen EM, et al. Aortic and mitral regurgitation: quantification
magnetic resonance velocity mapping. J Am Coll Cardiol. 1996;28:1827–1835. using moving slice velocity mapping. J Magn Reson Imaging. 2001;14:106–112.
208. Grothoff M, Spors B, Abdul-Khaliq H, et al. Evaluation of postoperative pulmonary 239. Hundley WG, Li HF, Willard JE, et al. Magnetic resonance imaging assessment of the
regurgitation after surgical repair of tetralogy of Fallot: comparison between Doppler severity of mitral regurgitation. Comparison with invasive techniques. Circulation.
echocardiography and MR velocity mapping. Pediatr Radiol. 2008;38:186–191. 1995;92:1151–1158.
209. Geva T. Repaired tetralogy of Fallot: the roles of cardiovascular magnetic resonance in 240. Kizilbash AM, Hundley WG, Willett DL, et al. Comparison of quantitative Doppler with
evaluating pathophysiology and for pulmonary valve replacement decision support. J magnetic resonance imaging for assessment of the severity of mitral regurgitation. Am
Cardiovasc Magn Reson. 2011;13:9. J Cardiol. 1998;81:792–795.
210. Sun AM, AlHabshan F, Cheung M, et al. Delayed onset of tricuspid valve flow in repaired 241. Kon MW, Myerson SG, Moat NE, et al. Quantification of regurgitant fraction in mitral
tetralogy of Fallot: an additional mechanism of diastolic dysfunction and interventricular regurgitation by cardiovascular magnetic resonance: comparison of techniques. J Heart
dyssynchrony. J Cardiovasc Magn Reson. 2011;13:43. Valve Dis. 2004;13:600–607.
LWBK1209-ch05_p73-92.indd 91 17/05/13 4:55 PM

242. Chan KM, Wage R, Symmonds K, et al. Towards comprehensive assessment of mitral regur- 262. Galjee MA, van Rossum AC, Doesburg T, et al. Quantification of coronary artery bypass graft
gitation using cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2008;10:61. flow by magnetic resonance phase velocity mapping. Magn Reson Imaging. 1996;14:485–493.
243. Tanaka K, Makaryus AN, Wolff SD. Correlation of aortic valve area obtained by the 263. Bedaux WL, Hofman MB, Vyt SL, et al. Assessment of coronary artery bypass graft dis-
velocity-encoded phase contrast continuity method to direct planimetry using cardiovas- ease using cardiovascular magnetic resonance determination of flow reserve. J Am Coll
cular magnetic resonance. J Cardiovasc Magn Reson. 2007;9:799–805. Cardiol. 2002;40:1848–1855.
244. Garcia J, Kadem L, Larose E, et al. Comparison between cardiovascular magnetic reso- 264. Salm LP, Langerak SE, Vliegen HW, et al. Blood flow in coronary artery bypass vein grafts:
nance and transthoracic Doppler echocardiography for the estimation of effective orifice volume versus velocity at cardiovascular MR imaging. Radiology. 2004;232:915–920.
area in aortic stenosis. J Cardiovasc Magn Reson. 2011;13:25. 265. Klipstein RH, Firmin DN, Underwood SR, et al. Blood flow patterns in the human aorta
245. Garcia J, Marrufo OR, Rodriguez AO, et al. Cardiovascular magnetic resonance evalu- studied by magnetic resonance. Br Heart J. 1987;58:316–323.
ation of aortic stenosis severity using single plane measurement of effective orifice area. 266. Kilner PJ, Yang GZ, Mohiaddin RH, et al. Helical and retrograde secondary flow pat-
J Cardiovasc Magn Reson. 2012;14:23. terns in the aortic arch studied by three-directional magnetic resonance velocity mapping.
246. Houlind K, Eschen O, Pedersen EM, et al. Magnetic resonance imaging of blood velocity distri- Circulation. 1993;88:2235–2247.
bution around St. Jude medical aortic valves in patients. J Heart Valve Dis. 1996;5:511–517. 267. Bogren HG, Buonocore MH. Blood flow measurements in the aorta and major arteries
247. Ringgaard S, Botnar RM, Djurhuus C, et al. High-resolution assessment of velocity fields with MR velocity mapping. J Magn Reson Imaging. 1994;4:119–130.
and shear stresses distal to prosthetic heart valves using high-field magnetic resonance 268. Bogren HG, Buonocore MH. 4D magnetic resonance velocity mapping of blood flow pat-
imaging. J Heart Valve Dis. 1999;8:96–103. terns in the aorta in young vs. elderly normal subjects. J Magn Reson Imaging. 1999;10:861–
248. Hasenkam JM, Ringgaard S, Houlind K, et al. Prosthetic heart valve evaluation by mag- 869.
netic resonance imaging. Eur J Cardiothorac Surg. 1999;16:300–305. 269. Markl M, Draney MT, Hope MD, et al. Time-resolved 3-dimensional velocity mapping in
249. Botnar R, Nagel E, Scheidegger MB, et al. Assessment of prosthetic aortic valve perfor- the thoracic aorta: visualization of 3-directional blood flow patterns in healthy volunteers
mance by magnetic resonance velocity imaging. MAGMA. 2000;10:18–26. and patients. J Comput Assist Tomogr. 2004;28:459–468.
250. Kozerke S, Hasenkam JM, Nygaard H, et al. Heart motion-adapted MR velocity mapping 270. Kim WY, Walker PG, Pedersen EM, et al. Left ventricular blood flow patterns in normal
of blood velocity distribution downstream of aortic valve prostheses: initial experience. subjects: a quantitative analysis by three-dimensional magnetic resonance velocity map-
Radiology. 2001;218:548–555. ping. J Am Coll Cardiol. 1995;26:224–238.
251. Kvitting JP, Dyverfeldt P, Sigfridsson A, et al. In vitro assessment of flow patterns and tur- 271. Kilner PJ, Yang GZ, Wilkes AJ, et al. Asymmetric redirection of flow through the heart.
bulence intensity in prosthetic heart valves using generalized phase-contrast MRI. J Magn Nature. 2000;404:759–761.
Reson Imaging. 2010;31:1075–1080. 272. Fyrenius A, Wigstrom L, Ebbers T, et al. Three dimensional flow in the human left atrium.
252. von Knobelsdorff-Brenkenhoff F, Dieringer MA, Greiser A, et al. In vitro assessment of heart Heart. 2001;86:448–455.
valve bioprostheses by cardiovascular magnetic resonance: four-dimensional mapping of 273. Urchuk SN, Plewes DB. MR measurements of pulsatile pressure gradients. J Magn Reson
flow patterns and orifice area planimetry. Eur J Cardiothorac Surg. 2011;40:736–742. Imaging. 1994;4:829–836.
253. van Rossum AC, Visser FC, Hofman MB, et al. Global left ventricular perfusion: noninva- 274. Yang GZ, Kilner PJ, Wood NB, et al. Computation of flow pressure fields from magnetic
sive measurement with cine MR imaging and phase velocity mapping of coronary venous resonance velocity mapping. Magn Reson Med. 1996;36:520–526.
outflow. Radiology. 1992;182:685–691. 275. Urchuk SN, Fremes SE, Plewes DB. In vivo validation of MR pulse pressure measurement
254. Kawada N, Sakuma H, Yamakado T, et al. Hypertrophic cardiomyopathy: MR measure- in an aortic flow model: preliminary results. Magn Reson Med. 1997;38:215–223.
ment of coronary blood flow and vasodilator flow reserve in patients and healthy subjects. 276. Tasu JP, Mousseaux E, Delouche A, et al. Estimation of pressure gradients in pulsatile
Radiology. 1999;211:129–135. flow from magnetic resonance acceleration measurements. Magn Reson Med. 2000;44:66–72.
255. Lund GK, Watzinger N, Saeed M, et al. Chronic heart failure: global left ventricular per- 277. Tyszka JM, Laidlaw DH, Asa JW, et al. Three-dimensional, time-resolved (4D) rela-
fusion and coronary flow reserve with velocity-encoded cine MR imaging: initial results. tive pressure mapping using magnetic resonance imaging. J Magn Reson Imaging.
Radiology. 2003;227:209–215. 2000;12:321–329.
256. Davis CP, Liu PF, Hauser M, et al. Coronary flow and coronary flow reserve measure- 278. Ebbers T, Wigstrom L, Bolger AF, et al. Estimation of relative cardiovascular pressures
ments in humans with breath-held magnetic resonance phase contrast velocity mapping. using time-resolved three-dimensional phase contrast MRI. Magn Reson Med. 2001;45:
Magn Reson Med. 1997;37:537–544. 872–879.
257. Grist TM, Polzin JA, Bianco JA, et al. Measurement of coronary blood flow and flow 279. Thompson RB, McVeigh ER. Fast measurement of intracardiac pressure differences with
reserve using magnetic resonance imaging. Cardiology. 1997;88:80–89. 2D breath-hold phase-contrast MRI. Magn Reson Med. 2003;49:1056–1066.
258. Hundley WG, Hamilton CA, Clarke GD, et al. Visualization and functional assessment of 280. Robson MD, Constable RT. Three-dimensional strain-rate imaging. Magn Reson Med.
proximal and middle left anterior descending coronary stenoses in humans with magnetic 1996;36:537–546.
resonance imaging. Circulation. 1999;99:3248–3254. 281. Arai AE, Gaither CC 3rd, Epstein FH, et al. Myocardial velocity gradient imaging by
259. Hundley WG, Hillis LD, Hamilton CA, et al. Assessment of coronary arterial restenosis phase contrast MRI with application to regional function in myocardial ischemia. Magn
with phase-contrast magnetic resonance imaging measurements of coronary flow reserve. Reson Med. 1999;42:98–109.
Circulation. 2000;101:2375–2381. 282. Selskog P, Heiberg E, Ebbers T, et al. Kinematics of the heart: strain-rate imaging from
260. Saito Y, Sakuma H, Shibata M, et al. Assessment of coronary flow velocity reserve using time-resolved three-dimensional phase contrast MRI. IEEE Trans Med Imaging. 2002;21:
fast velocity-encoded cine MRI for noninvasive detection of restenosis after coronary stent 1105–1109.
implantation. J Cardiovasc Magn Reson. 2001;3:209–214. 283. Zhu Y, Drangova M, Pelc NJ. Fourier tracking of myocardial motion using cine-PC data.
261. Nagel E, Thouet T, Klein C, et al. Noninvasive determination of coronary blood flow Magn Reson Med. 1996;35:471–480.
velocity with cardiovascular magnetic resonance in patients after stent deployment. 284. Zhu Y, Drangova M, Pelc NJ. Estimation of deformation gradient and strain from cine-PC
Circulation. 2003;107:1738–1743. velocity data. IEEE Trans Med Imaging. 1997;16:840–851.
LWBK1209-ch05_p73-92.indd 92 17/05/13 4:55 PM

Chapter
6
Rob J. van der Geest
Boudewijn P.F. Lelieveldt
Johan H.C. Reiber
Quantification in Cardiac Magnetic

Resonance Imaging and Computed
Tomography
■■ Quantification of Ventricular Dimensions Magnetic resonance imaging (MRI) and multislice computed
and Global Function tomography (MSCT) have become indispensable imaging
Accuracy and Reproducibility of Volumetric modalities for evaluation of the cardiac system. MRI pro-
Measurements from Multislice Short-axis vides unique capabilities for studying many aspects of cardiac
anatomy, function, perfusion, and viability in a single imaging
Acquisitions
session. Nevertheless, MSCT is rapidly gaining widespread
Impact of Slice Thickness and Slice Gap clinical use for the evaluation of coronary anatomy and car-
Global Function Assessment Using Radial diac function. The recent technologic developments in MSCT
Long-axis Views have resulted in a tremendous improvement of image quality
■■ Myocardial Mass and high sensitivity and specificity for the detection of coro-
Left Ventricular Mass nary stenosis.
Right Ventricular Mass Both MRI and MSCT are three-dimensional (3D) imaging
modalities that are used for the evaluation of ventricular func-
■■ Quantification of Ventricular Volumes and tion. Volumetric measurement of the ventricular cavities and
Global Function myocardium can be performed at high accuracy and precision
■■ Quantification of Regional Wall Motion and as has been demonstrated in many experimental and clinical
research studies. The 3D nature of these imaging modalities
Wall Thickening Using the Centerline
also provides detailed information of the cardiac system at
Method from Dynamic Short-Axis Images a regional level. Among others, regional end-diastolic wall
■■ Regional Function Analysis Using Magnetic thickness and systolic wall thickening provide useful informa-
Resonance Imaging Tagging and Velocity- tion for the assessment of the location, extent, and severity
Encoded Magnetic Resonance Imaging of ventricular abnormalities in ischemic heart disease. MRI
can also be used to study blood flow and myocardial perfu-
■■ Automated Contour Detection in Short-Axis
sion. Velocity-encoded cine MRI (VEC-MRI) is often used for
Multislice Cine Magnetic Resonance the quantification of blood flow through the aortic and pul-
Imaging monary valves and atrioventricular valve planes, which has
Automated Segmentation of Multislice shown to be clinically valuable in the evaluation of patients
Computed Tomography with complex congenital heart disease.
Automated Contour Detection Optimization for Typical cardiac MRI and MSCT examinations generate
Different Magnetic Resonance Pulse Sequences large data sets of images. To optimally and efficiently extract
New Automated Segmentation Methods the relevant clinical information from these data sets, dedi-
cated software solutions featuring automated image segmen-
■■ Magnetic Resonance Imaging Flow tation and optimal quantification and visualization methods
Quantification are needed. Quantitative image analysis requires the defini-
■■ Automated Quantification of Aortic Flow tion of contours describing the inner and outer boundaries
of the ventricles, which is a laborious and tedious task when
■■ Delayed Contrast-Enhanced Magnetic
based on manual contour tracing. Reliable automated or
Resonance Imaging semiautomated image analysis software would be required
■■ Summary to overcome these limitations. This chapter focuses on the
93
LWBK1209-ch06_p093-107.indd 93 17/05/13 4:56 PM

state-of-the-art postprocessing techniques for quantitative A Overestimation B Underestimation

assessment of global and regional ventricular functions from
cardiac MRI and MSCT.
Quantification of Ventricular
Dimensions and Global Function
Accuracy and Reproducibility of
Volumetric Measurements from
Multislice Short-Axis Acquisitions Figure 6.1. Left ventricular geometry intersected by multiple paral-
lel short-axis sections.Given the same section thickness and intersec-
MRI and MSCT allow imaging of anatomic objects in mul-
tion gap, both over- and underestimation of left ventricular volume
tiple parallel sections, enabling volumetric measurements
can occur dependent on the position of the left ventricle (LV) with
using the Simpson’s rule. According to Simpson’s rule, the respect to the imaging slices. In situation (A) the most basal slice will
volume of an object can be estimated by summation of the be included in the volumetric assessment, whereas in (B), the most
cross-sectional areas in each section multiplied by the sec- basal slice will not be taken into account because it intersects by less
tion thickness. When there is a gap in between slices, this than 50% with the left ventricular myocardium.
must be corrected for and the formula for volume becomes:
V = ∑{Areai × (Thickness + Gap)} between 4 and 8 mm usually without any intersection gap.
where V is the volume of the 3D object and Areai the area of Quantification of volumes and mass requires the definition of
the cross section in section number i. contours in the images describing the endocardial and epicar-
The short-axis orientation is the most commonly applied dial boundaries of the myocardium in several phases of the
image orientation for the assessment of left ventricular cham- cardiac cycle. Although an image voxel may contain several
ber size and mass. Although MRI is capable of directly ac- tissues, because of the partial volume effect, it is assumed that
quiring images in any orientation, MSCT images are always the traced contours represent the geometry of the ventricle at
generated in the axial orientation and image reformatting is the center of the imaged section. As shown in Figure 6.1, the
required to obtain short-axis images. The short-axis orienta- partial volume effect may lead to both over-and underestima-
tion has advantages over other slice orientations because it tion in the assessment of ventricular volume.
yields cross-sectional slices almost perpendicular to the myo- With a simple experiment using synthetically created
cardium for the largest part of the left ventricle (LV). This left ventricular shapes and short-axis cross sections it can
results in minimal partial volume effect at the myocardial be shown how the partial volume problem may affect mea-
boundaries and subsequently provides optimal depiction of surement accuracy and reproducibility. For this purpose a
the myocardial boundaries. However, the curvature of the LV computer-generated average left ventricular geometry with a
at the apical level leads to significant partial volume averag- fixed size was constructed and short-axis cross sections were
ing. The image voxels in this area simultaneously intersect automatically derived while varying the position of the ven-
with blood and myocardium yielding indistinct myocardial tricular geometry along the long-axis direction. The shape
boundaries. By minimizing the slice thickness, while keeping used and its dimensions are presented in Figure 6.2. In this
sufficient signal to noise, this partial volume effect at the apex
can be reduced. Given the relatively small cross-sectional area
of the LV in the apical section, the error introduced because
of the partial volume effect will be minimal. However, partial
volume averaging at the basal level of the heart has a much
greater impact because at this level the cross-sectional area of
the LV is largest. The base of the LV exhibits a through-plane
motion in the apical direction during systole on the order of
1.3 cm (1,2). Therefore, the significance of partial volume
varies over the cardiac cycle. Additional long-axis views may
prove helpful in determining more accurately how a basal
short-axis slice intersects with the various anatomic regions.
Impact of Slice Thickness and Slice Gap

It is a prerequisite for the accurate assessment of the ventricular
volumes that the stack of short-axis slices covers the complete
Figure 6.2. Left ventricular geometry used for the simulation
ventricle from base to apex. With MRI, a section thickness experiments. The phantom consists of a half ellipsoid with a length
between 6 and 10 mm is used, and the gap between slices of 70 mm and an outer diameter at the base of 60 mm; at the base
varies from no gap (consecutive slices) to 4 mm. With MSCT, the shape is extended with a cylinder with a diameter of 60 mm and a
any value can be chosen for the slice thickness and gap when length of 30 mm. The thickness of the phantom was set to 5 mm. In the
generating a short-axis data set by reformatting the original experiments the size of the object was varied between the dimensions
images. Typical parameters for MSCT are a slice thickness shown and 80% of this size.
LWBK1209-ch06_p093-107.indd 94 17/05/13 4:56 PM

Chapter 6 ■ Quantification in Cardiac Magnetic Resonance Imaging and Computed Tomography 95
10
Slice thickness: 6 mm
Slice thickness: 10 mm
8
Percent variation
6
Figure 6.3. Results of volume

calculation experiments using 4
synthetically constructed left
ventricular geometries. The vari-
ability of left ventricular volume 2
estimates increases with increasing
slice thickness and slice gap. For
a setting of the imaging param- 0
eters, such as a thickness of 6 mm
0 2 4 6 8 10
and a gap of 4 mm, the measure-
ment variability is 5%. Slice gap (mm)
experiment it is assumed that the contours in a short-axis MRI sequence. With a radial long-axis acquisition, mul-
slice will only be drawn in case more than 50% of the slice tiple long-axis views are acquired sharing a common axis
thickness intersects with myocardium. The results of the sim- of rotation (the LV long-axis) at equiangular intervals. This
ulations as depicted in Figure 6.3 demonstrate that the mea- orientation has intrinsic advantages over short-axis imaging
surement precision (or measurement variability) degrades because it allows clear visualization of the mitral and aortic
with increasing distance between the slices. For typical im- valve planes. In addition, long-axis views have less partial
aging parameters (section thickness 6 mm, gap 4 mm) the volume effect near the apex. After definition of the ventri
measurement variability is between 4% and 5%. The mea- cular contours, calculation of LV volume is performed by
surement accuracy is not dependent on the slice thickness or adding pie-shaped volume elements defined by the location
slice gap used. A section thickness of 10 mm with no gap of the axis of rotation, position of the contour, and angular
results in the same variability as a section thickness of 6 mm interval between the image sections. Bloomer et al. demon-
with a gap of 4 mm. strated a good agreement between multislice short-axis and
The result of this experiment has two important implica- radial long-axis acquisitions. As a result of the improved
tions. First, variations between successive scans of the same visualization of the myocardial boundaries and definition of
patient may result in volumetric differences of up to 5%, the base, interobserver agreement was better using the radial
which are inherent to the imaging technique used. Second, long-axis method. Figure 6.4 illustrates examples of mag-
because the base of the heart has a significant through-plane netic resonance images acquired using radial long-axis ori-
motion component, the measurement variability of up to entation. It clearly shows that the definition of the base and
5% will also be present over the cardiac cycle. Therefore, the contrast between blood and myocardium are excellent.
ejection fraction measurements will also be affected. By Further research is needed to evaluate whether radial long-
reducing the section thickness and the intersection gap, the axis acquisitions also prove to be valuable for the assess-
variability in volumetric measurements can be reduced. For ment of regional function.
MSCT, in which short-axis reformats are derived from high-
resolution near-isotropic volumetric data, it is useful to use
small values for the slice thickness without applying any gap. Myocardial Mass
The measurement of heart muscle weight is of clinical impor-
Global Function Assessment Using
tance to properly diagnose and understand a patient’s illness
Radial Long-Axis Views
and condition, and to estimate the effects of treatment. To
The accuracy of volumetric measurements from multislice detect small changes in mass it is of paramount importance
short-axis acquisitions is mainly determined by the accurate to use an accurate and reproducible measurement technique.
identification of the most basal slice level and the accurate Several validation studies have been performed comparing
definition of the endocardial and epicardial contours in this mass estimates as derived from MRI with postmortem mass
slice level. However, because of the relatively large section measurements. In a study by Florentine et al. (4) a stack of
thickness used, this is often difficult. The origin of the prob- axial slices was used to quantify left ventricular mass using
lem is the highly anisotropic nature of a typical short-axis the Simpson rule. In this early study, they found good agree-
examination in which the resolution in the Z-direction is ment (r = 0.95, standard error of estimate = 13 g). Maddahi
much worse than the in-plane resolution. To overcome this et al. (5) carried out extensive studies in a dog model com-
limitation, Bloomer et al. (3) investigated the use of multiple paring several slice orientations and measurement tech-
radial long-axis views for quantification of left ventricular niques for quantifying left ventricular mass. It was shown
volumes and mass using a steady-state free precession (SSFP) that in vivo estimates of left ventricular myocardial mass are
LWBK1209-ch06_p093-107.indd 95 17/05/13 4:56 PM

Figure 6.4. End-diastolic (top) and end-systolic (middle) magnetic resonance images acquired in radial
long-axis views using steady-state free precession (SSFP) magnetic resonance imaging (MRI). Note the excel-
lent conspicuity of the LV myocardial wall from base to apex. Orientation of the radial long-axis views
(white lines in short-axis images, bottom). (Image data courtesy of M. Friedrich.)
most accurate when the images are obtained in the short- images. The muscle volume is assessed from these contours
axis plane (r = 0.98, standard error of estimate = 4.9 g). by applying Simpson’s rule. The myocardial mass is derived
by multiplying the muscle volume with the specific density of
myocardium (1.05 g/cm3).
Left Ventricular Mass
Typically, a section thickness between 6 and 10 mm is
For the LV it is generally believed and also supported by lit- used with an intersection gap between 0 and 4 mm. At the
erature that the short-axis orientation is the most appropriate apex and basal sections significant partial volume averaging
imaging plane. To obtain optimal accuracy and reproduci will occur because of the section thickness used and tracing
bility it is important to cover the complete ventricle from apex of the myocardial boundaries may not be trivial. Similarly,
to base with a sufficient number of slices. Quantification of partial volume averaging will cause significant difficulties in
mass requires the definition of contours in the images describ- interpreting sections with a highly trabeculated myocardial
ing the endocardial and epicardial boundaries in the stack of wall and papillary muscles (6). There is no general consensus
LWBK1209-ch06_p093-107.indd 96 17/05/13 4:56 PM

on whether to include or exclude papillary muscles and tra- A vast amount of reports describe the applicability of MRI
beculae in the left ventricular mass. Although it is evident for accurate and reproducible quantification of left and right
that inclusion of these structures would result in more accu- ventricular volumes using various MRI strategies (4,10).
rate myocardial mass measurements, for regional wall thick- Cardiac-gated MSCT also allows quantification of global
ening analysis it is important to exclude these structures to function (11). Sufficient temporal resolution is required to
avoid artifacts in the quantification. Whether to use an end- properly capture the end-systolic phase. Generally a tem-
diastolic or end-systolic time frame for the measurement is poral resolution, or phase interval, on the order of 40 to
also a subject of ongoing debate. Most likely, optimal accu- 50 ms is assumed to be sufficient. Although MRI can provide
racy and reproducibility are obtained by averaging multiple such temporal resolution, the temporal resolution of MSCT
time frames, but this will have practical objections in case does not yet fulfill this requirement. However, global func-
the contours are derived by manual tracing. tion results obtained by MSCT are in good agreement with
results derived from MRI (12,13). For geometrically normal
LVs, one could rely on geometric models to derive the vol-
Right Ventricular Mass
umes from one or two long-axis imaging sections. In a group
For the right ventricle (RV) and geometrically abnormally of 10 patients with LV hypertrophy and 10 healthy subjects,
shaped LVs, multiple sections are required for an accurate Dulce et al. (14) demonstrated a good agreement between
volume assessment (7). MRI experiments with different slice biplane volumetric measurements using either the modified
orientations in phantoms and ventricular casts have shown Simpson’s rule of an ellipsoid model or true 3D volumetric
that no significant difference can be observed in accuracy measurements using a multislice MRI approach. In another
and reproducibility between slice orientations (8). However, study by Chuang et al. (15), 25 patients with dilated cardio-
in a clinical situation the choice of slice orientation also myopathies were evaluated using both biplane and 3D mul-
depends on the availability of a clear depiction of anatomic tislice approach. They reported a poor correlation between
features that are needed to define the myocardial bound- the two measurement methods. At the present state, a single
aries. Volumetric quantification of the RV may be better section with sufficient temporal resolution can be acquired
performed on the basis of axial views (9). This view shows within a single breath-hold on most available MRI systems.
improved anatomic detail and allows better differentiation The total duration of acquiring the 8 to 12 sections required
between the right ventricular and atrial lumen. Nevertheless, to image the entire ventricular cavity is on the order of
for practical reasons, the right ventricular mass is often mea- 5 minutes (16,17). All sections should be acquired at the
sured using a stack of short-axis slices, which is also used for same end-expiration or end-inspiration phase; otherwise reli-
measuring the left ventricular dimensions. able 3D quantification of volumes from the obtained images
is not possible. In MSCT, the complete volumetric data set
is acquired during a single breath-hold of 10 to 30 seconds.
Quantification of Ventricular Quantitative analysis starts with manual or (semi-)
Volumes and Global Function automated segmentation of the myocardium and blood pool
in the images. Once contours have been defined in the stack
Assessment of global ventricular function requires volumetric of images describing the endocardial and epicardial bound-
measurement of the ventricular cavities in at least two points aries of the myocardium, volumetric measurements includ-
in the cardiac cycle: The end-diastolic and end-systolic phases. ing stroke volume and ejection fraction can be obtained by
20
15
10
Figure 6.5. Comparison of LV dimensions mea-

Precent difference
5
(SSFP - GRE)
sured with an SSFP sequence or segmented gradient-

echo technique. (Data derived from: Alfakih K, Plein 0
S, Thiele H, et al. Normal human left and right
ventricular dimensions for MRI as assessed by turbo −5
gradient echo and steady-state free precession imaging
sequences. J Magn Reson Imaging. 2003;17(3):323– −10
329; Lee VS, Resnick D, Bundy JM, et al. Cardiac
function: MR evaluation in one breath-hold with −15
real-time True Fisp imaging with steady-state preces-
sion. Radiology. 2002;222:835–842; and Wei LI, Stern
−20
JS, Mai VM, et al. MR assessment of left ventricular Philips/TFE (%) GE/FastGRE (%) Siemens/FLASH (%)
function: Quantitative comparison of fast imaging EDV 8.8 9.8 13
employing steady-state acquisition (FIESTA) with fast
LV mass −13.8 −13.9 −16.5
gradient echo cine technique. J Magn Reson Imaging.
2002;16(5):559–564.) EF 3.8 −5.3 −10
LWBK1209-ch06_p093-107.indd 97 17/05/13 4:56 PM

A B
Figure 6.6. Four-chamber long-axis view and three basal level short-axis views acquired within the same
examination. Short-axis and long-axis imaging planes intersect with each other (white lines indicate intersec-
tion of long-axis view with short-axis slices; colored lines indicate intersection of the three short-axis slices
with the long-axis image). The movement of the base toward the apex in systole can easily be appreciated.
Review of how long-axis and short-axis planes intersect may facilitate the interpretation of basal level short-
axis images and may be valuable during tracing of the contours.
a pplying Simpson’s rule. Normal values for global ventricular Quantification of Regional Wall
function and mass have been reported by several investiga- Motion and Wall Thickening
tors for different populations and pulse sequences (18–20). Using the Centerline Method
It is important to note that normal values obtained using the From Dynamic Short-Axis Images
newer SSFP-type sequences differ significantly from values
obtained with previous techniques. The improved contrast The excellent depiction of the endocardial and epicardial
between blood and myocardium in SSFP is associated with boundaries of the left ventricular myocardium forms the
larger end-diastolic and end-systolic cavity volumes, smaller basis of quantitative analysis of regional myocardial func-
wall thickness values, and lower LV mass (20–22). In direct tion. Quantitative analysis methods for endocardial wall
comparisons of SSFP with conventional fast gradient-echo motion are hampered by the presence of rigid body motion
techniques within the same individuals, differences in LV of the heart. A floating centroid, based on the center of grav-
mass of up to 16.5% were reported (Fig. 6.5). ity of the endocardial or epicardial contours, can be used
At the basal imaging sections, a clear visual separation to isolate the rigid body motion from the actual endocar-
between the LV and the left atrium is often absent because dial deformation. On the other hand, quantification of wall
the imaging section may contain both ventricular and atrial thickness and thickening does not have this disadvantage. It
cavity and muscle. It is important to realize that while the has been demonstrated that wall-thickening analysis is more
imaging sections are fixed in space, the left ventricular sensitive in the detection of dysfunctional myocardium than
annulus exhibits a motion in the apical direction on the wall-motion analysis (23,24). The optimal slice orientation
order of 1.3 cm in normal hearts (1). Consequently, myo- for wall-thickness analysis of the LV is the short-axis plane
cardium that is readily visible in an end-diastolic time frame because in this orientation the major part of the myocardial
may be replaced by left ventricular atrium in the end-systolic wall is perpendicular to the imaging section (25–28). Local
time frame. Additional long-axis views may be helpful to wall thickness can be derived in these acquisitions from man-
more reliably analyze the most basal and apical slice levels ually or automatically defined endocardial and epicardial
of a multislice short-axis study (2). Figure 6.6 displays end- boundaries in each short-axis image. Radial wall thickness
diastolic and end-systolic time frames in a long-axis view quantification methods use an approximate center point in
and three basal short-axis sections obtained during a single the LV to measure the distance between the endocardial and
MRI examination. The lines, indicating the intersections epicardial contours along radial lines starting from this cen-
of the imaging planes, provide helpful additional informa- ter point. This approach may result in significant overesti-
tion for interpreting the structures seen in the short-axis mation of wall thickness in case the ventricular cross section
images. deviates much from a circular shape. The centerline method
LWBK1209-ch06_p093-107.indd 98 17/05/13 4:57 PM

A B
Regional wall tickness

18
16
14
Wall thickness
12
10 ED
8 ES Figure 6.7. Midventricular end-diastolic (A) and end-systolic (B)
6 short-axis images of the LV with endocardial and epicardial contours
defined. Wall thickness chords are constructed for measurement of
4
wall thickness in six myocardial segments. Six myocardial segments are
2 defined starting at the posterior junction of the right ventricular wall
0 with the LV. Segments are numbered from one to six in clockwise order.
1 2 3 4 5 6 As shown in (C) wall thickening is present in each of the six defined
C Segment number myocardial segments.
has advantages over the radial methods because it can be method to correct for this error by estimating the local angle
applied for a wide variety of shapes. In fact, it was originally between the imaging plane and the myocardial wall for each
developed for wall-motion analysis in x-ray angiograms and centerline chord. They demonstrated the improved accuracy
later modified for wall-motion and wall-thickness analyses of this method in phantom studies and also showed that the
of left ventricular short-axis views (29,30). As depicted in overestimation of wall thickness near the apex of the heart
Figure 6.7, the centerline method uses a path in between the in short-axis studies can be minimized.
endocardial and epicardial contours (the “centerline”) and
perpendicular to that path at evenly spaced intervals, start-
ing at a clearly visible anatomic reference point, chords are Regional Function Analysis Using
constructed from endocardium to epicardium. The length Magnetic Resonance Imaging
of such a chord represents the local wall thickness, and the Tagging and Velocity-Encoded
ratio between the end-systolic and the end-diastolic chord Magnetic Resonance Imaging
length equals the local end-systolic wall thickening. A suf-
ficient number of chords should be chosen so as to not Wall-thickening and wall-motion analyses from conventional
miss small anatomic abnormalities. Normal values for end- cine MRI have to take into account the through-plane motion
systolic wall thickening can be used for comparison to deter- of the heart. Furthermore, conventional cine MRI only allows
mine which myocardial regions are abnormal; subsequently, quantification of radial myocardial deformation; 3D myocar-
the size, extent, and severity of a wall-thickening abnormality dial tagging is a powerful MRI technique that allows quan-
can be quantified (31). In case the MRI slice is not exactly tification of myocardial strain in all three dimensions. With
perpendicular to the local myocardial orientation, the nor- the use of MRI tagging, cine MR images are acquired with
mal two-dimensional (2D) centerline method may lead to a a superimposed parallel, rectangular, or radial grid of dark
wall thickness overestimation. Buller et al. (32) described a saturation lines. These tagging lines are induced by a special
LWBK1209-ch06_p093-107.indd 99 17/05/13 4:57 PM

prepulse sequence immediately after the R-wave of the electro- on an expected shape or contraction pattern to fail in such
cardiogram and can subsequently be followed over the cardiac circumstances. Ideally, an automated algorithm should be in-
cycle. Dedicated computer algorithms have been developed to sensitive to variations in image characteristics and be appli-
automatically track the intersection points of the tagging lines cable to magnetic resonance images obtained from different
over the cardiac cycle to be able to quantify intramural myo- MRI scanners. In case the algorithm can operate without any
cardial deformations (33). By applying this technique in mul- user-interaction, the actual computation time is not of major
tiple slices in both short- and long-axis directions, 3D strain importance. If user interaction is required to control the al-
measurements can be performed (34). Alternatively, these gorithm, such as providing seed points or initial contours for
measurements can be derived directly by direct processing of each of the imaging slices, the actual algorithm should be
the Fourier spectrum of the tagged MRI data, a technique cur- must faster to increase the time efficiency of the operator.
rently known as harmonic phase imaging (35,36). In the next section, a short description is given of the un-
VEC-MRI may be used to quantitatively assess the 3D derlying methods and validation results from the algorithms
velocity of the myocardium over the cardiac cycle. With this developed at our laboratory, which have been integrated in
acquisition technique the myocardial velocities can be mea- the software package MASS (42). Our contour detection
sured in three orthogonal directions for each pixel within method follows a model-based approach and is directed to
the imaging plane. In contrast with MRI tagging methods, the definition of the endocardial and epicardial contours in
this technique can be used in combination with retrospective all the phases and slices of an imaging study. The amount of
gating such that data over the complete cardiac cycle are ob- user interaction required to obtain reliable contours is lim-
tained. For quantitative analysis, two different approaches ited and is minimal in case the images are of good quality.
can be followed. Motion-tracking techniques, which are The algorithm accommodates for anatomic and MRI-related
based on velocity integration, can be applied to obtain 2D variations in image appearance by providing a certain learn-
or 3D trajectories of myocardial sample points (37). An al- ing behavior. Manually traced or edited contours are assumed
ternative approach is the direct quantification of myocardial to be correct, and the contour detection algorithm was de-
strain rate by calculating the spatial velocity gradient along signed to generate a consistent set of contours for the total
different directions (38). Both approaches are sensitive to image data set using the manually defined contours as mod-
imperfections of the images such as insufficient temporal or els. The contour detection starts by searching for circular ob-
spatial resolution, blood-related artifacts, and beat-to-beat jects in the imaging slices to find the approximate long axis
variations. These problems may well be resolved by future of the LV, which results in an approximate left ventricular
improvements in image acquisition techniques. center point in each image. With this center point, epicardial
contours are found in the first phase and subsequently in the
remaining phases using a frame-to-frame contour detection
Automated Contour Detection procedure. This frame-to-frame epicardial contour detection
in Short-Axis Multislice Cine procedure is based on matching of line profiles that are po-
Magnetic Resonance Imaging sitioned perpendicularly to the model contour (derived from
the first phase) and then automatically positioned at the cor-
Despite the fact that the time required for image acquisition responding tissue transitions in other phases within the same
has been reduced tremendously over the last few years, a slice level. With this approach, the algorithm is able to deal
cardiac evaluation based on cardiovascular magnetic reso- with the fact that the epicardial boundary of the myocardium
nance (CVMR) including quantitative analysis remains time is adjacent to regions having different gray value character-
consuming because of the required postprocessing of the istics. A first estimate of the endocardial contour is found
large amount of images. With a ventricle that is imaged in 10 using an optimal thresholding technique within the region
imaging sections, 20 endocardial contours need to be defined described by the epicardial contour. The final endocardial
for the assessment of basic global function parameters such contour is found by using a model-based edge-detection tech-
as the ejection fraction and stroke volume. Additional epi- nique, known as the minimum cost algorithm (50). Figure 6.8
cardial contours are needed for quantification of left ventric- illustrates the algorithmic steps that are carried out to detect
ular mass. Manual image analysis requires tracing of these an endocardial contour given an image with an available epi-
myocardial outlines and is a time-consuming procedure that cardial contour.
takes between 10 minutes and 1 hour depending on the soft- The automated contour detection algorithm was evaluated
ware used. It also may introduce undesirable interobserver on MRI studies acquired on two different MRI systems of
and intraobserver variabilities. 10 patients with myocardial infarction and 10 healthy volun-
A considerable number of groups, including ours, have teers. In each of these studies, the endocardial and epicardial
contributed to the development of algorithms for the auto- contours were obtained by manual tracing and by using the
mated extraction of the left ventricular myocardial outlines semiautomated contour detection algorithm. The user was al-
from short-axis cine MRI studies (39–49). The development lowed to correct clear failures of the automated contour de-
of automated contour detection algorithms is a challenging tection, which was required for two epicardial contours per
problem because variations in gray value in MRI depend on study on an average. Manual editing of automatically detected
many factors such as the imaging parameters used, spatial de- endocardial contours was not allowed. In addition, the con-
pendency in case surface coils are used, and flow dependency. tours in the most basal slice were obtained by manual tracing.
In addition, the geometry of the cardiac chambers and its con- Manual tracing or editing of a contour was followed by a new
traction pattern may be abnormal in pathologic situations, iteration of contour detection such that a consistent series of
causing automated segmentation methods that rely too much contours was obtained. Manual tracing of the endocardial
LWBK1209-ch06_p093-107.indd 100 17/05/13 4:57 PM

A B C
D E F
Figure 6.8. Automated detection of the endocardial contour. (A) Original image with epicardial contour.
(B) Search region for the endocardial contour. The region outside the epicardial contour and a small region at the
inside of the epicardial contour are masked out from the original image. (C) Result after determination of the opti-
mal threshold. (D) Contour around the thresholded region serves as a starting contour for the subsequent edge-
based contour detection. (E) When papillary muscle needs to be excluded from the myocardium, a smooth convex
hull contour around the initial contour is determined. (F) Final result after minimum cost contour detection.
and epicardial contours in all the phases and slices of an imag- isotropic image resolution of MSCT, contour detection
ing study required 2 to 3 hours; the analysis time was reduced methods that operate on the complete 3D volumetric data
to less than 20 minutes using the semiautomated approach. In are expected to have advantages over multislice 2D meth-
Table 6.1 the random and systematic differences are reported ods. A good example is the use of 3D active shape models
for a number of global left ventricular function parameters. (ASMs) for the detection of the endocardial and epicardial
Some underestimation of ventricular volumes can be observed surfaces in MSCT (51). An ASM is a statistical description
in the group of healthy volunteers. No statistically significant of the shape and shape variation of the heart derived from a
differences between the two analysis methods were found in set of training images. By automatically fitting and deform-
the group of patients with infarction. The random differences ing an ASM, it can be used for contour detection in 3D volu-
are relatively small and on the same order of magnitude for metric MSCT image data (Fig. 6.9).
both study populations.
Automated Segmentation of Multislice

Computed Tomography
Short-axis multiplanar reformatted MSCT images have a lot
of similarities with images obtained with MRI. In practice we
have observed that because of the well-understood physical
properties of computed tomography, there is less varia-
tion in image characteristics between patients and different
MSCT systems. The contour detection algorithm described Figure 6.9. Application of a 3D active shape model (ASM) for the
previously was successfully adapted to operate on short- simultaneous detection of left and right ventricular surfaces in 3D mul-
axis MSCT images. To fully take advantage of the almost tislice computed tomography (MSCT).
LWBK1209-ch06_p093-107.indd 101 17/05/13 4:57 PM

esults of Semiautomated Image Analysis Compared

R
Table 6.1
with Manual Analysis
Healthy Subjects (n = 10) Patients (n = 10) Overall (n = 20)

Parameter Mean SD Mean SD Mean SD
EDV (mL) −13.4a 5.7 2.4 5.7 −5.5 9.7

ESV (mL) −5.9a 4.9 −1.4 7.1 −3.6a 6.5
SV (mL) −7.5a 4.4 3.8 7.6 −1.9a 8.4
EF (%) 1.4 3 2.1 4.9 1.7a 4.1
Mass (g) 22.8a 10.2 −8.2 16.2 7.3a 20.6
Systematic and random differences (auto/manual) in the assessment of left ventricular dimensions and func-
tion parameters using semiautomated or manual image analysis.
a
Statistically significant difference between semiautomated and manually determined parameter (p < 0.05).
EDV, end-diastolic volume; EF, ejection fraction; ESV, end-systolic volume; SD, standard deviation; SV,
stroke volume.
Automated Contour Detection parameters that control the smoothness of the detected
Optimization for Different Magnetic contours, and many others. The automatically detected con-
Resonance Pulse Sequences tours generated using a specific parameter setting are com-
pared with the reference contours by computing the degree
A major challenge when designing and implementing a reli- of similarity between both contour sets. A genetic algorithm
able automated contour detection algorithm is to deal with is used to generate new sets of parameter settings based on
the large variations in image characteristics because of the the results of the previously evaluated sets. By iterating this
differences in magnetic resonance pulse sequences, receiver procedure numerous times, an optimal set of parameters can
coils, and magnetic resonance scanners from different ven- be found for a specific set of images. The degree of simi-
dors. Consequently, for optimal performance, an automated larity, which is used to evaluate the quality of the detected
segmentation method needs to be optimized for a specific contours, is defined as the percentage of contour points that
type of acquisition procedure. We recently developed a con- lie within a 2-mm distance of the corresponding reference
tour detection optimization procedure that enables tuning contour. It was shown that for SSFP type acquisitions, the
the different parameters that control the automated contour degree of similarity for manually traced endocardial con-
detection (52). Figure 6.10 illustrates the mechanism of the tours obtained by repeated analysis of the same observer
contour detection tuning method. On the basis of a set of was 77%. Therefore, because the reference contours are
short-axis examinations with expert drawn reference con- generated manually, the similarity between automatically
tours available, automated contour detection is performed detected contours and the reference contours has a theoretic
using different settings of the contour detection procedure. upper bound of 77%. The described optimization approach
Contour detection settings that are varied are the convolu- was evaluated on a set of 30 SSFP examinations from the
tion kernels that are used for edge detection in the images, three main magnetic resonance scanner vendors to assess
Reference
contours Compute
MR studies
degree of
similarity
Detected
MASS
contours
Figure 6.10. Automated optimization procedure to
find the optimal contour detection settings for a specific
Setting Similarity pulse sequence. In an iterative procedure, MASS per-
1 Sim #1 forms automated contour detection in a set of MRI stud-
ies using a number of different parameter settings. The
2 Sim #2
Genetic detected contours are compared with manually defined
Settings 3 Sim #3
algorithm reference contours, and the average degree of similarity
4 Sim #4 is computed for each parameter setting. A genetic algo-
. . rithm is used to generate new parameter settings based
. . on the results of the parameter settings from the previous
iteration.
LWBK1209-ch06_p093-107.indd 102 17/05/13 4:57 PM

the improvement in the performance of automated contour Therefore, a decision about the exact location of the contour
detection. In all 30 studies endocardial contours were care- cannot always be made on the basis of the strongest image
fully traced in the end-diastolic and end-systolic phases, evidence but should be learned from the examples and pref-
which were used as reference. Automated contour detection erences provided by expert observers.
was performed in all studies with and without optimized set- To overcome these problems, prior knowledge about the
tings. The average degree of similarity was 49.5% when the image appearance, spatial organ embedding, and charac-
unoptimized settings were used, which increased to 63.3% teristic organ shape and its anatomic and pathologic shape
when the optimized settings were used. variations should form an integral part of a contour detec-
tion approach. Moreover, it should be adaptive to accom-
modate the preferences of an observer and to easily adjust
New Automated Segmentation Methods
to image characteristics of various pulse sequences and mag-
Reliable fully automated contour detection, not requiring netic resonance and MSCT systems.
any user interaction, would clearly be an important step to Recently, Cootes et al. (55) introduced the concept of
further improve the clinical utility of CVMR. Despite a lot active appearance models (AAMs), which are trainable math-
of research in this area, two major problems limit the suc- ematic models that can learn the shape and appearance of an
cess rate of many of the previously described contour detec- imaged object from a set of example images. This method
tion strategies for cardiovascular structures. First, because was originally developed for facial recognition and later op-
of the presence of noise and image acquisition artifacts, timized for the detection of the LV in CVMR (56). An AAM
image information can be ill defined, unreliable, or missing. consists of two components: A statistical model of the shape
In these cases, a human observer is still capable of tracing of an object and a statistical model of the image appearance
the myocardial contours in the image data based on experi- of the object. The combined model is trained to learn the
ence and prior knowledge, whereas many automated tech- shape and image structure of an organ from a representa-
niques fail. Second, a contour as drawn by an expert human tive set of example images from different subjects. The AAM
observer may not always correspond to the location of the can be automatically matched to a new study image by mini-
strongest local image evidence. In particular, in short-axis mizing an error function expressing the difference between
images the papillary muscles and trabeculae pose a problem. the model and the underlying image evidence. During this
For example, many experts prefer to draw the left ventricu- matching process, the model is constrained to only resemble
lar endocardial border as a convex hull around the blood statistically plausible shapes and appearances. Consequently,
pool, at a location somewhat “outside” of the strongest edge AAMs are able to capture the association between the ob-
(53,54). A second example is the epicardial boundary, which server preference and the underlying image evidence, mak-
may be embedded in fatty tissue, as a result of which the ing the AAMs highly suitable to model the expert observer’s
edge is strongest at the fat–air transitions. However, often analysis behavior. Moreover, AAMs can model multiple ob-
the contour should be drawn on the inside of this fatty layer, jects (in our case the left and right cardiac ventricles) in their
an intensity transition that is marked by only a faint edge. spatial embedding. In a study by Mitchell et al. (57), this
Figure 6.11. Detection results of

left and right ventricular contours
using the active appearance model
(AAM) contour detection method.
LWBK1209-ch06_p093-107.indd 103 17/05/13 4:57 PM

AAM technique showed excellent agreement with manually contour-detection algorithm was integrated in the FLOW
defined contours, both for the LV and RV simultaneously. software package.
Figure 6.11 shows examples of automatically detected con- The only user interaction required is the manual defini-
tours for the LV and RV obtained by this new approach. tion of an approximate center in one of the available im-
van der Geest et al. (58) investigated the value of incor- ages. In this first image an initial model contour is detected
porating image information of complete time series in an using gray value and edge information. The position of the
AAM-based contour detection method. The advantage of same vessel at another time frame can be estimated by shift-
this approach lies in the fact that information from a coming the model contour in a limited region around the ini-
plete time series is used during training and detection, which tial location and examining the edge values measured in the
results in consistent time-continuous segmentation results, modulus image along the contour points. An algorithm was
even in the presence of image frames with poor image quality. developed that finds the most likely contour position for
each time frame, with the restriction that a contour is only
allowed to displace 2 pixels (1.6 mm) from phase to phase,
Magnetic Resonance Imaging thereby imposing a temporal continuity of the motion. After
the correct contour location was found, a final optimized
Flow Quantification
contour was detected by allowing small deformations of the
model contour such that it would follow the edges in the
VEC-MRI also plays an important role in the evaluation
modulus image. For this purpose a 2D graph searching tech-
of global ventricular function. The accuracy of this imag-
nique was used. The resulting contour was dilated by 1 pixel
ing technique has been demonstrated in in vitro experiments
to be sure to encompass the complete region with flowing
using flow phantoms and comparison against other imaging
blood. The total contour detection process takes less than
techniques such as Doppler echocardiography and invasive
10 seconds for a study with 30 cardiac phases.
oximetry (59,60).
Validation was performed on flow velocity maps from a
Because flow measurements are obtained at high tempo-
study population of 12 healthy volunteers. Two independent
ral resolution over the complete cardiac cycle, VEC-MRI is
observers performed manual and automated image analy-
especially useful in the evaluation of left and right ventric-
ses. The first observer repeated the automated and manual
ular diastolic function parameters by measuring flow over
analyses after a 2-week interval to avoid learning effects.
the atrioventricular valves. Application of this technique to
The time required for manual analysis was 5 to 10 min-
the proximal portion of the ascending aorta or pulmonary
utes. During automated analysis the user had to identify the
artery allows the assessment of left and right ventricular sys-
approximate location of the center of the aorta in one of
tolic functions. After the cross section of a vessel is identi-
the available images. The total analysis time for automated
fied in the image by manual or automated contour detection,
analysis was less than 10 seconds. Stroke volume measure-
the instantaneous flow rate within the vessel cross section is
ments were obtained by integrating the flow over the com-
obtained by multiplying the average velocity within the con-
plete cardiac cycle. The mean left ventricular stroke volume
tour by its area. Ventricular stroke volume measurements
obtained by VEC-MRI in the group of 12 volunteers was
are derived by integrating the flow over a complete cardiac
86.4 mL (standard deviation (SD): 13.6 mL). No statisti-
cycle (61). The presence of aortic or pulmonary regurgita-
cally significant differences were found between the results
tion can be easily identified and quantified from the derived
of manual and automated analyses. The mean difference be-
flow curve. VEC-MRI has an established role in the evalua-
tween automated and manually assessed stroke volume was
tion of patients with congenital heart disease (62,63).
0.78 mL (SD: 1.99 mL). The intraobserver variability was
0.65 mL for manual analysis and 0.58 mL for automated
analysis; the intraobserver variability was 0.99 mL for man-
Automated Quantification ual analysis and 0.9 mL for automated analysis. From this
of Aortic Flow study, it can be concluded that the automated contour detec-
tion algorithm performs equally well as the manual method
Application of VEC-MRI to the proximal portion of the in the determination of left ventricular stroke volume de-
ascending aorta allows the assessment of left ventricular sys- rived from VEC-MRI studies of the ascending aorta.
tolic function by evaluating the flow over a complete car-
diac cycle. Such a study requires a VEC-MRI acquisition
in the transversal plane crossing the ascending aorta. The Delayed Contrast-Enhanced
left ventricular stroke volume can be measured by integrat- Magnetic Resonance Imaging
ing the flow over a complete cardiac cycle. For an accurate
assessment of volume flow, contours describing the lumen of Delayed contrast-enhanced MRI has become part of a stan-
the vessels have to be obtained in the images. The in-plane dard MRI examination because it is extremely valuable
motion of the greater vessels and changes in shape of the for the assessment of viable and nonviable myocardium in
vessel cross section over the cardiac cycle would require the infarcted and poor contractile areas (65,66). The excellent
user to trace the luminal border of the vessel in each indi- resolution of MRI enables the depiction of both transmural
vidual phase of the MRI examination. To overcome these and nontransmural regions of infarction. It was shown that
practical limitations, an automated analysis method was the transmural extent of enhancement is inversely related
developed in our department to automatically detect the to the likelihood of recovery of function after revascular-
required contours in each of the cardiac phases (64). This ization. Therefore, large nontransmural infarcts may have
LWBK1209-ch06_p093-107.indd 104 17/05/13 4:57 PM

Figure 6.12. Assessment of regional

myocardial scar transmurality from a
multislice short-axis delayed contrast-
enhanced acquisition (A). After defining
an appropriate intensity threshold, the
regional transmurality of scar can be
computed and visualized using a bull’s-
eye plot (B). This patient has a large
infarction with complete transmural
enhancement in the lateral region and
subendocardial enhancement in the pos-
B
terior region.
a better prognosis than relatively small transmural infarcts. Half Maximum criterion to obtain a threshold value. Amado
Quantification of the size of the infarction involves defin- et al. demonstrated in an animal experimental study that
ing a signal intensity threshold that separates normal myo- myocardial infarct size measurements using a Full Width
cardium from enhanced tissue. One approach to define this Half Maximum criterion agreed very well with pathology.
threshold is to assess the mean and SD of the signal inten- With the defined threshold, the enhanced regions within the
sity in a remote normal region and use a number of SDs myocardium are objectively defined and the regional degree
above the mean as the threshold value (65,67). Although of transmurality can be defined as illustrated in Figure 6.12.
this method has shown to work in a specific setting, it highly When delayed contrast-enhanced MRI is combined with cine
depends on the characteristics of the actual images. Schuijf MRI information, infarct transmurality can be related to the
et al. (68) and Amado et al. (69) propose to use a Full Width regions with poor contraction. This enables quantification
LWBK1209-ch06_p093-107.indd 105 17/05/13 4:57 PM

4. Florentine MS, Grosskreutz CJ, Chang W, et al. Measurement of left ventricular mass
in vivo using gated nuclear magnetic resonance imaging. J Am Coll Cardiol. 1986;8:
107–112.
5. Maddahi J, Crues J, Berman DS, et al. Noninvasive quantification of left ventricular mass
by gated proton magnetic resonance imaging. J Am Coll Cardiol. 1987;10(3):682–692.
6. Matheijssen NA, Baur LH, Reiber JH, et al. Assessment of left ventricular volume and mass
by cine-magnetic resonance imaging in patients with anterior myocardial infarction intra-
observer and inter-observer variability on contour detection. Int J Card Imaging. 1996;12:
11–19.
7. Niwa K, Uchishiba M, Aotsuka H, et al. Measurement of ventricular volumes by cine
magnetic resonance imaging in complex congenital heart disease with morphologically
abnormal ventricles. Am Heart J. 1996;131:567–575.
8. Jauhiainen T, Järvinen VM, Hekali PE, et al. MR gradient echo volumetric analysis of
the human cardiac casts: Focus on the right ventricle. J Comput Assist Tomogr. 1998;
22(6):899–903.
9. Alfakih K, Thiele H, Plein S, et al. Comparison of right ventricular volume measurement
between segmented k-space gradient-echo and steady-state free precession magnetic reso-
nance imaging. J Magn Reson Imaging. 2002;16(3):253–258.
10. Semelka RC, Tomei E, Wagner S, et al. Normal left ventricular dimensions and function:
Interstudy reproducibility of measurements with cine MR imaging. Radiology. 1990;174:
763–768.
11. Dirksen MS, Bax JJ, de Roos A, et al. Usefulness of dynamic multislice computed tomogra-
phy of left ventricular function in unstable angina pectoris and comparison with echocar-
diography. Am J Cardiol. 2002;90:1157–1160.
12. Mahnken AH, Spuentrup E, Niethammer M, et al. Quantitative and qualitative assessment
of left ventricular volume with ECG-gated multislice spiral CT: Value of different image
reconstruction algorithms in comparison to MRI. Acta Radiol. 2003;44(6):604–611.
13. Juergens KU, Grude M, Maintz D, et al. Multi-detector row CT of left ventricular func-
tion with dedicated analysis software versus MR imaging: initial experience. Radiology.
2004;230:403–410.
14. Dulce MC, Mostbeck GH, Friese KK, et al. Quantification of left ventricular volumes and
function with cine MR imaging: Comparison of geometrical models with three-dimen-
sional data. Radiology. 1993;188:371–376.
15. Chuang ML, Hibberd MG, Salton CJ, et al. Importance of imaging method over imaging
Figure 6.13. Data fusion between wall thickening information modality in noninvasive determination of left ventricular volumes and ejection fraction:
Assessment by two- and three-dimensional echocardiography and magnetic resonance
derived from cine MR and scar transmurality derived from delayed
imaging. J Am Coll Cardiol. 2000;35(2):477–484.
contrast-enhancement MRI visualized using bull’s-eye displays.The 16. Sakuma H, Fujia N, Foo TK, et al. Evaluation of left ventricular volume and mass with
red area in the bottom bull’s-eye represents the nonviable myocardium breath-hold cine MR imaging. Radiology. 1993;188:377–380.
where the wall thickening where is <2 mm while scar transmurality 17. Lamb HJ, Singleton RR, van der Geest RJ, et al. MR imaging of regional cardiac function:
Low-pass filtering of wall thickness curves. Magn Res Med. 1995;34:498–502.
is >50%.
18. Lorenz CH, Walker ES, Morgan VL, et al. Normal human right and left ventricular mass,
systolic function and gender differences by cine magnetic resonance imaging. J Cardiovasc
Magn Reson. 1999;1:7–21.
of the location and size of regions of poor contraction that 19. Rominger MB, Bachmann GF, Pabst W, et al. Right ventricular volumes and ejection frac-
are still viable (Fig. 6.13). tion with fast cine MR imaging in breath-hold technique: Applicability, normal values from
52 volunteers, and evaluation of 325 adult cardiac patients. J Magn Reson Imaging. 1999;
10(6):908–918.
20. Alfakih K, Plein S, Thiele H, et al. Normal human left and right ventricular dimensions
Summary for MRI as assessed by turbo gradient echo and Steady-State Free Precession imaging
sequences. J Magn Reson Imaging. 2003;17(3):323–329.
21. Lee VS, Resnick D, Bundy JM, et al. Cardiac function: MR evaluation in one breath
Cardiovascular MRI and MSCT are valuable techniques for hold with real-time True Fisp imaging with steady-state precession. Radiology. 2002;222:
835–842.
noninvasive quantitative assessment of global and regional 22. Wei LI, Stern JS, Mai VM, et al. MR assessment of left ventricular function: Quantitative
ventricular functions. Computed tomography analysis tech- comparison of fast imaging employing steady-state acquisition (FIESTA) with fast gradient
niques can help reduce the time required for quantifica- echo cine technique. J Magn Reson Imaging. 2002;16(5):559–564.
23. Lieberman An, Weiss JL, Jugdutt BI, et al. Two-dimensional echocardiography and infarct
tion and interpretation of the many images. In this chapter, size: Relationship of regional wall motion and thickening to the extent of myocardial
analytic methods for left ventricular function and vascular infarction in the dog. Circulation. 1981;63(4):739–746.
24. Azhari H, Sideman S, Weiss JL, et al. Three-dimensional mapping of acute ischemic regions
flow measurements based on automated contour detection using MRI: Wall thickening versus motion analysis. Am J Physiol. 1990;259:H1492–H1503.
approaches have been described. Validation studies of these 25. van Rugge FP, van der Wall EE, Spanjersberg SJ, et al. Magnetic resonance imaging during
methods have confirmed their accuracy, precision, robust- dobutamine stress for detection of coronary artery disease; quantitative wall motion analy-
sis using a modification of the centerline method. Circulation. 1994;90:127–138.
ness, and usefulness for clinical research studies. Fully auto- 26. Haag UJ, Maier SE, Jakob M, et al. Left ventricular wall thickness measurements by mag-
mated contour detection methods that operate reliably in a netic resonance: A validation study. Int J Card Imaging. 1991;7:31–41.
routine clinical environment are needed and may become 27. Baer FM, Smolarz K, Theissen P, et al. Regional 99mTc-methoxyisobutyl-isonitrile-uptake
at rest in patients with myocardial infarcts: Comparison with morphological and func-
available in the near future. tional parameters obtained from gradient-echo magnetic resonance imaging. Eur Heart J.
1994;15:97–107.
28. Holman ER, Vliegen HW, van der Geest RJ, et al. Quantitative analysis of regional left
ventricular function after myocardial infarction in the pig assessed with cine magnetic reso-
References nance imaging. Magn Reson Med. 1995;34:161–169.
29. Sheehan FH, Bolson EL, Dodge HT, et al. Advantages and applications of the centerline
1. Rogers WJ Jr, Shapiro EP, Weiss JL, et al. Quantification and correction for left ventricu- method for characterizing regional ventricular function. Circulation. 1986;74:293–305.
lar systolic long-axis shortening by magnetic resonance tissue tagging and slice isolation. 30. von Land CD, Rao SR, Reiber JHC. Development of an improved centerline wall motion
Circulation. 1991;84:721–731. model. Comput Cardiol. 1990;687–690.
2. Marcus JT, Götte MJ, de Waal LK, et al. The influence of through-plane motion on left ven- 31. Holman ER, Buller VG, de Roos A, et al. Detection and quantification of dysfunctional
tricular volumes measured by magnetic resonance imaging: Implications for image acquisi- myocardium by magnetic resonance imaging: A new three-dimensional method for quanti-
tion and analysis. J Cardiovasc Magn Res. 1998;1(1):1–6. tative wall-thickening analysis. Circulation. 1997;95:924–931.
3. Bloomer TN, Plein S, Radjenovic A, et al. Cine MRI using steady state free precession in 32. Buller VG, van der Geest RJ, Kool MD, et al. Assessment of regional left ventricular wall
the radial long axis orientation is a fast accurate method for obtaining volumetric data of parameters from short-axis MR imaging using a 3D extension to the improved centerline
the left ventricle. J Magn Reson Imaging. 2002;14(6):685–692. method. Invest Radiol. 1997;32(9):529–539.
LWBK1209-ch06_p093-107.indd 106 17/05/13 4:57 PM

33. Guttman MA, Prince JL, McVeigh ER. Tag and contour detection in tagged MR images of 52. Angelie A, de Koning PJH, Danilouchkine M, et al. Optimizing the automated segmenta-
the left ventricle. IEEE Trans Med Imaging. 1993;13(1):74–88. tion of the left ventricle in magnetic resonance images. Med Phys. 2005;32(2):1–7.
34. Moore CC, McVeigh ER, Zerhouni EA. Quantitative tagged magnetic resonance imaging 53. Pattynama PMT, Lamb HJ, van der Velde EA, et al. Left ventricular measurements with
of the normal human left ventricle. Top Magn Reson Imaging. 2000;11(6):359–371. cine and spin-echo MR imaging: A study of reproducibility with variance component anal-
35. Garot J, Bluemke DA, Osman NF, et al. Fast determination of regional myocardial strain ysis. Radiology. 1993;187:261–268.
fields from tagged cardiac images using harmonic phase MRI. Circulation. 2000;101(9): 54. Lamb HJ, Doornbos J, van der Velde EA, et al. Echo-planar MRI of the heart on a standard
981–988. system: Validation of measurement of left ventricular function and mass. J Comput Assist
36. Osman NF, McVeigh ER, Prince JL. Imaging heart motion using harmonic phase MRI. Tomogr. 1996;20(6):942–949.
IEEE Trans Med Imaging. 2000;19(3):186–202. 55. Cootes TF, Beeston C, Edwards GJ, et al. A unified framework for atlas matching using
37. Pelc NJ, Drangova M, Pelc LR, et al. Tracking of cyclic motion with phase-contrast cine active appearance models. LNCS. 1999;1613:322–333.
MR velocity data. J Magn Reson Imaging. 1995;5:339–345. 56. Mitchell SC, Lelieveldt BP, van der Geest RJ, et al. Multistage hybrid active appearance
38. Hennig J, Schneider B, Peschl S, et al. Analysis of myocardial motion based on velocity model matching: Segmentation of left and right ventricles in cardiac MR images. IEEE
measurements with a black blood prepared segmented gradient-echo sequence: Methodology Trans Med Imaging. 2001;20(5):415–423.
and applications to normal volunteers and patients. J Magn Reson Imaging. 1998;8(4): 57. Mitchell SC, Lelieveldt BP, van der Geest RJ, et al. Segmentation of cardiac MR images: An
868–877. active appearance model approach. Proc. SPIE Medical Imaging. 2000;3979:224–234.
39. McInerney T, Terzopoulos D. A dynamic finite element surface model for segmentation 58. van der Geest RJ, Lelieveldt BP, Angelie A, et al. Evaluation of a new method for auto-
and tracking in multidimensional medical images with application to cardiac 4D image mated detection of left ventricular boundaries in time series of magnetic resonance images
analysis. Comput Med Imaging Graph. 1995;19:69–83. using an active appearance motion model. J Cardiovasc Magn Reson. 2004;6(3):609–617.
40. Goshtasby A, Turner DA. Segmentation of cardiac cine MR images for right and left ven- 59. Karwatowski SP, Brecker SJ, Yang GZ, et al. Mitral valve flow measured with cine MR
tricular chambers. IEEE Trans Med Imaging. 1995;14:56–64. velocity mapping in patients with ischemic heart disease: Comparison with Doppler echo-
41. Baldy C, Doueck P, Croisille P, et al. Automated myocardial edge detection from breath- cardiography. J Magn Reson Imaging. 1995;5:89–92.
hold cine-MR images: Evaluation of left ventricular volumes and mass. Magn Reson 60. Beerbaum P, Korperich P, Barth P, et al. Noninvasive quantification of left-to-right shunt in
Imaging. 1994;12:589–598. pediatric patients. Phase-contrast cine magnetic resonance imaging compared with invasive
42. van der Geest RJ, Buller VG, Jansen E, et al. Comparison between manual and automated oximetry. Circulation. 2001;103, 2476–2482.
analysis of left ventricular volume parameters from short axis MR images. J Comput Assist 61. Kondo C, Caputo GR, Semelka R, et al. Right and left ventricular stroke volume measure-
Tomogr. 1997;21(5):756–765. ments with velocity-encoded cine MR imaging: In vitro and in vivo validation. AJR Am
43. Singleton HR, Pohost GM. Automatic cardiac MR image segmentation using edge detec- J Roentgenol. 1991;157:9–16.
tion by tissue classification in pixel neighborhoods. Magn Reson Med. 1997;37:418–424. 62. de Roos A, Helbing WA, Niezen RA, et al. Magnetic resonance imaging in adult con-
44. Furber A, Balzer P, Cavaro-Menard C, et al. Experimental validation of an automated genital heart disease. In: Higgins CB, Ingwall JS, Pohost GM, eds. Current and Future
edge-detection method for a simultaneous determination of the endocardial and epicardial Applications of Magnetic Resonance in Cardiovascular Disease. Armonk, NY: Futura
borders in short-axis cardiac MR images: Application in normal volunteers. J Magn Reson Publishing Company, Inc.; 1988:163–172.
Imaging. 1998;8:1006–1014. 63. Powel AJ, Geva T. Blood flow measurement by magnetic resonance imaging in congenital
45. Nachtomy E, Cooperstein R, Vaturi M, et al. Automatic assessment of cardiac function heart disease. Pediatr Cardiol. 2000;21:47–58.
from short-axis MRI: Procedure and clinical evaluation. Magn Reson Imaging. 1998;16(4): 64. van der Geest RJ, Niezen RA, van der Wall EE, et al. Automated measurement of volume
365–376. flow in the ascending aorta using MR velocity maps: Evaluation of inter- and interobserver
46. Üzümcü M, van der Geest RJ, Swingen C, et al. Time continuous tracking and segmenta- variability in healthy volunteers. J Comput Assist Tomogr. 1998;22(6):904–911.
tion of cardiovascular magnetic resonance images using multidimensional dynamic pro- 65. Kim RJ, Fieno DS, Parrish TH, et al. Relationship of MRI delayed contrast enhancement
gramming. Invest Radiol. 2006;41:52–62. to irreversible injury, infarct age, and contractile function. Circulation. 1999;100(19):
47. Codella NC, Weinsaft JW, Cham MD, et al. Left ventricle: Automated segmentation by 1992–2002.
using myocardial effusion threshold reduction and intravoxel computation at MR imaging. 66. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance imaging
Radiology. 2008;248:1004–1012. to identify reversible myocardial dysfunction. New Engl J Med. 2000;343:1445–1453.
48. Cocosco CA, Niessen WJ, Netsch T, et al. Automatic image-driven segmentation of the 67. Oshinski JN, Yang ZQ, Jones JR, et al. Imaging time after Gd-DTPA injection is critical in
ventricles in cardiac cine MRI. J Magn Reson Imaging. 2008;28:366–374. using delayed enhancement to determine infarct size accurately with magnetic resonance
49. Lynch M, Ghita O, Whelan PF. Segmentation of the left ventricle of the heart in 3-D+t imaging. Circulation. 2001;104(23):2838–2842.
MRI data using an optimized nonrigid temporal model. IEEE Trans Med Imaging. 68. Schuijf JD, Kaandorp TA, Lamb HJ, et al. Quantification of myocardial infarct size and
2008;27(2):195–203. transmurality by contrast-enhanced magnetic resonance imaging in men. Am J Cardiol.
50. Amini AA, Weymouth TE, Jain RC. Using dynamic programming for solving variational 2005;94(3):284–288.
problems in vision. IEE Trans PAMI. 1990;12(9):855–867. 69. Amado LC, Gerber BL, Gupta VK, et al. Accurate and objective infarct sizing by contrast-
51. van Assen HC, Danilouchkine MG, Behloul F, et al. Cardiac LV segmentation using a 3D enhanced magnetic resonance imaging in a canine myocardial infarction model. J Am Coll
active shape model driven by fuzzy inference. LNCS. 2003;2878:533–540. Cardiol. 2004;44(12):2383–2389.
LWBK1209-ch06_p093-107.indd 107 17/05/13 4:57 PM

Chapter
7
Harrie C.M. van den Bosch
Jos J.M. Westenberg
Albert de Roos
CVMR: Imaging Planes

and Anatomy
■■ Cardiac Axis Imaging Planes high-resolution imaging without using ionizing radiation.
Furthermore, the morphology of the right ventricle (RV) is
■■ Body Axes Imaging Planes
excellent delineated by CVMR, whereas in echocardiogra-
■■ Anatomy phy the assessment of RV geometry and function is chal-
lenging because of the particular crescentic shape of the RV
■■ (Ultra-) High-Field Imaging
wrapping around the left ventricle (LV) (2).
The choice for a specific scan protocol is mainly deter-
Cardiac magnetic resonance (CVMR) imaging has the abil- mined by the diagnostic question which has to be answered. In
ity to provide arbitrary views of the cardiac structures which CVMR imaging, both static and dynamic images of the heart
can be chosen freely since this modality is not hampered by can be acquired. Therefore, it is important to be adequately
the availability of acoustic windows, as in echocardiogra- informed by the referring clinician prior to the CVMR exami-
phy. Even though echocardiography, x-ray LV angiography, nation. Standardized nomenclature for cross-sectional anatomy
and cardiac computed tomography (CT) are nowadays com- has been described (3), facilitating comparison between differ-
monly used techniques for evaluating cardiac disease in clin- ent techniques and proper communication in clinical practice.
ical practice, CVMR has evolved to become the preferential The 17-segment model of the LV, proposed by the American
technique for anatomic and functional cardiac imaging. Heart Association (AHA), is nowadays widely used and ac-
Two-dimensional (2D), single-plane, multiple-2D, or cepted in clinical CVMR imaging, as in other cross-sectional
three-dimensional (3D) imaging is possible with CVMR. imaging modalities (e.g., cardiac CT and nuclear techniques).
Furthermore, temporal information of the dynamics of the The recommended model comprehends six basal segments, six
heart can be provided as imaging is synchronized to the car- mid-ventricular segments, four apical (distal) segments, and
diac frequency, using either prospective triggering or retro- one true apex (Fig. 7.1). These 17 segments are routinely evalu-
spective gating (1). With prospective triggering, the operator ated when regional LV performance is questioned.
will set the expected heart rate before the acquisition and Another important issue in clinical CVMR imaging is the
triggering will be performed according to this chosen heart ability of the patient to cooperate during the examination
rate. With retrospective gating, imaging is performed con- and to perform breath-holding. If a patient is capable to
tinuously and additionally the ECG signal will be stored. perform breath-holding, successive scan planes are obtained
In retrospect, k-space filling is synchronized to the stored with accelerated imaging, with the patient usually perform-
ECG. This synchronization enables time-resolved imaging ing breath-holding in expiration. Preferably, image planes in
and multiple phases of the cardiac cycle can be obtained. CVMR imaging are acquired in mid- or end-expiration, as
The acquired views in multiple phase of the heart can be the anatomic level may be obtained more reproducible com-
presented in cine mode, providing functional information on pared to planes which are scanned in inspiration (4).
the temporal behavior of the cardiac structures. For planning purposes, new generation clinical MR scan-
Imaging planes in CVMR are usually obtained in the ori- ners provide the possibility to plan the various scan planes
entation to the axes of the heart, or oriented to the major interactively with real-time imaging. During free-breathing
axes of the body. Therefore, the standard CVMR planes CVMR, planning can be performed accurately. After all scan
of the heart are comparable to the standard cardiac views planes are defined and obtained interactively, the acquisition
known and established in other noninvasive imaging modali- of the cine long-axis (LA) and short-axis (SA) views may
ties such as echocardiography, cardiac CT, x-ray LV angiog- be performed during breath-holding. This interactive ap-
raphy, and nuclear techniques (e.g., single-photon emission proach for planning is fast, reliable, and patient friendly, and
computed tomography [SPECT] and positron emission to- essential in patients who are not capable to hold their breath
mography [PET]). Compared to cardiac CT, x-ray angiog- consecutively, for example, patients with respiratory dis-
raphy, and nuclear techniques, CVMR allows noninvasive, ease or with heart failure. Fully automated CVMR planning
108
LWBK1209-ch07_p108-115.indd 108 18/05/13 2:47 PM

Chapter 7 ■ CVMR: Imaging Planes and Anatomy 109
A B
A B
Figure 7.2. Black-blood (A) and bright-blood (B) short-axis acqui-
sition illustrating a cardiac sarcoma (arrow heads) in the inferior wall,
acquired in a 19-year-old female.
cation. Typically, SSFP images should be acquired with slice

thickness of 6 to 8 mm and temporal resolution better than
45 milliseconds to obtain optimal accuracy in ventricular
function assessment (6,7).
In addition, cardiac morphology can be evaluated by
double-inversion, black-blood, spin-echo sequences with fat
C D suppression, providing static images of the heart with high
spatial resolution (optimally, in-plane acquired resolution of
better than 2 × 2 mm and slice thickness of 5 to 8 mm) in the
orientation of the heart or the patient’s body axes. These im-
ages are, for example, obtained in the work-up of congenital
heart disease or cardiac tumors (Fig. 7.2).
In the remainder of this chapter, the planning of the spe-
cific imaging planes will be discussed, as well as the normal
cardiac structures that are visualized. Furthermore, the as-
pects of cardiac imaging on (ultra-) high-field MRI will be
addressed.
E F
Figure 7.1. The standardized 17 segments of LV as proposed by the

Cardiac Axis Imaging Planes
American Heart Association. At basal (B) and mid-ventricular (D) level,
the myocardium is divided into six segments each, and at apical (F) level, To acquire imaging planes in the direction of the cardiac
the myocardium is divided into four segments. Nomenclature: Segment axes, multi-stack, single-shot SSFP scout views are used for
1, basal anterior; 2, basal anteroseptal; 3, basal inferoseptal; 4, basal planning. If available, free-breathing real-time scanning can
inferior; 5, basal inferolateral; 6, basal anterolateral; 7, mid-anterior; be used instead that is advantageous for patient’s comfort.
8, mid-anteroseptal; 9, mid-inferoseptal; 10, mid-inferior; 11, mid-inferolat- Perpendicular to an anatomical transverse image, display-
eral; 12, mid-anterolateral; 13, apical anterior; 14, apical septal; 15, apical ing the heart’s four chambers, an acquisition plane is cho-
inferior; 16, apical lateral. Segment 17 (not presented) is the true apex, sen through the middle of the atrioventricular junction at
which can be evaluated on a long-axis view. Dashed lines on the four-cham- the level of the mitral valve and running through the apex
ber view (A,C and E) indicate the planning of the acquisition level.
(Fig. 7.3). This plane is the so-called vertical long-axis (VLA)
plane. On this VLA view, a plane is defined intersecting the
ethods have also been described and can provide accurate
m apex and the middle of the mitral valve, resulting in the hori-
and reproducible measurements of LV dimensions (5). zontal long-axis (HLA) view. The HLA view is almost com-
With CVMR, the choice of scanning technique is aimed parable to the four-chamber view; however, in this HLA view
at the choice between bright-blood and black-blood im- often a part of the LV outflow tract (LVOT) is visualized. On
aging, which essentially determines the contrast between the acquired HLA plane, the SA views covering the entire LV
myocardium and the intra-cardiac blood pool. For the as- are planned parallel to the ring of the mitral valve and per-
sessment of left and right ventricular function, fast gradient- pendicular to the line intersecting the apex.
echo sequences are usually performed in combination with For reasons of reproducibility and comparison, the true
steady-state free precession (SSFP) technique (balanced- two- and four-chamber view can still be obtained. The two-
TFE, True-FISP, Fiesta) for optimal contrast. On these im- chamber view is planned perpendicular to the anterior and
ages, the blood pool is presented with bright signal whereas inferior wall of the LV through the center of the LV cav-
the myocardium is represented dark with low signal. This ity on a mid-ventricular SA image intersecting the apex. On
results in an excellent definition of the LV endocardial and the two-chamber view, the apex, anterior and inferior wall
epicardial borders, which is required for accurate image of the LV, the mitral valve, and left atrium can be analyzed
segmentation during cardiac volume and function quantifi- (Fig. 7.4A). The four-chamber view is planned also on
LWBK1209-ch07_p108-115.indd 109 16/05/13 9:42 PM

A B
Figure 7.3. Planning acquisition of standard

cardiac views. On two transverse slices (A) and
(B), the left ventricular vertical long-axis (VLA)
E F (C) is planned by a plane transecting the mitral
valve and the apex. The horizontal long-axis
(HLA) (D) is obtained by acquiring a plane
transecting the VLA through the mitral valve
and apex. A short-axis image can be obtained
perpendicular to HLA, at mid-ventricular (E)
and basal level (F). The four-chamber (G) of the
LV is obtained as indicated from a plane tran-
secting both LV and RV. The two-chamber (H)
of the LV is acquired perpendicular to the four-
chamber. The three-chamber LV (I) is obtained
G H I from a plane transecting the LV through the
LVOT.
Figure 7.4. Normal cardiac anatomy on

two-(A), four-(B), and three-chamber (C) views.
LA, left atrium; LV, left ventricle; MV, mitral
valve; TV, tricuspid valve; RV, right ventricle;
RA, right atrium; pm, papillary muscle; AV,
A B C
Vein aortic valve; Ao, aorta.
LWBK1209-ch07_p108-115.indd 110 16/05/13 9:42 PM

A B
C D
Figure 7.5. Bright-blood acquisition of the right ventricle, illustrat-
ing the right ventricular outflow tract (RVOT). PA, pulmonary artery;
RVOT, right ventricular outflow tract; Ao, aorta; RA, right atrium;
RV, right ventricle.
a mid-ventricular SA image by a plane through the cen-

ter of the LV cavity and the acute margin of the RV, also
intersecting the apex. The four-chamber view depicts the
inferior interventricular septum, the anterior lateral wall of
the LV, the free wall of the RV, left and right atrium as the
interatrial septum, and both the mitral and tricuspid valves E F
(Fig. 7.4B).
Routinely, the three-chamber or the so-called LVOT view is Figure 7.6. Normal cardiac anatomy on black-blood and bright-
blood acquisitions, in transverse (A and B), sagittal (C and D), and
planned perpendicular to a basal SA plane. This view also in-
coronal (E and F) views. RV, right ventricle; LV, left ventricle; RA,
tersects the apex. The LVOT view (Fig. 7.4C) depicts the apex, right atrium; LA, left atrium; Ao, aorta; rPA, right pulmonary artery;
the anterior interventricular wall, the LVOT, the inferior lateral Ao Asc, ascending aorta; LA, left atrium; RV, right ventricle; Ao Desc,
wall, as the aortic and mitral valve, respectively. The standard descending aorta; PA, pulmonary artery; AV, aortic valve.
SSFP cine CVMR protocol for assessing LV function should in-
clude the two-, four-, and three-chamber views in combination
CT. Black- and bright-blood sequence approaches (Fig. 7.6)
with SA images covering the entire LV, resulting in scans cover-
can be used in optimally adjusted planes to answer specific
ing all described 17 LV segments in two directions.
clinical questions. These imaging sequences provide static
In addition, the RV outflow tract can be obtained. This
images in single phase, not suitable for quantification of LV
view can be planned on a coronal image, depicting the out-
or RV dimensions, as end-diastolic diameters or wall thick-
flow tract of the RV. Alternatively, an optimized view of the
ness. For this analysis, SSFP multiphase images with appro-
RV outflow tract view can be obtained from a plane outlin-
priate temporal resolution are more suitable.
ing the tricuspid valve plane and the outflow tract. On this
Transversely oriented planes (Fig. 7.7) are especially use-
plane, the outflow tract, pulmonary valve, tricuspid valve,
ful for the evaluation of thoracic vascular structures as the
and the basal (diaphragmatic) part of the RV wall are all
ascending and descending thoracic aorta, the superior and
visualized (Fig. 7.5).
inferior vena cava, the pulmonary trunk, and right and left
pulmonary artery. The right and left pulmonary veins (PVs)
Body Axes Imaging Planes entering the left atrium are also well depicted. Images in
transverse orientation through the heart allow studying mor-
For the evaluation of cardiac morphology, the pericardium, phology of the ventricles and atria, as their internal relation-
the great thoracic vessels, and (para-) cardiac masses imag- ship. Also the RV free wall, the RV outflow tract (infundibu-
ing planes can be used oriented to the main body axes. lum), the pericardium, and mediastinum are well depicted.
Also, the transverse (or axial), coronal (frontal), and sagittal It has been described that RV volume and function quantifi-
planes are well known to surgeons and other clinicians, as cation by planimetry can be performed more accurately on
these anatomic orientations are similar to clinical (cardiac) transversely oriented images instead of SA images (8).
LWBK1209-ch07_p108-115.indd 111 16/05/13 9:42 PM

A B C
Figure 7.7. Normal cardiac anatomy on

transverse black-blood acquisitions. SVC, supe-
rior vena cava; T, trachea; Ao Asc, ascending
aorta; Ao Desc, descending aorta; C, carina;
rPA, right pulmonary artery; PA, pulmonary
D E F artery; lPA, left pulmonary artery; LA, left
atrium; RVOT, right ventricular outflow tract;
LA, left atrium; laa, left atrial appendage; AV,
aortic valve; RV, right ventricle; RA, right
atrium; TV, tricuspid valve; LV, left ventricle;
pm, papillary muscle; MV, mitral valve; LAD,
left arterial descending coronary artery; RCA,
right coronary artery; cs, coronary sinus; ct,
G H I
crista terminalis; Es, esophagus; pc, pericard.
Coronal or frontal anatomic views can be very instruc- Anatomy

tive to analyze the connection between the heart and the
great vessels. In the heart, the LVOT and the left atrium CVMR images present distinct anatomic features of both
with its branches are clearly imaged. An advantage of the atria and ventricles. New generation MRI scanners pro-
frontal view is the similarity to the chest x-ray, well known vide morphologic characteristics in great detail and CVMR
to the clinicians. On sagittal images, the RV outflow tract images are used to evaluate cardiac anatomy in patients with
in relation to the pulmonary valve is well outlined and the complex cardiac anomalies (see Chapter 31). For evalua-
connection of the right atrium with the superior and inferior tion, either transverse (Fig. 7.7) or longitudinal planes, SA
vena cava can be studied. (Fig. 7.8) or LA cardiac views can be chosen.
A B C
D E F
Figure 7.8. Normal cardiac anatomy on short-
axis, black-blood acquisitions. LA, left atrium;
lPA, left pulmonary artery; Ao Asc, ascending
aorta; RA, right atrium; Ao Desc, descending
aorta; IVC, inferior vena cava; PA, pulmonary
Apex artery; RCA, right coronary artery; RVOT, right
ventricular outflow tract; LVOT, left ventricular
outflow tract; LAD, left arterial descending coro-
nary artery; LV, left ventricle; RV, right ventricle;
G H I PDA, posterior arterial descending coronary
artery; pm, papillary muscle.
LWBK1209-ch07_p108-115.indd 112 16/05/13 9:42 PM

Figure 7.9. Transverse black-blood acquisition illustrating the

Eustachian valve. RV, right ventricle; Ev, Eustachian valve; RA, right Figure 7.10. Four-chamber, bright-blood acquisition illustrating
atrium; LV, left ventricle; Ao Desc, descending aorta. the crista terminalis (arrow head) in the right atrium.
The pericardial sac encloses the heart and the roots of tal remnant of the right valve of the sinus venosus, in litera-
the great vessels. The pericardium consists of two layers. ture a prevalence of 1.5% to 2% has been described (9).
The outer fibrous pericardium is attached anteriorly to the The appendage of the left atrium has a narrow attachment
sternum, posteriorly to the thoracic spine, and inferiorly to to the atrium and is more tubular shaped. Characteristically,
the diaphragm. The inner, serous pericardium can be divided the left atrium receives in total four PVs, two on both sides,
into a parietal and a visceral layer. The parietal layer of the although several variations of this occur. Nowadays, imaging
inner pericardium is closely attached to the outer fibrous the venous anatomy of the heart is becoming more relevant.
pericardium. The visceral layer forms the epicardium, cov- Moreover, in the preablation work-up the referring clinician
ering the epicardial surface of the heart and the epicardial needs to be informed about the exact anatomy of the left
fat and coronary arteries. The pericardial cavity is outlined atrium and spatial orientation of the PV ostia. In patients
by the parietal and visceral layer of the inner pericardium. with atrial fibrillation, MR or CT images of the left atrium
Normal pericardium presents a low-signal intensity on MRI, and PVs are used to guide the interventional procedure, and
and can be well visualized by surrounding epicardial and provide indispensable information regarding PV anatomy,
pericardial fat. Normally, the pericardium measures less ostial dimensions, and shape (10,11). Three-dimensional
than 4 mm on CVMR images. Posteriorly to the ascending MR or CT reconstructions of the left atrium can be superim-
aorta reaches the superior pericardial recess. Effusion in this posed on fluoroscopy images during the interventional pro-
recess has to be differentiated from mediastinal pathology, cedure, thereby facilitating optimal catheter positioning and
for example, lymphadenopathy. improving procedural results (12).
In normal cardiac anatomy the right atrium can be recog- The interatrial septum separates the two atria and can be
nized by a broad-based triangular appendage. At the base, appreciated as a thin line. As part of the interatrial septum,
the tricuspid valve, positioned between the right atrium and the fossa ovalis is very thin and can hardly be depicted on
the RV, is located closer to the apex when compared to the CVMR images due to the limited spatial acquisition reso-
mitral valve on the left side. The right atrium receives venous lution. In some patients, the septum may be infiltrated by
blood from the superior and inferior vena cava, and the lipomatous tissue and thereby thickened (Fig. 7.11) or show
coronary sinus. The coronary sinus enters the right atrium localized bulging (aneurysm).
in the posterior atrioventricular groove. In the right atrium, The RV is normally triangular in shape and anteriorly ori-
the Eustachian valve (Ev) (Fig. 7.9) can be recognized at the entated to the right. Morphologically, the RV has typical fea-
orifice of the inferior vena cava. The crista terminalis sepa- tures that can be depicted on CVMR images. The RV shows
rates the anterior and posterior part, two embryologic dis- a muscular moderator band (Fig. 7.12), carrying branches
tinctive parts of the right atrium. The crista terminalis can of the conducting system. Also, the RV contains a muscular
present diverse, as prominent, broad-based or thin, valve-like outflow tract (infundibulum or conus arteriosus) and typi-
(Fig. 7.10). Chiari’s network is another normal anatomic cally, the RV wall is more trabeculated as compared to the
variant that can be identified. Chiari’s network is a congeni- LV. In normal anatomy, the LV is positioned posteriorly and
Figure 7.11. Four-chamber, bright-blood (A and B)

and black-blood (C) acquisitions, illustrating lipomatous
hypertrophy of the intra-atrial septum (arrow heads). A B C
The fossa ovalis is indicated by star.
LWBK1209-ch07_p108-115.indd 113 16/05/13 9:42 PM

muscle, respectively. The LV reveals two larger papillary

muscles, the anterior and posterior papillary muscle.
Cine SSFP, LA and SA images, as well as transverse images
are all well suited for depicting morphology and function of
the valvular apparatus. The valve leaflets can be depicted if
spatial resolution is adequate. Dedicated acquisitions of spe-
cific valvular planes are used to image the valve area, which
is especially useful when studying aortic valve competence.
Both SSFP as well as fast gradient-echo sequences are used.
A B
Papillary muscles are well visualized on both cine bright-
Figure 7.12. Transverse black-blood (A) and bright-blood (B) blood as well as black-blood sequences. Chordae tendineae,
acquisitions illustrating the moderator band (mb) in the right ventricle. on the other hand, are usually not sufficiently visualized by
mb, moderator band; RV, right ventricle; TV, tricuspid valve; RA, right MRI due to the limited spatial resolution.
atrium; LV, left ventricle; pm, papillary muscle; LA, left atrium; MV,
mitral valve; Ao Desc, descending aorta.
(Ultra-) High-Field Imaging
A decade ago, high-field 3-T whole-body MRI scanners have
to the left. The septum is smooth with no trabeculae and the been introduced in clinical practice. Past studies reported 20%
LVOT lacks a muscular wall. The interventricular septum to 150% improvement in signal-to-noise ratio (SNR) for SSFP
consists of a muscular and membranous part. Especially, the acquisitions at 3-T MRI when compared to 1.5 T (13), but
membranous part is very thin and sometimes not depicted on with the increase in field strength an increasing effect of imag-
CVMR images. ing artifacts also occurred. For CVMR imaging, SSFP techniques
At the outlet of each of the four chambers of the heart, have been implemented at 3 T and optimized (14), but espe-
one-way valves are positioned that ensure blood flow in the cially SSFP sequences are more prone to field inhomogeneities,
proper direction. The blood flow through the atria into the ven- resulting in artifacts that may hamper image evaluation. The
tricles is regulated by the atrioventricular valves: The tricuspid major source of artifact is off-resonance effects that become
valve on the right side and the mitral valve on the left side, more pronounced at higher field strength (15). Effective shim-
respectively. The pulmonary valve guards the outflow tract of ming of the B0 magnetic field is paramount, but the heart is a
the RV toward the pulmonary trunk, and the aortic valve con- difficult organ to shim owing to the complex field patterns in
nects the LVOT to the thoracic aorta. The tricuspid valve com- that region of the body (e.g., due to the lungs and liver) (16).
prises three cusps, whereas the mitral valve exists of two cusps. Dedicated shim systems providing higher-order, cardiac-
Both the pulmonary and the aortic valve (Fig. 7.13) consist of phase–specific shimming have reported improved field homo-
three cusps. geneity across the heart (17), but currently, SSFP at high field
Opening of the atrioventricular valves is predominantly remains not 100% reliable for use in CVMR imaging. Alternative
determined by pressure differences between the atria and to SSFP sequences, multiple-phase fast gradient-echo sequences
ventricles. These differences are the result of the isovolumic without SSFP may be used at high-field 3-T MRI or beyond.
relaxation of the ventricles. Furthermore, the motion of the Feasibility of CVMR imaging at ultra-high–field 7-T
valves is regulated by papillary muscles, which originate MRI was first demonstrated by Snyder et al. (18). Besides
from the inferior myocardial wall and are connected to the the above-mentioned susceptibility artifacts, that are even
valve leaflets by chordae tendineae. During contraction of more pronounced at 7 T, radio-frequency (RF) heating ef-
the ventricle the papillary muscles contract as well, pulling fects will limit the application of SSFP at ultra-high–field
on the chordae tendineae, closing the valves and preventing even further. The specific absorption rate (SAR) level
blood flow from the ventricles into the atria (i.e., regurgita- increases by the square of the field strength, and there-
tion). Normally, in the RV three papillary muscles can be fore, the use of SSFP has not yet been demonstrated at
depicted: The anterior, the posterior, and the septal papillary 7 T. Brandts et al. showed the feasibility of assessing LV
A B C
Figure 7.13. A segmented gradient-echo acquisition illustrating the aortic valve area at peak systole.
In (A), the planning of the acquisition plane is presented (dotted line). In (B), a normal valve with three
cusps is presented (lcc, left coronary cusp; rcc, right coronary cusp; ncc, noncoronary cusp), while in (C),
a bicuspid aortic valve is presented, with a fused noncoronary and left coronary cusp.
LWBK1209-ch07_p108-115.indd 114 16/05/13 9:42 PM

volumes, function, and atrial–ventricular flow at 7-T MRI 7. Kramer CM, Barkhausen J, Flamm SD, et al. Standardized cardiovascular magnetic reso-
nance imaging (CMR) protocols, society for cardiovascular magnetic resonance: Board
using standard multislice, multiphase cine gradient-echo of trustees task force on standardized protocols. J Cardiovasc Magn Reson. 2008;10:
and phase-contrast techniques, and they demonstrated 10–35.
8. Alfakih K, Plein S, Bloomer T, et al. Comparison of right ventricular volume measurements
similar quantitative results as compared to the gold stan- between axial and short axis orientation using steady-state free precession magnetic reso-
dard of 1.5-T CVMR (19). Still, without the availability nance imaging. J Magn Reson Imaging. 2003;18:25–32.
of dedicated hardware such as dedicated transmit/receive 9. Schneider B, Hofmaan T, Justen MH, et al. Chiari’s network: Normal variant or risk factor
for arterial embolic events. J Am Coll Cardiol. 1995;26:203–210.
coils, and optimized imaging techniques, CVMR at 7 T 10. Mansour M, Holmvang G, Sosnovik D, et al. Assessment of pulmonary vein anatomic
will remain an area of research. variability by magnetic resonance imaging: Implications for catheter ablation techniques
for atrial fibrillation. J Cardiovasc Electrophysiol. 2004;15(4):387–393.
11. Van der Voort PH, van den Bosch HCM, Post JC, et al. Determination of the spatial ori-
entation and shape of pulmonary vein ostia by contrast-enhanced magnetic resonance
References angiogrpahy. Europace. 2006;8(1):1–6.
12. Stevenhagen J, van der Voort PH, Dekker LRC, et al. Three-dimensional CT over-
1. Lenz GW, Haacke EM, White RD. Retrospective cardiac gating: A review of technical lay in comparison to Cartomerge for pulmonary vein antrum isolation. J Cardiovasc
aspects and future directions. Magn Reson Imaging. 1989;7(5):445–455. Electrophysiol. 2010;21:634–639.
2. Ho SY, Nihoyannopoulos P. Anatomy, echocardiography, and normal right ventricular 13. Hinton DP, Wald LL, Pitts J, et al. Comparison of cardiac MRI on 1.5 and 3.0 Tesla clinical
dimensions. Heart. 2006;92:I2–I13. whole body systems. Invest Radiol. 2003;38:436–442.
3. Cerqueira MD, Weismann NJ, Dilsizian V, et al. Standardized myocardial segmentation 14. Schär M, Kozerke S, Fischer S, et al. Cardiac SSFP imaging at 3 Tesla. MRM. 2004;51:
and nomenclature for tomographic imaging of the heart: A statement for healthcare profes- 799–806.
sionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the 15. Tyler DJ, Hudsmith LE, Petersen SE, et al. Cardiac cine MR-imaging at 3 T: FLASH vs.
American Heart Association. Circulation. 2002;105:539–542. SSFP. J Cardiovasc Magn Reson. 2006;8(5):709–715.
4. Uribe S, Muthurangu V, Boubertakh R, et al. Whole-heart cine MRI using real-time respi- 16. Atalay MK, Poncelet BP, Kantor HL, et al. Cardiac susceptibility artifacts arising from the
ratory self-gating. MRM. 2007;57:606–613. heart–lung interface. Magn Reson Med. 2001;45:341–345.
5. Danilouchkine M, Westenberg J, Reiber J, et al. Accuracy of short-axis cardiac MRI auto- 17. Kubach MR, Bornstedt A, Hombach V, et al. Cardiac phase-specific shimming (CPSS) for
matically derived from scout acquisitions in free-breathing and breath-holding modes. SSFP MR cine imaging at 3 T. Phys Med Biol. 2009;54(20):N467–N478.
MAGMA. 2005;18:7–18. 18. Snyder CJ, DelaBarre L, Metzger GJ,et al. Initial results of cardiac imaging at 7 Tesla.
6. Miller S, Simonetti OP, Carr J, et al. MR imaging of the heart with cine true fast imaging Magn Reson Med. 2009;61:517–524.
with steady-state precession: Influence of spatial and temporal resolutions on left ventricu- 19. Brandts A, Westenberg JJ, Versluis MJ, et al. Quantitative assessment of left ventricular
lar functional parameters. Radiology. 2002;223:263–269. function in humans at 7 T. Magn Reson Med. 2010;64(5):1471–1477.
LWBK1209-ch07_p108-115.indd 115 16/05/13 9:42 PM

Chapter
8
Karen G. Ordovas
Maria Clara N. Lorca
Charles B. Higgins
Coronary CTA: Display

Techniques and Planes
■■ Introduction define what dataset has the least amount of cardiac motion
artifact affecting the structures of interest. Prior coronary
■■ Selecting the Best Dataset to Perform the CTA studies have described that the ideal phase of the cardiac
Reconstructions cycle for visualization of the coronary arteries vary. Choi et al.
■■ An Overview of the Most Common (3) have shown that the best phase for right coronary artery
Postprocessing Tools and Its Clinical Utilities depiction is approximately 40% to 45% of the R–R inter-
Multiplanar Reconstruction val, whereas 70% of the cycle is the ideal phase for the left
Curved Multiplanar Reconstruction coronary system. When the heart rate is below 60 to 65 bpm,
usually 70% to 75% of the R–R interval provides good image
Maximum Intensity Projection
quality for visualization of all the coronary segments. In some
Volume Rendering cases, particularly when the heart rate is elevated, review of
Vessel Wall Imaging multiple cardiac phases is needed to accomplish interrogation
of the entire coronary tree without motion artifacts.
Selection of a specific phase of the R–R interval for im-
Introduction aging reconstruction is also necessary when cardiac valves
are being evaluated. For evaluation of valvular stenosis,
Although most of the findings on coronary computed tomog- reformations are ideally obtained throughout the cardiac
raphy angiography (CTA) can be identified on axial images cycle, but the maximum opening of the valvular orifice will
of the heart, the use of different reformation techniques be depicted during end-systole for the aortic valve and end-
can be helpful for diagnosis and for determining severity of diastole for the mitral valve.
coronary stenoses (1). In addition, reformations can be very
useful when the relationship between the coronary arteries
and the adjacent mediastinal structures must be defined. An Overview of the Most Common
The acquisition of isotropic or near-isotropic voxels with
Postprocessing Tools and Its
multidetector CT allows for precise visualization of the im-
ages through reformations in innumerable planes (2).
Clinical Utilities
An organized approach for evaluation of coronary imag-
Multiplanar Reconstruction
ing studies is essential for optimization of the interpretation
sessions. The selection of postprocessing tools should be tai- Multiplanar reconstruction (MPR) is typically used in an
lored to each specific case, with the ultimate goal to address interactive way, allowing the imager to navigate through
the main clinical question. Review of the axial images is very sagittal, coronal, axial, and the several planes oriented
helpful to identify possible coronary lesions and to guide an to the heart itself rather than the thorax. This technique
optimal approach to a time-efficient evaluation of the stud- can be understood as a two-dimensional slab of the three-
ies (2). Clinical-based strategies for the postprocessing of dimensional dataset. Although this technique provides
cardiac imaging studies will be discussed in this chapter. flexibility for understanding the spatial relationship of the
cardiac chamber, great vessels, and other mediastinal struc-
tures, it is somewhat limited for evaluation of long and tortu-
Selecting the Best Dataset to ous vascular segments such as the coronary arteries. Short
Perform the Reconstructions and nontortuous segments of the coronary arteries can be
well evaluated with MPR (4).
When performing postprocessing for visualization of the The most useful applications of MPR are in cases where
coronary arteries and cardiac structures, it is important to the relationship of the coronary arteries and the great vessels
116
LWBK1209-ch08_p116-124.indd 116 17/05/13 6:43 PM

Chapter 8 ■ Coronary CTA: Display Techniques and Planes 117
Figure 8.1. Multiplanar reconstruction of a left internal mammary Figure 8.2. Multiplanar reconstruction aligned to the main pulmo-
(LIMA) graft (arrows) in a parasagittal plane displays the relationship nary artery (MPA) and origin of the left main coronary artery (LCA).
of the graft with the inner surface of the sternum. This reformation shows clearly the inferior displacement of the LCA
(arrow), without significant luminal stenosis.
and chest wall structures must be examined. A common clini- morphology and function, especially when retrospective car-
cal indication for coronary CTA is visualization of coronary diac gating is obtained and cine loops can be played encom-
bypass grafts and its relationship to the sternum in patients passing the entire cardiac cycle. This approach allows for
undergoing reoperation, with the goal of avoiding graft injury visualization of leaflet excursion, particularly in the setting
during sternotomy (5). MPR in the sagittal plane can typically of aortic or mitral stenosis. For interrogation of the aortic
depict the relationship of anterior venous grafts or internal valve, MPR reformat is obtained aligned parallel to the left
mammary grafts to the inner face of the sternum (Fig. 8.1). ventricular outflow tract and proximal ascending aorta,
Another clinical scenario suitable for MPR reformation is which allows for visualization of the leaflet thickness and
the investigation of coronary artery compression by adjacent excursion and poststenotic dilatation of the ascending aorta
structures, such as paracardiac masses or dilated pulmonary (Fig. 8.4A). An orifice view, which is parallel to and at the
arteries in patients with pulmonary hypertension (Fig. 8.2). level of the valvular annulus, can be useful for measurement
MPR is the most suitable method for visualization of car- of maximum opening at end-systole (Fig. 8.4B). Similarly,
diac anatomy and function. For assessment of regional cardiac MPR in the two-chamber view is optimal for visualization of
contractility, three standard cardiac planes are typically used mitral valve leaflet excursion. Although rarely an indication
for MPR reformats. The horizontal long-axis plane, also called for cardiac CTA, evaluation of the pulmonary and tricuspid
four-chamber plane, is aligned to the long axis of the heart, valves can also be obtained by aligning the MPR parallel to
and displays the four cardiac chambers, the mitral and tricus- the outflow and inflow of the right ventricle respectively.
pid valves and the atrial and ventricular septum in alignment.
The vertical long-axis plane, also called two-chamber plane, Curved Multiplanar Reconstruction
can be obtained from the four-chamber plane by prescribing
a plane between the left ventricular apex and the center of the Curved MPR is a postprocessing technique that allows
mitral valve. Finally the short-axis plane is prescribed perpen- for delineation of a plane along the course of a curvilinear
dicular to the horizontal long-axis plane (Fig. 8.3). structure (6). It is very commonly used for coronary CTA
Most postprocessing softwares offer a tool for automatic evaluation, as it is the most precise technique for quantifi-
depiction of the standard cardiac planes. In retrospectively cation of luminal stenosis (Fig. 8.5). Most postprocessing
gated cardiac studies, a cine loop in the standard cardiac softwares have tools for building curved MPR reformats
planes allows for a comprehensive depiction of global and aligned to the course of the main coronary vessels. Since
regional left ventricular wall motion. In addition, volumetric these curved planes are obtained from a –3D dataset, they
analysis can be performed using the Simpson’s method for can be reviewed dynamically, with a 360-degree overview of
quantification of end-diastolic volume, end-systolic volume, the vessel lumen, allowing for identification of the maximum
and ejection fraction of both ventricles. degree of stenosis (Fig. 8.6). In general, the curved MPR
MPR is also the best technique for assessment of valvular reformats are reviewed with a focused approach, based on
function. In general, CTA is useful for evaluation of valvular the initial identification of suspicious areas of stenosis on the
LWBK1209-ch08_p116-124.indd 117 17/05/13 6:43 PM

A B
Figure 8.3. Multiplanar reconstructions in the

standard cardiac planes. A: Vertical long-axis plane
is aligned to the left atrium (LA) and left ventricle
(LV). B: Horizontal long-axis plane shows the four
cardiac chambers and the atrial and ventricular septa
in alignment. C: Short-axis plane at the level of the
C
ventricles. RV, right ventricle; RA, right atrium.
A B
Figure 8.4. Multiplanar reconstruction in a patient with aortic stenosis. A: The reformation is obtained in
alignment with the left ventricular outflow tract, showing in a single image, the thickened aortic valve leaflets
(arrow) and the poststenotic dilatation of the ascending aorta (Ao). LV, left ventricle. B: A similar reformation
technique is used to obtain an orifice view of the aortic valve at end-diastole for measurement of the maximum
valve opening. An aortic orifice area of 1.8 cm2 is consistent with a mild degree of aortic stenosis. Note the
thickened aortic leaflets (arrow) with associated coarse calcification (arrowhead).
LWBK1209-ch08_p116-124.indd 118 17/05/13 6:43 PM

A B
Figure 8.5. Axial image from a coronary CTA (A) depicts a calcified plaque at the proximal circumflex
(arrow), which suggests a severe stenosis. The curved multiplanar reconstruction (B) shows that the stenosis
(arrow) is in fact less than 50%. Ao, aorta.
A B C
Figure 8.6. Curved multiplanar reconstruction aligned to the course of the left anterior descending coro-
nary artery shows an eccentric-mixed atherosclerotic plaque. As shown on the left upper quadrant diagram,
the severity of the stenosis is clearly depicted as reformations are rotated clockwise (black line), with minimal
stenosis on (A), less than 50% stenosis on (B) and more than 50% stenosis on (C).
LWBK1209-ch08_p116-124.indd 119 17/05/13 6:43 PM

A A
A B B
C C
Figure 8.7. Axial images from coronary CTA (A and B) depict a complex atherosclerotic plaque (arrow)
on a tortuous segment of the proximal left anterior descending coronary artery. Curved multiplanar reformat
image (C) demonstrates a low degree of luminal stenosis caused by this lesion (arrow). Ao, aorta.
axial images (Fig. 8.7). However, some eccentric soft plaques soft atherosclerotic plaque is present (4). Values of 250 HU and
and lesions in very small, tortuous, or angled coronary seg- above are ideal for generating accurate curved reformats (9,10).
ments can be very difficult to visualize in axial images and
therefore only depicted when curved MPRs are performed
Maximum Intensity Projection
(Fig. 8.8). Another useful application for curved MPR is
evaluation of the presence and significance of a myocardial This technique averages the pixel density to the maximum
bridge. This technique can delineate the length and depth of density within each particular pixel. Therefore, it enhances the
the intramyocardial coronary course, which has been associ- depiction of high-density structures, such as vessels filled with
ated with dynamic obstructive disease (Fig. 8.9) (7,8). iodinated contrast media, and suppresses the visualization of
A very useful way of quantifying the degree of luminal lower-density structures (11). Maximum intensity projection
stenosis causedC by an atherosclerotic plaque is to obtain a (MIP) can be a useful technique for overall visualization of the
cross-sectional view of the coronary vessel. Most postpro- anatomy of the coronary tree and depiction of high-grade ste-
cessing softwares offer this tool when reviewing curved noses. However, it limits significantly the visualization of subtle
MPR images (Fig. 8.10). lesions, particularly soft plaques. In addition, blooming artifact
Optimal coronary enhancement (adequate contrast-to- from coronary wall calcification are accentuated and can lead
noise) is essential for generating a reliable curved MPR. Poor to overestimation of the degree of stenosis with this technique
opacification may lead to misdiagnosis, particularly when a (Fig. 8.11) (4).
LWBK1209-ch08_p116-124.indd 120 17/05/13 6:43 PM

A B
A B C
Figure 8.8. Small vessel stenosis. Axial images from coronary CTA (A and B) partially show an eccentric
plaque (arrow) at the origin of the second diagonal branch of the left anterior descending coronary artery. A
thicker curved multiplanar reconstruction (C) provides depiction of the location and degree of stenosis. Ao, aorta.
Volume Rendering
Volume rendering is an aesthetically appealing reformation
technique that displays the heart and/or coronary arteries
and adjacent structures with a 3D perspective. Volume-
rendered images can facilitate significantly the commu-
nication of complicated 3D structures to surgeons and
interventional specialists before stent placement or revas-
cularization surgery (4). However, the use of volume ren-
dering for detection and quantification of coronary stenosis
is quite limited.
The most important application of this technique is the
understanding of the spatial relationship of the coronary
arteries with the great vessels and chest wall. In particular,
volume rendering is a very useful technique for understan
ding the complicated course of an anomalous coronary ar-
C tery and its relationship to the ascending aorta and main
pulmonary artery (MPA) (Fig. 8.12). Another application
of volume rendering for coronary CTA is the visualization
of structures surrounding bypass graft prior to reoperation,
with special focus on the sternotomy site.
Vessel Wall Imaging

Although not used in clinical practice, postprocessing tools
are available for display of the coronary wall thickness and
tissue characteristics of plaques. Coronary wall imaging is
currently considered a research technique. Typically, the
Figure 8.9. Curved multiplanar reconstruction precisely depicts density of the wall components is displayed in color codes,
the length and depth of an intramyocardial bridge (arrows) in the mid- depicting calcification, soft tissue density, and fat density
segment of the left anterior descending coronary artery. (Fig. 8.13) (12).
LWBK1209-ch08_p116-124.indd 121 17/05/13 6:43 PM

A B
Figure 8.10. Curved multiplanar reconstruction (A) along the left anterior descending coronary artery
shows multiple atherosclerotic plaques. An eccentric calcified plaque in the proximal segment is better inter-
rogated with a cross-sectional (white line) reformat. Cross-sectional image (B) demonstrates the eccentric and
nonobstructive characteristic of the lesion (arrow).
A B
Figure 8.11. Maximum intensity projection of the coronary arterial tree (A) shows coarse calcific plaques
along the proximal segment of the left anterior descending (LAD) coronary artery (arrowheads), which seem
to cause severe coronary obstruction due to blooming artifact. Curved multiplanar reconstruction of the LAD
(B) shows that, in reality, the stenoses are of approximately 50% diameter reduction (arrowheads).
LWBK1209-ch08_p116-124.indd 122 17/05/13 6:43 PM

A B
Figure 8.12. Three-dimensional volume-rendering image (A) of the heart shows an anomalous course of
the left coronary artery (LCA), along the anterior aspect of the proximal main pulmonary artery (MPA). A
curved maximum intensity projection (B) aids in depicting the anomalous origin (arrowhead) of the LCA from
the right coronary sinus and confirms the anterior pulmonary course (arrow). RCA, right coronary artery; Ao,
aorta.
A B
Figure 8.13. Curved multiplanar reconstruction (A) of the left anterior descending coronary artery shows
a long and complex atherosclerotic plaque with soft (arrowheads) and calcified (arrow) components. A color
mapping (B) of the vessel wall highlights the different densities of the plaque components including calcium
(red), soft plaque (yellow), and contrast-enhanced lumen (green). Ao, aorta.
LWBK1209-ch08_p116-124.indd 123 17/05/13 6:43 PM

References 7. Ishikawa Y, Kawawa Y, Kohda E, et al. Significance of the anatomical properties of a

myocardial bridge in coronary heart disease. Circ J. 2011;75:1559–1566.
8. Ishii T, Asuwa N, Masuda S, et al. The effects of a myocardial bridge on coronary athero-
1. Ferencik M, Ropers D, Abbara S, et al. Diagnostic accuracy of image postprocessing meth-
sclerosis and ischaemia. J Pathol. 1998;185:4–9.
ods for the detection of coronary artery stenoses by using multidetector CT. Radiology.
9. Johnson PT, Pannu HK, Fishman EK. IV contrast infusion for coronary artery CT angi-
2007;243:696–702.
ography: literature review and results of a nationwide survey. AJR Am J Roentgenol.
2. Kerl JM, Hofmann LK, Thilo C, et al. Coronary CTA: image acquisition and interpreta-
2009;192:W214–W221.
tion. J Thorac Imaging. 2007;22:22–34.
10. Husmann L, Gaemperli O, Valenta I, et al. Impact of vessel attenuation on quantitative
3. Choi HS, Choi BW, Choe KO, et al. Pitfalls, artifacts, and remedies in multi-detector row
coronary angiography with 64-slice CT. Br J Radiol. 2009;82:649–653.
CT coronary angiography. Radiographics. 2004;24:787–800.
11. Fishman EK, Ney DR, Heath DG, et al. Volume rendering versus maximum intensity pro-
4. Johnson PT, Fishman EK. Postprocessing techniques for cardiac computed tomographic
jection in CT angiography: what works best, when, and why. Radiographics. 2006;26:
angiography. Radiol Clin North Am. 2010;48:687–700.
905–922.
5. van Ooijen PM, Ho KY, Dorgelo J, et al. Coronary artery imaging with multidetector CT:
12. Jinzaki M, Yamada M, Sato K, et al. Overview image of the lumen and vessel wall in coro-
visualization issues. Radiographics. 2003;23:e16.
nary CT angiography. Circ J. 2008;72:671–673.
6. Sirineni GK, Kalra MK, Pottala KM, et al. Visualization techniques in computed tomo-
graphic coronary angiography. Curr Probl Diagn Radiol. 2006;35:245–257.
LWBK1209-ch08_p116-124.indd 124 17/05/13 6:43 PM

Chapter
Mark A. Fogel 9
Cardiovascular MR in Infants
and Children
■■ Introduction ventricles, and great vessels) may not be in the proper posi
tion and connections between them may be altered (as in
■■ Technical Considerations in CVMR transposition of the great arteries, where the aorta is con
of Infants and Children nected to the right ventricle and the pulmonary artery is
Spatial Resolution connected to the left ventricle). Venous anatomy too may
Temporal Resolution not be connected properly (such as total anomalous pul
Coronary Imaging monary venous connection, where the pulmonary veins
Inability of Pediatric Patients to Breath-Hold are connected to any structure other than the left atrium).
Gadolinium-Based Techniques Communications may present where they should not be
(e.g., ventricular septal defect where there is a communica
■■ Generalized Protocol of CVMR for Infants
tion between the right and left ventricles), and some cardiac
and Children structures might not even be present or present in a mark
Anatomic Imaging edly hypoplastic form (e.g., hypoplastic left heart syndrome,
Cine CVMR where the left ventricle is markedly hypoplastic). As surgery
Gadolinium Imaging and catheter-based therapies play such an important role in
Velocity Mapping CHD, the postinterventional anatomy must be familiar to
Delayed Enhancement the physician performing the CVMR examination.
Other Techniques 2. Physiology/function: In the world of the infant and child
with CHD, there are many times that a unique physiology
■■ A Worked Example—Single Ventricles
is present and a result of the altered anatomy. Evaluation of
■■ Conclusion shunts (e.g., in a patient with a truncus arteriosus referred
to CVMR, where one great vessel arises from the heart to
give rise to the systemic, pulmonary, and coronary systems)
Introduction or the complex physiology of transposition of the great
arteries plays a critical role in the assessment of any disease
Cardiovascular magnetic resonance (CVMR) had been
entity in pediatrics and has little parallel in the adult world.
applied to infants and children for nearly 30 years and is
In addition, assessment of ventricular function in ventricles
firmly established in the evaluation of anatomy, physiology,
with strange and unusual shapes are routine in the prac
and function in congenital heart disease (CHD) (1–11). In
tice of CVMR in children (e.g., L-looped ventricles which
this age range, CVMR is generally used as a complement
have a “left sided” right ventricle and a “right sided” left
to echocardiography and angiography; it nevertheless is a
ventricle) and demand unique solutions. Assessment of the
gold standard in a number of areas including ventricular vol
postoperative physiology where the surgeon, for example,
umes, mass (12,13) and vascular rings (14–16).
creates an anastamosis between the cava and the pulmo
CVMR in infants and children utilizes nearly all the vari
nary artery (when repairing a single ventricle) must be
ous techniques discussed in this textbook; however, it is a
dealt with. CVMR must be adapted to fit these needs and
free standing discipline for numerous reasons, divided into a
the physician interpreting the study needs to understand
few broad categories:
the complex physiology. Even without gross anatomical
1. Anatomy: A rigorous and systematic approach to ana alterations, function can be severely altered such as in an
tomy of native CHD and of repaired CHD is very different infant with a left ventricular aneurysm or with a dysplastic
from heart disease in the adult. Cardiac segments (atria, tricuspid valve (Figs. 9.1 and 9.2).
125
LWBK1209-ch09_p125-134.indd 125 16/05/13 9:40 PM

A B C
D E F
Figure 9.1. Three-day-old neonate with left ventricular (LV) aneurysm (*) performed with the feed and
swaddle technique. Four-chamber (A), two short-axis (B and C), and one long-axis view (D) by cine SSFP of
the aneurysm. Viability imaging in the four-chamber (E) and short-axis view (F) demonstrating endocardial
scar (red arrows).
3. Technical challenges: Infants and children present a unique 1. As alluded to above, sedation and general anesthesia for
challenge in the world of CVMR; they require increased young patients must be administered to allow a child to
spatial resolution because of their small size and increased remain motionless for a 45- to 60-minute scan; this has
temporal resolution because of their high heart rates. recently become less of an issue with the advent of the
When imaging adults, tradeoffs can be made between “feed and swaddle” technique which has been success
those two competing parameters to make imaging quicker fully used in infants at 6 months of age (17) (Fig. 9.1). In
and simpler without sacrificing diagnostic accuracy. As addition, even though they may have the cognitive ability
pulse sequences were generally designed to image adults, to hold still, cooperation by the preteen or teen may be
these tradeoffs can be performed in infants and children problematic (e.g., breath-holding).
to a limited degree. In addition, because of the need for 2. As surgery and catheter-based interventions play an impor
increased spatial resolution, signal-to-noise ratio (SNR) tant role in CHD, wires, stents, coils, and clips may all
considerations play an important role and are always cause artifacts if they are near the structure(s) of interest.
taken into account. Further, children under 7 to 9 years 3. As many critically ill infant and children are treated in
of age usually require sedation which can make sequences the intensive care unit, the lack of portability of CVMR
designed for breath-holding useless. Since breath-holding is certainly disadvantageous; moving the patient to the
is the mainstay of adult CVMR, “work-arounds” have CVMR suite can be a challenge.
been developed to successfully image the pediatric patient. 4. Children with arrhythmias, especially after surgery, may
These CVMR modifications must be understood for chil not allow proper triggering while other patients may have
dren to be imaged properly. bizarre T waves or bundle branch blocks which may not
have the same effect. Although this can be problematic,
Limitations of CVMR in infants and children must be it is less of an issue now that “single shot” CVMR, “real
nderstood by the imager; some to keep in mind are the
u time” cine CVMR, and sequences with “arrhythmia
following: rejection” are realities.
LWBK1209-ch09_p125-134.indd 126 16/05/13 9:40 PM

Chapter 9 ■ Cardiovascular MR in Infants and Children 127
A B
C D E
F G
Figure 9.2. Five-year-old with dysplastic tricus

pid valve and severe insufficiency. The four-chamber
view at end-diastole (A) and in systole (B) demon
strat severe tricuspid insufficiency (red arrow).
Short-axis magnitude (C) and phase-contrast
(D) images of the coaptation defect (red arrow) of the
valve. Volume-rendered 3D gadolinium images in
a coronal (E) and off-axis sagittal (F) view demon
strating the dilated right ventricle (RV); left ventricle
(LV) is labeled for comparison. Four-dimensional
flow imaging of the regurgitant jet (red arrow) into
the right atrium (RA) is demonstrated in (G).
5. Pacemakers and ICDs are still a relative contraindication “work-arounds” are required for CVMR sequences, which
and patients with them usually do not undergo CVMR, are designed for adults, to work.
although there is some data to suggest that this may be
safe (18).
Spatial Resolution
To increase the spatial resolution, small voxels are needed;
Technical Considerations in however, this can be problematic because of the poor SNR.
CVMR of Infants and Children Employing parallel imaging generally makes this even worse.
Imaging without SNR considerations will result in a poor-
As there is a need for high spatial and temporal resolution in quality image which may not be diagnostic. To increase
infants and children along with the inability to breath-hold, SNR, strategies have been employed either in isolation or in
LWBK1209-ch09_p125-134.indd 127 16/05/13 9:40 PM

combination, such as utilizing phase oversampling, decreas In retrospective gating, it is important to remember that
ing the bandwidth, employing multiple averages (excita the MR scanner is continuously acquiring data and record
tions), and avoiding techniques such as parallel imaging or ing the ECG. After all the lines of k-space are acquired, each
partial phase Fourier. These increase imaging time but are line of k-space is then placed into “bins” in the closest phase
worth it for high quality. of the cardiac cycle using interpolation. If one is performing
At times, for the smallest voxels, even these approaches interpolation between phases, it is important that the number
are not enough. In those instances, the imager must use of calculated phases should not exceed twice the measured
larger voxels and this is done by either decreasing the matrix phases (essentially the RR interval (milliseconds)/TR (milli
size (e.g., from 256 or 192 to 128 in the frequency encoding seconds) where TR is the line TR × lines of k-space obtained);
direction) or increasing the field of view—both of which can there should be two measured points between each interpo
nevertheless yield diagnostic images yielding pixels of 1.5 to lated point to obtain robust data. The formula used is:
2 mm or less. Slice thicknesses generally do not go below 2.5
2 × RR interval (msec)/number of calculated phases
to 3 mm because of SNR considerations. With gadolinium
= TR (msec)
images, spatial resolution can be submillimeter (see gado
linium section). where TR is defined as the line TR × lines of k-space
When imaging small children, small fields of view may obtained. This is especially important in velocity mapping
be needed (Fig. 9.1—a 3-day-old) and at certain settings, for accurate data to be obtained.
sequences are written that do not allow for the fields of At times, “single shot” imaging in children can be used
view to decrease very low (e.g., 150 mm in a 3-kg baby). because of respiratory motion or arrhythmia, where all the
Similarly, smaller slice thicknesses than adults are needed and lines of k-space are acquired in one heartbeat. A typical
once again, the sequence parameters may not allow for that “single shot” imaging sequence is the half fourier acquisition
change at a specific setting. However, relaxing one param single shot turbo spin echo (HASTE) sequence mentioned
eter may allow another to be changed. The imager must find above, but other sequences such as static steady-state free
combinations by trial and error of changing, for example, precession (SSFP) and cine SSFP can be performed in “single
bandwidths, TR, matrix, segments of k-space (“views” on shot” mode; cine SSFP utilizing single shot imaging is “real
some scanners), field of view, and slice thickness to allow for time” cine. Perfusion sequences are another example of this
optimal imaging for the size of the patient. There are idiosyn type of imaging. If the heart rate is too fast, it is sometimes
crasies in each manufacturer’s code and it would be prohibi advantageous to obtain the image over two heartbeats (i.e.,
tive to list all the combinations here. By relaxing some pa imaged at the end of two heartbeats [2 × RR interval]); this is
rameters (e.g., longer TR), other parameters may be allowed commonly used in HASTE, perfusion or single shot viability
to be optimized to image the smaller patients (e.g., smaller imaging in children with heart rates in the triple digits.
field of view).
Coronary Imaging
Temporal Resolution
When CVMR for coronaries in small children are performed,
The high heart rates in infants and children (not uncom timing is very important. A high temporal resolution four-
monly, 120 to 170 beats/minute) must be taken into account chamber and left ventricular outflow tract cines should be
for a successful CVMR scan in this age group. As for the obtained first and the quiescent phase of the cardiac cycle should
number of phases to be acquired, as a general rule of thumb: be identified; generally, the imager focuses on the atrioventricu
lar valve and aortic annular motion in the four-chamber and left
•• For an RR interval of <500 milliseconds, 18 to 20 phases
ventricular outflow tract views, respectively. It is not uncom
should be obtained
mon to have the quiescent phase be at end-systole and not end-
•• For an RR interval of 500 to 750 milliseconds, 20 to 25
diastole in children. In addition, because the quiescent phase is
phases should be obtained
generally short, the “shot time” for coronary imaging needs to
•• For an RR interval of 750 to 1,000 milliseconds, 25 to 30
be short (decreased number of segments are needed in general
phases should be obtained
as decreasing the resolution in a small child will not work).
•• For an RR interval of >1,000 milliseconds, 30 to 35 phases
should be obtained.
Inability of Pediatric Patients
To obtain this type of temporal resolution at fast heart
to Breath-Hold
rates, the TR needs to be low and to accomplish this, in
general, the number of segments (views) needs to be low as To perform CVMR in patients who cannot breath-hold or
well. This does not add very much time to the scan, since de remain motionless in the scanner, some institutions advocate
spite the small number of segments obtained per heartbeat, general anesthesia in all patients; in this way the anesthesi
each heartbeat comes quicker at high heart rates relative to ologist can do the breath-holding. This takes a noninvasive
slow ones. If the TR and the number of segments are not low imaging modality and makes it invasive, adds to the expense
enough, in retrospectively gated imaging, the walls of the and is less physiologic than free breathing. Deep sedation
heart will appear blurry and with a double or triple shadow with free breath during CVMR has been successfully utilized
(does not “freeze” the motion of the wall). Adequate tempo for many years in infants and children with little untow
ral resolution can be obtained at fast heart rates with three ard effect (19), and in infants <6 months of age, using no
segments at a minimum and TRs in the teens and 20s mil medication (17). General anesthesia should be reserved for
liseconds. those with cardiorespiratory compromise, those who failed
LWBK1209-ch09_p125-134.indd 128 16/05/13 9:40 PM

Figure 9.3. Time-resolved dynamic gadolinium imaging in a Fontan patient. After injection in the left arm, the
right-sided structures are seen (A) such as the right (RPA) and left pulmonary arteries (LPA) and the Fontan baffle
(B). Note the right upper lobe hypoperfusion (red arrow) due to either pulmonary embolus, right upper pulmonary
artery atresia or competitive flow from aortic to pulmonary collaterals. In (B), the left-sided structures are seen such
as the ventricle and the aorta (Ao). In the recirculation phase (C), the right-sided structures can be seen again.
deep sedation or patients who must undergo more than one c. Delayed enhancement (DE) imaging also may need to be
examination at the same visit where deep sedation will not performed over two RR intervals to obtain high-quality
last for that extended period of time. images (Figs. 9.1, 9.4, and 9.5).
Imaging the free-breathing infant or child with CVMR
(Figs. 9.1–9.8) utilizes (a) multiple excitations to “average
out” the respiratory motion (generally three to five excita Generalized Protocol of CVMR
tions are needed which also increases SNR), (b) navigator- for Infants and Children
based techniques, or (c) single shot or real time cine imaging.
In addition to these strategies, placing a saturation band CVMR in infants and children requires fine tuning during
over the chest wall with dark blood imaging can also aid in the examination as unexpected findings are not uncommon.
minimizing any respiratory artifact which does occur. These Nevertheless, certain basic principles exist and can be for
are the most common strategies which allow high-quality mulated into a generalized protocol. Although not the only
images to be created albeit at the cost of a little extra time way to approach the small child, it is in the author’s opinion,
which is not prohibitive. the most efficient and complete. Protocols of course need to
be individualized to the patient and the disease.
Gadolinium-Based Techniques
The above applies to all types of imaging; however, gadolin Anatomic Imaging
ium-based techniques have some additional considerations.
As physiology and function must be interpreted in light of the
a. When performing “static” three-dimensional (3D) gado prevailing anatomy, this must be the first step. A full contigu
linium imaging and a “tracking” or “CARE” bolus tech ous set of axial images from the diaphragm to the thoracic inlet
nique is used, the quick circulatory time of infants and is obtained (typically 40 to 50 images) and must be extended
children must be taken into account. The imager must outside the thorax for special cases (e.g., neck, if arch anatomy
be aware that the gadolinium will reach the structure of is being evaluated); the author prefers diastolic “static” SSFP.
interest much faster than in adults and if not careful, may The purpose of this is not only to survey the cardiovascular
miss the first pass of contrast agent, especially important anatomy and make the “anatomic diagnosis,” but also acts
when trying to image the right side of the circulation. as a localizer themselves for future physiologic and functional
With the advent of dynamic gadolinium imaging, this is imaging. One disadvantage to “static” SSFP is that if diastolic
less of a consideration (Fig. 9.3). turbulence is present, the structure may demonstrate signal
b. The perfusion technique obtains images at a single slice loss and not be visualized (e.g., a Blalock–Taussig shunt); this
position at different time points in the cardiac cycle after drawback can be compensated for by performing HASTE or
a delay. With heart rates so high in infants and children, it dark blood imaging. HASTE imaging can be performed while
may only be possible to fit two slices within an RR inter the stack of static SSFP are being manipulated with multipla
val. If this occurs, the imager can perform the sequence nar reconstruction (MPR) (see below). Alternatively, a set of
over two RR intervals without degradation of diagnostic axial cines can be performed.
quality and obtain more slices (three or four) during the MPR is then used to reformat the stack of SSFP and
scan. Alternatively, decreasing the phase-encoding steps HASTE images, when needed, to obtain the exact slice ori
in the matrix (by, for example, increasing the rectangular entation and position for any future imaging of the patient
field of view) can also be used to obtain more slices. during the study. If the study is terminated prematurely
LWBK1209-ch09_p125-134.indd 129 16/05/13 9:41 PM

Figure 9.4. Myocardial tagging

and delayed enhancement in a single
ventricle (SV) patient. Myocardial
tagging at end-diastole (A) and in
systole (B) in the “three-chamber”
view using spatial modulation of
magnetization (SPAMM) to divide
the myocardium into “cubes of
magnetization”—note how the
deformation can be visualized.
(C) Demonstrates this in long axis.
Myocardial scarring (red arrow) is
seen in the “three-chamber” (D) and
short-axis views (E).
Figure 9.5. Cine SSFP, volume rendering, and delayed enhancement imaging after a bidirectional Glenn (supe
rior vena cava to pulmonary artery anastomosis). The patient has dextrocardia and severe atrioventricular valve
insufficiency (red arrow) which can be seen in the “three-chamber” view at end-diastole (A) and mid-systole (B).
Without a right superior vena cava, the left superior vena cava (LSVC) to left pulmonary artery (LPA) anastomosis
can be seen alone (C) or with the single ventricle (SV) and the aorta (Ao) in place as well (D). Delayed enhancement
imaging (E) shows increased signal intensity in the atrioventricular valve en face (red arrows) in this short-axis view.
LWBK1209-ch09_p125-134.indd 130 16/05/13 9:41 PM

A Cine CVMR
B
This is tailored to the lesion under study and along with
velocity mapping, are the major sequences utilized to delin
eate physiology and function in CVMR of infants and chil
dren. (Figs. 9.1, 9.2, 9.5, and 9.6). In addition, it is used
to obtain higher resolution images of the anatomy (either
systole or diastole) and to confirm the resultant physiology
(e.g., turbulence and loss of signal through a stenotic blood
vessel). A typical use of cine is to determine right ventricular
volume overload in a patient with an atrial septal defect,
where a stack of ventricular short-axis images are acquired
C D to measure ventricular volumes, mass, ejection fraction,
stroke volume, and cardiac index. A typical use for anat
omy would be to determine the degree of branch pulmonary
artery stenosis or the anatomy of valves, viewing them en
face. Even though SSFP may be used, utilizing gradient-echo
imaging with a high flip angle (e.g., 25 degrees) is very effec
tive as the high signal of the blood flowing into the imaging
plane outlines the valve leaflets very well.
Figure 9.6. Cine SSFP images of a single ventricle after Fontan. Gadolinium Imaging
The “three-chamber” (A) and short-axis views (B) demonstrate
ventricular function as well as the baffle (B) and right-to-left flow
To delineate the anatomy further and to obtain high-
across the fenestration (red arrow). The (B) is outlined in long-axis resolution 3D images, gadolinium imaging is then per
in (C) and the superior vena cava (SVC) to right pulmonary artery formed (Figs. 9.3, 9.5, and 9.7). It is done at this point of
(RPA) connection is outlined in (D). LPA, left pulmonary artery. the examination to allow for 10 to 15 minutes to elapse for
DE imaging; in the interim, velocity mapping is performed.
Dynamic 3D gadolinium imaging (e.g., TWIST) is espe
cially useful in pediatrics with the quick circulation; besides
ecause of patient instability or technical reasons, not only
b the creation of 3D volume-rendered images, because of
can the stack of axial images demonstrate the anatomy but the higher resolution, smaller vessels can be imaged. This
the salient points can also be demonstrated with MPR. is important in, for example, diseases such as tetralogy of
If needed, a set of high resolution double inversion recov Fallot with pulmonary atresia where tiny aortic to pulmo
ery dark blood images can be obtained for the regions of in nary collaterals are generally present and are important to
terest after delineated via MPR. This is used judiciously since identify. Dynamic, time-resolved gadolinium imaging can
it requires relatively long scanning times for just a few images. also be used to image lung perfusion; if there is a pulmonary
A B C
D E
Figure 9.7. Volume-rendered 3D images from
the gadolinium sequence in Figure 9.3. Right-sided
structures are seen such as the right (RPA) and left
pulmonary arteries (LPA) and the Fontan baffle (B)
from the anterior (A) and posterior (B) views while the
hypoplastic native (nAo) and neoaorta (native main
pulmonary artery) are visualized in (C). Both right- and
left-sided structures are combined in (D) while the
lateral view of the Ao is seen in (E).
LWBK1209-ch09_p125-134.indd 131 16/05/13 9:41 PM

embolus or competitive flow to one lung segment from aor Velocity mapping can even be utilized to assess anatomy
tic to pulmonary collaterals such as occurs in patients with of valves. Through-plane velocity mapping can outline the
single ventricle, decreased signal intensity to that part of the leaflets of valves when imaged en face and can be successful
lung will be seen (Fig. 9.3). when routine cine imaging is not.
Velocity Mapping Delayed Enhancement

Another workhorse for physiology and function in CVMR Detection of myocardial scar (viability) is not just the province
of infants and children is blood flow assessment via through- of adult heart disease (Figs. 9.4 and 9.5) (20–22). Patients with
plane, retrospectively acquired, velocity mapping and is tai congenital coronary artery abnormalities (e.g., anomalous left
lored to the lesion under study (Figs. 9.2 and 9.8). For exam coronary artery from the pulmonary artery), those who have
ple, in a patient with a ventricular septal defect where assessing had surgical manipulation of the coronaries (e.g., transposi
Qp/Qs is important, velocity mapping across the main pulmo tion of the great arteries after arterial switch or a Ross pro
nary artery is used to measure Qp with internal confirmation cedure) as well as those who have had specific inflammatory
by measuring flow across both branch pulmonary arteries. Qs diseases (e.g., Kawasaki’s disease) can all benefit from myocar
is measured by velocity mapping across the aortic root with dial scar assessment. In addition, hearts that have undergone
internal confirmation by measuring flow in the superior and surgical reconstruction with cardiopulmonary bypass and
inferior vena cava (Fig. 9.8). Qp/Qs by through-plane velocity deep hypothermic circulatory arrest are candidates for DE
mapping has been validated against oximetry (10). In com imaging. DE imaging can also be used to delineate anatomy
bination with cine imaging, velocity mapping can measure as it has been demonstrated that surgically placed patches and
atrioventricular valve insufficiency by utilizing the total stroke valves can become bright with this technique (21) (Fig. 9.5).
volume of the ventricle and forward flow out the semilunar
valve (e.g., for repaired atrioventricular canal with a leaking
other Techniques
left atrioventricular valve). Limited use of in-plane velocity
mapping can be used as well to assess lesions such as shunts There are special techniques that do not fall into this gener
(atrial septal defects), narrowed blood vessels (coarctation), alized protocol, which may be inserted at different points,
and valve regurgitation (truncal valve insufficiency). depending on the lesion:
A B C
mL/s
MPA = 7 L/min/m2
Ao = 3.4 L/min/m2
D E
Figure 9.8. Phase-contrast velocity mapping of the main pulmonary artery (MPA) and aorta (Ao) of a
patient with a ventricular septal defect. Flow–time curves (A) are created from the phase–velocity maps of the
MPA (magnitude image in (B), phase map in (C)), and the Ao (magnitude image in (D) and phase map in (E)).
The pulmonary to systemic flow ratio (Qp/Qs) was 2.1 L/min/m2.
LWBK1209-ch09_p125-134.indd 132 16/05/13 9:41 PM

•• Myocardial and blood tagging: In patients where there impossible, the heart falls into the category of functional
is a question of regional wall motion abnormalities (e.g., single ventricle (3,6,7,11,26) (Figs. 9.3–9.7). The aorta may
dilated cardiomyopathy or single ventricle [6,11,23]) or be hypoplastic and there may be other associated anoma
whether or not a tissue is even contracting (e.g., tumor), lies such as anomalous venous connections. Single ventri
myocardial tagging (6,8) can be used to assess this both cles undergo staged surgical reconstruction (in two or three
qualitatively and quantitatively (Fig. 9.4). In addition, to stages, depending upon anatomy and physiology) culminat
outline atrial or ventricular septal defect flow, a saturation ing in the Fontan procedure. The person performing CVMR
band can be laid down on the blood to “tag” it as dark on these patients must be familiar with the various forms
on a gradient-echo image and this dark blood can be fol of single ventricles, physiologic/functional sequelae, asso
lowed so shunting can be visualized. ciated anomalies, and the various surgical reconstructive
•• Coronary artery imaging (22): Using navigators and for techniques. A comprehensive treatment of this evaluation is
some patients, breath-hold imaging, this technique is used beyond the scope of this chapter and the reader is referred to
in infants and children who have had coronary manipula a more specialized text or articles (3,6,7,11,26).
tion (e.g., transposition of the great arteries after arterial At different stages, specific targets of the examination
switch [22]) or with native or acquired coronary disease change, but the overall goal of assessing anatomy, physiol
(e.g., anomalous coronary from the opposite sinus or in ogy, and function remains. At all stages of surgical recon
Kawasaki’s disease). Whole heart SSFP coronary imaging struction, including the single ventricle in the native state,
has been successfully applied in infants and children. the following is the minimum assessment:
•• Perfusion: As with viability, myocardial perfusion is not
•• Aortic arch imaging, aimed mostly at patients with an
solely the province of adult heart disease (20). In pediat
aortic to pulmonary anastomosis (Fig. 9.7)
rics, it is also performed with “adenosine” (140 μg/kg/min
•• Pulmonary artery imaging assessing for stenosis, hypopla
for 4 to 6 minutes) as well as rest. Coronary abnormalities
sia, discontinuity (Figs. 9.3, 9.5, 9.6, and 9.7)
both preoperatively (e.g., pulmonary atresia with intact
•• Anatomic assessment of the Fontan baffle (Fig. 9.7), atrial
ventricular septum) and postoperatively (e.g., transposi
septal defect, ventricular outflow tract obstruction, aortic
tion of the great arteries after arterial switch) are exam
to pulmonary collaterals, anomalous venous structures,
ples where it would be used.
and pulmonary or systemic venous obstruction.
•• Real-time interactive cine imaging: Small septal defects may
•• Ventricular function (Figs. 9.4–9.6) including regional wall
be difficult to visualize using the standard imaging techniques
motion, ejection fraction, ventricular volumes and mass,
of CVMR. Real-time interactive cine imaging can be used
cardiac index, and atrioventricular valve regurgitation.
to “sweep” the atrial or ventricular septum for these defects
•• Velocity mapping (Fig. 9.8) for cardiac index, Qp/Qs, rel
with the imager controlling the imaging plane in “real-time.”
ative flows to both lungs, regurgitation of semilunar and
Similar to adults, there are special protocols for the fol atrioventricular valves (Figs. 9.2 and 9.5), and to assess
lowing diseases which also occur in infants and children: aortic to pulmonary collateral flow. For all stages, the dif
ference between pulmonary venous and pulmonary arte
•• Tumor/mass characterization: Many cardiac tumors occur
rial flow as well as the difference between aortic flow and
in pediatrics, most commonly rhabdomyomas, fibromas,
caval return represent aortic to pulmonary collateral flow
and myxomas. SSFP and T1-weighted images with and
and is an essential evaluation at each stage (26).
without fat saturation and after gadolinium administra
tion, T2-weighted images, gradient-echo imaging, myo In the single ventricle’s native state, a detailed anatomic
cardial tagging, perfusion, and viability are all used to assessment is needed. Anomalous venous structures such as
characterize the mass (24). Functional imaging is used to a decompressing vein from the left atrium, the presence of a
assess the effect of the mass on ventricular performance. left superior vena cava, pulmonary venous drainage, and so
•• Myocarditis: The Lake Louise criteria and protocol are on are all important details to sort out. In addition, because
used in children just as in adults (without data to back some patients may have needed resuscitation, assessment of
this up), including T2 imaging, T1 imaging before and after ventricular function and valve insufficiency via cine and ve
gadolinium, and DE. locity mapping are also extremely important.
•• Arrhythmogenic right ventricular dysplasia: Similar criteria After the Stage I procedure (e.g., for hypoplastic left heart
for adults are used in infants and children and although syndrome), aortic arch imaging via cine and 3D gadolinium
CVMR has been successfully utilized in adults for this sequences is important to evaluate the initial repair because
disease, in pediatrics, there is a question to how useful it of the aortic to pulmonary anastomosis. Visualization of
truly is (25). The protocol includes T1-weighted imaging, the pulmonary artery shunt (either Blalock–Taussig or
cine for ventricular function, one-dimensional right ventric Sano shunt) via dark blood imaging or gadolinium is im
ular myocardial tagging if needed, phase-encoded velocity portant. Qp/Qs is obtained by velocity mapping across the
mapping and DE which has recently shown to be helpful. native aorta and native pulmonary valve (neoaorta); be
cause of aortic to pulmonary collaterals, flows in the pul
monary veins represent Qp. Nevertheless, flow by velocity
A Worked Example—Single mapping across the shunt accompanied by velocity maps in
Ventricles each branch pulmonary artery (using a high velocity encod
ing (VENC) such as 3 m/s and a low TE) is also obtained.
When only one usable ventricle is present to pump blood The atrial septal defect should be assessed and since this
or when both ventricles are linked in such a way as sepa is a volume-loaded stage, ventricular function is also key
ration of the circulations into two pumping chambers are (Figs. 9.5 and 9.6).
LWBK1209-ch09_p125-134.indd 133 16/05/13 9:41 PM

After the bidirectional Glenn or hemi-Fontan stage (supe 3. Fogel MA, Donofrio MT, Ramaciotti C, et al. Magnetic resonance and echocardiographic
imaging of pulmonary artery size throughout stages of Fontan reconstruction. Circulation.
rior vena cava to pulmonary artery anastomosis) (Fig. 9.5), 1994;90:2927–2936.
imaging this connection is the major difference with the Stage 4. Fogel MA, Hubbard A, Weinberg PM. A simplified approach for assessment of intracardiac
baffles and extracardiac conduits in congenital heart surgery with two- and three-dimen
I patients. This is generally done with either cine or gadolin sional magnetic resonance imaging. Am Heart J. 2001;142(6):1028–1036.
ium sequences. Qp/Qs utilizes flow in the superior and inferior 5. Fogel MA, Rychik J, Chin AJ, et al. Evaluation and follow-up of patients with left ventricular
vena cavae and because of aortic to pulmonary collaterals, apical to aortic conduits using two and three - dimensional magnetic resonance imaging
and doppler echocardiography: A new look at an old operation. Am Heart J. 2001;141:
flows in the pulmonary veins represent Qp (26). If a hemi- 630–636.
Fontan was performed, imaging should assess whether any 6. Fogel MA, Weinberg PM, Gupta KB, et al. Mechanics of the single left ventricle: a study in
leak was present from the superior vena cava–pulmonary ar ventricular-ventricular interaction ii. Circulation. 1998;98:330–338.
7. Fogel MA, Weinberg PM, Rychik J, et al. Caval contribution to flow in the branch pulmo
tery anastomosis into the atrium. nary arteries of Fontan patients using a novel application of magnetic resonance presatura
After the Fontan procedure (Figs. 9.3, 9.6, and 9.7), the tion pulse. Circulation. 1999;99:1215–1221.
8. Fogel MA, Weinberg PM, Hubbard A, et al. Diastolic biomechanics in normal infants
most important structure to image is the entire systemic utilizing MRI tissue tagging. Circulation. 2000;102:218–224.
venous pathway (Fontan baffle) for obstruction or clot. 9. Rebergen SA, Chin JGJ, Ottenkamp J, et al. Pulmonary regurgitation in the late postopera
Visualizing the fenestration flow (using cine, in-plane veloc tive follow-up of tetralogy of Fallot. Volumetric quantification by MR velocity mapping.
Circulation. 1993;88:2257–2266.
ity mapping, or blood tagging) and assessment for thrombus 10. Beerbaum P, Korperich H, Barth P, et al. Non-invasive quantification of left-to-right shunt
in this structure are important points to determine. Since it is in pediatric patients. Phase-contrast cine magnetic resonance imaging compared with inva
known that Fontan patients have poor ventricular function, sive oxymetry. Circulation. 2001;10:2476–2482.
11. Fogel MA, Weinberg PM, Chin AJ, et al. Late ventricular geometry and performance
cine for this assessment is an essential part of the examina changes of functional single ventricle throughout staged Fontan reconstruction assessed by
tion. In addition, gadolinium imaging can help determine magnetic resonance imaging. J Am Coll Cardiol. 1996;28(1):212–221.
12. Mor-Avi V, Sugeng L, Weinert L, et al. Fast measurement of LV mass with real-time 3-dimen
the presence of collaterals and to assess the aortic arch. DE sional echocardiography: comparison with MRI. Circulation. 2004;110:1814–1818.
imaging can be used to visualize any myocardial scarring 13. Bu L, Munns S, Zhang H, et al. Rapid full volume data acquisition by real-time 3-dimen
which has resulted and both the reconstructed aorta as well sional echocardiography for assessment of LV indexes in children: A validation study com
pared with MRI. J Am Soc Echocardiogr. 2005;18:299–305.
as the atrioventricular valves will “light up.” Perfusion may 14. Fleenor JT, Weinberg PM, Kramer SS, et al. Vascular rings and their effect on tracheal
also be performed to evaluate for defects. Qp/Qs utilizes geometry. Pediatr Cardiol. 2003;24:430–435.
flow in the superior and inferior vena cavae and because of 15. Beekman R, Hazekamp M, Sobotka M, et al. A new diagnostic approach to vascular rings
and pulmonary slings: the role of MRI. Magn Reson Imaging. 1998;16(2):137–145.
aortic to pulmonary collaterals, flows in the pulmonary veins 16. van Son J, Julsrud P, Hagler D, et al. Imaging strategies for vascular rings. Ann Thorac
represent Qp. Nevertheless, flow in both branch pulmonary Surg. 1994;57(3):604–610.
17. Fogel MA, Pawlowski TW, Harris MA, et al. Comparison and usefulness of cardiac mag
arteries are also obtained via velocity mapping. Flow in the netic resonance versus computed tomography in infants six months of age or younger
Fontan baffle just above the hepatic veins and just below with aortic arch anomalies without deep sedation or anesthesia. Am J Cardiol. 2011;108:
the branch pulmonary artery connection can be used to as 120–125.
18. Martin ET, Coman JA, Shellock FG, et al. Magnetic resonance imaging and cardiac pace
sess fenestration flow (the difference between the two is the maker safety at 1.5-Tesla. J Am Coll Cardiol. 2004;43:1315–1324.
right-to-left shunt). 19. Fogel MA, Weinberg P, Parave E, et al. Safety and efficacy of deep sedation in infants,
children and adolescents undergoing cardiac magnetic resonance. J Pediatr. 2008;152:
534–539.
Conclusion 20. Prakash A, Powell AJ, Krishnamurthy R, et al. Magnetic resonance imaging evaluation of
myocardial perfusion and viability in congenital and acquired pediatric heart disease. Am
J Cardiol. 2004;93:657–661.
CVMR of CHD in infants and children is a unique field 21. Harris M, Johnson T, Weinberg P, et al. Delayed enhancement cardiovascular magnetic
in the world of CVMR. The imager needs to understand resonance identifies fibrous tissue in children after congenital heart surgery. J Thorac
Cardiovasc Surg. 2007;133:676–681.
the anatomy, function, and physiology of native as well as 22. Taylor AM, Dymarkowski S, Hamaekers P, et al. MR coronary angiography and late-
treated CHD (both surgical and catheter based) to perform enhancement myocardial MR in children who underwent arterial switch surgery for trans
position of great arteries. Radiology. 2005;234:542–547.
the examination correctly in addition to overcoming the 23. Fogel MA. Cardiac magnetic resonance of single ventricles. J Cardiovasc Magn Reson.
myriad of technical issues associated with imaging in pediat 2006;8(4):661–670.
rics. Advances in hardware and software are bringing about 24. Beroukhim RS, Prakash A, Buechel ER, et al. Characterization of cardiac tumors in chil
dren by cardiovascular magnetic resonance imaging: a multicenter experience. J Am Coll
new CVMR fields such as functional fetal CVMR (27) and Cardiol. 2011;58:1044–1054.
XMR (the combination of catheterization and CVMR) (28) 25. Fogel MA, Weinberg PM, Rhodes L. Usefulness of magnetic resonance imaging for the
in infants and children. diagnosis of right ventricular dysplasia in children. Am J Cardiol. 2006;97(8):1232–1237.
26. Whitehead KK, Gillespie MJ, Harris MA, et al. Noninvasive quantification of systemic to
pulmonary collateral flow: A major source of inefficiency in patients with superior cavopul
monary connections. Circ Cardiovasc Imaging. 2009;2:405–411.
References 27. Fogel MA, Wilson DR, Flake A, et al. A new method of functional assessment of the fetal
heart using a novel application of “real time” cardiac magnetic resonance imaging. Fetal
Diagn Ther. 2005;20:475–480.
1. Fletcher BD, Jacobstein MD, Nelson AD, et al. Gated magnetic resonance imaging of con
28. Dori Y, Sarmiento M, Glatz AC, et al. X-ray magnetic resonance fusion to internal
genital cardiac malformations. Radiology. 1984;150:137–140.
markers and utility in congenital heart disease catheterization.Circ Cardiovasc Imaging.
2. Higgins CB, Byrd BF, Farmer DW, et al. Magnetic resonance imaging in patients with
2011;4:415–424.
congenital heart disease. Circulation. 1984;70:851–860.
LWBK1209-ch09_p125-134.indd 134 16/05/13 9:41 PM

PART
2
Acquired Heart Disease
LWBK1209-ch10_p135-148.indd 135 16/05/13 9:40 PM

LWBK1209-ch10_p135-148.indd 136 16/05/13 9:40 PM
C h apt e r
10
Gobinath Nadeshalingam
Iacopo Carbone
Matthias G. Friedrich
Magnetic Resonance of
Cardiomyopathies and Myocarditis
■■ The Role of other Diagnostic Modalities in toxic injury or infiltration, with or without a genetic predis
Cardiomyopathies position of the myocardium may contribute to the develop
ment of DCM (3), ARVC (4), and RCM.
■■ Cardiovascular Magnetic Resonance in In clinical settings, cardiomyopathies are typically dia
Cardiomyopathies: General Aspects gnosed after exclusion of other cardiovascular causes, mainly
■■ Cardiovascular Magnetic Resonance explaining coronary artery disease. Importantly, the type of
Approach to the Patient with cardiomyopathy has a strong impact on therapy and progno
sis. Therapy can be guided by the etiology, the disease stage,
Cardiomyopathy
and severity. Thus, imaging techniques are of paramount
Morphology and Function importance for both diagnosis and therapy. The modalities
Tissue frequently used in these diseases are echocardiography, con
Metabolism ventional angiography, radionuclide ventriculography, and
Dilated Cardiomyopathy cardiovascular magnetic resonance (CVMR) imaging.
Cardiovascular Magnetic Resonance
Myocarditis
Hypertrophic Cardiomyopathy The Role of other Diagnostic
Takotsubo Cardiomyopathy Modalities in Cardiomyopathies
Arrhythmogenic Right Ventricular
Cardiomyopathy In clinical routine, transthoracic echocardiography serves as
Restrictive Cardiomyopathy the standard imaging technique. It is easily accessible, nonin
vasive, and fast. But although there are a variety of applica
■■ Infiltrative Secondary Cardiomyopathies and tions of echocardiography in cardiomyopathies (5), results
Endomyocardial Diseases feature substantial interstudy and interobserver variability,
Sarcoidosis which is a limiting factor in their use (6–8). Thus, the reli
Amyloidosis ability of phenotyping a patient with a cardiomyopathy may
Transplant Cardiomyopathy be limited and follow-up data may be inconsistent, especially
Hemochromatosis when different observers are involved (9,10). The use of
Endomyocardial Diseases transesophageal echocardiography (TEE) or techniques such
as acoustic quantification (11), automated border detection
Diabetic Cardiomyopathy
(12), application of contrast (13), and three-dimensional
■■ Noncompaction Cardiomyopathy (3D) postprocessing (14) may improve image quality and
output, yet the endocardial border often remains difficult
Cardiomyopathies are chronic, progressive myocardial dis to detect, especially in the apex (15). Comparative studies
eases with distinct morphologic, functional, and electro indicate that 3D postprocessed echocardiography may be as
physiologic characteristics and have been classified into four accurate as CVMR (16,17). Still, the reliability of volumetric
categories: dilated cardiomyopathy (DCM), hypertrophic results relies on the quality of the raw image data set, which
cardiomyopathy (HCM), restrictive cardiomyopathy (RCM), depending on the disease may be nondiagnostic in approxi
and arrhythmogenic right ventricular cardiomyopathy/ mately 15% of patients. With respect to cardiomyopathies,
dysplasia (ARVC/D) (1). one of the most important limitations of echocardiography
While genetic defects are considered primarily respon (as well as other “ray” modalities) is the lack of techniques
sible for HCM (2), other factors such as inflammatory or to analyze tissue pathology itself. Despite the initial hope to
137
LWBK1209-ch10_p135-148.indd 137 16/05/13 9:40 PM

138 Part 2 ■ Acquired Heart Disease
identify specific pathology-related changes of echogenicity substantial shortening of the ventricular long axis (33) lead
(18), results to date have generally been disappointing. Thus, ing to a smaller number of slices covering the heart in systole
echocardiography can be considered the modality of choice compared with diastole. Commercially available software
for the initial diagnostic workup, but has limited value for for automated edge detection may facilitate the evaluation
a more detailed disease myocardium or myocardial disease. process in clinical routine (34,35), but is so far not capable
Computed tomography has been applied to assess left ven of including papillary muscles and may, therefore, not be
tricle (LV) function and mass (19–21); its value for advanced sufficient for precise measurements.
tissue characterization however is limited, since there are A frequent finding in patients with cardiomyopathies is
no relevant applications of CT for markers for myocardial mild-to-moderate mitral regurgitation. Promoting factors
inflammation and edema, iron, amyloid, or diffuse fibrosis. are dilatation of the mitral valve ring (DCM, infiltrative car
Furthermore, CT is limited by radiation and the necessity diomyopathies) and papillary muscle dysfunction caused by
for contract agents. infiltration (sarcoidosis, amyloidosis, hemochromatosis, or
tumor). If quantification of mitral regurgitation is required
for therapeutic decision making, an established technique
Cardiovascular Magnetic using flow analysis should be performed (36). “Eyeball”
Resonance in Cardiomyopathies: quantification of the regurgitant jet, as used in ventriculog
General Aspects raphy or echocardiography (37), may be misleading and
should be used with great caution, in particular because the
CVMR noninvasively and accurately visualizes left and right application of SSFP sequences results in a greater homogene
ventricular morphology and function (22,23). It is consid ity of the blood pool and might therefore induce underesti
ered the in vivo gold standard for phenotyping patients with mation of valvular turbulences.
suspected or known cardiomyopathies.
As a unique feature of CVMR, it allows obtaining in Tissue
formation on tissue pathology in cardiomyopathies. Since
proton relaxivity depends on the chemical environment, “Black blood” T1-weighted spin-echo techniques are often
pathologic processes with a more or less distinct local chem used for cardiac anatomy because of the excellent contrast
istry may allow a specific identification of diseased tissue. between the myocardium and adjacent structures such
Consequently, disease-specific standardized protocols have as epicardial fat and intracavital blood. Slice orientation
been proposed (24). depends on the question being posed to the MR study; how
ever, the orientation should include views orthogonal to the
anatomic axis of the heart. Gadolinium (Gd) administration
Cardiovascular Magnetic followed by a repeat T1 study may be helpful in infiltrative
Resonance Approach to the and inflammatory myocardial disease. T2-weighted image
Patient with Cardiomyopathy quality has markedly improved with short T1 inversion
recovery (STIR) techniques, and fluid accumulation, such as
Morphology and Function edema and effusion in inflammatory or malignant diseases,
may be sensitively visualized. Visualization of intramyocar
Cardiomyopathies are characterized by specific altera dial fibrosis is certainly desirable for the workup of cardio
tions of ventricular and myocardial geometry, and/or func myopathies. Newer T2-weighted protocols may improve
tion. To assess volumes and mass, generally white blood overall image quality (38). Beyond imaging focal lesions,
steady-state free precession (SSFP) gradient-echo sequences the quantification of absolute T1 or T2 is a reliable marker
are applied during a breath-hold with approximately 30 to detect diffuse myocardial changes (39–42). T1 mapping
phases per heartbeat. A stack of short-axis slices covering has recently emerged as a powerful tool for clinical assess
the entire LV from the mitral plane to the apex can be con ment of myocardial diffuse fibrosis, further building on
sidered the gold standard to assess left ventricular volumes information already provided by late gadolinium enhance
and mass (25,26). Recent developments in parallel imaging ment (LGE)-CVMR. The prospective use of T1 mapping will
allow complete left ventricular coverage in one breath-hold provide much more precise myocardial tissue characteriza
but are not yet considered routine (27–29). Under routine tion (43), such as the quantification of diffuse myocardial
clinical circumstances, a biplanar approach (long- and short- fibrosis in heart failure (44), as well as use in the detection
axis views) may be sufficient (30,31) to estimate mass and of acute myocardial edema (45). T2 mapping may provide a
systolic function. To cover the whole diastolic phase, tech better delineation in the extent of disease when compared to
niques have been developed with continuous data acquisition signal-intensity–based methods (46,47).
and retrospective gating. The inclusion of the end-diastolic
phase is important for the analysis of time–volume curves
Metabolism
with respect to late diastole (32). Reliable angulation of
the images by a series of angulated scouts obtained during Magnetic resonance spectroscopy (MRS) has generally
breath-hold is crucial because the anatomic axis of the heart relied on 1H and 31P and has been applied in several stud
is not perpendicular to any of the orthogonal planes of the ies of cardiomyopathy. Changes of high-energy phosphates
magnetic field. The slice thickness should be 8 to 10 mm; in as studied by 31P-MRS in cardiomyopathy were reported
case of circumscribed or subtle global changes, it should be for DCM (48,49) and HCM (50). However, MRS remains
reduced adequately. It is important to notice that there is a an experimental approach for several reasons: 1H-MRS is
LWBK1209-ch10_p135-148.indd 138 16/05/13 9:40 PM

Chapter 10 ■ Magnetic Resonance of Cardiomyopathies and Myocarditis 139
limited by a strong signal from water-bound protons and

difficulties in spectral interpretation, and 31P-MRS is lim
ited by the weakness of the phosphorus signal. Thus, voxels
must be sufficiently large to cover circumscribed myocar
dial regions, and spectra are often altered by blood or adja
cent tissue (e.g., skeletal muscle). Newer techniques feature
irregularly shaped voxels and a significantly lower degree
of spectral contamination (51). This approach may allow
reproducible acquisition of reliable and highly informa
tive myocardial spectra and even identify local pathology.
Buchthal et al. (52) demonstrated changes in the ratio of
myocardial phosphocreatine to adenosine triphosphate in
women with chest pain but normal angiograms. However,
MRS techniques require extensive experience and are prone
to motion artifacts (53). Thus, the number of centers with
access to this promising tool is currently limited.
The increasing use of higher field strengths such as 3 T in
recent trials will lead to the development of CVMR spec
troscopy tools with high spatial resolution.
Dilated Cardiomyopathy
DCM is characterized by a progressive dilatation of the
heart with loss of contractile function (54). The typical pat Figure 10.1. Diastolic steady-state free precession (SSFP) image
tern may be the final result of a disease process with multiple in a patient with severe left ventricular dilatation and dysfunction in
possible triggers, such as infectious organisms, toxic agents dilated cardiomyopathy.
like anthracyclines, autoantibodies, or genetic disorders.
The histologic hallmark of DCM is a progressive interstitial brosis in patients with DCM, with a remarkable strong re
fi
fibrosis with a numeric decrease of contractile myocytes. In lationship with diastolic function and symptomatology (67).
advanced stages, DCM is also associated with at least rela Recently, CVMR has shown the ability to accurately de
tive wall thinning. tect immunohistologically confirmed myocardial inflamma
tion in patients with DCM (68).
LGE-CVMR is a strong and independent predictor of all-
cause mortality, independent of contractile function (69).
CVMR nowadays recognized as a gold standard for func
tional imaging and assessment and with the state-of-the-art
sequences is considered appropriate for many clinical indica
tions (55). Main targets of CVMR studies in DCM are LV
morphology and function using SSFP gradient-echo sequences
(Fig. 10.1). CVMR has been proven to have low interobserver
and intraobserver variabilities of left ventricular mass and vol
ume measurements (56,57) with a good correlation to results
obtained with positron emission tomography (58). CVMR
was also used to analyze wall thickening in DCM (59), visual
ize impaired fiber shortening (60), and calculate end-systolic
wall stress, which may be a very sensitive parameter for
changes of LV function (61). The right ventricle (RV) is also
frequently affected in DCM, and its morphology and function
are accurately and reproducibly assessed by CVMR (62,63).
CVMR is the method of choice for a longitudinal follow-
up in patients with DCM under pharmacologic interven
tions (64), and its reproducibility allows for a substantial re
duction of the required sample size of clinical trials in DCM
(65). Thus, costs could be reduced markedly and time could
be saved in clinical research.
CVMR has also been shown to successfully detect fibrotic
patterns, which allow for differentiating dilated from isch
emic cardiomyopathy (66). A specific pattern involving in
tramural and especially subepicardial areas is often found Figure 10.2. Inversion-recovery prepared gradient-echo image
in patients with myocarditis (Fig. 10.2). T1 mapping offers (“delayed enhancement”) in a patient with recent myocarditis showing
a completely new approach to detect diffuse myocardial focal necrosis of typical subepicardial localization (arrows).
LWBK1209-ch10_p135-148.indd 139 16/05/13 9:40 PM

Spectroscopic studies have shown that high-energy phos

phate metabolism is altered in DCM (70), and a low ratio
of phosphocreatine to adenosine triphosphate as assessed
by MRS was shown to be of prognostic value in DCM.
Recent data indicate a role in the early detection of meta
bolic changes of the myocardium (71). Future studies will
shed more light on this exciting field of research, and further
clinical studies are warranted.
Myocarditis
Myocarditis, caused by infectious agents, toxicity, or auto
immune processes, may be an important pathophysiologic
component of cardiomyopathies including DCM (72) and
ARVD/C (73). Common features of myocarditis include cel
lular infiltration, edema, necrosis, and fibrotic scars in late
stage (74).
While echocardiography is still performed in cases of
myocarditis to qualitatively assess functional abnormalities,
wall thickness, and pericardial effusion (75), its diagnostic
value may be limited in less severe cases, where function
typically is normal.
One of the main advantages of CVMR over other imaging
modalities is its ability to assess changes in tissue. In myocar Figure 10.3. Short T1 Inversion Recovery (STIR) spin-echo image
ditis, this has been used to assess edema, capillary leakage, of myocardial edema in myocarditis (arrow).
hyperemia, and in severe cases, cellular necrosis and fibrosis
(76). T2-weighted imaging, utilizing triple inversion-recovery
turbo spin-echo sequences with inversion pulses for fat and (87). Edema as studied by T2-weighted CVMR was associ
blood suppression, has shown tremendous contrast for the ated with a transient increase of LV mass in myocarditis (88).
differentiation of edema from myocardium (77). This tech The ability to identify several tissue markers for inflam
nique has displayed high diagnostic accuracy for acute in mation or inflammatory damage renders CVMR the most
flammatory or ischemic injury, which is significant as edema comprehensive noninvasive diagnostic modality for myocar
is an important hallmark of inflammatory injury (78–80). ditis (74).
Signal-to-noise ratio of T2-weighted images still remains an Further, to assess myocardial tissue markers, CVMR cine
issue but newly developed sequences may lead to better image images can be used for examining distribution and hemo
quality and diagnostic accuracy (81). Contrast-enhanced fast dynamic significance of pericardial effusion in patients with
spin-echo T1-weighted CVMR has also been used to assess myocarditis (89–91).
hyperemia and capillary leak in myocarditis (82). This is
done using an early gadolinium enhancement ratio (EGEr) to
Hypertrophic Cardiomyopathy
observe the distribution of contrast in the early distribution
period. As opposed to EGE, Late Gadolinium Enhancement HCM is classified as an autosomal dominant disease of the
(LGE) of the myocardium specifically indicates irreversible cardiac sarcomere and is characterized by generalized or
injury, due to necrosis and fibrosis, thus making it an im localized left ventricular hypertrophy (92). Histologically,
portant diagnostic tool for late stages of myocarditis (74) areas of hypertrophy reveal a pattern of myofibrillar disar
(Fig. 10.3). Clinical studies have demonstrated that LGE has ray, which are not exclusively but predominantly seen in
a high specificity for detection of irreversible injury in pa HCM and patchy areas of necrotic tissue caused by relative
tients with myocarditis (Fig. 10.3) (78,83–86). coronary insufficiency. HCM features inappropriate myo
CVMR has become the primary diagnostic tool for assess cardial hypertrophy with loss of diastolic function. Almost
ment of myocardial inflammation in patients with suspected two-thirds of patients with HCM show a Left ventricular
myocarditis (74). In 2006, The International Consensus outflow tract (LVOT) obstruction (93,94).
Group on CVMR Diagnosis of Myocarditis was founded Endomyocardial biopsy is mainly performed for investiga
in order to achieve a consensus amongst CVMR experts to tive reasons such as analysis of gene expression. However, for
develop and standardize diagnostic protocols for CVMR clinical decision making, the sensitivity and specificity in terms
in myocarditis. One of the significant conclusions reached of exclusion of RCM and pressure-induced hypertrophy are
by this consensus group was for the comprehensive use of not satisfying for an invasive procedure with inherent risks.
CVMR criteria. Concluding that, based on two studies con CVMR studies have been applied to assess mass, func
ducted comparing all three tissue-based techniques (75,76), tion, morphology, tissue characterization, and hemodynamic
presence of at least two positive criteria out of three is to be relevance of obstruction. Because of its high sensitivity for
defined as a positive CVMR study for myocarditis (74). detecting regional morphologic changes and its noninvasive
Persisting CVMR markers for capillary leaks and hyper character, CVMR may be of special importance in screen
emia have been associated with reduced functional recovery ing families of index patients. In addition, there is evidence
LWBK1209-ch10_p135-148.indd 140 16/05/13 9:40 PM

showing the ability of CVMR to predict spontaneous clini data have been published. The detection of fibrotic lesions
cal arrhythmias and sudden cardiac death (92). in HCM might have a huge clinical impact because they
CVMR cine sequences are suitable for functional stud might represent a risk factor for overall risk, especially heart
ies, visualization of turbulent flow in LVOT obstruction, and failure (102,103). Focal T2 abnormalities in approximately
mass quantification. In severe forms, the contracting ventricu one-third of HCM patients may indicate a dynamic disease
lar walls may oppose each other and the end-systolic volume process with acute events leading to irreversible injury (104).
may remain less than 10 mL. Thus, a very careful contour def CVMR evidence for myocardial fibrosis in HCM patients
inition and a contiguous set of short-axis slices are necessary was also associated with an increased frequency of tachyar
to prevent underestimation of end-systolic volume, which is rhythmia (105) and with small intramural coronary arteriole
likely to occur when only long-axis views are used. dysplasia (106).
Mitral valve regurgitation, probably caused by a patho MR techniques have been established to investigate
logic change of leaflet geometry, occurs frequently and phosphate metabolism in HCM. Myocardial phosphocre
should be included in a CVMR workup. atine/adenosine triphosphate ratio and the signal of phos
Diastolic function (or dysfunction) is a powerful clinical phomonoesters were found to be changed in patients with
and prognostic factor in hypertrophy but does not yet be HCM (47) regardless of the presence of hypertrophy (107).
long to the routine measurements of CVMR. Preliminary Spindler et al. (108) were able to correlate diastolic dysfunc
clinical results suggest that tagging analysis of the early tion to a decrease of myocardial energy reserve related to
untwisting motion of the apical myocardium may be a high-energy phosphate metabolism. Phosphorus metabolism
helpful tool to assess diastolic dysfunction in hypertro was also found to be altered in skeletal muscles of patients
phic heart diseases (95,96). Other functional changes with HCM (109).
detected by the use of myocardial tagging are reduced The turbulent jet during systolic LVOT obstruction is eas
posterior rotation, reduced radial displacement of the in ily detected when suitable echo times (∼4 milliseconds for
ferior septal myocardium (97), heterogeneities of regional typical blood flow velocities) with standard cine gradient-
function (98), and reduced 3D myocardial shortening echo sequences are used. Current SSFP gradient-echo images
(99,100). These findings may be more sensitive in the de are often impaired by flow artifacts. The systolic anterior
tection and quantification of functional impairment than motion of the anterior mitral valve leaflet may contribute
conventional parameters such as mitral valve inflow pat significantly to the LVOT obstruction and is a typical fea
terns in echocardiography. ture of obstructive HCM. The phenomenon is detectable by
Beyond quantification of left ventricular mass and func CVMR (110) and best visualized in the four-chamber view
tion, CVMR is capable in characterizing hypertrophic tis or a short-axis view through the valvular plane perpendicu
sue. LGE imaging reveals myocardial heterogeneity in HCM lar to the outflow tract.
(Fig. 10.4) (101). On the basis of similar lesions in myo A promising approach to assess the hemodynamic rele
cardial infarctions, these areas most likely represent zones vance of LVOT obstruction may be the noninvasive measure
of focal fibrosis, although only few histologic comparison ment of the effective outflow tract area by CVMR planimetry
of transplanar flow in the LVOT during systole (111). Schulz-
Menger et al. (112) used this method to show that CVMR
was able to accurately differentiate obstructive from nonob
structive HCM. This method may overcome limitations of
pressure gradient measurements by echocardiography.
CVMR may also be very important in the follow-up of
patients after surgical (107) or pharmacologic intervention.
A long-term follow-up is sensitive to morphologic changes
during the natural course of the disease (113) and after in
tervention (111,114). The acute and chronic morphologic
and functional changes caused by the interventional ablation
of a septal artery in obstructive HCM are well shown by
T1-weighted images.
CVMR is helpful in distinguishing HCM from Anderson–
Fabry disease (AFD), a genetic X-linked sphingolipid meta
bolic disorder, frequently associated with left ventricular hy
pertrophy. CVMR shows a specific pattern of diffuse, mainly
subendocardial enhancement in patients with AFD (115).
Larger scale prospective study is underway to elucidate key
identifying features of AFD and optimal treatment protocol.
Takotsubo Cardiomyopathy
Stress-induced or Takotsubo cardiomyopathy (TTC) is char
Figure 10.4. Inversion-recovery prepared gradient-echo image acterized by severe, transitory LV dysfunction with a specific
(“delayed enhancement”) in a patient with hypertrophic cardiomyopa regional distribution pattern, most frequently involving apical
thy (HCM) showing focal fibrosis (arrows). segments in the absence of evidence for ischemic injury or
LWBK1209-ch10_p135-148.indd 141 16/05/13 9:40 PM

matching coronary artery disease (116). CVMR provides the

ability to detect the presence of typical regional wall motion
abnormalities, associated with reversible myocardial injury
(edema), in the absence of areas of irreversible injury (117).

Cardiomyopathy
ARVC is characterized by a progressive degeneration of
the right ventricular myocardium and to a lesser extent
the left ventricular myocardium with regional dysfunction.
Morphologic features include fibrous and/or fatty replace
ment of myocardial tissue, extensive wall thinning, and
atypical arrangement of trabecular muscles. In 1994, a set of
criteria was defined to establish the diagnosis (118). CVMR
is a very important tool for noninvasively examining and
assessing RV anatomy and function. The task force criteria
have been recently modified (119). The impact of the modi
fication is under debate (120). The associated morphologic
spectrum ranges from subtle changes to extensive fibrofatty
dysplasia of the RV (118,121) leading to RV congestive heart
failure in rare cases. However, the clinical course is generally
determined by the occurrence of severe ventricular arrhyth
mias with a substantial risk of sudden cardiac death (122). Figure 10.5. SSFP image in a patient with arrhythmogenic right
Endomyocardial biopsy of affected myocardium shows ventricular cardiomyopathy (ARVC). Note the marked dilatation of
fibrous and/or fatty replacement of myocytes and is accepted the right ventricle (RV).
as strongly suggestive for the diagnosis of ARVC. However,
fatty replacement is also found in other cardiac diseases Restrictive Cardiomyopathy
like DCM, myocarditis, or alcoholic myocardial injury.
Therefore, the specificity of this finding has been questioned. Primary infiltration of the myocardium by fibrosis or other
Furthermore, the correlation of morphology and electro tissues leads to the rare entity of RCM associated with a
physiologic findings in ARVC is not well understood. grave prognosis (131). The condition is characterized by a
Several studies have confirmed the accuracy and re severe diastolic dysfunction, biatrial dilatation, preserved
producibility of CVMR in the assessment of the RV (63). LV size, and usually normal systolic function (132). Atrial
Furthermore, CVMR studies have helped to elucidate that the thrombi may occur in patients with this disease. The main
degree of associated LV dysfunction in fact parallels the de differential diagnostic consideration is constrictive pericar
gree of RV dysfunction in ARVC/D (123). Quantitative MRI, ditis, which must be excluded in a patient with suspected
in addition to qualitative assessment, has strong clinical impli RCM. Pericardial thickening as an important diagnostic clue
cations in increasing the sensitivity to detect ARVC/D (124). in constrictive pericarditis may not be present in all cases
With the prone position, a dedicated cardiac-phased (133) or may not be adequately visualized by ultrasound.
array coil, and a slice thickness equal to or less than 6 mm, CVMR studies in RCM should focus on myocardial mor
black blood T1-weighted spin-echo studies may accurately phology and function, as well as on the exclusion of con
visualize fatty infiltration and wall thinning as well as dys strictive pericardial disease. Biatrial dilatation is easily visu
plastic trabecular structures. Orthogonal image planes (axial alized and CVMR volumetry of the enlarged atria may be
and sagittal) and additional short-axis view reveal the best useful (134). A combined approach using contrast-enhanced
results. Additional studies with fat suppression may help and nonenhanced CVMR techniques may allow to detect/
to differentiate fat from fibrous tissue. Schick et al. (125) assess atrial thrombi (135).
proposed water- and fat-selective gradient-echo imaging for
better differentiation of fat, fibers, and myocardium. SSFP Infiltrative Secondary
cine sequences are used to detect RV dilatation (Fig. 10.5), Cardiomyopathies and
global or local hypokinesia, localized early diastolic bulging, Endomyocardial Diseases
or, more specifically, circumscribed saccular aneurysmatic
outpouchings. Abnormalities have been most frequently de Infiltrative cardiomyopathies include sarcoidosis, amyloido
scribed in infundibular, apical, and subtricuspid areas (tri sis, and hemochromatosis. The myocardium may be infil
angle of dysplasia) (126,127). Although many centers use trated in these systemic diseases leading to impairment of
black blood fast spin-echo CVMR to detect myocardial fat function and/or conduction abnormalities. Compared with
as a hallmark of ARVC (128), the detection of fat by CVMR other forms of cardiomyopathy, infiltrative myocardial
generally has an insufficient diagnostic accuracy (129) and disease was found to have a worse prognosis (131). Since
consequently is not part of the task force criteria. infiltration of the tissue is accompanied by changes of myo
Acute, reversible RV dysfunction has been observed in cardial signal properties, CVMR may become the most pow
endurance athletes (130). erful diagnostic tool, although the specificity may be limited.
LWBK1209-ch10_p135-148.indd 142 16/05/13 9:40 PM

Figure 10.6. T1-weighted spin-echo image in a patient with cardiac Figure 10.7. Morphologic changes in amyloid heart disease in an
sarcoidosis. Lateral area of marked contrast enhancement indicates end-diastolic frame (SSFP). Note the marked thickening of right and
focal inflammation. left ventricular walls and interatrial septum.
Sarcoidosis is diffuse, endomyocardial biopsy is very sensitive in the

detection of amyloid infiltrates. Echocardiography reveals
Sarcoidosis, a granulomatous inflammation of unknown
increased wall thickness, a marked increase of myocardial
cause, frequently involves the myocardium (136), with up
echogenicity, the combination of small ventricles with large
to 50% of deaths in sarcoidosis related to cardiac involve
atria, and diastolic dysfunction (Fig. 10.7). Thrombi can be
ment (137), primarily because of sudden death and conges
found in the atrial appendages. Attempts have been made
tive heart failure. Histologic evaluation shows granulomas,
to visualize myocardial changes in amyloidosis using ultra
which are often transmural in distribution. The sensitivity of
sound with cycle-dependent backscatter (152).
endomyocardial biopsy is less than 30%, probably because
CVMR is an excellent tool to identify cardiac amyloido
of the focal appearance (138,139). As in other organs, gad
sis based on its characteristic to substantially increase the
olinium agents accumulate in sarcoid lesions of the heart
extracellular volume distribution of gadolinium, which in
(140,141). T2-weighted sequences followed by T1-weighted
fact is removed from the blood pool much faster than in
spin-echo techniques in short-axis and long-axis orientation
normal individuals. This leads to the specific combination
before and after a standard dose of gadolinium are useful
of findings, that is, a diffuse rapid uptake of gadolinium
to detect or exclude suspected granulomas (Fig. 10.6) (142)
in the myocardium with an equally fast clearance of the
or more general myocardial inflammation (143). Contrast-
blood pool from the agent (153). Ruberg et al. (154) dem
enhanced CVMR findings seem to correlate with clinical sta
onstrated the high sensitivity and specificity of LGE imag
tus and steroid therapy (144–146). CVMR findings can be
ing for the identification of cardiac involvement in patient
used to “guide” endomyocardial biopsy (147).
with systemic amyloidosis. The use of CVMR may prove to
Recently, Patel et al. (148) found that LGE imaging is
be a key component in determining prevalence of cardiac
more sensitive for identifying cardiac involvement in pa
involvement in systemic amyloidosis patients when mor
tients with sarcoidosis than other clinical protocols. In ad
phologic changes are not apparent through echocardio
dition, abnormal areas in LGE images were associated with
graphy (155).
future adverse events, including cardiac death, although this
association requires further confirmation in larger cohort
studies. Transplant Cardiomyopathy
Histologic evaluation of biopsy samples is still considered
Amyloidosis
the gold standard for the diagnosis of transplant rejection
Infiltration of the heart by amyloid deposits is found in (156). Noninvasive detection of allograft rejection might
almost all cases of primary amyloidosis and in approxi guide biopsy and finally replace it. Clinically moderate
mately one-fourth of familial amyloidosis (149), leading to forms of rejection (grades 2 and 3) require intravenous drug
a loss of atrial (150) or left ventricular function and conges therapy, so imaging does have implications for clinical man
tive left ventricular failure (151). Since the infiltration often agement.
LWBK1209-ch10_p135-148.indd 143 16/05/13 9:40 PM

Several animal studies revealed the potential of calculat followed by extensive subendocardial fibrosis and frequent
ing myocardial relaxation times to detect edema (39,157). apical thrombus formation. Both ventricles may be affected
Clinical studies were published on several CVMR methods with progressive diastolic dysfunction and decreased stroke
like T2 quantification (40) or a combination of several tech volumes.
niques (158). 31P MRS was successfully applied to assess The morphologic and functional features can be well vi
donor hearts (159). Transplant rejection and perfusion ab sualized and quantified by CVMR (171,172). Fibrosis or
normalities associated with transplant arteriopathy can be calcifications may be visible as a dark rim in bright blood
detected by contrast-enhanced CVMR (160). Sufficient clini prepared gradient-echo sequences, but may also reveal an
cal experience and standard protocols, however, still have to intermediate signal intensity (173). T1 quantification may
be defined. detect diffuse changes that are not large enough to be detect
Gadolinium-enhanced CVMR allows for identification of able by visual evaluation. Differentiation from apical infarc
silent myocardial infarction in heart transplantation patients tion is easy by visualization of the preserved (or increased)
with absent or mild angiographic transplant-related coro wall thickness and the V-shaped outer form of the apex in
nary artery disease (161). the long-axis views.
Hemochromatosis Diabetic Cardiomyopathy

Hemochromatosis of the myocardium is sometimes Diabetic cardiomyopathy is characterized by systolic and
characterized by extensive iron deposits leading to wall diastolic ventricular dysfunction as well as hypertrophy. In
thickening, ventricular dilatation, progressive loss of func patients with diabetes mellitus type 2 (DM2), the risk for
tion with congestive heart failure, and subsequent death. heart failure is much greater when compared with traditional
Because of a predominantly subepicardial deposition of risk factors such as hypertension and coronary artery dis
iron, endomyocardial biopsy may fail to confirm the diag ease. McGavock et al. (174) used 1H-MRS to show that
nosis (162). Hemochromatosis can also occur in diseases of impaired glucose tolerance (IGT) was accompanied with
the hemoglobin metabolism, as in thalassemia. In patients cardiac steatosis in patients with DM2. The study’s clinical
with myocardial siderosis, a toxic DCM, the effects can be significance demonstrated that cardiac steatosis actually pre
reversed with aggressive chelation, if diagnosed and treated ceded DM2 and eventual LV systolic dysfunction.
early (163). As with other cardiomyopathies, comprehensiveness is
The CVMR approach is directed toward the detection of a specific strength of CVMR, which allows for efficiently
iron deposits as the specific marker for the disease. Iron has scanning functional and metabolic changes as well as
very strong paramagnetic properties, and myocardial depos morphologic changes in patients with suspected diabetic
its imply an extensive signal loss in native T1-weighted, but cardiomyopathy (175).
also T2- and T2*-weighted imaging (164–166). The pattern
of focal signal loss in a dysfunctional myocardium combined
with a hypointense liver signal may be sufficient to confirm Noncompaction Cardiomyopathy
the diagnosis of cardiac hemochromatosis by CVMR alone.
T2* quantification has successfully been applied to assess the In recent years, noncompaction cardiomyopathy has gained
relationship of myocardial iron content to liver iron (167) attention as a rare entity distinct from other cardiomyopa
and to monitor myocardial iron and ventricular function thies (176). The endocardial rim of the myocardium appears
during therapy (168). T2* measurements have been shown as a loose meshwork because of the incomplete intrauterine
to be reproducible and strong predictors of clinical outcome, development of the myocardial layers (Fig. 10.8). Therefore,
especially in asymptomatic beta-thalassemia patients with the myocardium looks thickened but with deep trabecula
myocardial siderosis (169). In fact, cardiac T2*-weighted tions. The end-diastolic volume can be increased, the systolic
imaging in patients with beta-thalassemia major was more function can be depressed, and the RV may also be involved.
predictive for the development of heart failure and arrhyth Isolated noncompaction cardiomyopathy is defined in the
mia than measures of liver iron and serum ferritin. Similarly, absence of other defects such as congenital heart disease.
Pennell et al. (170) assessed the combined treatment of che Mutations with a wide phenotypic spectrum have been
lation and cardiac T2* for treatment of myocardial siderosis described. Transitions between DCM and noncompaction
in patients with beta-thalassemia major. The study showed cardiomyopathy may exist. Patients who have congestive
that the combination of early detection of myocardial sid heart failure, arrhythmias, and thromboembolic events have
erosis by T2* CVMR and treatment with deferiprone was a poor prognosis (177).
more effective than current protocol of long-term deferox CVMR not only diagnoses this disease entity accurately
amine use. but also differentiates it from other forms of left ventricular
hypertrophy (178). Petersen et al. (178) established a thick
ness ratio of noncompacted over compacted myocardium of
Endomyocardial Diseases
≥2.3 predictive for LV noncompaction; yet the sole use of
Endomyocardial fibrosis is related to two forms: one this criterion may lead to “over-calling” (179).
occurring in the tropics and one occurring in a temperate Of note, the improved spatial resolution of CVMR may
climate, termed “Löffler endocarditis.” Both conditions allow for a better understanding of the potentially dynamic
lead to primarily posterobasal concentric wall thickening process of LV hypertrabeculation.
LWBK1209-ch10_p135-148.indd 144 16/05/13 9:40 PM

15. Moon JC, Fisher NG, McKenna WJ, et al. Detection of apical hypertrophic cardiomyopathy
by cardiovascular magnetic resonance in patients with non-diagnostic echocardiography.
Heart. 2004;90(6):645–649.
16. Kim WY, Sogaard P, Egeblad H, et al. Three-dimensional echocardiography with tissue
harmonic imaging shows excellent reproducibility in assessment of left ventricular vol
umes. J Am Soc Echocardiogr. 2001;14(6):612–617.
17. Chuang ML, Beaudin RA, Riley MF, et al. Three-dimensional echocardiographic
measurement of left ventricular mass: Comparison with magnetic resonance imaging
and two-dimensional echocardiographic determinations in man. Int J Card Imaging.
2000;16(5):347–357.
18. Skorton DJ, Collins SM. Clinical potential of ultrasound tissue characterization in cardio
myopathies. J Am Soc Echocardiogr. 1988;1(1):69–77.
19. Gerber TC, Behrenbeck T, Allison T, et al. Comparison of measurement of left ventri
cular ejection fraction by Tc-99m sestamibi first-pass angiography with electron beam
computed tomography in patients with anterior wall acute myocardial infarction. Am J
Cardiol. 1999;83(7):1022–1026.
20. Kivelitz DE, Enzweiler CN, Wiese TH, et al. Determination of left ventricular function
parameters and myocardial mass: Comparison of MRI and EBT. Rofo. 2000;172(3):244–
250.
21. Mahnken AH, Koos R, Katoh M, et al. Sixteen-slice spiral CT versus MR imaging for
the assessment of left ventricular function in acute myocardial infarction. Eur Radiol.
2005;15(4):714–720.
22. Benjelloun H, Cranney GB, Kirk KA, et al. Interstudy reproducibility of biplane cine
nuclear magnetic resonance measurements of left ventricular function. Am J Cardiol.
1991;67(16):1413–1420.
23. Pattynama PM, Lamb HJ, Van der Velde EA, et al. Reproducibility of MRI-derived
measurements of right ventricular volumes and myocardial mass. Magn Reson Imaging.
1995;13(1):53–63.
24. Taylor AJ, Cerqueira M, Hodgson JM, et al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/
SCAI/SCMR 2010 appropriate use criteria for cardiac computed tomography. A report of
the American College of Cardiology Foundation Appropriate Use Criteria Task Force, the
Society of Cardiovascular Computed Tomography, the American College of Radiology,
Figure 10.8. Noncompaction cardiomyopathy. End-diastolic still the American Heart Association, the American Society of Echocardiography, the
frame of an SSFP data set showing a thick noncompacted myocardial American Society of Nuclear Cardiology, the North American Society for Cardiovascular
Imaging, the Society for Cardiovascular Angiography and Interventions, and the Society
layer filling large middle and apical parts of the left ventricle (LV) for Cardiovascular Magnetic Resonance. J Am Coll Cardiol. 2010;56(22):1864–1894.
(arrows). 25. Bellenger NG, Burgess MI, Ray SG, et al. Comparison of left ventricular ejection frac
tion and volumes in heart failure by echocardiography, radionuclide ventriculogra
phy and cardiovascular magnetic resonance; are they interchangeable? Eur Heart J.
2000;21(16):1387–1396.
References 26. van der Geest RJ, Reiber JH. Quantification in cardiac MRI. J Magn Reson Imaging.
1999;10(5):602–608.
1. Richardson P, McKenna W, Bristow M, et al. Report of the 1995 World Health 27. Sodickson DK. Tailored SMASH image reconstructions for robust in vivo parallel MR
Organization/International Society and Federation of Cardiology Task Force on the imaging. Magn Reson Med. 2000;44(2):243–251.
Definition and Classification of cardiomyopathies. Circulation. 1996;93(5):841–842. 28. Jakob PM, Griswold M, Hillenbrand C, et al. High speed and high resolution cardiac
2. Spirito P, Seidman CE, McKenna WJ, et al. The management of hypertrophic cardiomy MRI (parallel acquisition techniques & modular imaging). MAGMA. 2000;11(1–2):52–
opathy [see comments]. N Engl J Med. 1997;336(11):775–785. 54.
3. Bender JR. Idiopathic dilated cardiomyopathy. An immunologic, genetic, or infectious 29. Pruessmann KP, Weiger M, Boesiger P. Sensitivity encoded cardiac MRI. J Cardiovasc
disease, or all of the above? [editorial; comment].Circulation. 1991;83(2):704–706. Magn Reson. 2001;3(1):1–9.
4. Burke AP, Farb A, Tashko G, et al. Arrhythmogenic right ventricular cardiomyopathy and 30. Lawson MA, Blackwell GG, Davis ND, et al. Accuracy of biplane long-axis left ventricu
fatty replacement of the right ventricular myocardium: Are they different diseases? [see lar volume determined by cine magnetic resonance imaging in patients with regional and
comments]. Circulation. 1998;97(16):1571–1580. global dysfunction. Am J Cardiol. 1996;77(12):1098–1104.
5. Wood MJ, Picard MH. Utility of echocardiography in the evaluation of individuals with 31. Friedrich MG, Schulz-Menger J, Strohm O, et al. The diagnostic impact of 2D- versus
cardiomyopathy. Heart. 2004;90(6):707–712. 3D-left ventricular volumetry by MRI in patients with suspected heart failure. MAGMA.
6. Germain P, Roul G, Kastler B, et al. Inter-study variability in left ventricular mass 2000;11(1–2):16–19.
measurement. Comparison between M-mode echography and MRI. Eur Heart J. 32. Feinstein JA, Epstein FH, Arai AE, et al. Using cardiac phase to order reconstruc
1992;13(8):1011–1019. tion (CAPTOR): A method to improve diastolic images. J Magn Reson Imaging.
7. van Royen N, Jaffe CC, Krumholz HM, et al. Comparison and reproducibility of visual 1997;7(5):794–798.
echocardiographic and quantitative radionuclide left ventricular ejection fractions. Am J 33. Rogers WJ Jr, Shapiro EP, Weiss JL, et al. Quantification of and correction for left ven
Cardiol. 1996;77(10):843–850. tricular systolic long-axis shortening by magnetic resonance tissue tagging and slice isola
8. Jensen-Urstad K, Bouvier F, Hojer J, et al. Comparison of different echocardiographic meth tion. Circulation. 1991;84(2):721–731.
ods with radionuclide imaging for measuring left ventricular ejection fraction during acute 34. Singleton HR, Pohost GM. Automatic cardiac MR image segmentation using edge
myocardial infarction treated by thrombolytic therapy. Am J Cardiol. 1998;81(5):538–544. detection by tissue classification in pixel neighborhoods. Magn Reson Med. 1997;37(3):
9. Gottdiener JS, Livengood SV, Meyer PS, et al. Should echocardiography be performed to 418–424.
assess effects of antihypertensive therapy? Test-retest reliability of echocardiography for 35. Graves MJ, Berry E, Eng AA, et al. A multicenter validation of an active contour-based
measurement of left ventricular mass and function. J Am Coll Cardiol. 1995;25(2):424– left ventricular analysis technique. J Magn Reson Imaging. 2000;12(2):232–239.
430. 36. Fujita N, Chazouilleres AF, Hartiala JJ, et al. Quantification of mitral regurgitation by
10. Gardin JM. How reliable are serial echocardiography measurements in detecting velocity-encoded cine nuclear magnetic resonance imaging [see comments]. J Am Coll
regression in left ventricular hypertrophy and changes in function? J Am Coll Cardiol. Cardiol. 1994;23(4):951–958.
1999;34(5):1633–1636. 37. McCully RB, Enriquez Sarano M, Tajik AJ, et al. Overestimation of severity of ischemic/
11. Chandra S, Bahl VK, Reddy SC, et al. Comparison of echocardiographic acoustic quanti functional mitral regurgitation by color Doppler jet area. Am J Cardiol. 1994;74(8):790–
fication system and radionuclide ventriculography for estimating left ventricular ejection 793.
fraction: Validation in patients without regional wall motion abnormalities. Am Heart J. 38. Cocker MS, Shea SM, Strohm O, et al. A new approach towards improved visualization
1997;133(3):359–363. of myocardial edema using T2-weighted imaging: A cardiovascular magnetic resonance
12. Lucariello RJ, Sun Y, Doganay G, et al. Sensitivity and specificity of left ventricular ejec (CMR) study. J Magn Reson Imaging. 2011;34:286–292.
tion fraction by echocardiographic automated border detection: Comparison with radio 39. Lund G, Morin RL, Olivari MT, et al. Serial myocardial T2 relaxation time measure
nuclide ventriculography. Clin Cardiol. 1997;20(11):943–948. ments in normal subjects and heart transplant recipients. J Heart Transplant. 1988;7(4):
13. Malm S, Frigstad S, Sagberg E, et al. Accurate and reproducible measurement of left ven 274–279.
tricular volume and ejection fraction by contrast echocardiography: A comparison with 40. Walker PM, Marie PY, Danchin N, et al. Comparison of T1 estimation techniques in
magnetic resonance imaging. J Am Coll Cardiol. 2004;44(5):1030–1035. cardiac MRI. Magn Reson Imaging. 1994;12(1):43–50.
14. Kupferwasser I, Mohr-Kahaly S, Stahr P, et al. Transthoracic three-dimensional echo 41. Marie PY, Angioi M, Carteaux JP, et al. Detection and prediction of acute heart trans
cardiographic volumetry of distorted left ventricles using rotational scanning. J Am Soc plant rejection with the myocardial T2 determination provided by a black-blood mag
Echocardiogr. 1997;10(8):840–852. netic resonance imaging sequence. J Am Coll Cardiol. 2001;37(3):825–831.
LWBK1209-ch10_p135-148.indd 145 16/05/13 9:40 PM

42. Anderson LJ, Holden S, Davis B, et al. Cardiovascular T2-star (T2*) magnetic resonance 69. Cheong BY, Muthupillai R, Wilson JM, et al. Prognostic significance of delayed-
for the early diagnosis of myocardial iron overload. Eur Heart J. 2001;22(23):2171–2179. enhancement magnetic resonance imaging: Survival of 857 patients with and without
43. Mewton N, Liu CY, Croisille P, et al. Assessment of myocardial fibrosis with cardiovascu ventricular dysfunction. Circulation. 2009;120(21):2069–2076.
lar magnetic resonance. J Am Coll Cardiol. 2011;57(8):891–903. 70. Hardy CJ, Weiss RG, Bottomley PA, et al. Altered myocardial high-energy phosphate
44. Iles L, Pfluger H, Phrommintikul A, et al. Evaluation of diffuse myocardial fibrosis in metabolites in patients with dilated cardiomyopathy. Am Heart J. 1991;122(3 Pt 1):795–
heart failure with cardiac magnetic resonance contrast-enhanced T1 mapping. J Am Coll 801.
Cardiol. 2008;52(19):1574–1580. 71. Scheuermann-Freestone M, Madsen PL, Manners D, et al. Abnormal cardiac and
45. Ferreira V, Piechnik SK, Dall’Armellina E, et al. T1-mapping accurately detects acute skeletal muscle energy metabolism in patients with type 2 diabetes. Circulation.
myocardial edema: A comparison to T2-weighted cardiovascular magnetic resonance 2003;107(24):3040–3046.
imaging. J Cardiovasc Magn Reson. 2012;14:42. 72. Kawai C. From myocarditis to cardiomyopathy: Mechanisms of inflammation and cell
46. Thavendiranathan P, Walls M, Giri S, et al. Improved detection of myocardial involve death: Learning from the past for the future. Circulation. 1999;99:1091–1100.
ment in acute inflammatory cardiomyopathies using T2 mapping. Circ Cardiovasc 73. Calabrese F, Basso C, Carturan E, et al. Arrhythmogenic right ventricular cardiomyopa
Imaging. 2012;5(1):102–110. thy/dysplasia: Is there a role for viruses? Cardiovasc Pathol. 2006;15:11–17.
47. Ghugre NR, Ramanan V, Pop M, et al. Quantitative tracking of edema, hemorrhage, and 74. Friedrich MG, Sechtem U, Schulz-Menger J, et al. Cardiovascular magnetic resonance in
microvascular obstruction in subacute myocardial infarction in a porcine model by MRI. myocarditis: A JACC white paper. J Am Coll Cardiol. 2009;53:1475–1487.
Magn Reson Med. 2011;66(4):1129–1141. 75. Felker GM, Boehmer JP, Hruban RH, et al. Echocardiographic findings in fulminant and
48. Neubauer S, Krahe T, Schindler R, et al. 31P magnetic resonance spectroscopy in dilated acute myocarditis. J Am Coll Cardiol. 2000;36:227–232.
cardiomyopathy and coronary artery disease. Altered cardiac high-energy phosphate 76. Kishimoto C, Hiraoka Y. Clinical and experimental studies in myocarditis. Curr Opin
metabolism in heart failure.Circulation. 1992;86(6):1810–1818. Cardiol. 1994;9:349–356.
49. Neubauer S, Horn M, Cramer M, et al. Myocardial phosphocreatine-to-ATP ratio 77. Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging technique for the visu
is a predictor of mortality in patients with dilated cardiomyopathy. Circulation. alization of myocardial infarction. Radiology. 2001;218:215–223.
1997;96(7):2190–2196. 78. Abdel-Aty H, Boye P, Zagrosek A, et al. Diagnostic performance of cardiovascular mag
50. Jung WI, Sieverding L, Breuer J, et al. 31P NMR spectroscopy detects metabolic abnor netic resonance in patients with suspected acute myocarditis: Comparison of different
malities in asymptomatic patients with hypertrophic cardiomyopathy. Circulation. approaches. J Am Coll Cardiol. 2005;45:1815–1822.
1998;97(25):2536–2542. 79. Gutberlet M, Spors B, Thoma T, et al. Suspected chronic myocarditis at cardiac MR:
51. Lóffler R, Sauter R, Kolem H, et al. Localized spectroscopy from anatomically matched Diagnostic accuracy and association with immunohistologically detected inflammation
compartments: Improved sensitivity and localization for cardiac 31P MRS in humans. and viral persistence. Radiology. 2008;246:401–409.
J Magn Reson. 1998;134(2):287–299. 80. Abdel-Aty H, Zagrosek A, Schulz-Menger J, et al. Delayed enhancement and T2-weighted
52. Buchthal SD, den Hollander JA, Merz CN, et al. Abnormal myocardial phosphorus-31 cardiovascular magnetic resonance imaging differentiate acute from chronic myocardial
nuclear magnetic resonance spectroscopy in women with chest pain but normal coronary infarction. Circulation. 2004;109:2411–2416.
angiograms. N Engl J Med. 2000;342(12):829–835. 81. Aletras AH, Kellman P, Derbyshire JA, et al. ACUT2E TSE-SSFP: A hybrid method for
53. Cannon RO 3rd, Balaban RS. Chest pain in women with normal coronary angiograms. T2-weighted imaging of edema in the heart. Magn Reson Med. 2008;59:229–235.
N Engl J Med. 2000;342(12):885–887. 82. Miller DD, Holmvang G, Gill JB, et al. MRI detection of myocardial perfusion changes
54. Kasper EK, Agema WR, Hutchins GM, et al. The causes of dilated cardiomyopathy: A by gadolinium-DTPA infusion during dipyridamole hyperemia. Magn Reson Med.
clinicopathologic review of 673 consecutive patients [see comments]. J Am Coll Cardiol. 1989;10:246–255.
1994;23(3):586–590. 83. Mahrholdt H, Goedecke C, Wagner A, et al. Cardiovascular magnetic resonance assess
55. Hendel RC, Patel MR, Kramer CM, et al. ACCF/ACR/SCCT/SCMR/ASNC/NASCI/SCAI/ ment of human myocarditis: A comparison to histology and molecular pathology.
SIR 2006 appropriateness criteria for cardiac computed tomography and cardiac magnetic Circulation. 2004;109:1250–1258.
resonance imaging: A report of the American College of Cardiology Foundation Quality 84. Laissy JP, Hyafil F, Feldman LJ, et al. Differentiating acute myocardial infarction from
Strategic Directions Committee Appropriateness Criteria Working Group, American College myocarditis: Diagnostic value of early- and delayed perfusion cardiac MR imaging.
of Radiology, Society of Cardiovascular Computed Tomography, Society for Cardiovascular Radiology. 2005;237:75–82.
Magnetic Resonance, American Society of Nuclear Cardiology, North American Society for 85. Ingkanisorn WP, Paterson DI, Calvo KR, et al. Cardiac magnetic resonance appearance
Cardiac Imaging, Society for Cardiovascular Angiography and Interventions, and Society of of myocarditis caused by high dose IL-2: Similarities to community-acquired myocarditis.
Interventional Radiology. J Am Coll Cardiol. 2006;48(7):1475–1497. J Cardiovasc Magn Reson. 2006;8:353–360.
56. Strohm O, Schulz-Menger J, Pilz B, et al. Measurement of left ventricular dimen 86. Mahrholdt H, Wagner A, Deluigi CC, et al. Presentation, patterns of myocardial damage,
sions and function in patients with dilated cardiomyopathy. J Magn Reson Imaging. and clinical course of viral myocarditis. Circulation. 2006;114:1581–1590.
2001;13(3):367–371. 87. Wagner A, Schulz-Menger J, Dietz R, et al. Long-term follow-up of patients with acute
57. Semelka RC, Tomei E, Wagner S, et al. Interstudy reproducibility of dimensional and myocarditis by magnetic resonance imaging. MAGMA. 2003;16:17–20.
functional measurements between cine magnetic resonance studies in the morphologically 88. Zagrosek A, Wassmuth R, Abdel-Aty H, et al. Relation between myocardial edema and
abnormal left ventricle. Am Heart J. 1990;119(6):1367–1373. myocardial mass during the acute and convalescent phase of myocarditis—a CMR study.
58. Schaefer WM, Lipke CS, Nowak B, et al. Validation of an evaluation routine for left J Cardiovasc Magn Reson. 2008;10:19.
ventricular volumes, ejection fraction and wall motion from gated cardiac FDG PET: A 89. Ammann P, Naegeli B, Schuiki E, et al. Long-term outcome of acute myocarditis is inde
comparison with cardiac magnetic resonance imaging. Eur J Nucl Med Mol Imaging. pendent of cardiac enzyme release. Int J Cardiol. 2003;89:217–222.
2003;30(4):545–553. 90. Carniel E, Sinagra G, Bussani R, et al. Fatal myocarditis: Morphologic and clinical fea
59. Buser PT, Auffermann W, Holt WW, et al. Noninvasive evaluation of global left ven tures. Ital Heart J. 2004;5:702–706.
tricular function with use of cine nuclear magnetic resonance. J Am Coll Cardiol. 91. Karjalainen J, Heikkila J. “Acute pericarditis”: Myocardial enzyme release as evidence for
1989;13(6):1294–1300. myocarditis. Am Heart J. 1986;111:546–552.
60. MacGowan GA, Shapiro EP, Azhari H, et al. Noninvasive measurement of shortening in 92. Nazarian S, Lima JA. Cardiovascular magnetic resonance for risk stratification of arrhyth
the fiber and cross-fiber directions in the normal human left ventricle and in idiopathic mia in hypertrophic cardiomyopathy. J Am Coll Cardiol. 2008;51(14):1375–1376.
dilated cardiomyopathy. Circulation. 1997;96(2):535–541. 93. Maron MS, Olivotto I, Zenovich AG, et al. Hypertrophic cardiomyopathy is predominantly
61. Fujita N, Duerinekx AJ, Higgins CB. Variation in left ventricular regional wall stress with a disease of left ventricular outflow tract obstruction. Circulation. 2006;114(21):2232–2239.
cine magnetic resonance imaging: Normal subjects versus dilated cardiomyopathy. Am 94. Maron BJ, McKenna WJ, Danielson GK, et al. American College of Cardiology/European
Heart J. 1993;125(5 Pt 1):1337–1345. Society of Cardiology clinical expert consensus document on hypertrophic cardiomyopa
62. Lytrivi ID, Ko HH, Srivastava S, et al. Regional differences in right ventricular systolic thy. A report of the American College of Cardiology Foundation Task Force on Clinical
function as determined by cine magnetic resonance imaging after infundibulotomy. Am J Expert Consensus Documents and the European Society of Cardiology Committee for
Cardiol. 2004;94(7):970–973. Practice Guidelines. J Am Coll Cardiol. 2003;42(9):1687–1713.
63. Grothues F, Moon JC, Bellenger NG, et al. Interstudy reproducibility of right ventricu 95. Stuber M, Scheidegger MB, Fischer SE, et al. Alterations in the local myocardial motion
lar volumes, function, and mass with cardiovascular magnetic resonance. Am Heart J. pattern in patients suffering from pressure overload due to aortic stenosis. Circulation.
2004;147(2):218–223. 1999;100(4):361–368.
64. Osterziel KJ, Strohm O, Schuler J, et al. Randomised, double-blind, placebo-controlled 96. Nagel E, Stuber M, Burkhard B, et al. Cardiac rotation and relaxation in patients with
trial of human recombinant growth hormone in patients with chronic heart failure due to aortic valve stenosis. Eur Heart J. 2000;21(7):582–589.
dilated cardiomyopathy. Lancet. 1998;351(9111):1233–1237. 97. Maier SE, Fischer SE, McKinnon GC, et al. Evaluation of left ventricular segmental
65. Bellenger N, Davies L, Francis J, et al. Reduction in sample size for studies of remodeling wall motion in hypertrophic cardiomyopathy with myocardial tagging. Circulation.
in heart failure by the use of cardiovascular magnetic resonance. J Cardiovasc Magn Res. 1992;86(6):1919–1928.
2000;2(4):271–278. 98. Kramer CM, Reichek N, Ferrari VA, et al. Regional heterogeneity of function in hypertro
66. McCrohon JA, Moon JC, Prasad SK, et al. Differentiation of heart failure related to phic cardiomyopathy. Circulation. 1994;90(1):186–194.
dilated cardiomyopathy and coronary artery disease using gadolinium-enhanced cardio 99. Dong SJ, MacGregor JH, Crawley AP, et al. Left ventricular wall thickness and regional
vascular magnetic resonance. Circulation. 2003;108(1):54–59. systolic function in patients with hypertrophic cardiomyopathy. A three-dimensional
67. Iles L, Pfluger H, Phrommintikul A, et al. Evaluation of diffuse myocardial fibrosis in tagged magnetic resonance imaging study. Circulation. 1994;90(3):1200–1209.
heart failure with cardiac magnetic resonance contrast-enhanced T1 Mapping. J Am Coll 100. Young AA, Kramer CM, Ferrari VA, et al. Three-dimensional left ventricular deformation
Cardiol. 2008;52:1574–1580. in hypertrophic cardiomyopathy. Circulation. 1994;90(2):854–867.
68. Voigt A, Elgeti T, Durmus T, et al. Cardiac magnetic resonance imaging in dilated car 101. Choudhury L, Mahrholdt H, Wagner A, et al. Myocardial scarring in asymptomatic or
diomyopathy in adults–towards identification of myocardial inflammation. Eur Radiol. mildly symptomatic patients with hypertrophic cardiomyopathy. J Am Coll Cardiol.
2011;21(5):925–935. 2002;40(12):2156–2164.
LWBK1209-ch10_p135-148.indd 146 16/05/13 9:40 PM

102. O’Hanlon R, Grasso A, Roughton M, et al. Prognostic significance of myocardial fibrosis 128. Castillo E, Tandri H, Rodriguez ER, et al. Arrhythmogenic right ventricular dysplasia:
in hypertrophic cardiomyopathy. J Am Coll Cardiol. 2010;56(11):867–874. Ex vivo and in vivo fat detection with black-blood MR imaging.Radiology. 2004;232(1):
103. Bruder O, Wagner A, Jensen CJ, et al. Myocardial scar visualized by cardiovascular mag 38–48.
netic resonance imaging predicts major adverse events in patients with hypertrophic car 129. Bluemke DA, Krupinski EA, Ovitt T, et al. MR imaging of arrhythmogenic right ven
diomyopathy. J Am Coll Cardiol. 2010;56(11):875–887. tricular cardiomyopathy: Morphologic findings and interobserver reliability. Cardiology.
104. Abdel-Aty H, Cocker M, Strohm O, et al. Abnormalities in T2-weighted cardiovascu 2003;99(3):153–162.
lar magnetic resonance images of hypertrophic cardiomyopathy: Regional distribution 130. La Gerche A, Burns AT, Mooney DJ, et al. Exercise-induced right ventricular dysfunction
and relation to late gadolinium enhancement and severity of hypertrophy. J Magn Reson and structural remodeling in endurance athletes. Eur Heart J. 2012;33(8):998–1006.
Imaging. 2008;28(1):242–245. 131. Felker GM, Thompson RE, Hare JM, et al. Underlying causes and long-term survival in
105. Adabag AS, Maron BJ, Appelbaum E, et al. Occurrence and frequency of arrhythmias patients with initially unexplained cardiomyopathy. N Engl J Med. 2000;342(15):1077–
in hypertrophic cardiomyopathy in relation to delayed enhancement of cardiovascular 1084.
magnetic resonance. J Am Coll Cardiol. 2008;51(14):1369–1374. 132. Kushwaha SS, Fallon JT, Fuster V. Restrictive cardiomyopathy. N Engl J Med. 1997;
106. Kwon DH, Smedira NG, Rodriguez ER, et al. Cardiac magnetic resonance detection 336(4):267–276.
of myocardial scarring in hypertrophic cardiomyopathy: Correlation with histopa 133. Talreja DR, Edwards WD, Danielson GK, et al. Constrictive pericarditis in 26 patients
thology and prevalence of ventricular tachycardia. J Am Coll Cardiol. 2009;54(3): with histologically normal pericardial thickness. Circulation. 2003;108(15):1852–1857.
242–249. 134. Jarvinen VM, Kupari MM, Poutanen VP, et al. A simplified method for the determina
107. Crilley JG, Boehm EA, Blair E, et al. Hypertrophic cardiomyopathy due to sarcomeric tion of left atrial size and function using cine magnetic resonance imaging. Magn Reson
gene mutations is characterized by impaired energy metabolism irrespective of the degree Imaging. 1996;14(3):215–226.
of hypertrophy. J Am Coll Cardiol. 2003;41(10):1776–1782. 135. Barkhausen J, Hunold P, Eggebrecht H, et al. Detection and characterization of intracar
108. Spindler M, Saupe KW, Christe ME, et al. Diastolic dysfunction and altered energetics diac thrombi on MR imaging. AJR Am J Roentgenol. 2002;179(6):1539–1544.
in the alphaMHC403/+ mouse model of familial hypertrophic cardiomyopathy. J Clin 136. Sugie T, Hashimoto N, Iwai K. Clinical and autopsy studies on prognosis of sarcoidosis.
Invest. 1998;101(8):1775–1783. Nihon Rinsho. 1994;52(6):1567–1570.
109. Jung WI, Sieverding L, Breuer J, et al. Detection of phosphomonoester signals in proton- 137. Perry A, Vuitch F. Causes of death in patients with sarcoidosis. A morphologic study of 38
decoupled 31P NMR spectra of the myocardium of patients with myocardial hypertro autopsies with clinicopathologic correlations. Arch Pathol Lab Med. 1995;119(2):167–
phy. J Magn Reson. 1998;133(1):232–235. 172.
110. White RD, Obuchowski NA, Gunawardena S, et al. Left ventricular outflow tract 138. Ratner SJ, Fenoglio JJ Jr, Ursell PC. Utility of endomyocardial biopsy in the diagnosis of
obstruction in hypertrophic cardiomyopathy: Presurgical and postsurgical evaluation by cardiac sarcoidosis. Chest. 1986;90(4):528–533.
computed tomography magnetic resonance imaging. Am J Card Imaging. 1996;10(1): 139. Uemura A, Morimoto S, Hiramitsu S, et al. Histologic diagnostic rate of cardiac sarcoid
1–13. osis: Evaluation of endomyocardial biopsies. Am Heart J. 1999;138(2 Pt 1):299–302.
111. Schulz-Menger J, Strohm O, Waigand J, et al. The value of magnetic resonance imaging 140. Riedy K, Fisher MR, Belic N, et al. MR imaging of myocardial sarcoidosis. AJR Am J
of the left ventricular outflow tract in patients with hypertrophic obstructive cardiomy Roentgenol. 1988;151(5):915–916.
opathy after septal artery embolization. Circulation. 2000;101(15):1764–1766. 141. Smedema JP, van Kroonenburgh MJ, Snoep G, et al. Images in cardiovascular medicine.
112. Schulz-Menger J, Abdel-Aty H, Busjahn A, et al. Left ventricular outflow tract planim Cardiac sarcoidosis in a patient with hypertrophic cardiomyopathy demonstrated by
etry by cardiovascular magnetic resonance differentiates obstructive from non-obstructive magnetic resonance imaging and single photon emission computed tomography dual-
hypertrophic cardiomyopathy. J Cardiovasc Magn Reson. 2006;8(5):741–746. isotope scintigraphy. Circulation. 2004;110(24):e529–e531.
113. Suzuki J, Shimamoto R, Nishikawa J, et al. Morphological onset and early diagnosis in 142. Schulz-Menger J, Strohm O, Dietz R, et al. Visualization of cardiac involvement in
apical hypertrophic cardiomyopathy: A long term analysis with nuclear magnetic reso patients with systemic sarcoidosis applying contrast-enhanced magnetic resonance imag
nance imaging. J Am Coll Cardiol. 1999;33(1):146–151. ing. MAGMA. 2000;11(1–2):82–83.
114. Amano Y, Takayama M, Amano M, et al. MRI of cardiac morphology and function 143. Schulz-Menger J, Wassmuth R, Abdel-Aty H, et al. Patterns of myocardial inflamma
after percutaneous transluminal septal myocardial ablation for hypertrophic obstructive tion and scarring in sarcoidosis as assessed by cardiovascular magnetic resonance. Heart.
cardiomyopathy. AJR Am J Roentgenol. 2004;182(2):523–527. 2006;92:399–400.
115. Moon JC, McKenna WJ, McCrohon JA, et al. Toward clinical risk assessment in hyper 144. Doherty MJ, Kumar SK, Nicholson AA, et al. Cardiac sarcoidosis: The value of magnetic
trophic cardiomyopathy with gadolinium cardiovascular magnetic resonance. J Am Coll resonance imaging in diagnosis and assessment of response to treatment. Respir Med.
Cardiol. 2003;41(9):1561–1567. 1998;92(4):697–699.
116. Maron BJ, Towbin JA, Thiene G, et al. Contemporary definitions and classification of the 145. Shimada T, Shimada K, Sakane T, et al. Diagnosis of cardiac sarcoidosis and evaluation of
cardiomyopathies: An American Heart Association Scientific Statement from The Council the effects of steroid therapy by gadolinium-DTPA-enhanced magnetic resonance imag
on Clinical Cardiology, Heart Failure and Transplantation Committee; Quality of Care and ing. Am J Med. 2001;110(7):520–527.
Outcomes Research and Functional Genomics and Translational Biology Interdisciplinary 146. Vignaux O, Dhote R, Duboc D, et al. Clinical significance of myocardial magnetic
Working Groups; and Council On Epidemiology And Prevention. Circulation. 2006; resonance abnormalities in patients with sarcoidosis: A 1-year follow-up study. Chest.
113(14):1807–1816. 2002;122(6):1895–1901.
117. Eitel I, von Knobelsdorff-Brenkenhoff F, Berhardt P, et al. Clinical characteristics and car 147. Dupuis JM, Victor J, Furber A, et al. Value of magnetic resonance imaging in cardiac
diovascular magnetic resonance findings in stress (Takotsubo) cardiomyopathy. JAMA. sarcoidosis. Apropos of a case. Arch Mal Coeur Vaiss. 1994;87(1):105–110.
2011;306(3):277–286. 148. Patel MR, Crawley PJ, Heitner JF, et al. Detection of myocardial damage in patients with
118. McKenna WJ, Thiene G, Nava A, et al. Diagnosis of arrhythmogenic right ventricu sarcoidosis. Circulation. 2009;120(20):1969–1977.
lar dysplasia/cardiomyopathy. Task Force of the Working Group Myocardial and 149. Gertz MA, Kyle RA, Thibodeau SN. Familial amyloidosis: A study of 52 North
Pericardial Disease of the European Society of Cardiology and of the Scientific Council American-born patients examined during a 30-year period. Mayo Clin Proc. 1992;67(5):
on Cardiomyopathies of the International Society and Federation of Cardiology. Br Heart 428–440.
J. 1994;71(3):215–218. 150. Plehn JF, Southworth J, Cornwell GG 3rd. Brief report: Atrial systolic failure in primary
119. Marcus FI, McKenna WJ, Sherrill D, et al. Diagnosis of arrhythmogenic right ventricular amyloidosis [see comments]. N Engl J Med. 1992;327(22):1570–1573.
cardiomyopathy/dysplasia: Proposed modification of the task force criteria. Circulation. 151. Kyle RA, Spittell PC, Gertz MA, et al. The premortem recognition of systemic senile
2010;121(13):1533–1541. amyloidosis with cardiac involvement. Am J Med. 1996;101(4):395–400.
120. Vermes E, Strohm O, Otmani A, et al. Impact of the revision of arrhythmogenic right ven 152. Koyama J, Ray-Sequin PA, Falk RH. Prognostic significance of ultrasound myocardial
tricular cardiomyopathy/dysplasia task force criteria on its prevalence by CMR criteria. tissue characterization in patients with cardiac amyloidosis. Circulation. 2002;106(5):
JACC Cardiovasc Imaging. 2011;4(3):282–287. 556–561.
121. Corrado D, Basso C, Thiene G, et al. Spectrum of clinicopathologic manifestations of 153. Maceira AM, Joshi J, Prasad SK, et al. Cardiovascular magnetic resonance in cardiac
arrhythmogenic right ventricular cardiomyopathy/dysplasia: A multicenter study. J Am amyloidosis. Circulation. 2005;111(2):186–193.
Coll Cardiol. 1997;30(6):1512–1520. 154. Ruberg FL, Appelbaum E, Davidoff R, et al. Diagnostic and prognostic utility of cardio
122. Committees UI-US. Survivors of out-of-hospital cardiac arrest with apparently normal vascular magnetic resonance imaging in light-chain cardiac amyloidosis. Am J Cardiol.
heart. Need for definition and standardized clinical evaluation. Consensus Statement 2009;103(4):544–549.
of the Joint Steering Committees of the Unexplained Cardiac Arrest Registry of Europe 155. Selvanayagam JB, Hawkins PN, Paul B, et al. Evaluation and management of cardiac
and of the Idiopathic Ventricular Fibrillation Registry of the United States. Circulation. amyloidosis. J Am Coll Cardiol. 2007;50(22):2101–2110.
1997;95(1):265–272. 156. Billingham ME, Cary NR, Hammond ME, et al. A working formulation for the stan
123. Jain A, Shehata ML, Stuber M, et al. Prevalence of left ventricular regional dysfunc dardization of nomenclature in the diagnosis of heart and lung rejection: Heart rejection
tion in arrhythmogenic right ventricular dysplasia: A tagged MRI study. Circ Cardiovasc study group. The International Society for Heart Transplantation. J Heart Transplant.
Imaging. 2010;3:290–297. 1990;9(6):587–593.
124. Tandri H, Macedo R, Calkins H, et al. Role of magnetic resonance imaging in arrhythmo 157. Nishimura T, Sada M, Sasaki H, et al. Identification of cardiac rejection with magnetic
genic right ventricular dysplasia: Insights from the North American arrhythmogenic right resonance imaging in heterotopic heart transplantation model. Heart Vessels. 1987;3(3):
ventricular dysplasia (ARVD/C) study. Am Heart J. 2008;155(1):147–153. 135–140.
125. Schick F, Miller S, Hahn U, et al. Fat- and water-selective MR cine imaging of the human 158. Almenar L, Igual B, Martinez-Dolz L, et al. Utility of cardiac magnetic resonance imag
heart: Assessment of right ventricular dysplasia. Invest Radiol. 2000;35(5):311–318. ing for the diagnosis of heart transplant rejection. Transplant Proc. 2003;35(5):1962–
126. Ricci C, Longo R, Pagnan L, et al. Magnetic resonance imaging in right ventricular dys 1964.
plasia. Am J Cardiol. 1992;70(20):1589–1595. 159. Kober F, Caus T, Riberi A, et al. Objective and noninvasive metabolic characterization
127. Blake LM, Scheinman MM, Higgins CB. MR features of arrhythmogenic right ventricular of donor hearts by phosphorous-31 magnetic resonance spectroscopy.Transplantation.
dysplasia. AJR Am J Roentgenol. 1994;162(4):809–812. 2002;74(12):1752–1756.
LWBK1209-ch10_p135-148.indd 147 16/05/13 9:40 PM

160. Muehling OM, Wilke NM, Panse P, et al. Reduced myocardial perfusion reserve and 170. Pennell DJ, Berdoukas V, Karagiorga M, et al. Randomized controlled trial of deferi
transmural perfusion gradient in heart transplant arteriopathy assessed by magnetic reso prone or deferoxamine in beta-thalassemia major patients with asymptomatic myocardial
nance imaging. J Am Coll Cardiol. 2003;42(6):1054–1060. siderosis. Blood. 2006;107(9):3738–3744.
161. Steen H, Merten C, Refle S, et al. Prevalence of different gadolinium enhancement pat 171. Celletti F, Fattori R, Napoli G, et al. Assessment of restrictive cardiomyopathy of amy
terns in patients after heart transplantation. J Am Coll Cardiol. 2008;52(14):1160–1167. loid or idiopathic etiology by magnetic resonance imaging. Am J Cardiol. 1999;83(5):
162. Olson LJ, Edwards WD, McCall JT, et al. Cardiac iron deposition in idiopathic hemo 798–801, A10.
chromatosis: Histologic and analytic assessment of 14 hearts from autopsy. J Am Coll 172. D’Silva SA, Kohli A, Dalvi BV, et al. MRI in right ventricular endomyocardial fibrosis.
Cardiol. 1987;10(6):1239–1243. Am Heart J. 1992;123(5):1390–1392.
163. Kirk P, Roughton M, Porter JB, et al. Cardiac T2* magnetic resonance for prediction of 173. Huong DL, Wechsler B, Papo T, et al. Endomyocardial fibrosis in Behcet’s disease. Ann
cardiac complications in thalassemia major. Circulation. 2009;120(20):1961–1968. Rheum Dis. 1997;56(3):205–208.
164. Mavrogeni SI, Gotsis ED, Markussis V, et al. T2 relaxation time study of iron overload in 174. McGavock JM, Lingvay I, Zib I, et al. Cardiac steatosis in diabetes mellitus: A
b-thalassemia. MAGMA. 1998;6(1):7–12. 1H-magnetic resonance spectroscopy study. Circulation. 2007;116(10):1170–1175.
165. Ooi GC, Chen FE, Chan KN, et al. Qualitative and quantitative magnetic resonance imaging 175. Schmidt AR, Friedrich MG. Imaging targets in diabetic cardiomyopathy: Current status
in haemoglobin H disease: Screening for iron overload. Clin Radiol. 1999;54(2):98–102. and perspective. Canadian Journal of Diabetes. 2011;35(4):353–362.
166. Papanikolaou N, Ghiatas A, Kattamis A, et al. Non-invasive myocardial iron assessment 176. Oechslin EN, Attenhofer Jost CH, Rojas JR, et al. Long-term follow-up of 34 adults with
in thalassaemic patients. T2 relaxometry and magnetization transfer ratio measurements. isolated left ventricular noncompaction: A distinct cardiomyopathy with poor prognosis.
Acta Radiol. 2000;41(4):348–351. J Am Coll Cardiol. 2000;36(2):493–500.
167. Jensen PD, Jensen FT, Christensen T, et al. Evaluation of myocardial iron by mag 177. Ritter M, Oechslin E, Sutsch G, et al. Isolated non-compaction of the myocardium in
netic resonance imaging during iron chelation therapy with deferrioxamine: Indication adults. Mayo Clin Proc. 1997;72(1):26–31.
of close relation between myocardial iron content and chelatable iron pool. Blood. 178. Petersen S, Selvanayagam J, Wiesmann F, et al. Left ventricular non-compaction: Insights
2003;101(11):4632–4639. from cardiovascular magnetic resonance imaging. J Am Coll Cardiol. 2005;46(1):
168. Anderson LJ, Wonke B, Prescott E, et al. Comparison of effects of oral deferiprone and 101–105.
subcutaneous desferrioxamine on myocardial iron concentrations and ventricular func 179. Vogel-Claussen J, Santaularia TM, Newatia A, et al. Cardiac MRI evaluation of
tion in beta-thalassaemia. Lancet. 2002;360(9332):516–520. hypertrophic cardiomyopathy: Left ventricular outflow tract/aortic valve diameter
169. Wood JC. History and current impact of cardiac magnetic resonance imaging on the ratio predicts severity of LVOT obstruction. J Magn Reson Imaging. 2012;36(3):
management of iron overload. Circulation. 2009;120(20):1937–1939. 598–603.
LWBK1209-ch10_p135-148.indd 148 16/05/13 9:40 PM

Chapter
11
Peter T. Buser
Michael Zellweger
Jens Bremerich

Cardiomyopathy/Dysplasia
■■ Clinical Background and fibrous tissue and numerous structural abnormalities of

the right ventricle including areas of wall thinning with for-
■■ Cardiac Magnetic Resonance Techniques
mation of single or multiple aneurysms. These abnormali-
Black Blood Imaging ties are most often found in the subtricuspid area, the right
Bright Blood Imaging ventricular apex, and the right ventricular outflow tract—
■■ Interpretation of Cardiac Magnetic the so-called triangle of ARVC. However, left ventricular in-
Resonance Findings volvement has been described especially in the posterolateral
Right Ventricular Volumes, Global and Regional wall (6), may be age dependent and associated with arrhyth-
mic events, cardiomegaly, and heart failure. Twelve genes
Dysfunction
have been identified which are linked to ARVC, encoding
Intramyocardial Fat several components of the cardiac desmosome. Dysfunction
Right Ventricular Wall Thinning of desmosomes causes defective cell adhesion, loss of electri-
Right Ventricular Wall Hypertrophy cal myocyte coupling leading to apoptotic myocyte death,
Trabecular Disarray of the Right Ventricle fibrofatty replacement, and arrhythmias. Clinically, four
Enlargement of the Right Ventricular Outflow stages of the disease have been proposed: (a) A concealed
Tract phase, in which anatomic changes are subtle and arrhyth-
Detection of Myocardial Fibrosis mias may be minor, but in which sudden cardiac death may
Diastolic Dysfunction of the Right Ventricle be the first indication of the disease; (b) an overt electrical
disorder presenting with ventricular tachycardias or sudden
■■ Summary
cardiac death, in which structural and functional abnormali-
ties of the right ventricle are more apparent; (c) progressive
right ventricular failure with preserved left ventricular func-
Clinical Background tion; and (d) biventricular dysfunction (7).
The diagnosis of ARVC can be made according to the
Arrhythmogenic right ventricular cardiomyopathy (ARVC) International Task Force criteria, originally published in
is a predominantly genetically determined and inherit- 1994 (8) and modified in 2010 (9). Abnormalities are di-
able disease characterized by fibrofatty replacement of the vided into major and minor categories according to their
right ventricular myocardium with ventricular arrhythmias specificity for ARVC. Criteria for global and regional dys-
and sudden death (1,2). Its prevalence has been estimated function and/or structural alterations of the right ventricle
to range from 1 in 2,000 to 1 in 5,000, and men are more are based on findings with transthoracic echocardiography
frequently affected than women with a ratio of 3:1. It has or cardiac magnetic resonance (CMR) imaging. Further
been reported to account for 3% to 10% of unexplained criteria include tissue characterization of myocardial walls
sudden cardiac death in patients less than 65 years old by endomyocardial biopsy and histology, repolarization
(3,4). Patients with symptomatic ARVC may therefore be and depolarization abnormalities by ECG, arrhythmias by
candidates for an active therapeutic management, including (Holter-) monitoring, and family history with pathologically
antiarrhythmic medication, invasive electrophysiologic pro- proven ARVC. The modified ARVC Task Force criteria pro-
cedures, implantation of a cardioverter-defibrillator, and in vide detailed cutoffs for abnormal RV size and wall motion
some cases, cardiac transplantation (5). as assessed by CMR. RV ejection fraction ≤40%, or regional
It is a rare disorder, but it is the most common cause of akinesia, dyssynchronous RV contraction, RV end-diastolic
sudden cardiac death in younger populations. Major patho- volume/body surface area ≥110 mL/m2 (male) or ≥100 mL/
logic findings are replacement of the myocardium by fatty m2 (female) are considered major criteria for ARVC.
149
LWBK1209-ch11_p149-155.indd 149 17/05/13 6:04 PM

CMR has the unique advantage to be an absolutely non- Black Blood Imaging
invasive technique, providing information on dimensions,
For black blood techniques, breath-hold imaging with dou-
geometry, shape, volumes, and function of all cardiac cham-
ble-inversion recovery fast spin-echo (DIR-FSE) techniques
bers, tissue characterization such as differentiation of fat
is preferred to traditional spin-echo (SE) imaging. These
from myocardium, and identification of fibrous formations
techniques substantially shorten imaging time and virtually
such as myocardial scar. CMR, therefore, has been increas-
eliminate respiratory motion artifacts. Black blood inversion
ingly used for the evaluation of right ventricular disease and
prepared, half Fourier single-shot turbo spin-echo (HASTE)
has evolved as the noninvasive imaging modality of choice
imaging currently is not recommended because of blurring
in ARVC.
of subtle anatomic details. Frequent ventricular ectopy can
substantially deteriorate image quality. In patients with sus-
Cardiac Magnetic Resonance pected ARVC and frequent ventricular ectopy, a low dose
of oral beta-blocker (e.g., 50 mg metoprolol) can be given
Techniques
1 hour before the CMR examination to reduce ventricular
ectopy. However, contraindications for beta-blockade must
There is no widely accepted standard imaging protocol for
be considered.
the investigation of patients with suspected ARVC using
Right ventricular morphology is best shown in axial and
CMR. The CMR imaging protocol is aimed at recognizing
sagittal imaging planes. Axial imaging planes should encom-
both, typical features of ARVC such as RV akinesia, dys-
pass the whole heart starting from the pulmonary artery to
kinesia, dyssynchrony, and reduced RV ejection fraction
the diaphragm. Axial imaging planes provide the best view
(modification of ARVC/D Task Force criteria) as well as
of the right ventricular anterior wall up to the proximal
features of other conditions with similar clinical presenta-
parts of the right ventricular outflow tract. Since the ana-
tion, for example, granulomas in cardiac sarcoidosis. We
tomic course of the right ventricular outflow tract toward
suggest an imaging protocol that comprises cine steady-state
the pulmonary valve and the common pulmonary artery
free precession (Cine-bSSFP) for morphology and function,
progresses from ventral to dorsal, axial planes will not be
T1-weighted turbo-spin-echo (T1-TSE) for morphology, and
optimal for the assessment of dimensions and morphology
late gadolinium enhancement inversion recovery gradient
of the subpulmonary part of the right ventricular outflow
recalled echo (LGE-IR-GRE) sequences acquired in axial
tract. Sagittal planes through the right ventricular outflow
and short-axis orientation. The right ventricle has a complex
tract will depict this portion optimally.
geometry, is asymmetric, and is highly trabeculated. The
mean right ventricular free wall thickness in healthy individ-
uals is 2.7 ± 0.4 mm and only 1.9 ± 1.1 mm at the right ven-
Bright Blood Imaging
tricular apex (8). Epicardial fat is usually present, especially
in association with the right coronary artery within the right For bright blood imaging, balanced steady-state free preces-
atrioventricular groove and with the left anterior descending sion imaging (FIESTA, true FISP) is the preferred technique
coronary artery in the apical region. Tongues of epicardial because it allows better endocardial definition when com-
fat may extend into the myocardium in healthy individuals pared with spoiled gradient-echo images (FLASH or FAST-
(10,11). The identification of small morphologic alterations CARD). Cine imaging in the axial plane is optimal to assess
of the thin right ventricular free wall, the analysis of regional right ventricular global and regional function. Cine imaging
wall-motion abnormalities of a complex geometric struc- in the axial plane is optimal to assess right ventricular global
ture, and the identification of intramural or transmural fatty and regional function. Again, the whole heart is covered
infiltrates of the right ventricular myocardium require high- from the pulmonary artery to the diaphragm. Contraction
resolution images without significant artifacts. Retrospective abnormalities of the right ventricular outflow tract, areas of
analysis of static magnetic resonance images of 39 patients wall thinning and reduced contraction, aneurysm formation,
from a ARVC registry revealed an excellent image quality regions of focal hypokinesia, and akinesia or dyskinesia of
in less than 10% of cases (12). Thus, basic requirements for the right ventricular free wall can be best depicted in axial
equipment, software, imaging protocol, readout, and docu- imaging planes. However, at least one additional imaging
mentation should be fulfilled for the CMR investigation of plane should be used to assess function of the diaphrag-
patients with suspected ARVC. A cardiac phased array coil matic right ventricular wall segments, function of the right
is preferred, and gradient coil strength should minimally ventricular outflow tract, and left ventricular volumes and
be 20 mT/m. Cardiac gating and breath-hold imaging are function. In sagittal imaging planes, the right ventricular
required (13). Black blood imaging is used to depict mor- outflow tract is usually well depicted along its long axis as
phologic abnormalities of the right ventricle and intramyo- well as diaphragmatic segments of the right ventricular wall.
cardial fatty infiltration. Bright blood cine imaging is used In addition, to assess left ventricular regional and global
for visualizing global and regional ventricular function and function accurately and to measure right ventricular and left
to measure end-diastolic and end-systolic volumes for cal- ventricular volumes, proper left ventricular short-axis planes
culation of stroke volume and ejection fraction. In addi- encompassing the whole left ventricle from above the atrio-
tion, significant valvular disease can be demonstrated or ventricular plane to the apex should be used. Quantification
excluded. Diastolic dysfunction of the right ventricle can be of ventricular volumes is performed by contouring the end-
assessed by measurement of blood-flow velocities across the diastolic and end-systolic frames of the entire right ventri-
tricuspid valve during the whole cardiac cycle by application cle and left ventricle, using a summation of disks method
of MR velocity mapping (14). (Simpson’s rule) with integration over the image slices. For
LWBK1209-ch11_p149-155.indd 150 17/05/13 6:04 PM

Chapter 11 ■ Arrhythmogenic Right Ventricular Cardiomyopathy/Dysplasia 151
further evaluation of left ventricular abnormalities, espe-

cially in the anterior and apical segments, two-chamber and/
or four-chamber long-axis imaging planes should be added.
Interpretation of Cardiac
Magnetic Resonance Findings
Right Ventricular Volumes, Global and
Regional Dysfunction
Regional akinesia or dyskinesia of the right ventricular wall,
or dyssynchronous contraction of the right ventricle coin-
ciding with an enlarged right ventricular end-diastolic vol-
ume (Figs. 11.1 and 11.2) indexed by body surface area of
≥110 mL/m2 for male and ≥100 mL/m2 for female, or an
impaired right ventricular ejection fraction ≤40%, are con-
sidered as major criteria for the diagnosis of ARVC accord-
ing to the revised Task Force criteria (9). Minor criteria are
regional akinesia or dyskinesia of the right ventricular wall
or dyssynchronous contraction of the right ventricle coincid-
ing with an enlarged right ventricular end-diastolic volume Figure 11.2. Bright blood image from a cine loop in an axial plane
indexed by body surface area of 100 to 110 mL/m2 for male at end-systole in a patient with ARVC. Aneurysm in the subtricuspid
and 90 to 100 mL/m2 for female or an impaired right ven- area (arrow), microaneurysms in the RV apex (triangular mark).
tricular ejection fraction between 40% and 45%.
CMR is the gold standard imaging modality for the as-
sessment of cardiac ventricular volumes. Minor changes in found that 75% of patients with Task Force criteria for
ventricular volumes overtime can be detected and provide ARVC have some degree of right ventricular enlargement
insight into an evolving disease process. Small increases in and dysfunction. These abnormalities were confirmed dur-
right ventricular end-diastolic volumes may be an early sign ing re-evaluation of 14 patients in whom right ventricular
of ARVC. Auffermann et al. (15) reported an increased right functional abnormalities were noted on CMR performed at
ventricular end-diastolic volume index in 10 patients with initial evaluation. Furthermore, all of these patients met the
ARVC and inducible ventricular arrhythmias. Ventricular Task Force criteria for ARVC during re-evaluation (16).
volumes of those patients who were not inducible did not High-resolution CMR cine imaging with state of the art
differ from those of control subjects. Tandri et al. (13) SSFP sequences is considered as gold standard for the assess-
ment of ventricular volumes, myocardial mass, and systolic
function with high intra- and interobserver agreement and
accuracy (17,18). Regional abnormalities of right ventricular
function are thought to precede global dysfunction. These
include regional hypokinesia, defined as systolic wall thick-
ening less than 40%, and akinesia, defined as systolic wall
thickening less than 10% because of myocyte loss, fibrosis,
and impaired myocardial contraction. Dyskinesia, defined
as systolic outward movement, and bulging during dias-
tole are signs of aneurysm formation. Several studies have
consistently reported a high incidence of right ventricular
regional dysfunction in ARVC (13,15,19–23). Tandri et al.
(13) found regional contraction abnormalities in 67% of pa-
tients with Task Force criteria of ARVC, which correlated to
the area of fatty myocardial infiltrates. Fifty percent of those
patients had aneurysm formation that was localized to the
region of high T1 signals. Overall, aneurysm formation was
observed in 25% of patients with a final diagnosis of ARVC.
A recent retrospective analysis comparing the original with
the revised Task Force criteria for CMR in a cohort of 294
patients found a significantly reduced prevalence of major
and minor criteria, when applying the revised Task Force
criteria (24). With application of the revised Task Force
Figure 11.1. Bright blood image from a cine loop in an axial plane criteria, a high diagnostic specificity may be maintained.
at end-diastole in a patient with ARVC. Enlargement of the RV, thin- However, an increase in sensitivity remains to be proven in
ning of the RV free myocardial wall. larger populations.
LWBK1209-ch11_p149-155.indd 151 17/05/13 6:04 PM

Intramyocardial Fat nine patients supported the diagnosis of ARVC. Midiri et

al. (22) found fatty replacement of myocardium in 14 of 30
The depiction of intramyocardial fatty tissue by CMR is not patients with clinically suspected ARVC. When ARVC was
a diagnostic component of the modified Task Force criteria highly probable, fatty replacement was described in seven
(9). On black blood images, normal myocardium shows an of eight patients; when ARVC was only probable, fatty re-
intermediate signal intensity similar to that of skeletal mus- placement was described in four of four patients; and when
cle. Fat appears as a bright signal. In healthy individuals, ARVC was not likely, fatty replacement was described in
epicardial fat is usually seen in the atrioventricular groove three of seven patients. A final diagnosis of ARVC based
and around the right ventricular apex. However, distribu- on the Task Force criteria was made in 12 of 128 patients
tion of epicardial fat varies considerably in healthy individu- who were evaluated by CMR for possible ARVC. High T1
als and may cover the whole aspect of the right ventricular signal within the right ventricular myocardium indicative of
free wall as a thick bright tissue layer. There is often a clear fat was observed in 9 of 12 patients (75%) and in none of
demarcation line between the myocardium and the epicar- the patients without ARVC. The region of fatty infiltration
dial fat. Disruption of this demarcation line and extension was in the right ventricular anterior wall in eight patients
of the bright signals into the myocardium are frequently and in the apex in one patient. The same region showed re-
seen. This is not a specific sign of ARVC, and it can also be duced wall thickening and contraction in eight of nine pa-
observed in healthy individuals (11,25) and in lipomatous tients. No patients with ARVC had involvement of the left
metaplasia of chronically infarcted myocardium (26). It is ventricular myocardium or the ventricular septum (13). The
important to realize that findings of increased intramyocar- reliability and intraobserver variability of intramyocardial
dial fat signals on CMR are not a part of the Task Force cri- fat on T1-weighted images in patients with suspected ARVC
teria for the diagnosis of ARVC and that experts in the field were tested in one study (12). Hardcopy films of 39 patients
do not recommend equating of intramyocardial fat signal with suspected or definite ARVC and of six CMR exami-
on CMR with fatty infiltration found on endomyocardial nations not performed for suspected ARVC were reviewed
biopsy. independently by eight radiologists and five cardiologists
Hyperintense intramyocardial signals on T1-weighted with expertise in CMR and blinded to clinical data of these
images have been described in 20% to 100% of patients patients. The readers reported the presence of high-signal in-
with ARVC (12,13,15,19,20–22). Auffermann et al. (15) intensity areas in the right ventricular myocardium in 41% of
vestigated 36 patients with biopsy-proven ARVC and found controls, 41% of patients with suspected ARVC, and 46%
intramyocardial hyperintense signals on T1-weighted SE im- of patients with definite ARVC, which was not significantly
ages in 22%. In this study, fatty infiltration on CMR, but not different. The location of the high T1 signal areas was not
on endomyocardial biopsy, predicted the inducibility of ven- related to the diagnosis. However, when a high T1 signal in
tricular tachycardias on electrophysiologic testing. In 8 of the right ventricular myocardium was detected, it was highly
15 patients with the clinical diagnosis of ARVC, Ricci et al. associated with the final diagnosis. Percentage reader agree-
(19) found myocardial areas with hyperintense signals in an- ment for identification of a high T1 signal in the right ven-
atomic sites of the right ventricle usually affected by the dis- tricular myocardium showed a large variability, and it was
ease. In addition, in seven of these eight patients these areas concluded that identification of fat signal using MR scanners
showed an overlap with dyskinetic areas assessed with gra- without dedicated cardiovascular software is not a reliable
dient-echo cine MR or echocardiography. In the explanted predictor of ARVC and is not reproducibly identified by ex-
heart of one patient undergoing heart transplantation, an pert MR readers. In a group of 19 patients with a definite
excellent correlation between the localizations of the fatty clinical diagnosis of ARVC and 17 patients with suspected
infiltrates described in CMR and histologic examination was but clinically excluded ARVC, CMR showed a specificity of
found. Molinari et al. (20) investigated 124 patients with 82%, sensitivity of 89%, negative predictive value of 88%,
ventricular arrhythmias and left bundle branch block (LBBB) and positive predictive value of 88% for the diagnosis of
morphology and 38 control subjects with one of three differ- ARVC (23). Univariate analysis of all assessed CMR param-
ent MR scanners. In one system, fat-suppression techniques eters revealed right ventricular myocardial fatty infiltration
were applied for identification of intramyocardial fat. In this as the only single parameter that was significantly correlated
retrospective study with three different MR systems and pa- with the diagnosis of ARVC. In this study, patients without
tients with a wide range of LBBB ventricular tachycardias, CMR diagnosis of ARVC had a significantly better arrhyth-
an increased rate of myocardial replacement by adipose tis- mia-free survival over a period of 3 years compared with
sue was found with increasing complexity of arrhythmias. those with CMR diagnosis of ARVC.
Intramyocardial hyperintense signals were also described in One of the major original Task Force criteria for the di-
the left ventricular myocardium in 5% of the subjects with agnosis of ARVC was fibrofatty replacement of myocardium
sustained ventricular tachycardias and polymorphic prema- on endomyocardial biopsy (8). The modified Task Force cri-
ture beats. Menghetti et al. (21) described findings with SE teria (9) include now a more quantitative analysis of tissue
imaging in 15 patients with ARVC according to the original from endomyocardial biopsy (27). The presence of myocar-
Task Force criteria (8). They found intramyocardial hyper- dial fat identified by CMR has been reported to be strongly
intense signals in 10 of 15 patients (67%). In five patients, associated with ARVC (13,15,20,22) and its prognosis (23).
the MR signal was believed to be inappropriate for iden- However, the prevalence of these findings showed a wide
tification of intramyocardial fat. In 9 of 15 patients, both range depending on the patient selection, MR hardware and
endomyocardial biopsy and CMR were performed. Biopsy software used, interpreter’s clinical and MR expertise, and
in eight of nine patients and CMR examination in five of the extent of fibrofatty replacement of normal myocytes
LWBK1209-ch11_p149-155.indd 152 17/05/13 6:04 PM

as assessed by CMR has never been quantified. Specificity, are not part of the modified Task Force criteria for the diag-
interobserver variability, and reproducibility are not suffi- nosis of ARVC.
cient for the diagnosis of ARVC on the basis of this. Fatty
replacement of myocardium may be a prominent finding in
Right Ventricular Wall Hypertrophy
some and absent in other cases with ARVC. Since ARVC is
a rare disease, reported single-center studies with CMR ex- The assessment of RV wall hypertrophy by CMR is not a
aminations of patients with ARVC report on a small number diagnostic component of the modified Task Force criteria
of patients, thereby reflecting this center’s experience with its (9). Hypertrophy of the right ventricular wall is defined as
specific patient selection. Large multicenter registries have wall thickness of more than 8 mm. In ARVC, this finding
therefore been started in the United States and Europe to is rarely observed at autopsy because right ventricular wall
improve our knowledge on specific findings and definition of thickness is measured excluding epicardial fat (11). In CMR
ARVC. Until these data are published, the CMR diagnosis images with fibrofatty infiltration of the right ventricular
of ARVC should not be made on the basis of high T1 signals myocardium, it is sometimes difficult to define the epicar-
within the right ventricular myocardium, suggesting fatty in- dial edge of the myocardium and to differentiate epicardial
filtration. In addition, minimal technical quality standards fat from the fatty infiltrated myocardium. In such cases, the
(12) should be met and images should be interpreted by phy- right ventricular free wall may appear thickened. This find-
sicians who are experts in CMR reading and have clinical ing was described in 5 of 12 patients with Task Force criteria
knowledge of ARVC. of ARVC by Tandri et al. (30). With the use of chemical-shift
fat suppression, multiple signal voids within the myocar-
dium can be shown in areas with high T1 signals in nonfat-
Right Ventricular Wall Thinning
suppressed images (13).
The assessment of RV wall thickness by CMR is not a diag-
nostic component of the modified Task Force criteria (9).
Trabecular Disarray of
In healthy individuals, the mean right ventricular free wall
the Right Ventricle
thickness is 2.7 ± 0.4 mm, and at the right ventricular apex,
it is 1.9 ± 1.1 mm (8). Right ventricular wall thinning is The depiction of trabecular disarray of the RV by CMR is
defined as a focal abrupt reduction in wall thickness of less not a diagnostic component of the modified Task Force cri-
than 2 mm, surrounded by regions of normal wall thickness teria (9). Molinari et al. (20) described thickened Y-shaped
(13). Although the exact pathogenesis of wall thinning is not trabeculae and hypertrophy of the moderator band in
known, it is thought to be caused by apoptotic loss of right patients with ARVC. Mild hypertrophy could be observed in
ventricular myocytes. Right ventricular wall thinning in 2 of 20 healthy volunteers, and it was moderate in 7 of 12
ARVC was often reported at autopsy (28,29). Auffermann and severe in 5 of 12 patients with ARVC. Tandri et al. (30)
et al. (15) reported wall thinning in 67% of patients with found a prevalence of trabecular hypertrophy and disarray
biopsy-proven ARVC on SE images. In the series reported by in 40% of patients with ARVC. Abnormal right ventricular
Tandri et al. (13) on 12 patients who met the Task Force cri- morphology, including muscle bundles, trabeculation, and
teria for the diagnosis of ARVC, right ventricular wall thin- aneurysm formation, was found in 16% of healthy controls,
ning using a DIR-FSE technique was found in less than 25%. 21% of patients with suspected ARVC, and 47% of patients
Bomma et al. (16) re-evaluated CMR findings in 89 patients with a definite ARVC diagnosis (12). Such findings may
with ARVC. Thirty patients had undergone implantation of represent the angiographic findings of deep fissures with a
a cardioverter-defibrillator because of the primary diagno- “stack of plates” appearance. However, this is not a specific
sis. The remaining 59 patients were reinvestigated by CMR finding for ARVC and may represent any condition with
imaging applying DIR-FSE with and without chemical-shift right ventricular hypertrophy or enlargement.
fat suppression and bright blood imaging. The most com-
mon abnormality of the right ventricular free wall, which
Enlargement of the Right Ventricular
was believed to be consistent with the diagnosis of ARVC
Outflow Tract
during the primary CMR examination, was wall thinning
and/or focal intramyocardial fat in 46 of 60 patients; 84% The assessment of an enlarged RV outflow tract by CMR
of those patients underwent re-evaluation by CMR. The is not a diagnostic component of the modified Task Force
findings were not confirmed on repeat CMR in any of the criteria (9). In adults, the diameter of the right ventricular
patients who initially were diagnosed with ARVC based on outflow tract is usually equal to or slightly smaller than
wall thinning and intramyocardial fat alone. ARVC was not the left ventricular outflow tract at the level of the aortic
ultimately diagnosed in any of these patients. valve. Ricci et al. (19) reported enlarged dimensions of the
Although CMR is a useful tool for the assessment of the right ventricular outflow tract in 15 patients with the clini-
right ventricular myocardium, the study by Bomma et al. cal diagnosis of ARVC compared with patients with dilated
(16) showed a poor agreement among CMR physicians in cardiomyopathy. Papavassiliu et al. (31) investigated 20
interpretation and qualitative findings such as wall thinning. patients with proven Brugada syndrome, which may have
Limited spatial resolution, motion artifacts, high epicardial some linkage to ARVC at least in some patients. With the
fat signal adjacent to the area of interest, and lack of stan- use of T1-weighted turbo SE and gradient-echo sequences,
dardized CMR protocols for ARVC may contribute to this the right ventricular outflow tract was noted to be signifi-
fact. In addition, findings of right ventricular free wall thin- cantly enlarged in patients with Brugada syndrome com-
ning and/or increased intramyocardial fat signal on CMR pared with controls. High intramyocardial T1 signal was
LWBK1209-ch11_p149-155.indd 153 17/05/13 6:04 PM

ARVC. Endomyocardial biopsy was performed in 9 of 12

patients with ARVC and revealed fibrofatty infiltrations in
four patients. Each of these four patients had increased right
ventricular signal on delayed enhancement images. Delayed
contrast-enhanced CMR may have the potential to improve
overall diagnostic accuracy of CMR for ARVC. However, it
may be difficult to exclude the disease in early stages when
ventricular arrhythmias are present but histologic alterations
are still subtle. The presently available imaging techniques
may not have the necessary spatial and contrast resolution
to demonstrate very small amounts of fibrous tissue within
thinned segments of the right ventricular and anterior wall
during early stages of ARVC.
Diastolic Dysfunction of
the Right Ventricle
The assessment of RV diastolic dysfunction by CMR is not
a diagnostic component of the modified Task Force crite-
ria (9). Abnormal right ventricular diastolic function can be
detected in patients with ventricular tachycardias that origi-
nated from the right ventricle using CMR, even if systolic
function is preserved (32). Kayser et al. (14) applied MR
Figure 11.3. Bright blood image from a cine loop in a RV long- velocity mapping to measure blood flow across the tricuspid
axis plane through the RV outflow tract in a patient with ARVC. valve during the whole cardiac cycle for the assessment of
Thinning and outward bulging of the myocardial wall in the anterior diastolic right ventricular function in 25 patients with LBBB
region of the RV outflow tract (arrow). ventricular tachycardias. Fourteen patients with anatomic
and morphologic right ventricular abnormalities suggestive
for ARVC had diastolic right ventricular dysfunction com-
observed in 4 of 20 patients and in none of the controls. pared with 11 patients without ARVC. The diastolic func-
Right ventricular dimensions were not significantly larger in tional abnormalities correlated well with morphologic alter-
patients with Brugada syndrome. In contrast, in the series ations. Thus, analysis of diastolic right ventricular function
reported by Ricci et al. (19), patients with ARVC also had in patients with ARVC may improve the identification of
increased right ventricular end-diastolic diameters and vol- individuals at risk for arrhythmic events during follow-up.
umes. Tandri et al. (30) noted that dysmorphic appearance of
the right ventricular outflow tract may be even more impor-
tant than simple enlargement. In the absence of pulmonary Summary
hypertension enlargement, dysmorphic appearance of the
right ventricular outflow tract with dyskinetic segments dur- The diagnosis of ARVC is presently based on the fulfill-
ing systole is highly suspicious for ARVC (Fig. 11.3). ment of the modified ARVC Task Force criteria (9). CMR
is an excellent imaging modality to delineate the morphol-
ogy of the right ventricle and is considered to be the gold
Detection of Myocardial Fibrosis
standard method for the assessment of RV volumes, global,
The assessment of myocardial fibrosis by CMR is not a diag- and regional function. These features make CMR imaging
nostic component of the modified Task Force criteria (9). the most useful imaging modality for the diagnosis of ARVC
In endomyocardial biopsy specimens, replacement of myo- according to the modified Task Force criteria.
cardium by fibrous and fatty tissue is one of the key find- However, further developments in spatial, temporal, and
ings suggesting ARVC. Delayed contrast-enhanced CMR contrast CMR resolution may enhance the diagnostic accu-
imaging of the myocardium is widely used to depict myo- racy of CMR for ARVC by improvement of the detection
cardial scars in patients with ischemic heart disease and and delineation of subtle morphologic and functional altera-
fibrous intramyocardial tissue in various cardiomyopathies. tions of the RV myocardium and the replacement of myocar-
These alterations are most often found in the left ventricular dium by fibrofatty tissue.
myocardium. Papavassiliu et al. (31) prospectively evalu-
ated 30 patients with suspected ARVC by using myocar-
dial delayed-enhancement CMR for noninvasively detect- References
ing fibrosis in ARVC. Diagnosis of ARVC was based on
1. Marcus FI, Fontaine G, Guiraudon G, et al. Right ventricular dysplasia: A report of 24
original Task Force criteria (8) and did not include CMR adult cases. Circulation. 1982;65:384–398.
findings. A total of 12 of 30 patients met the original Task 2. Fontaine G, Fontaliran F, Frank R. Arrhythmogenic right ventricular cardiomyopathies:
Force criteria, and 8 of 12 patients demonstrated increased Clinical forms and main differential diagnoses. Circulation. 1998;97:1532–1535.
3. Tabib A, Loire R, Chalabreysse L, et al. Circumstances of death and gross and microscopic
signals on delayed enhancement images in the right ventric- observations in a series of 200 cases of sudden death associated with arrhythmogenic right
ular myocardium compared with 0 of 18 patients without ventricular cardiomyopathy and/or dysplasia. Circulation. 2003;108:3000–3005.
LWBK1209-ch11_p149-155.indd 154 17/05/13 6:04 PM

4. Maron BJ, Shirani J, Polac L, et al. Sudden death in young competitive athletes: Clinical, 18. Tandri H, Macedo R, Calkins H, et al. Role of magnetic resonance imaging in arrhythmo-
demographic and pathological profiles. JAMA. 1996;276:199–204. genic right ventricular dysplasia: Insights from the North American Arrhythmogenic Right
5. Thiene G, Nava A, Corrado D, et al. Right ventricular cardiomyopathy and sudden death Ventricular Dysplasia (ARVD/C) study. Am Heart J. 2008;155:147–153.
in young people. N Engl J Med. 1988;318:129–133. 19. Ricci C, Longo R, Pagnan L, et al. Magnetic resonance imaging in right ventricular dyspla-
6. Corrado D, Basso C, Thiene G, et al. Spectrum of clinicopathologic manifestations of sia. Am J Cardiol. 1992;70(20):1589–1595.
arrhythmogenic right ventricular cardiomyopathy/dysplasia: A multicenter study. J Am 20. Molinari G, Sardanelli F, Gaita F, et al. Right ventricular dysplasia as a generalized car-
Coll Cardiol. 1997;30:1512–1520. diomyopathy? Findings on magnetic resonance imaging. Eur Heart J. 1995;16(11):1619–
7. Thiene G, Nava A, Angelini A, et al. Anatomoclinical aspects of arrhythmogenic right 1624.
ventricular cardiomyopathy. In: Baroldi G, Camerini F, Goodwin JF, eds. Advances in 21. Menghetti L, Basso C, Nava A, et al. Spin-echo nuclear magnetic resonance for tissue char-
Cardiomyopathy. Berlin, Germany: Springer Verlag; 1990:397–408. acterisation in arrhythmogenic right ventricular cardiomyopathy. Heart. 1996;76:467–
8. McKenna WJ, Thiene G, Nava A, et al. Diagnosis of arrhythmogenic right ventricu- 470.
lar dysplasia/cardiomyopathy. Task Force of the Working Group Myocardial and 22. Midiri M, Finazzo M, Brancato M, et al. Arrhythmogenic right ventricular dysplasia: MR
Pericardial Disease of the European Society of Cardiology and of the Scientific Council on features. Eur Radiol. 1997;7(3):307–312.
Cardiomyopathies of the International Society and Federation of Cardiology. Br Heart J. 23. Keller DI, Osswald S, Bremerich J, et al. Arrhythmogenic right ventricular cardiomyopa-
1994;71:215–218. thy: Diagnostic and prognostic value of the cardiac MRI in relation to arrhythmia-free
9. Marcus FI, McKenna WJ, Sherrill D, et al. Diagnosis of arrhythmogenic right ventricular survival. Int J Cardiovasc Imaging. 2003;19:537–543; discussion 545–547.
cardiomyopathy/dysplasia. Proposed modification of the Task Force Criteria. Eur Heart J. 24. Vermes E, Strohm O, Otmani A, et al. Impact of the revision of Arrhythmogenic Right
2010;31:806–814. Ventricular Cardiomyopathy/Dysplasia Task Force criteria on its prevalence by CMR crite-
10. Fontaine G, Fontaliran F. Arrhythmogenic right ventricular dysplasia masquerading as ria. J Am Coll Cardiology Img. 2011;4:282–287.
dilated cardiomyopathy. Am J Cardiol. 1999;84:1143. 25. Fontaine G, Fontaliran F, Hebert JL, et al. Arrhythmogenic right ventricular dysplasia.
11. Burke AP, Farb A, Tashko G, et al. Arrhythmogenic right ventricular cardiomyopathy Annu Rev Med. 1999;50:17–35.
and fatty replacement of the right ventricular myocardium: are they different diseases? 26. Goldfarb JW, Roth M, Han J. Myocardial fat deposition after left ventricular myocar-
Circulation. 1998;97:1571–1580. dial infarction: Assessment by using MR water-fat separation imaging. Radiology.
12. Bluemke DA, Krupinski EA, Ovitt T, et al. MR imaging of arrhythmogenic right ven- 2009;253:65–73.
tricular cardiomyopathy: Morphologic findings and interobserver reliability. Cardiology. 27. Basso C, Ronco F, Marcus F, et al. Quantitative assessment of endomyocardial biopsy
2003;99:153–162. in arrhythmogenic right ventricular cardiomyopathy/dysplasia: An in vitro validation of
13. Tandri H, Calkins H, Nasir K, et al. Magnetic resonance imaging findings in patients diagnostic criteria. Eur Heart J. 2008;29:2760–2771.
meeting task force criteria for arrhythmogenic right ventricular dysplasia. J Cardiovasc 28. Tabib A, Loire L, Chalabreysse L, et al. Circumstances of death and gross and microscopic
Electrophysiol. 2003;14(5):476–482. observations in a series of 200 cases of sudden death associated with arrhythmogenic right
14. Kayser HW, de Roos A, Schalij MJ, et al. Usefulness of magnetic resonance imaging in ventricular cardiomyopathy and/or dysplasia. Circulation. 2003;108:3000–3005.
diagnosis of arrhythmogenic right ventricular dysplasia and agreement with electrocardio- 29. Lobo FV, Heggtveit HA, Butany J, et al. Right ventricular dysplasia: Morphological find-
graphic criteria. Am J Cardiol. 2003;91(3):365–367. ings in 13 cases. Can J Cardiol. 1992;8(3):261–268.
15. Auffermann W, Wichter T, Breithardt G, et al. Arrhythmogenic right ventricular disease: 30. Tandri H, Friedrich MG, Calkins H, et al. MRI of arrhythmogenic right ventricular cardio-
MR imaging vs angiography. AJR Am J Roentgenol. 1993;161(3):549–555. myopathy/dysplasia. J Cardiovasc Magn Reson. 2004;6:557–563.
16. Bomma C, Rutberg J, Tandri H, et al. Misdiagnosis of arrhythmogenic right ventricular 31. Papavassiliu T, Wolpert C, Fluchter S, et al. Magnetic resonance imaging findings in
dysplasia/cardiomyopathy. J Cardiovasc Electrophysiol. 2004;15:300–306. patients with Brugada syndrome. J Cardiovasc Electrophysiol. 2004;15(10):1133–1138.
17. Tandri H, Castillo E, Ferrari VA, et al. Magnetic resonance imaging of arrhythmogenic 32. Kayser H, Schalij M, Van der Wall E, et al. Biventricular function in patients with nonisch-
right ventricular dysplasia: Sensitivity, specificity, and observer variability of fat detection emic right ventricular tachyarrhythmias assessed with MR imaging. Am J Roentgenol.
versus functional analysis of the right ventricle. J Am Coll Cardiol. 2006;48:2277–2284. 1997;159:995–999.
LWBK1209-ch11_p149-155.indd 155 17/05/13 6:04 PM

Chapter
12 Karen G. Ordovas
Charles B. Higgins
MRI and CT of Pericardial Diseases
■■ Goals and Indications is also not very accurate for depiction of pericardial thick-
ening, because echogenicity of the pericardium is similar to
■■ Techniques
adjacent tissues.
■■ Normal Pericardium and Pericardial Recesses CT and MRI (magnetic resonance imaging) provide ex-
■■ Pericardial Effusion cellent delineation of the mediastinal anatomy, with high
soft-tissue contrast and spatial resolution (1–5). The visu-
■■ Acute Pericarditis alization of the entire chest can also give important infor-
■■ Cardiac (Pericardial) Tamponade mation for differential diagnosis and extent of pericardial
diseases. Both methods provide some information on tissue
■■ Constrictive Pericarditis characterization, which is useful for the diagnosis of pericar-
■■ Absence of Pericardium dial masses and pericardial inflammation.
CT is the best method for depiction of pericardial calcifi-
■■ Pericardial Cysts and Diverticula cation, a finding suggestive of constrictive pericarditis in the
■■ Pericardial Hematomas appropriate clinical setting. CT-guided pericardiocenteses is a
well-established technique for management of loculated peri-
■■ Pericardial Tumors
cardial effusions. Electrocardiogram-gated CT can eliminate
the artifacts related to cardiac motion and is superior to non-
gated studies for the evaluation of pericardial abnormalities.
Goals and Indications MRI has the advantage of avoiding iodinated contrast
media and ionizing radiation as well as providing imaging in
The goals and indications for tomographic imaging in peri- multiple planes. A major contribution for MRI is the mea-
cardial diseases are the following: surement of pericardial thickness, which is useful for diag-
1. Visualization of loculated pericardial effusions nosis of constrictive pericarditis in patients with constrictive/
2. Differentiation of hemorrhagic from nonhemorrhagic restrictive physiology. Cine MRI can usually differentiate a
effusion small pericardial effusion from pericardial thickening, which
3. Diagnosis of constrictive pericardial disease and differen- can be misinterpreted by CT.
tiation from restrictive cardiomyopathy
4. Evaluation and characterization of pericardial masses
Techniques
5. Identification of pericardial inflammation and calcifica-
tion.
Multidetector CT imaging with cardiac gating is ideally used
There are several imaging techniques that are effective for to assess pericardial anatomy and cardiac function with ade-
the diagnosis of pericardial diseases. The imaging modality quate temporal and high spatial resolution. However, non-
initially and most frequently utilized is transthoracic echo- gated multidetector CT is usually adequate for diagnosis of
cardiography. Asymptomatic abnormalities of the pericar- pericardial disease.
dium are occasionally identified on computed tomography Electrocardiogram-gated MRI studies with morphologic
(CT) of the thorax. and functional sequences allow a complete evaluation of the
The imaging modality of choice for initial evaluation of pericardium with high soft-tissue contrast and spatial reso-
patients with suspected pericardial effusion is transthoracic lution. Images of the entire heart can be acquired with a
echocardiography. This widely available and relatively low- few breath-holds or during free-breathing. New techniques
cost method has a high accuracy for detection of pericardial with a phased-arrayed coil can image the heart with a high
effusion and signs of tamponade. It is also a good method temporal resolution.
for guiding diagnostic or therapeutic pericardiocenteses. The Turbo spin-echo (SE) T1-weighted images are usually
main limitation of echocardiography in pericardial diseases acquired at least in the axial and coronal planes for mor-
is its inability to assess the entire pericardial extension. It phologic assessment of the pericardium. SE or inversion
156
LWBK1209-ch12_p156-175.indd 156 5/18/13 12:51 PM

Chapter 12 ■ MRI and CT of Pericardial Diseases 157
Figure 12.1. Normal pericardium on CT. Cardiac-

gated axial images before (left) and after (right) iodin-
ated contrast media show normal appearance of a thin
pericardium, best seen adjacent to the right ventricular
free wall (arrows).
recovery gradient echo (GRE) images after administration abnormal (7,8). CT can overestimate pericardial thickness
of gadolinium chelates may be used to detect enhancement in the presence of isodense effusion (2). Visualization of the
caused by inflammatory processes of the pericardium. Cine pericardium on both CT and MRI is possible because of its
MR images using GRE or steady-state free precession se- position between two fat planes (epicardial and pericardial).
quences are acquired to evaluate ventricular function, ven- Although the pericardium is visible over the right cardiac
tricular sliding motion over the pericardium, and diastolic chambers in most studies, it may not be visible over the lat-
ventricular septal motion. eral and posterior walls of the left ventricle (7).
The pericardial cavity normally contains up to 50 mL of
serous fluid that distributes diffusely around the heart sur-
Normal Pericardium and face and inside the pericardial recesses (6). The complex
Pericardial Recesses configuration of the pericardial cavity can be understood as
two connected complex spaces surrounding the heart and
The pericardium is a compliant sac that envelops the heart great vessels (9). The first region surrounds the proximal
and the origin of the great vessels. It is attached to the ster- two-thirds of the ascending aorta and pulmonary artery
num anteriorly and to the diaphragm inferiorly. The left out to its bifurcation, and is known as transverse sinus (Fig.
atrium is only partially covered by the pericardium. Parietal 12.3) (9,10). The second region surrounds the attachment
pericardium consists of fibrous tissue with an inner layer of the inferior vena cava, superior vena cava, and pulmo-
of mesothelial cells that reflect in the regions of pericar- nary veins, delineating a cul-de-sac behind the left atrium
dial attachment and cover the surface of the heart, forming called oblique sinus (Fig. 12.4). The pericardial cavity has
the visceral pericardium (6). On MRI and CT studies, the several recesses that can be recognized on CT and MR im-
normal pericardium is usually less than 2 mm thick (Figs. ages: These are the superior and inferior aortic recesses, left
12.1 and 12.2). A thickness greater than 4 mm is considered and right pulmonic recesses, posterior pericardial recess, left
A B
Figure 12.2. Normal pericardium. Computed tomography (CT) (A) and axial spin-echo (SE) images (B)
show the normal appearance of the pericardium (arrows).
LWBK1209-ch12_p156-175.indd 157 16/05/13 9:39 PM

nodes or abnormalities of adjacent mediastinal structures

(13–16). In the setting of malignancy, a misinterpretation
of a normal recess as lymphadenopathy could lead to inap-
propriate staging of tumors (17). Occasionally, one of the
recesses can be much more prominent than the others, in-
creasing the chances of misinterpretation (Fig. 12.6).
Normal pericardial fluid resides in the recesses of normal
individuals. Fluid is invariably observed in one or more re-
cesses on CT and MRI in the absence of a pericardial effu-
sion. One or more recesses may be particularly prominent in
some normal subjects.
The pericardium is composed of fibrous tissue and ap-
pears dark on T1- and T2-weighted SE images (Fig. 12.2). In
the area of the right ventricle, it is located between the pro
minent bright pericardial and epicardial fat, which provides
a natural high level of contrast, so that the sensitivity for the
visualization of the pericardium in this region is 100% (7). In
regions adjacent to the lung, the natural contrast is lower, so
that the sensitivity for visualization in the area of the lateral
wall of the left ventricle is lower. The thickness of the peri-
Figure 12.3. Transverse sinus of the pericardium. Axial cine mag- cardium on MR images ranges from 1 to 3 mm. Anatomic
netic resonance (MR) image at the level of the main pulmonary artery studies have shown the thickness of the pericardium to be
(MPA) and ascending aorta (Ao) shows the transverse sinus of the in the range of 0.4 to 1 mm. The overestimation on MRI
pericardium (arrow). Note that the pericardial fluid inside the sinus has been presumed to result from motion-induced signal loss
has the same signal intensity of the flowing blood. Also shown are the of the normal pericardial fluid, which cannot be readily dis-
oblique sinus (arrowhead) and left pulmonic vein recess (asterisk). tinguished from the pericardium itself. Pericardial thickness
depends also on the anatomic level and increases toward the
diaphragm. This effect results from the ligamentous insertion
and right pulmonary vein recesses, and postcaval recess (Fig. of the pericardium into the diaphragm and from its tangential
12.5) (9,11,12). direction to the imaging plane in axial images. Consequently,
Knowledge of the cross-sectional anatomy of the pericar- the measurement of pericardial thickness is most reliable at
dial recesses is essential for its differentiation from lymph the midventricular level.
A B
Figure 12.4. A: Oblique sinus of the pericardium. Contrast-enhanced CT image at the level of the right
pulmonary artery (RPA) shows fluid inside the oblique sinus of the pericardium (arrow). Note that a fat plane
separates this pericardial sinus from the anterior esophageal wall. B: Contrast-enhanced CT scan shows small
pericardial recess around right inferior pulmonary vein (arrow).
LWBK1209-ch12_p156-175.indd 158 16/05/13 9:39 PM

A B
C D
Figure 12.5. Pericardial recesses. Axial cine MRI (A) and axial SE (B) images at the level of the pulmonary
artery bifurcation show the anterior (black arrow) and posterior (arrowhead) portions of the superior aortic
recess and the left pulmonic recess (white arrow). Axial SE image (C) and contrast-enhanced CT (D) at the
level of the left pulmonary artery also show the anterior (arrow) and posterior (arrowhead) portions of the
superior aortic recess.
Pericardial Effusion echocardiography, CT, or MRI. The procedure of choice for

evaluation of the presence and size of pericardial effusion is
The pericardial space normally contains 10 to 50 mL of transthoracic echocardiography. Collections of fluid within
fluid. In most normal subjects, pericardial fluid can be iden- the pericardial cavity can represent transudate or exudates.
tified in the pericardial recesses on CT and MRI. The numer- The latter include serous, fibrinous, purulent, and hemor-
ous etiologies of pericardial effusion are listed in Table 12.1. rhagic or (rarely) chylous pericardial effusion (18).
For many pericardial effusions, the etiology is never estab- Simple pericardial effusion (transudate) has a density
lished. The effusion is recognized as an incidental finding on close to water on CT (− 20 to 20 Hounsfield units) (Fig. 12.7).
LWBK1209-ch12_p156-175.indd 159 16/05/13 9:39 PM

tiologies of Pericardial
E
Table 12.1
Effusions
1. Acute pericarditis (including myopericarditis)

2. Effusive-constrictive pericarditis
3. Congestive heart failure
4. Collagen vascular diseases (especially systemic lupus
erythematosus)
5. Uremia
6. Postmyocardial infarction (Dressler’s syndrome)
7. Postpericardiotomy syndrome
8. Trauma
9. Aortic dissection—type A
10. Drug reaction
11. Myxedema
12. Chylopericardium
13. Neoplastic (pericardial metastases, frequently hemorrhagic)
14. Idiopathic
increased fluid density and heterogeneity on CT suggest

exudates. On MRI, intermediate signal on T1-weighted
spin-echo images suggests exudates (Fig. 12.9) (4).
Figure 12.6. Prominent superior aortic recess. Contrast-enhanced The appearance of pericardial hematoma depends on
CT at the level of the carina shows more than usual volume of peri- the age of the pericardial blood. Acute hemopericardium
cardial fluid inside the posterior portion of the superior aortic recess has a high density on CT (greater than 40 to 50 Hounsfield
(arrow). units), whereas subacute hematomas present with a hetero-
geneous density (Fig. 12.10). MRI of acute pericardial he-
matoma shows a high signal intensity on SE T1-weighted
The typical appearance on MRI is low signal intensity on sequences and a low signal intensity on cine GRE images
T1-weighted SE and high signal intensity on T2-weighted (21,22). Subacute and chronic hematomas are usually het-
SE, cine GRE, and steady-state free precession sequences erogeneous with both high and low signal intensity regions
(Fig. 12.8) (4,19,20). on T1-weighted images (Fig. 12.11) (23,24). Distinguishing
Differentiation of transudate from exudates is usually features of pericardial hematoma compared to nonhemor-
not possible based on imaging appearance alone. However, rhagic effusion are summarized in Table 12.2.
A B
Figure 12.7. Pericardial effusion on CT. CT image (A) demonstrates homogeneous small pericardial effu-
sion with water density (asterisk). Contrast-enhanced CT (B) shows a large pericardial effusion (asterisk).
LWBK1209-ch12_p156-175.indd 160 16/05/13 9:39 PM

A B
Figure 12.8. Pericardial effusion on magnetic resonance imaging (MRI). Axial T1-weighted SE image (A)
shows the low signal intensity of a pericardial effusion (asterisk). High signal intensity of the effusion (asterisk)
is observed on axial cine MR image (B).
Chylous effusion is extremely rare. It has been described or myocardial tumor on T1-weighted SE images. CT and
as a low-density pericardial collection on CT. MRI are very sensitive for the detection of generalized or
Small amounts of pericardial fluid can be misinterpreted loculated pericardial effusions. CT and MRI can detect
on CT imaging because of similar appearance to pericardial pericardial fluid volume as small as 30 mL (4). Both provide
thickening (2,4). Cine MRI (steady-state free precession se- information on the location of pericardial effusion and are
quences) can distinguish between the high intensity of a especially effective for demonstrating loculated effusions
small simple effusion and the adjacent low signal of normal (Fig. 12.11A). Because of uneven distribution, the total
parietal pericardium. However, small volume of fluid and fluid volume cannot be calculated from the width of the
pericardium may be indistinguishable on spin-echo MRI pericardial space by applying a simple formula. However,
and thereby simulate pericardial thickening. Pericardial ef- a semiquantitative estimation can be obtained by measur-
fusion shows no increase in signal on T1-weighted images ing the width of the pericardial space in front of the right
after administration of gadolinium chelate. This feature aids ventricle. A moderate effusion is associated with a width
in distinguishing between pericardial fluid and pericardial of more than 5 mm. Loculated effusion occurs when fluid
A B
Figure 12.9. Pericardial exudate.SE MR images in the axial (A) and sagittal (B) planes show a heteroge-
neous pericardial effusion with areas of intermediate signal in the fluid (arrows). A thickened pericardium is
also visible (arrowhead), consistent with an inflammatory process.
LWBK1209-ch12_p156-175.indd 161 16/05/13 9:39 PM

A B
Figure 12.10. Pericardial hematoma on CT. Contrast-enhanced CT at the level of the right pulmonary
artery (A) shows a type-A aortic dissection (arrows) associated with a high-density pericardial effusion (aster-
isk) consistent with acute hemopericardium. CT image at the level of the base of the heart (B) in another
patient after cardiac surgery shows a large heterogeneous collection on the anterior aspect of the pericardium
(arrows). Note the compression of the right ventricle by the hematoma.
A B
C D
Figure 12.11. MR appearance of pericardial hematoma. Axial SE MR images at the midventricular level
show different ages of pericardial blood. The homogeneous high-intensity collection adjacent to the right
atrium (arrow) is an acute hematoma (A). Note the low intensity of simple pericardial effusion at the left side
of the cavity (asterisk). Subacute pericardial hemorrhage with areas of low and high signal intensity is shown
in loculated (B, C) pericardial hematomas (arrows). The homogeneous low-intensity collection compressing
the left ventricle (D) is a typical presentation of an old chronic hematoma (arrow).
LWBK1209-ch12_p156-175.indd 162 16/05/13 9:39 PM

emorrhagic Versus
H tiologies of Acute
E
Table 12.3
Table 12.2 Nonhemorrhagic Pericarditis
Pericardial Effusion
1. Infection
Hemorrhagic Effusion a. Viral (Coxsackie A, Coxsackie B, echovirus, adenovirus,
1. High density on CT (higher than 40–45 Hounsfield units) hepatitis B, herpes virus, parvovirus, AIDs, etc.)
2. High intensity on T1-weighted spin-echo MRI b. Bacterial (staphylococcal, streptococcal, pneumococcal, etc.)
3. Heterogeneous (high, medium, low) on T1-weighted spin-echo c. Fungal (histoplasmosis, coccidioidomycosis, candidosis,
MRI aspergillosis, cryptococcosis, etc.)
4. Heterogeneous or low signal on gradient-echo (cine MRI) and d. Tuberculosis
balanced steady-state free precession (cine MRI) images 2. Acute myocardial infarction
5. Low signal on all types of MR sequences (old loculated 3. Acute myocarditis (myopericarditis)
pericardial hematoma) 4. Neoplastic (especially lymphoma, melanoma, lung cancer,
breast cancer, leukemia)
Nonhemorrhagic Effusions 5. Radiation
1. Low (water density) on CT (usually −20 to 20 Hounsfield units) 6. Collagen vascular diseases (especially systemic lupus
2. Low signal (less than myocardium) on T1-weighted spin-echo erythematosus, rheumatoid arthritis, scleroderma, mixed
images connective tissue disease)
3. Homogeneous high signal on T2-weighted spin-echo images 7. Sarcoidosis
4. Homogeneous high signal on gradient-echo (cine MRI) and 8. Systemic diseases (amyloidosis, Behcet’s disease, inflammatory
balanced steady-state free precession (cine MRI) images bowel disease)
9. Drug reaction (adriamycin, hydralyzine, procainamide,
diphenylhydantoin, methysergide)
10. Postmyocardial infarction (Dressler’s syndrome)
is trapped by adhesions between the parietal and visceral 11. Postpericardiotomy syndrome
pericardium. MRI is more sensitive than echocardiography 12. Thoracic trauma
in detecting loculated effusion because of its wide field of 13. Idiopathic
view.
Acute Pericarditis
Acute pericarditis is an inflammation of the pericardium due Cardiac (Pericardial) Tamponade
to multiple etiologies (Table 12.3) of which the most fre-
quent is viral (18,25). The diagnostic features are effusion Cardiac tamponade causes life-threatening hypotension
and thickened pericardium. The thickening of the pericar- through compression of the right-sided cardiac chambers by
dium is due to fibrinous exudates and edema. accumulation of pericardial effusion. Tamponade can occur
The thickened pericardium and effusion can be identi- with pericardial fluid of any type but rapid bleeding into
fied on CT and MRI (Fig. 12.12). After contrast media the pericardium is likely the most frequent mechanism today
administration, the pericardium shows various levels of en- either as a complication of type-A aortic dissection or from
hancement (Fig. 12.12). Enhancement of the pericardium pericardial hematoma accumulation after cardiac surgery.
is demonstrated on early spin-echo MRI, especially using As little as 250 mL of fluid accumulating rapidly can cause
fat saturation technique and on delayed gadolinium-en- tamponade while chronic effusion of 1 to 2 L may be accom-
hanced (inversion recovery GRE) images (Fig. 12.13). The modated without tamponade.
diagnostic features of acute pericarditis are summarized in The diagnosis is most frequently made by echocardiogra-
Table 12.4. phy but is sometimes made by CT or MRI. On CT and MRI,
the diagnosis is suggested by loss of the convex contour of
the right atrium or even reversal of the convexity to a con-
cavity of the right atrial wall.
T and MRI Features of
C
Table 12.4
Acute Pericarditis
1. Pericardial effusion
Constrictive Pericarditis
a. Water density on CT
b. Low intensity on T1-weighted spin-echo MRI Constrictive pericarditis is due to progressive pericardial
c. High intensity on cine MRI fibrosis, which impedes diastolic ventricular expansion.
2. Thickened pericardium on CT and MRI (>3-mm thickness) It is a nonspecific reaction to various conditions, such as
3. Early contrast enhancement of pericardium on CT and MRI infectious pericarditis, connective tissue disease, neoplasm,
4. Delayed gadolinium enhancement of pericardium ± epicardial trauma, long-term renal dialysis, cardiac surgery, and radia-
myocardium on delayed inversion recovery gradient-echo MR tion therapy (Table 12.5) (18,26). Currently, constrictive
sequence (10–15 min after gadolinium administration) pericarditis is most often a sequela of cardiac surgery in
5. Possible early enhancement of pericardial inflammatory deposits
Europe and North America. However, tuberculous pericar-
and adhesions on gadolinium-enhanced T1-weighted spin-echo
MRI (usually with fat saturation)
ditis remains a leading cause of constrictive pericarditis in
developing countries.
LWBK1209-ch12_p156-175.indd 163 16/05/13 9:39 PM

A B
Figure 12.12. Acute pericarditis. Contrast-enhanced CT image

(A) shows enhancement of the pericardial layers (arrows) and a small
pericardial effusion. Note the large pleural effusions (asterisk). Axial
T1-weighted SE images before (B) and after (C) administration of
gadolinium chelate show thickening of the pericardium (arrows) and
pericardial effusion (asterisk). After administration of contrast mate-
C rial, the pericardium enhances (arrowheads). Note the intermediate
signal intensity of the effusion, consistent with an exudate.
Figure 12.13. Acute pericarditis. T1-weighted spin-echo images before (left) and after (right) gadolinium
chelate administration. Thickened pericardium (arrows) is markedly enhanced by the contrast.
LWBK1209-ch12_p156-175.indd 164 16/05/13 9:39 PM

tiologies of Chronic
E T and MRI Features of
C
Table 12.5
Constrictive Pericarditis Table 12.6 Constrictive Pericarditis—
Morphologic
1. Postcardiac surgery (most frequent cause in advanced
industrialized nations) 1. Acute inflammatory constrictive
2. Idiopathic (nearly as frequently as 1—probably due to previous 1. Pericardial thickness ≥4 mm
episode(s) of viral pericarditis) 2. Effusion
3. Postradiation 3. Early contrast enhancement of pericardium
4. Tuberculosis (still most frequent cause in developing nations) 2. Effusive constrictive
5. Fungal and parasitic infections 1. Pericardial thickness ≥4 mm
6. Chronic renal failure (complication of uremic pericarditis) 2. Effusion
7. Collagen vascular diseases (especially SLE and rheumatoid arthritis) 3. Adhesions between parietal and visceral pericardium
8. Neoplastic infiltration (encasement) of pericardium 4. Early contrast enhancement of parietal and visceral
9. Asbestosis (very rare) pericardial
3. Chronic fibrous constrictive
SLF, systemic lupus erythematosus. 1. Pericardial thickness ≥4 mm
2. No effusion
3. No early contrast enhancement
As a result of inflammatory processes, fibrosis of the peri- 4. Enhancement of thickened pericardium on delayed
cardium and/or pericardial adhesions may occur, sometimes (10–15 min after gadolinium) inversion recovery
with concomitant calcifications. The less compliant pericar- gradient-echo MR imagesa
dium impedes ventricular and/or atrial filling, depending on 4. Adhesive constrictive (very rare)
the extent of the disease. Another consequence of the rigid 1. Pericardial thickness >4 mm
2. Adhesions between visceral and parietal pericardium
pericardium is ventricular interdependence. Because the free
3. Adhesions between pericardium and other surrounding
walls of ventricular chambers have restricted expansion mediastinal tissues
during diastole, the volumetric capacity of one ventricle is 5. Cardiac chamber configuration (types 1–3)
strictly related to the filling of the opposite-sided ventricle. 1. Dilated right atrium and inferior vena cava
Therefore, minimal and instantaneous differences in ven- 2. Normal size or small right ventricle
tricular filling result in diastolic ventricular septal motion 3. Tunnel-shaped right ventricle
toward the left ventricle (18). 4. Sigmoid-shaped ventricular septum
There are four different presentations of constrictive peri-
a
carditis: Acute inflammatory, effusive, adhesive, and chronic Greater degree of enhancement may indicate residual inflammation
fibrous. The CT and MRI morphologic features are sum- which responds to anti-inflammatory therapy (23).
marized in Table 12.6. The classic picture is classified as
chronic fibrous constrictive pericarditis and consists of fi-
brotic changes with or without associated calcifications con-
sequent to a previous chronic inflammatory process (Fig. is crucial for the patient’s management because constrictive
12.14). Adhesive constrictive pericarditis is characterized pericarditis is treated surgically and has a good prognosis,
by the presence of fibrous bridges between the parietal and whereas restrictive cardiomyopathy is treated medically and
visceral pericardium and between parietal pericardium and has a poor outcome (27,28).
other mediastinal tissues, and is usually related to previ- CT and MRI are the methods of choice for measuring peri-
ous suppurative or caseous pericarditis (Fig. 12.15). Acute cardial thickness. MRI has a reported accuracy of 93% for
inflammatory constrictive pericarditis presents as an acute the diagnosis of constrictive pericarditis based on the presence
pericarditis with marked thickening of the pericardium asso- of pericardial thickness (≥4 mm) in the appropriate clinical
ciated with constrictive physiology (Fig. 12.16). The inflam- setting (29). It is important to emphasize that rarely constric-
matory process can resolve completely or become a chronic tive pericarditis can occur with normal pericardial thickness.
constriction. Effusive-constrictive pericarditis is subacute An advantage of CT is its high sensitivity for detection of
constrictive pericarditis mainly involving the visceral peri- pericardial calcification, a finding highly suggestive of con-
cardium and associated with a chronic pericardial effusion strictive pericarditis in the correct clinical setting (Fig. 12.14).
(Fig. 12.17). It is usually recognized when constrictive physi- Other morphologic characteristics that can be visualized in
ology remains after drainage of the pericardial effusion. constrictive pericarditis are the presence of adhesions between
Various patterns of distribution have been described in- visceral and parietal pericardium, tube-like configuration of
cluding global, right sided, left sided, and focal (Fig. 12.18). one or both ventricles, enlargement of one or both atria, and
The focal type usually involves the atrioventricular groove enlargement of superior and inferior vena cavae (4,5,29).
and thereby may impede right atrial, and less frequently, left Cine MRI with high temporal resolution can depict ab-
atrial emptying (Fig. 12.19). normal ventricular septal motion toward the left ventricle in
The distinction between constrictive pericarditis and re- early diastole (Fig. 12.20) (30,31). Previously described on
strictive cardiomyopathy is clinically difficult because both echocardiography, this finding is related to the characteristic
diseases present with symptoms of diastolic heart failure increased ventricular interdependence present in constrictive
and frequently have similar findings of increased and equal- pericarditis (32). This finding also has high sensitivity and
ized atrial and ventricular diastolic pressures on both sides specificity for the diagnosis of constrictive pericarditis in
of the heart on cardiac catheterization. This differentiation patients with diastolic dysfunction.
LWBK1209-ch12_p156-175.indd 165 16/05/13 9:39 PM

A B
Figure 12.14. Chronic fibrous constrictive pericarditis in three

patients. CT image at the midventricular level (A) shows thin layer
of calcification over right side of pericardium (arrows). CT images at
the base of the heart (B) and at the level of diaphragm (C) in another
patient demonstrate thick and irregular calcifications in the anterior
pericardial region (arrows) and the atrioventricular groove (arrow-
heads). Axial (D) and coronal (E) SE MR images in a third patient show
E
diffuse thickening of the pericardium (arrows).
LWBK1209-ch12_p156-175.indd 166 16/05/13 9:39 PM

A B
Figure 12.15. Adhesive constrictive pericarditis. Axial SE images at the midventricular level before (A) and
after (B) administration of contrast media show a diffusely thick and irregular pericardium (arrows) and peri-
cardial effusion with heterogeneous signal intensity. Note the presence of adhesions between the pericardial
layers (arrowheads). After the administration of gadolinium chelate, there is enhancement of the pericardium
and the adhesions in this patient with chronic pericardial tuberculosis.
A B
Figure 12.16. Acute inflammatory constrictive pericarditis.

Axial SE images before (A) and after (B) administration of gado-
linium chelate show thickening of the pericardium, with diffuse
pericardial enhancement after the administration of contrast
media (arrows). Cine MR image (C) at the same level shows a
C
thick pericardium with intermediate signal intensity (arrows).
LWBK1209-ch12_p156-175.indd 167 16/05/13 9:39 PM

Cine MRI usually shows normal right and left ventricu-

lar ejection fractions and normal wall motion in constrictive
pericarditis. It may also demonstrate the loss of normal slid-
ing motion of the ventricle over the pericardium. The func-
tional features of constrictive pericarditis on cine MRI are
summarized in Table 12.7.
Delayed gadolinium-enhanced (DGE) MRI shows enhance-
ment of the pericardium in constrictive pericarditis due to
fibrosis or residual inflammation (Fig. 12.21). This imaging
technique is an inversion recovery gradient-echo sequence
acquired at 10 to 15 minutes after administration of gad-
olinium chelate (0.1 to 0.2 mmol/kg) with inversion time
optimized to null signal of normal myocardium (typically
225 to 275 milliseconds). A different contrast-enhanced
method can be used also to document pericardial inflam-
mation (edema); this consists of fat-saturated, T1-weighted
spin-echo imaging within the first minute after injection of
gadolinium chelate (Figs. 12.12 and 12.13). A recent report
(23) showed that greater degree of DGE on MRI in patients
with constrictive pericarditis occurs in patients with ongoing
inflammation which frequently responds to anti-inflamma-
Figure 12.17. Effusive-constrictive pericarditis. Contrast-enhanced
CT at the midventricular level shows thickening and enhancement of tory therapy.
both the visceral (arrowheads) and parietal (arrows) pericardium asso-
ciated with a large pericardial effusion (asterisk).
A B
Figure 12.18. Different distributions of constrictive pericarditis.

SE MR images in the axial (A) and coronal (B) planes demonstrate
the global distribution of the pericardial thickening (arrows). Axial
SE MR image at the midventricular level (C) shows thickening of
the pericardium over the right atrium and right ventricle (arrows;
right-sided distribution). Note the tubular configuration of the right
C
ventricle (RV).
LWBK1209-ch12_p156-175.indd 168 16/05/13 9:39 PM

A B
RA
C D
Figure 12.19. Focal constrictive pericarditis. Axial SE (A) and axial cine MR (B) images at the midventric-
ular level show thickening of the pericardium at the atrioventricular groove (arrow). Note that the focal peri-
cardial constriction compresses the tricuspid annulus (arrowhead). The right atrium (RA) is enlarged, whereas
the right ventricle (RV) is small and has a tubular shape. Axial CT image (C) and axial cine MR image (D) of
another patient demonstrate calcification localized at the atrioventricular groove (arrow) with narrowing of
the tricuspid annulus and a very large RA.
unctional Features of
F Absence of Pericardium
Table 12.7 Constrictive Pericarditis on
Cine MRI Pericardial defects are rare and usually asymptomatic con-
genital abnormalities. The defect can be partial or complete
1. Early diastolic leftward bowing of inlet ventricular septum and usually on the left side. CT and MRI can depict hernia-
(cine MRI in axial or horizontal long-axis planes) tion of cardiac structure through the defect. Discontinuation
2. Ventricular septal bounce in diastole of the pericardial line can occasionally be detected in the
3. Sigmoid shape of ventricular septum in diastole partial form. The most reliable signs of complete absence
4. Normal right and left ventricular ejection fractions
of the left pericardium are interposition of lung between the
LWBK1209-ch12_p156-175.indd 169 16/05/13 9:39 PM

A B
Figure 12.20. Abnormal ventricular septal motion in constric-

tive pericarditis. Axial cine MR images in the systolic (A) and early
diastolic (B) phases demonstrate bending of the ventricular septum
toward the left ventricle in early diastole (arrow). Axial SE image (C)
shows a normal pericardial thickness (arrowheads), an unusual pre-
sentation of constrictive pericarditis. Diagnosis is possible because of
C abnormal diastolic ventricular motion characteristic for constrictive
pericarditis.
LWBK1209-ch12_p156-175.indd 170 16/05/13 9:39 PM

Figure 12.21. Delayed enhancement in constrictive pericarditis. Axial black blood image (left), inversion
recovery gradient-echo immediately after gadolinium (middle) and 15 minutes later (right). Note the thick
pericardium without early contrast enhancement but intense hyperenhancement on the delayed images (arrow-
heads) consistent with chronic fibrotic form of constrictive pericarditis. Hyperenhancement on delayed gado-
linium-enhanced images has been seen both with fibrosis and residual inflammation. Greater intensity of this
enhancement may indicate cases which will respond to anti-inflammatory therapy. (Feng D, Glockner J, Kim
K, et al. Cardiac magnetic resonance imaging pericardial late gadolinium enhancement and elevated inflamma-
tory markers can predict reversibility of constrictive pericarditis after anti-inflammatory therapy: A pilot study.
Circulation. 2011;124:1830–1837.)
aorta and main pulmonary artery, in the aortopulmonary the major coronary arteries are compressed against the edge
window, and a rotation of the cardiac axis to the left side of the defect.
(Fig. 12.22) (24,33).
The usual presentation is in an asymptomatic person with
an abnormal chest radiograph with a configuration of the Pericardial Cysts and Diverticula
chest suggestive of complete or partial absence of pericar-
dium. A possible complication is bulging of the left atrial ap- Pericardial cysts are caused by developmental abnormali-
pendage and base of the heart through the defect. The edges ties and are alleged to occur when a small portion of the
of the defect can compress the left coronary artery when the pericardium is pinched off during embryonic development.
base of the left ventricle herniates through the partial de- Pseudocyst of the pericardium may develop after surgi-
fect causing myocardial ischemia, ventricular arrhythmias, cal pericardiotomy. Ninety percent of pericardial cysts are
or sudden death. Because of the herniation, one or more of located in the cardiophrenic angles (70% on the right and
A B
Figure 12.22. Absence of pericardium. Axial SE images at the level of the great vessels (A) and ventricles
(B). There is interposition of the lung between the aortic knob and the main pulmonary artery (arrow) associ-
ated with a leftward shift of the heart. Ao, aorta; PA, pulmonary artery; RV, right ventricle; LV, left ventricle.
LWBK1209-ch12_p156-175.indd 171 16/05/13 9:39 PM

20% on the left side). However, they can occur anywhere the density may approach solid tissue if the cyst contains
in the pericardium, and a pericardial cyst at an unusual highly proteinaceous fluid (40%) (19).
location sometimes cannot be distinguished from a thymic
or bronchogenic cyst. Pericardial cysts usually do no com-
municate with the pericardial cavity. In contrast, diverticula Pericardial Hematomas
have a narrow communication with the pericardial cavity.
With increase or decrease in pericardial fluid, diverticula can Pericardial hematomas occur most frequently as a sequela
change in size. of cardiac or coronary surgery. Rarely, they are caused by
A cyst filled with simple fluid (low protein content) ap- transcatheter interventional procedures or blunt or penetrat-
pears dark on T1-weighted images and homogeneously ing thoracic trauma. Loculated subacute or chronic hemato-
bright on T2-weighted images (Fig. 12.23) (34). The cyst mas can mimic pericardial tumors on thoracic radiography
does not enhance after the administration of gadolinium che- or echocardiography. These masses have an external margin
late. If the cyst is filled with blood or highly proteinaceous of pericardium and can compress the adjacent cardiac cham-
fluid (almost 40% of cysts), it is bright on T1-weighted im- ber (Fig. 12.11).
ages. The pericardial rim is visible with the characteristic On T1-weighted spin-echo images the signal is character-
low signal intensity of the pericardium. On CT, cyst has istic (Fig. 12.11). There is mixed high-to-low intensity for
water density typically (Figs. 12.23 and 12.24). However, subacute hematomas due to various components of the clot.
A B
Figure 12.23. Pericardial cyst. CT image (A), axial T1-weighted

(B) and T2-weighted (C) SE images. Cystic mass (C) conforms to the
contour of the heart and main pulmonary artery. The cyst has water
density on CT, homogeneous low to intermediate signal intensity on
C the T1-weighted images, and homogeneous high signal intensity on the
T2-weighted images.
LWBK1209-ch12_p156-175.indd 172 16/05/13 9:39 PM

A B
Figure 12.24. Pericardial cyst. CT images before (A) and after (B) the administration of contrast media
show a homogeneous mass at the right cardiophrenic angle with water attenuation (arrow). No enhancement
is observed in the postcontrast image.
With old hematomas the signal is very low on T1 and T2 Most neoplasms have medium signal intensity on T1-
spin-echo images. Hematomas have low signal on cine MR weighted SE images and high signal intensity on T2-weighted
images and do not enhance on spin-echo images after gado- SE images. Melanoma is an exception with characteristic
linium. high intensity on T1-weighted images.
Pericardial hemorrhagic effusion caused by a malignant
primary or secondary tumor of the heart or pericardium
can be recognized by MRI and CT. A hemorrhagic effusion
Pericardial Tumors should raise suspicion for tumor involvement of the peri-
cardium. Pericardial metastases can also be recognized as
Primary pericardial tumors are rare, with mesothelioma enhancing nodules or masses on the visceral or parietal peri-
being the most frequent. Sarcoma, lipoma, hemangioma, cardium after administration of contrast media (Fig. 12.27).
dermoid, and teratoma can also occur in the pericardium The characteristic signal intensity of masses such as cyst,
(Fig. 12.25) (35). lipoma, hemangioma, and melanoma metastases allows a
Secondary tumors of the pericardium are far more com- high probability of a histologic diagnosis based on MRI.
mon. These occur from local invasion of lung and mediasti- However, detection, evaluation of the tumor extension, and
nal malignancies and from distant metastases. Lymphomas, differentiation between benign and malignant appearances
melanomas, lung, and breast carcinomas are the most com- are the most important roles of cross-sectional imaging in
mon primary tumors that involve the pericardium (Fig. this setting. Nevertheless, definitive differentiation between
12.26) (36,37). benign and malignant tumors is frequently not possible.
Figure 12.25. ECG-gated spin-echo coronal images

at the level of the right pulmonary artery (left) and 1
cm further ventral (right) show an extensive pericardial
sarcoma. The tumor infiltrates throughout the left side of
the pericardial cavity and the superior recesses (arrow).
The tumor is located between the left ventricle and the
pericardium (arrowhead), with an additional component
that extends outside the pericardium.
LWBK1209-ch12_p156-175.indd 173 16/05/13 9:39 PM

Figure 12.26. Pericardial neoplastic invasion in three patients.

Contrast-enhanced CT at the level of the great vessels (A) shows inva-
sion of the pericardium by a bronchogenic carcinoma (BC). Note
obliteration of the pericardial line (arrow) and pericardial effusion
(asterisk). Contrast-enhanced CT at the level of the right pulmonary
artery (B) in another patient shows a mediastinal lymphoma (L) involv-
ing the pericardial cavity. Contrast-enhanced CT at the midventricular
level (C) in a third patient demonstrates another invasive lymphoma (L)
obliterating the adjacent pericardial line, with an associated pericardial
C
effusion (asterisk).
A B
Figure 12.27. Pericardial metastases. CT images before (A) and after (B) administration of contrast media
show a loculated pericardial effusion (asterisks). There are enhancing pericardial nodules on the visceral peri-
cardium (arrows).
LWBK1209-ch12_p156-175.indd 174 16/05/13 9:39 PM

References 21. Seelos KC, Funari M, Chang JM, et al. Magnetic resonance imaging in acute and subacute
mediastinal bleeding. Am Heart J. 1992;123:1269–1272.
22. Vilacosta I, Gomez J, Dominguez J, et al. Massive pericardiac hematoma with severe con-
1. Mulvagh SL, Rokey R, Vick GWD, et al. Usefulness of nuclear magnetic resonance imaging
strictive pathophysiologic complications after insertion of an epicardial pacemaker. Am
for evaluation of pericardial effusions, and comparison with two-dimensional echocardiog-
Heart J. 1995;130:1298–1300.
raphy. Am J Cardiol. 1989;64:1002–1009.
23. Feng D, Glockner J, Kim K, et al. Cardiac magnetic resonance imaging pericardial late
2. Wang ZF, Reddy GP, Gotway MB, et al. CT and MR imaging of pericardial disease.
gadolinium enhancement and elevated inflammatory markers can predict reversibility
Radiographics. 2003;23:S167–S180.
of contrictive pericarditis after anti-inflammatory therapy: A pilot study. Circulation.
3. White CS. MR evaluation of the pericardium. Top Magn Reson Imaging. 1995;7:258–266.
2011;124:1830–1837.
4. Axel L. Assessment of pericardial disease by MRI and CT. J. Magn Reson Imaging.
24. Letanche G, Gayer C, Souquet PJ, et al. Agenesis of the pericardium: Clinical, echocardio-
2004;19:816.
graphic and MRI aspects. Rev Pneumol Clin. 1988;44:105–109.
5. Reinmiller R, Groll R, Lipton MJ. CT and MR imaging of pericardial disease. Radiol Clin
25. Troughton RW, Asher CR, Klein AL. Pericarditis. Lancet. 2004;363:717–727.
North Am. 2004;42:587–601.
26. Fowler NO. Constrictive pericarditis: Its history and current status. Clin Cardiol.
6. Edwards ED. Applied anatomy of the heart. In: Giuliani ER, Fuster V, eds. Cardiology:
1995;18:341–350.
Fundamentals and Practice. 2nd ed. St Louis, MO: Mosby-Year Books; 1991:47–112.
27. Hancock EW. Differential diagnosis of restrictive cardiomyopathy and constrictive pericar-
7. Sechtem U, Tscholakoff D, Higgins CB. MRI of the normal pericardium. AJR Am J
ditis. Heart. 2001;86:343–349.
Roentgenol. 1986;147:239–244.
28. Nishimura RA. Constrictive pericarditis in the modern era: A diagnostic dilemma. Heart.
8. Bull RK, Edwards PD, Dixon AK. CT dimensions of the normal pericardium. Br J Radiol.
2001;86:619–623.
1998;71:923–925.
29. Masui T, Finck S, Higgins CB. Constrictive pericarditis and restrictive cardiomyopathy:
9. Vesely TM, Cahill DR. Cross-sectional anatomy of the pericardial sinuses, recesses, and
Evaluation with MR imaging. Radiology. 1992;182:369–373.
adjacent structures. Surg Radiol Anat. 1986;8:221–227.
30. Giorgi B, Mollet NRA, Dymarkowski S, et al. Clinically suspected constrictive pericarditis:
10. Im JG, Rosen A, Webb WR, et al. MR imaging of the transverse sinus of the pericardium.
MR imaging assessment of ventricular septal motion and configuration in patients and
AJR Am J Roentgenol. 1988;150:79–84.
healthy subjects. Radiology. 2003;228:417–424.
11. McMurdo KK, Webb WR, von Schulthess GK, et al. Magnetic resonance imaging of the
31. Ordovas KG, Reddy GP, Sharp MJ, et al. MRI evaluation of ventricular septal motion
superior pericardial recesses. AJR Am J Roentgenol. 1985;145:985–988.
abnormality (septal bounce): Differentiation of constrictive pericarditis from restrictive
12. Black CM, Hedges LK, Javitt MC. The superior pericardial sinus: Normal appearance on
cardiomyopathy. Radiology. 2004;233(P):668.
gradient-echo MR images. AJR Am J Roentgenol. 1993;160:749–751.
32. Himelman RB, Lee E, Schiller NB. Septal bounce, vena cava plethora, and pericardial adhe-
13. Solomon SL, Brown JJ, Glazer HS, et al. Thoracic aortic dissection: Pitfalls and artifacts in
sion: Informative two-dimensional echocardiographic signs in the diagnosis of pericardial
MR imaging. Radiology. 1990;177:223–228.
constriction. J Am Soc Echo. 1988;1:333–340.
14. Batra P, Bigoni B, Manning J, et al. Pitfalls in the diagnosis of thoracic aortic dissection at
33. Gutierrez FR, Shackelford GD, McKnight RC, et al. Diagnosis of congenital absence of left
CT angiography. Radiographics. 2000;20:309–320.
pericardium by MR imaging. J Comput Assist Tomogr. 1985;9:551–553.
15. Chiles C, Baker ME, Silverman PM. Superior pericardial recess simulating aortic dissection
34. Oyama N, Oyama N, Komura K, et al. Computed tomography and magnetic resonance
on computed tomography. J Comput Assist Tomogr. 1986;10:421–423.
imaging of the pericardium: Anatomy and pathology. Magn Reson Med Sci. 2004;3:145–
16. Groell R, Schaffler GJ, Rienmueller R. Pericardial sinuses and recesses: Findings at electro-
152.
cardiographically triggered electron-beam CT. Radiology. 1999;212:69–73.
35. Grebenc ML, Rosado de Christenson ML, Burke AP, et al. Primary cardiac and pericardial
17. Truong MT, Erasmus JJ, Gladish GW, et al. Anatomy of pericardial recesses on multidetec-
neoplasms: Radiologic-pathologic correlation. Radiographics. 2000;20:1073–1103.
tor CT: Implications for oncologic imaging. AJR Am J Roentgenol. 2003;181:1109–1113.
36. Abraham KP, Reddy V, Gattuso P. Neoplasms metastatic to the heart: Review of 3314
18. Spodick DH. Pericardial diseases. In: Braunwald E, Zipes DP, Libby P, eds. Heart Disease:
consecutive autopsies. Am J Cardiovasc Pathol. 1990;3:195–198.
A Textbook of Cardiovascular Medicine. 6th ed. Philadelphia, PA: WB Saunders; 2001:
37. Klatt EC, Heitz DR. Cardiac metastases. Cancer. 1990;65:1456–1459.
1823.
38. Brown DL, Ivey TD. Giant organized pericardial hematoma producing constrictive pericar-
19. Smith WHT, Beacock DJ, Goddard AJP, et al. Magnetic resonance evaluation of the peri-
ditis: A case report and review of the literature. J Trauma. 1996;41:558–560.
cardium. Br J Radiol. 2001;74:384–392.
39. Ferguson ER, Blackwell GG, Murrah CP, et al. Evaluation of complex mediastinal masses
20. Sechtem U, Tscholakoff D, Higgins CB. MRI of the abnormal pericardium. AJR Am J
by magnetic resonance imaging. J Cardiovasc Surg (Torino). 1998;39:117–119.
Roentgenol. 1986;147:245–252.
This chapter is modified from Chapter 34 of Webb, WR and Higgins, CR: Thoracic Imaging, Second Edition. Lippincott Wilkiams and Wilkens, Philadelphia, 2011.
LWBK1209-ch12_p156-175.indd 175 16/05/13 9:39 PM

CHAPTER
13 Charles B. Higgins
Karen G. Ordovas
Cardiac and Paracardiac Masses
■■ Goals and Indications 3. Distinguish between benign and malignant tumor, if

possible.
■■ Techniques
4. Distinguish between primary and secondary tumor.
Computed Tomography 5. Distinguish between tumor and thrombus.
Magnetic Resonance Imaging 6. Differentiation of cardiac masses from normal anatomic
■■ Location of Cardiac and Paracardiac variants
Masses The term mass is used rather than tumor in this chapter
■■ Benign Primary Cardiac Tumors because the most frequent mass within a cardiac chamber
Myxoma is thrombus rather than tumor. Primary tumors are rare.
Secondary tumors, either metastatic or representing direct
Lipoma and Lipomatous Hypertrophy of the
extension of primary tumors of another organ, are about 40
Atrial Septum times more frequent than primary cardiac tumors.
Papillary Fibroelastoma Computed tomography (CT) and magnetic resonance
Rhabdomyoma imaging (MRI) can help determine the presence and extent
Fibroma of cardiac and paracardiac tumors (1–7). These modalities,
Pheochromocytoma especially MRI, can also sometimes characterize the mass.
Hemangioma Although CT may be adequate for the evaluation of cardiac
■■ Malignant Primary Cardiac Tumors and paracardiac masses, MRI is usually employed for this
purpose. Consequently, this chapter focuses on the findings
Angiosarcoma on MRI.
Rhabdomyosarcoma Because of a wide field of view that encompasses the
Other Sarcomas (Fibrosarcoma, Osteosarcoma, cardiovascular structures, mediastinum and adjacent lung
Leiomyosarcoma, Liposarcoma) simultaneously, CT and MRI can display the intracardiac
Lymphoma and extracardiac extent of tumors. In addition, the ability
■■ Secondary Cardiac Tumors to image in multiple planes makes MRI especially suited
for the demarcation of the spatial relationship of a mass to
Direct Extension from Adjacent Tumors
the various cardiac and mediastinal structures. Multiplanar
Metastasis images or reconstructions overcome the volume-averaging
Transvenous Extension into the Heart problem at the diaphragmatic interface encountered with a
■■ Intracardiac Thrombus solely transaxial imaging plane. These features permit a clear
delineation of the possible infiltration of a mass lesion into
■■ Differentiation between Tumor and
cardiac and adjacent mediastinal structures. MRI allows the
Blood Clot assessment of functional parameters such as ventricular wall
■■ Differentiation of Cardiac Masses from thickening, ejection fraction, or flow velocity in adjacent
Normal Amatonic Variants vessels. Therefore, the impact of a tumor on cardiac func-
tion can be evaluated.
In clinical practice, MRI and CT are most often used
Goals and Indications to verify or exclude a possible mass suggested initially by
echocardiography. Echocardiography clearly depicts car-
The goals of tomographic imaging in evaluation of cardiac
diac morphology and provides an assessment of functional
masses are the following:
parameters. The effectiveness of transthoracic echocardiog-
1. Identify an intracardiac or extracardiac mass. raphy is limited by the acoustic window, however, which
2. Demonstrate the extent and invasiveness of the mass varies considerably with patient habitus. Image quality of
(staging of tumor). echocardiography may be severely decreased by obesity or
176
LWBK1209-ch13_p176-195.indd 176 5/18/13 12:52 PM

Chapter 13 ■ Cardiac and Paracardiac Masses 177
chronic obstructive pulmonary disease. Transesophageal Collimation is usually 1.25 to 2.5 mm. Retrospective
echocardiography overcomes this problem but adds inva- reconstruction of volumetric data in the sagittal or coro-
siveness. The soft tissue contrast achieved with echocar- nal plane may be useful. Reconstructions in the sagittal
diography remains limited in comparison with that ob- and coronal planes are done routinely using the multislice
tained with MRI and CT. Usually, pericardial involvement axial data usually with slice thickness of 1.25 mm or less.
and infiltration of the myocardium can be better visualized Positron emission tomography/CT can also be used to indi-
with MRI and CT. cate the likely malignant nature of intracardiac and para-
Tissue characterization based on specific T1 and T2 re- cardiac tumors (Fig. 13.2).
laxation times is possible to a limited degree. Nevertheless,
definitive differentiation between benign and malignant tu-
Magnetic Resonance Imaging
mors is sometimes not possible. Most cardiac tumors have
low to intermediate signal intensity on T1-weighted images ECG-gated transaxial T1-weighted spin-echo (SE) images
and high signal intensity on T2-weighted images. However, of the entire thorax are initially acquired for the evaluation
the combination of imaging characteristics of a cardiac mass, of suspected cardiac or paracardiac masses. In addition,
such as location, signal intensity on T1- and T2-weighted such images are frequently acquired in the sagittal or coro-
images, possible hyperenhancement after the administration nal plane to delineate the regions that are displayed subop-
of paramagnetic contrast agents, and possible suppression timally in the transaxial plane, such as the diaphragmatic
of signal with the application of a fat saturation technique, surface of the heart. Coronal images facilitate the evaluation
may render a specific tissue diagnosis highly probable in of masses involving the aortopulmonary window and pul-
some cases. monary hili, and mediastinal masses that extend through the
cervicothoracic junction. The wide field of view afforded by
sagittal and coronal images can readily display the extent
of tumors (Figs. 13.3 and 13.4). Contrast between intra-
Techniques
mural tumor and normal myocardium may be low on non-
enhanced T1-weighted images. Transaxial T2-weighted SE
Computed Tomography
images are acquired to enhance the contrast between myo-
Multislice or spiral single-slice CT scans in the axial plane cardium and tumor tissue, which usually has a longer T2
after contrast enhancement are used to identify and deter- relaxation time, and to delineate possible cystic or necrotic
mine the extent of masses (Fig. 13.1). For this evaluation, components of a mass. The comparison of signal intensities
ECG-gated multislice CT acquisitions are not necessary. of a mass lesion on T1-weighted and T2-weighted images
However, ECG-gated multislice CT is now sometimes may to a certain degree allow for tissue characterization.
used especially for evaluation of an intracardiac mass. For example, lipomas have relatively high signal intensity
A B
Figure 13.1. (A) Contrast-enhanced CT in the axial plane shows a low-density left ventricular mass involving
the apex (arrows). (B) Coronal reconstruction from the CT better demonstrates the relationship of the mass (arrow-
heads) with the left ventricular apex. Note the peripheral rim of high density with some calcifications surround-
ing the lower-density central mass (fibroma). LV, left ventricle; LA, left atrium; RV, right ventricle; Ao, aorta.
LWBK1209-ch13_p176-195.indd 177 16/05/13 8:17 PM

A B
Figure 13.2. A: Axial contrast-enhanced CT image shows a right atrial mass (arrowheads) with a wide
point of attachment to the wall. Note a second mass (arrow) within the right ventricle (RV). B: PET scan in
the axial plane shows avid uptake of FDG in part of the atrial mass (large arrow) and in the ventricular mass
(small arrows) in patient with metastatic thyroid cancer. Note that part of the atrial mass (arrowhead) has no
uptake due to tumor necrosis. LV, left ventricle.
on T1-weighted images and moderate signal intensity on pentaacetic acid) usually improves the contrast between
T2-weighted images. Cystic lesions (filled with simple fluid) tumor tissue and myocardium on T1-weighted images and
have low signal intensity on T1-weighted images and high may facilitate tissue characterization. Hyperenhancement
signal intensity on T2-weighted images (Fig. 13.5). The of tumor tissue with MR contrast agents indicates either an
administration of Gd-DTPA (gadolinium diethylenetriamine enlarged extracellular space of tumor tissue in comparison
Figure 13.3. Angiosarcoma. ECG-gated spin-echo image in the

coronal plane shows a large tumor in the right atrium extending Figure 13.4. Angiosarcoma. ECG-gated spin-echo image in the
through the atrial wall (arrow). The wide field of view of the coronal coronal plane demonstrates a tumor (T) infiltrating through the right
plane demonstrates the extent of this angiosarcoma. atrial cavity and extending around the superior vena cava (arrows).
LWBK1209-ch13_p176-195.indd 178 16/05/13 8:17 PM

A B
Figure 13.5. Pericardial cyst. ECG-gated spin-echo T1-weighted (A) and T2-weighted (B) images of a
pericardial cyst (C). The simple fluid in the cyst has typical low signal on T1-weighted and homogeneous high
signal on T2-weighted images.
with normal myocardium (Fig. 13.6) or a high degree of Location of Cardiac and
vascularization of the mass. Application of a fat saturation Paracardiac Masses
sequence, which vitiates the bright signal of fat, is effective
for the tissue characterization of lipomas (Fig. 13.7). 1. Intracavitary
In patients with cardiac tumors, cine MRI provides valu- 2. Intramural
able information regarding the movement of the cardiac 3. Intrapericardial—outer contour of pericardium with
mass relative to cardiovascular structures. Since cine MR compression of adjacent cardiac chamber
images are acquired with gradient-echo or steady-state free 4. Paracardiac or mediastinal (Fig. 13.8)
precession (SSFP) sequences, a different contrast is obtained
than with the SE technique. On SE images, flowing blood
appears with low signal intensity, whereas gradient-echo or Benign Primary Cardiac Tumors
SSFP images display the blood pool with high signal inten-
sity. Most studies using white blood imaging now use some About 80% of primary cardiac tumors are benign (Table
form of SSFP sequence. 13.1). Although these tumors do not metastasize or invade
A B
Figure 13.6. Angiosarcoma. ECG-gated spin-echo T1-weighted images before (A) and after (B) gadolinium
chelate administration demonstrates hyperenhancement of the tumor (T) compared with the septal myocar-
dium. The postcontrast image uses fat saturation.
LWBK1209-ch13_p176-195.indd 179 16/05/13 8:17 PM

A B
Figure 13.7. Lipoma. ECG-gated spin-echo images in coronal plane, before (A) and after (B) fat satura-
tion, of a mass situated above the left atrium (LA). Signal of the mass is suppressed with fat saturation.
Figure 13.8. Location of masses on tomographic imaging.
LWBK1209-ch13_p176-195.indd 180 16/05/13 8:17 PM

rimary Benign Cardiac

P ypical Sites of Origin
T
Table 13.1 Table 13.2
Tumors of Cardiac Tumors
• Myxoma Tumor Site(s)
• usually attached to atrial septum
• most frequently in LA • Myxoma Left atrium 75%; right atrium 20%
• Lipoma or lipomatous hypertrophy of atrial septum • Lipoma Right atrium; atrial septum
• Papillary fibroelastoma • Papillary fibroelastoma Aortic valve 30%; mitral valve 2%;
• usually attach to valve (aortic valve most frequent) pulmonary valve 13%; tricuspid
• Rhabdomyoma valve 17%
• most common tumor in children • Rhabdomyoma Left and right ventricular
• Fibroma myocardium
• Pheochromocytoma • Fibroma Right ventricular wall and ventricular
• Hemangioma septum
• Pheochromocytoma Peri-left atrium; retroaortic;
aorticopulmonary window
locally, they may lead to significant morbidity and mortal- • Hemangioma Any cardiac chamber
• Angiosarcoma Right atrium, pericardial cavity
ity by causing arrhythmias, valvular obstruction, or embo-
• Rhabdomyosarcoma Ventricular myocardium
lism. An intramyocardial location can interfere with normal • Lymphoma Right atrium
conduction pathways and produce arrhythmias, obstruct
coronary blood flow, or diminish compliance or contractility
through replacement of myocardium. Both benign and malig-
nant tumors have characteristic sites of origin (Table 13.2). myxomas have a wide base of attachment to the atrial septum
(see Fig. 13.9B). However, a wide mural attachment is more
frequently encountered with malignant tumors. The extent of
Myxoma
attachment may be difficult to assess for large tumors, which
Myxoma is the most common benign cardiac tumor (8,9). fill nearly the entire cavity so that they are compressed against
It is located in the left atrium in 75% of cases and in the the septum (Fig. 13.9B). As a result, the tumor appears to have
right atrium in 20% of cases. Very rarely, myxoma occurs broad contact with the atrial septum on static MR images.
in the ventricles. Multiple atrial myxomas may occur rarely, Myxomas can grow through a patent foramen ovale and
especially in Carney’s syndrome (10). Myxomas are read- extend into both atria, a condition that has been described
ily shown by contrast-enhanced CT and MRI (11–15). This as a “dumbbell” appearance. Cine MRI permits an evalua-
tumor is usually spherical, but the shape may vary during tion of tumor motion and may help to identify the site and
the cardiac cycle because of its gelatinous consistency. Left length of attachment of the tumor to the wall or walls of the
atrial myxomas are typically attached by a narrow pedicle cardiac chambers (11,12,15). With this technique, myxo-
to the area of the fossa ovalis (Fig. 13.9A). Infrequently, mas have been shown to prolapse through the funnel of the
A B
Figure 13.9. Myxomas. ECG-gated spin-echo images display two left atrial myxomas with a narrow point
of attachment (pedicle; A) and a wide point of attachment (B) to the left side of the atrial septum.
LWBK1209-ch13_p176-195.indd 181 16/05/13 8:17 PM

A B
Figure 13.10. Myxoma. Cine MR images (balanced steady-state free precession) in the axial plane display
a right atrial myxoma in diastole (A) and systole (B). The motion of the tumor is evident with movement into
the tricuspid valve during diastole.
atrioventricular valve (Fig. 13.10) or into the corresponding intensity of the tumor on cine MR and spin-echo images.
ventricle during diastole. Rarely, myxoma can have a wide Rarely, myxomas have been reported to be invisible on SE
point of attachment to ventricular endocardium (Figs. 13.11 images because of a lack of contrast with the dark blood
and 13.12). pool. Such tumors can be delineated with cine MRI, on
Usually, myxomas display intermediate signal intensity which they appear with high contrast against the surround-
(isointense to the myocardium) on T1-weighted SE images. ing bright blood. Most myxomas show increased signal in-
On T2-weighted SE images, myxomas usually have higher tensity after the administration of Gd-DTPA on T1-weighted
signal intensity than myocardium. However, myxomas with images (5) (Fig. 13.13) and on delayed gadolinium-enhanced
very low signal intensity have also been observed. Fibrous images, which is probably secondary to an increased intersti-
stroma, calcification, and the deposition of paramagnetic tial space; and therefore, a larger distribution volume of the
iron following interstitial hemorrhage can vitiate the signal contrast agent within the tumor than in normal tissue.
A B
Figure 13.11. Left ventricular myxoma. T1-weighted axial spin-echo images before (A) and after (B)
the administration of gadolinium chelate show a tumor (arrow) with a wide point of attachment to the left
ventricular (LV) endocardium. Tumor markedly enhances after contrast media. Fat saturation was used after
gadolinium. RV, right ventricle.
LWBK1209-ch13_p176-195.indd 182 16/05/13 8:17 PM

are encountered most frequently in middle-aged and elderly

adults. Lipomas consist of encapsulated mature adipose cells
and fetal fat cells. The tumor consistency is soft, and lipomas
may grow to a large size without causing symptoms. Lipomas
are typically located in the right atrium (Fig. 13.14) or atrial
septum. Rarely, lipomas may occur in the ventricles (16). They
arise from the endocardial surface and have a broad base of
attachment. Lipomas have the same signal intensity as subcu-
taneous and epicardial fat on all MRI sequences (16,17,20).
Since fat has a short T1 relaxation time, lipomas have high
signal intensity on T1-weighted images, which can be sup-
pressed with fat-saturating pulse sequences (see Fig. 13.14).
Usually, they appear with homogeneous signal intensity but
may have a few thin septations. They do not enhance after the
administration of contrast. On T2-weighted images, lipomas
have intermediate signal intensity.
Lipomatous hypertrophy of the atrial septum is con-
sidered to be an entity distinct from intracavitary lipoma
(17,18,20). Lipomatous hypertrophy of the atrial septum is
more common and is alleged to be a cause of supraventricu-
lar arrhythmias. Lipomatous hypertrophy is defined as a de-
position of fat in the atrial septum around the fossa ovalis
that exceeds 2 cm in transverse diameter. It spares the fossa
Figure 13.12. Contrast-enhanced CT in the axial plane shows the ovalis, a characteristic feature that is clearly delineated with
left ventricular myxoma (arrow) with some enhancement of the central T1-weighted SE images (Fig. 13.15). Lipomatous hypertro-
part of the tumor. LV, left ventricle; RV, right ventricle. phy has the same cellular composition as lipoma but is not
encapsulated and infiltrates through the tissue of the atrial
Lipoma and Lipomatous Hypertrophy septum. It is not a true neoplasm. Fatty tissue may extend
of the Atrial Septum from the septum into both atria to a considerable degree.
Signal intensity on MRI is similar to that of lipomas.
Lipomas are reported to be the second most common benign
cardiac tumor in adults but may actually be the most common.
Papillary Fibroelastoma
If the mass projects into the right atrium, it is called a lipoma,
while lipomatous hypertrophy is abundant fat confined to Papillary fibroelastomas constitute about 10% of benign pri-
the atrial septum (16–20). They may occur at any age but mary cardiac tumors (21–25). These tumors consist of avascu-
A B
Figure 13.13. Myxoma. ECG-gated spin-echo images of a myxoma before (A) and after (B) gadolinium
chelate administration. Tumor (arrows) increases substantially in signal intensity.
LWBK1209-ch13_p176-195.indd 183 16/05/13 8:17 PM

A B
Figure 13.14. Lipoma. ECG-gated spin-echo images of a lipoma (arrows) of the right atrium without (A)
and with (B) fat saturation prepulse sequence. With fat saturation, signal is depleted from the right atrial mass.
lar fronds of connective tissue lined by endothelium. Papillary of valve motion during the cardiac cycle can be excellent. So,
fibroelastomas are attached to the valves by a short pedicle small masses attached to valves can be depicted with cine
in approximately 90% of cases. They usually do not exceed MRI (21,22). In most cases, signal intensity characteristics
1 cm in diameter. Papillary fibroelastomas have been found after the administration of Gd-DTPA are difficult to evalu-
on the aortic (29%), mitral (2%), pulmonary (13%), and ate, so the differential diagnosis between thrombus and tumor
tricuspid (17%) valves. Right-sided tumors usually remain may not be feasible. Cine MRI can be used to assess the effect
asymptomatic. Symptoms associated with fibroelastoma are of valvular tumors on valve function; it demonstrates jet flow
related to embolization of thrombi, which may accumulate caused by either obstruction or regurgitation.
on the tumor. Because of their high content of fibrous tissue,
they have low signal intensity on T2-weighted images. The
Rhabdomyoma
diagnosis of these valvular tumors is challenging because of
their small size, low contrast relative to the blood pool on SE Rhabdomyomas are the most common cardiac tumors in chil-
images, and location on the rapidly moving valves (21,22,24). dren, representing 40% of all cardiac tumors in this age group
On cine MR sequences with improved homogeneity of the (26–28). Thirty to fifty percent of rhabdomyomas occur in
bright blood pool signal (SSFP sequence), visualization patients with tuberous sclerosis. Rhabdomyomas may vary in
A B
Figure 13.15. Lipomatous hypertrophy (infiltration) of the atrial septum. ECG-gated T1-weighted spin-
echo images before (A) and after (B) fat saturation show the atrial septum thickened by fatty infiltration.
Signal of the fat is decreased on the fat-saturation image.
LWBK1209-ch13_p176-195.indd 184 16/05/13 8:17 PM

A B
Figure 13.16. Fibroma. ECG-gated T1-weighted spin-echo images before (A) and after (B) gadolinium
chelate administration show a fibroma (T) of the ventricular septum in a young child. There is hyperenhance-
ment of the periphery (P) of the mass but no enhancement of the center (C).
size and are frequently multiple. They are characterized by an composed of fibroblasts interspersed among collagen fibers.
intramural location and involve equally the left ventricle (LV) It arises within the myocardial walls. Unlike most other pri-
and right ventricle (RV). Small, entirely intramural tumors may mary cardiac tumors, fibromas usually do not display cys-
be difficult to identify. Larger tumors distort the shape of the tic changes, hemorrhage, or focal necrosis, but dystrophic
myocardial wall or may bulge into the cavity. Larger tumors can calcification is common. Fibromas may cause arrhythmias
also distort the epicardial contour of the heart. Rhabdomyomas and have been reported to be associated with sudden death.
may have a signal intensity similar to that of normal myocar- Approximately, 30% of these tumors remain asymptomatic
dium on SE images and may display hyperenhancement after and may be discovered incidentally because of heart mur-
the administration of gadolinium contrast (29). murs, ECG changes, or abnormalities on chest roentgen-
ography. Fibromas occur most often within the septum or
anterior wall of the RV and can reach a large size (29,30)
Fibroma
(Figs. 13.16 and 13.17). On T2-weighted MR images, they
Fibroma is the second most common benign cardiac tumor are characteristically hypointense to the surrounding myo-
in children (26–28). It is a connective tissue tumor that is cardium, which is compatible with the short T2 relaxation
B
A
Figure 13.17. Fibroma. ECG-gated T1-weighted spin-echo transaxial images before (A) and after (B) gado-
linium chelate administration in an infant. The periphery of the huge mass shows hyperenhancement. The center
of the mass shows low intensity on the cine MR image. The mass bulges off the free wall of the right ventricle.
LWBK1209-ch13_p176-195.indd 185 16/05/13 8:17 PM

Figure 13.19. Pheochromocytoma. ECG-gated T1-weighted

image in coronal plane. The mass originates from the aortic body and
is situated above the roof of the left atrium (LA). It has high signal
Figure 13.18. Pheochromocytoma. ECG-gated T1-weighted spin- intensity on T1-weighted images.
echo images show a mass of high signal intensity (M) adjacent to the
left atrium.
time of fibrous tissue. On T1-weighted images, fibromas broad interface with the heart. Cardiac pheochromocyto-
may appear isointense to the myocardium. Fibromas show mas are usually benign. Pheochromocytomas are gener-
hyperenhancement of the periphery of the tumor early ally highly vascularized. The average size ranges from 3 to
after the administration of Gd-DTPA. Administration of 8 cm at diagnosis. Pheochromocytomas are hyperintense to
Gd-DTPA has been effective for demarcating these intra- the myocardium on T2-weighted images and isointense or
mural tumors more clearly from normal myocardium (see hyperintense on T1-weighted images (see Figs. 13.18 and
Figs. 13.16 and 13.17). Hyperenhancement of compressed 13.19). After Gd-DTPA administration, they show strong
myocardium at the margin of the tumor facilitates delinea- signal enhancement because of their high vascularity.
tion of the borders of the tumor. Delayed (15 to 20 minutes Enhancement may be heterogeneous, with central nonen-
after contrast administration) hyperenhancement of the hancing areas, related to tumor necrosis. The combination
entire mass has also been observed on the inversion recov- of imaging findings, clinical symptoms, and biochemical
ery gradient-echo (viability) sequence. evidence of catecholamine overproduction usually permits
The differential diagnosis for intramural masses in chil- a confident diagnosis.
dren is rhabdomyoma versus fibroma. If the tumor is soli-
tary and has low signal intensity on T2-weighted images,
Hemangioma
fibroma is more likely. If multiple tumors are present with
high intensity on T2-weighted images, rhabdomyomas are Cardiac hemangiomas are composed of endothelial cells that
the likely diagnosis. line interconnecting vascular channels. These vascular cavi-
ties are separated by connective tissue. According to the size
of the vascular channels, hemangiomas are divided into cap-
Pheochromocytoma
illary, cavernous, or venous types. Calcification, which can
Pheochromocytomas arise from neuroendocrine cells clus- easily be identified on CT, is often present in these tumors.
tered in the visceral paraganglia in the wall of the left atrium, Hemangiomas may involve the endocardium, myocardium,
roof of the right and left atrium, atrial septum, behind the or epicardium. They have been found in all chambers and
ascending aorta, in the aorticopulmonary window, and also the pericardium (33–36). On T2-weighted images,
along the coronary arteries (31,32). Hypertension, the most hemangiomas are hyperintense. On T1-weighted images,
common symptom, is related to catecholamine overpro- they are of intermediate signal intensity but can have higher
duction by the mass. The average age at diagnosis is 30 to intensity than myocardium (33–36) (Fig. 13.20). Because of
50 years. interspersed calcifications and possible flow voids at areas of
Pheochromocytomas can be found at each of the above blood flow in the channels of hemangiomas, they may have
locations but are predominantly encountered in and inhomogeneous signal intensity. They usually show intense
around the left atrium (Figs. 13.18 and 13.19). Most are enhancement after the administration of gadolinium con-
located outside the cardiac chamber. They usually have a trast because of their rich vascularity.
LWBK1209-ch13_p176-195.indd 186 16/05/13 8:17 PM

Figure 13.20. Hemangioma.

T1-weighted (top left), T2-weighted fat
saturation (top right), and T1-weighted
after gadolinium chelate administration
(bottom left) spin-echo images and a
gradient-echo cine MR image (bottom
right) demonstrate a large mass (M)
originating in the septum and bulging
into the right ventricular cavity. The
mass has high signal on all sequences.
Malignant Primary Cardiac combined with pericardial effusion. Cardiac tamponade,

Tumors as a consequence of hemorrhagic pericardial effusion, may
be demonstrated as diastolic indentation of the right atrial
One-fourth of primary cardiac tumors are malignant; sarco- free wall. Features of malignant tumors are summarized in
mas represent the largest number, followed by primary car- Table 13.4.
diac lymphomas. The types of malignant cardiac tumor are Extension into the mediastinum and metastasis are also
indicated in Table 13.3. The features of malignant cardiac clear signs of malignancy. The organs most frequently in-
tumors are the following: Involvement of more than one car- volved are the lungs, pleurae, mediastinal lymph nodes,
diac chamber; extension into pulmonary veins, pulmonary and liver. The rapid growth of malignant cardiac tumors
arteries, or vena cavae; wide point of attachment to the wall may cause focal necrosis in the central part of the tumor.
of a chamber or chambers; necrosis within the tumor; exten- Necrotic areas are delineated as regions of lower signal in-
sion outside the heart; and hemorrhagic pericardial effusion tensity within a hyperenhancing mass after the administra-
(37). A combined intramural and intracavitary location is tion of Gd-DTPA (Fig. 13.23).
another suggestive feature of malignant tumors (Fig. 13.21).
MRI is effective for demonstrating invasion of the pericar- Angiosarcoma
dium and extension into the pericardial fat (Figs. 13.3 and
13.22). Pericardial infiltration is displayed on MRI as a dis- Angiosarcomas are the most common malignant cardiac
ruption, thickening, or nodularity of the pericardium, often tumors in adults and constitute one-third of malignant
I maging Features of
Table 13.4
Malignant Cardiac Tumors
rimary Malignant Cardiac
P
Table 13.3 1. Involvement of more than one cardiac chamber or chamber and
Tumors
attaching vessel.
• Angiosarcoma (most frequent) 2. Extension into pulmonary veins, vena cavae, pulmonary artery,
• Rhabdomyosarcoma or aorta.
• Leiomyosarcoma 3. Wide point of attachment to wall of cardiac chamber(s).
• Fibrosarcoma 4. Combined intramural and intracavitary components.
• Liposarcoma 5. Necrosis within tumor.
• Malignant histiosarcoma 6. Extension outside a cardiac chamber.
• Osteochrondrosarcoma 7. Hemorrhagic pericardial effusion.
• Lymphoma 8. Pulmonary metastasis(es).
LWBK1209-ch13_p176-195.indd 187 16/05/13 8:17 PM

A B
Figure 13.21. Lymphoma. Coronal spin-echo (A) and gradient-echo (cine MR; B) images demonstrate a
large mass (M), which involves the lateral wall of the left ventricle and extends into the left ventricular (LV)
and atrial (LA) cavities.
c ardiac tumors (37–41) (Figs. 13.3, 13.4, 13.24, and 13.25). is associated with the acquired immunodeficiency syndrome.
They occur predominantly in men between 20 and 50 years Angiosarcomas consist of ill-defined anastomotic vascu-
of age. This entity has been divided into two clinicopatho- lar spaces that are lined by endothelial cells and avascular
logic forms. Most frequently, angiosarcomas are found in clusters of moderately pleomorphic spindle cells surrounded
the right atrium (see Figs. 13.3, 13.4, 13.24, and 13.25). In by collagen stroma. T1-weighted SE imaging usually dem-
this form, no evidence of Kaposi sarcoma is found. Another onstrates heterogeneous signal intensity of the tumor with
form is characterized by involvement of the epicardium or focal areas of high signal intensity, which presumably repre-
pericardium in the presence of Kaposi sarcoma. This form sent hemorrhage (see Fig. 13.24). However, angiosarcomas
A B
Figure 13.22. Rhabdomyosarcoma. Coronal (A) and transaxial after gadolinium chelate administra-
tion (B) spin-echo images. Coronal images show a loculated pericardial effusion (E) with high intensity on
T1-weighted imaging (hemorrhagic effusion). The epicardial fat line (arrows) is disrupted by the tumor extend-
ing into the pericardial space. Transaxial image after contrast administration shows the tumor (T) demon-
strated by the marked enhancement, whereas the pericardial fluid is not enhanced.
LWBK1209-ch13_p176-195.indd 188 16/05/13 8:17 PM

Figure 13.24. Angiosarcoma. ECG-gated spin-echo image

demonstrates a mass filling most of the right atrium and extending
through the atrial wall. The tumor has heterogeneous intensity on this
T1-weighted image. High signal region (arrow) represents intratu-
Figure 13.23. Left ventricular myocardial sarcoma. Sagittal image moral hemorrhage.
after gadolinium chelate administration demonstrates enhancement of
the mass (arrows) involving the diaphragmatic wall of the left ventricle
(LV). The central necrotic region does not enhance. There is a large
loculated pericardial effusion below the diaphragmatic wall. Rhabdomyosarcoma
Rhabdomyosarcomas are the most common malignant car-
can also have homogeneous signal intensity. After contrast diac tumors in children. They can arise anywhere in the
administration, angiosarcomas show hyperenhancement. myocardium and are often multiple. Their signal intensity
Some of the tumors show regions of low signal intensity on on MRI is variable (37,41). Rhabdomyosarcomas may be
both T1- and T2-weighted images. These central regions isointense to the myocardium on T1- and T2-weighted
have high signal intensity on cine gradient-echo images and images, but areas of necrosis can exhibit heterogeneous sig-
represent vascular channels (see Fig. 13.25). This finding is nal intensity and patchy hyperenhancement after Gd-DTPA
often described as a “cauliflower” appearance. Cases with administration (see Fig. 13.22). Extracardiac extension
diffuse pericardial infiltration have been found to show lin- into the pulmonary arteries and descending aorta has been
ear hyperenhancement along vascular spaces. clearly delineated with MRI.
A B
Figure 13.25. Angiosarcoma. T1-weighted spin-echo (A) and gradient-echo (B) images show components
of the mass along the posterior right atrial wall and in the pericardial cavity. Note the sinuous regions in the
pericardial mass, which represent vascular channels. These channels have high signal because of blood flow on
the gradient-echo images (arrows). RA, right atrium; RV, right ventricle.
LWBK1209-ch13_p176-195.indd 189 16/05/13 8:17 PM

A B
Figure 13.26. Lymphoma. T1-weighted spin-echo image with fat saturation after gadolinium chelate
administration at the level of the right atrium (A) and the right ventricle (B). A mass (M) invades the right
atrial cavity and right ventricular wall. RA, right atrium; RV, right ventricle.
Other Sarcomas (Fibrosarcoma, Secondary Cardiac Tumors

Osteosarcoma, Leiomyosarcoma,
Liposarcoma) Secondary tumors of the heart and pericardium are about
Other possible primary sarcomas are fibrosarcomas, osteo- 40-fold more frequent than primary tumors (48). Three
sarcomas, leiomyosarcomas, and liposarcomas (36,40). routes of spread to the heart can be discerned: (1) Direct
These all are rare tumors. The signal intensity characteris- extension from intrathoracic tumors (mediastinum and
tics of these entities are nonspecific. Most of these tumors lungs); (2) extension of abdominal malignancies through the
show signal intensity isointense to normal myocardium inferior vena cava into the right atrium (renal, adrenal, and
on T1-weighted images and hyperintense on T2-weighted hepatic carcinomas); and (3) metastasis (Table 13.5).
images. Most of these tumors show increased signal inten-
sity on T1-weighted images after Gd-DTPA administration, Direct Extension from Adjacent Tumors
so that the lesions are more conspicuous and the delineation
of the tumor margins is increased. Tumors of the lung and mediastinum can infiltrate the peri-
cardium and heart directly (Figs. 13.27 and 13.28). It is
important to recognize invasion of the heart because such
Lymphoma a tumor is usually nonresectable. In mediastinal lymphoma,
Primary cardiac lymphoma is less common than secondary possible invasion of the pericardium can change the staging
lymphoma involving the heart, which usually represents the of the tumor. CT and MRI are effective for delineating para-
spread of non-Hodgkin’s lymphoma (42–47). Primary lym- cardiac tumors and possible extension into the heart because
phoma of the heart most often occurs in immunocompro- of its wide field of view. CT and MRI clearly show extension
mised patients and is highly aggressive. Almost all primary
cardiac lymphomas are B-cell lymphomas. Although primary Table 13.5 Secondary Cardiac Tumors
cardiac lymphoma is rare, it is mandatory to consider this
entity in the diagnosis of malignant cardiac tumors because 1. Direct extension from primary lung and mediastinal malignancy
early chemotherapy is usually effective. These tumors arise • lung cancer
• malignant thymoma
most often on the right side of the heart, especially in the right
• malignant mesothelioma
atrium, but have also been found in the other chambers (see 2. Inferior vena caval extension of abdominal malignancy
Fig. 13.21). A large pericardial effusion is frequently present. • hepatocellular carcinoma
Variable morphology of the masses has been described; both • renal cell carcinoma
circumscribed polypoid and ill-defined infiltrative lesions have • adrenal carcinoma
been reported. Lymphomas may appear hypointense to the 3. Metastasis
myocardium on T1-weighted images and hyperintense on • lung cancer
T2-weighted images. After Gd-DTPA administration, homo- • breast cancer
geneous or heterogeneous enhancement of the tumor, depend- • malignant melanoma
ing on the presence of necrosis, may be seen. Some lymphomas • lymphoma, leukemia
• sarcoma
consist of a large extracardiac mediastinal mass as well as an
• other
invasive mass in the cardiac chambers (Fig. 13.26).
LWBK1209-ch13_p176-195.indd 190 16/05/13 8:17 PM

Figure 13.27. Mediastinal tumor (T) invading the pericardium

and atrial walls. LA, left atrium; RA, right atrium. Figure 13.29. Metastasis to the left atrium. ECG-gated
T1-weighted spin-echo transaxial images.
of these tumors to the cardiac structures (Figs. 13.27, and at autopsy. The mechanism of metastatic spread of tumors
13.28) and possible evidence of hemorrhagic or nonhemor- to the heart is direct seeding to the endocardium, passage of
rhagic pericardial effusion. These modalities are effective in tumor emboli through the coronary arteries, or retrograde
demonstrating invasion of the pericardium and myocardium lymphatic spread through bronchomediastinal lymphatic
in advanced lung cancer. channels. MRI is highly effective for delineating the location
and extent of metastatic tumors in cardiac chambers (Figs.
13.29 and 13.30) and assessing potential respectability.
Metastasis
Melanomas, leukemias, and lymphomas (see Fig. 13.26) are
the tumors that most frequently metastasize to the heart,
but cardiac metastases can arise from almost any malignant
tumor in the body. Melanomas have the highest frequency of
seeding into the heart and are frequently found in the heart
Figure 13.30. Metastasis to right atrial appendage. ECG-gated

T1-weighted spin-echo sagittal image shows a large mass (M) filling
Figure 13.28. Extension of lung cancer through left atrial wall. and expanding the right atrial appendage.
LWBK1209-ch13_p176-195.indd 191 16/05/13 8:17 PM

Figure 13.31. Renal cell carcinoma. Direct extension of mass (M)

into the right atrium via inferior vena cava. CS, coronary sinus; RA,
right atrium.
Figure 13.32. Left ventricular thrombus after myocardial infarc-
Transvenous Extension into the Heart tion. Transaxial gradient-echo (cine MR) image shows a low-intensity
intracavitary mass (arrow).
Another site for the entry of secondary tumors into the heart
is tumor extension through the large veins connecting with
cardiac chambers. Tumor thrombus arising from carcinoma Intracardiac Thrombus
of the kidney, liver, or adrenal gland can extend through the
inferior vena cava into the right atrium (Fig. 13.31), and pri- Thrombus is the most common intracardiac mass, involv-
mary carcinoma of the lung and thymus can extend through ing most frequently the LV or left atrium. Atrial thrombus
the superior vena cava. Lung cancer can extend into the left is encountered in patients with mitral valve disease or atrial
atrium through a pulmonary vein. The evaluation of the fibrillation. LV mural thrombus is associated with akinetic
possible attachment of such tumors to the atrial wall is man- or dyskinetic regions of the ventricle. It is most often located
datory for surgical planning. If the atrial walls are not infil- in the LV at the site of myocardial infarction (Figs. 13.32–
trated, complete resection of the tumor may still be possible. 13.34) or at the apex in dilated cardiomyopathy. However,
A B
Figure 13.33. A: Cine MR image in the vertical long-axis plane shows apical wall thinning and mural
thrombus (arrow). B: Delayed-enhancement image in the same plane shows transmural hyperenhancement
(arrowheads) consistent with prior myocardial infarction and a low-intensity mural thrombus (arrow). LA,
left atrium; LA, left ventricle. Compliments of H. Sakuma, Mie, Japan.
LWBK1209-ch13_p176-195.indd 192 16/05/13 8:17 PM

composition of the thrombus. The signal intensity of throm-

bus can vary with time since it is influenced by paramag-
netic hemoglobin breakdown products, such as intracellu-
lar methemoglobin and hemosiderin, or superparamagnetic
substances, such as ferritin. Fresh thrombus usually shows
high signal intensity on T1- and T2-weighted SE images,
whereas older thrombus has low signal intensity on T1- and
T2-weighted images. Intracavitary high signal on SE images
caused by slowly flowing blood may be difficult to distin-
guish from thrombus. However, this problem can be over-
come by using cine MRI.
Differentiation between Tumor

and Blood Clot
The distinction between clot and tumor is usually made
with gradient-echo sequences (49–51). The gradient-echo
technique is more sensitive to susceptibility and T2* effects
than is the SE technique (Table 13.6). As the various blood
Figure 13.34. Delayed-enhancement image in the axial plane degeneration products pass through the different stages of
shows a low signal intensity multilayered organized thrombus in the magnetic susceptibility, they cause shortening of T2* relax-
left ventricular apex (arrow). Note the transmural area of delayed
ivity. The result is low signal intensity of the thrombus on
enhancement in the apical and apical–septal regions (arrowheads).
cine MR images (see Figs. 13.32, 13.33, and 13.35). An
exception to this generalization is fresh thrombus, which
any region of the ventricular cavity with static blood is prone can have high signal intensity. Tumor tissue usually is
to thrombus formation. CT and MRI are especially advanta- hyperintense in comparison with myocardium and skeletal
geous for detecting thrombus in the left atrial appendage, muscle on T2-weighted SE images and cine MR images.
which may be difficult to assess using transthoracic echocar- However, some myxomas containing substantial iron pro-
diography (49,50) (Figs. 13.35 and 13.36). duce low signal and so mimic thrombus (50). Another
On SE images, the signal intensity of thrombus can vary method for differentiating between tumor and clot is to use
from low to high depending on age-related changes in the Gd-DTPA–enhanced T1-weighted images. Thrombus does
A B
Figure 13.35. Left atrial appendage thrombus (A) and metastatic tumor of the right ventricle (B).
Transaxial gradient-echo (cine MR) image shows a mass (M) filling the left atrial appendage with low
intensity. The mass (M) in the right ventricle has intermediate signal intensity.
LWBK1209-ch13_p176-195.indd 193 16/05/13 8:17 PM

such as a prominent crista terminalis, eustachian valve, or

Chiari network. The crista terminalis, a fibromuscular band
extending between the ostia of the superior and inferior
vena cavae on the posterior right atrial wall, represents a
residuum of the septum spurium where the sinus venosus
was incorporated into the right atrial wall. The Chiari net-
work, a reticulum situated in the right atrium, is attached to
the region of the crista terminalis and extends to the valves
of the inferior vena cava and coronary sinus, or sometimes
to the floor of the right atrium near the ostium of the coro-
nary sinus. The Chiari network is derived from the valvulae
venosi. These structures regress to variable degrees, and a
nodule-like residue in the right atrium is visible on MRI in
some patients. Awareness of these variants can prevent mis-
interpretation as mass lesions.
References
1. Link KM, Lesko NM. MR evaluation of cardiac/juxtacardiac masses. Top Magn Reson
Imaging. 1995;7:232–245.
2. Kaminaga T, Takeshita T, Kimar I. Role of magnetic resonance imaging for evaluation of
tumors in the cardiac region. Eur Radiol. 2003;13(suppl 4):L1–L10.
3. Hoffman U, Globits S, Schima W, et al. Usefulness of magnetic resonance imaging of car-
diac and paracardiac masses. Am J Cardiol. 2003;92:890–895.
Figure 13.36. Axial image from a contrast-enhanced CT shows a 4. Fujita N, Caputo GR, Higgins CB. Diagnosis and characterization of intracardiac masses
filling defect within the left atrial appendage (arrow), with lobulated by magnetic resonance imaging. Am J Card Imaging. 1994;8:69–80.
contour, consistent with a thrombus. Note the enlargement of the left 5. Funari M, Fujita N, Peck WW, et al. Cardiac tumors: assessment with Gd-DTPA enhanced
atrium (LA) in this patient with atrial fibrillation. MR imaging. J Comput Assist Tomogr. 1991;15:953–958.
6. Araoz PA, Mulvagh SL, Tazelaar HD, et al. CT and MR imaging of benign primary cardiac
neoplasms with echocardiographic correlation. Radiographics. 2000;20:1303–1319.
7. Semelka RC, Shoenut JP, Wilson ME, et al. Cardiac masses: Signal intensity features on
not enhance after the administration of Gd-DTPA, whereas spin-echo, gradient-echo, gadolinium-enhanced spin-echo, and TurboFLASH images. J
tumors show enhancement (Fig. 13.13). Tumor can usually Magn Reson Imaging. 1992;2:415–420.
8. MacGowan SW, Sidhu P, Aherne T, et al. Atrial myxoma: National incidence, diagnosis
be differentiated from thrombus by using cine MR images and surgical management. Ir J Med Sci. 1993;162:223–226.
and T1-weighted SE images after Gd-DTPA administration. 9. Reynen, K. Cardiac myxomas. N Engl J Med. 1995;333(24):1610–1617.
Another approach is using the delayed gadolinium-enhanced 10. Carney JA. Carney complex: the complex of myxomas, spotty pigmentation, endocrine
overactivity, and schwannomas. Semin Dermatol. 1995:14,90–98.
sequence (52). At an inversion delay that nulls myocardial 11. Grebenc ML, Rosado-de-Christenson ML, Green CE, et al. Cardiac myxoma: imaging fea-
signal, the thrombus has even lower signal than myocardium tures in 83 patients. Radiographics. 2002:22:673–689.
(Figs. 13.33 and 13.34). 12. Sievers B, Fritzsche D, Bias-Franken R, et al. Magnetic resonance imaging in the detection
of a large left atrial myxoma. Am J Med. 2003;115:155–157.
13. Roberts-Thomson KC, Teo KS, Stuklis R, et al. Left atrial myxoma: magnet or echo? Intern
Med J. 2004;34:210–211.
Diferentiation of Cardiac Masses 14. Masui T, Takahashi M, Miura K, et al. Cardiac myxoma: identification of intratumoral
hemorrhage and calcification on MR images. AJR Am J Roentgenol. 1995;164:850–852.
from Normal Anatomic Variants 15. Matsuoka H, Hamada M, Honda T, et al. Morphologic and histologic characterization of
cardiac myxomas by magnetic resonance imaging. Angiology. 1996;47:693–698.
16. Oyama N, Oyama N, Komatsu H, et al. Images in cardiovascular medicine. Left ventricu-
A potential pitfall in the evaluation of intracardiac masses lar asynchrony caused by intramuscular lipoma: Computed tomographic and magnetic
may arise from the misdiagnosis of normal anatomic variants, resonance detection. Circulation. 2003;107:e200–e201.
17. Salanitri JC, Pereles FS. Cardiac lipoma and lipomatous hypertrophy of the interatrial septum:
cardiac magnetic resonance imaging findings. J Comput Assist Tomogr. 2004;28:852–856.
18. Heyer CM, Kagel T, Lemburg SP, et al. Lipomatous hypertrophy of the interatrial sep-
tum: a prospective study of incidence, imaging findings, and clinical symptoms. Chest.
istinguishing MR
D 2003;124:2068–2073.
Table 13.6 Imaging Features of 19. Gaerte SC, Meyer CA, Winer-Muram HT, et al. Fat-containing lesions of the chest.
Radiographics. 2002;22(spec no):S61–S78.
Tumors Versus Clot 20. Hananouchi GI, Goff WB. Cardiac lipoma: six-year follow-up with MR characteristics,
and a review of the literature. Magn Reson Imaging. 1990;8:825–828.
TUMOR 21. Jahnke C, Hamdan A, Fleck E, et al. Tissue characterization of a suspected aortic valve fibro-
elastoma with cardiac magnetic resonance imaging. Circ Cardiovasc Imaging. 2008;1:87–88.
• Medium signal on T1 and high signal on T2 SE
22. de Arenaza DP, Pietrani M, Moon JC, et al. Cardiac fibroelastoma: cardiovascular mag-
• Early enhancement after gadolinium chelate on T1 SE netic resonance characteristics. J Cardiovasc Magn Reson. 2007;9:621.
• Shorter T1 nulling time than myocardium on delayed 23. Gowda RM, Khan IA, Nair CK, et al. Cardiac papillary fibroelastoma: a comprehensive
gadolinium-enhanced sequence (IRGRE) analysis of 725 cases. Am Heart J. 2003;146:404–410.
24. Shirashi J, Tagawa M, Yamada T, et al. Papillary fibroelastoma of the aortic valve:
• Higher signal than myocardium on cine GRE or cine SSFP
Evaluation with transesophageal echocardiography and magnetic resonance imaging. Jpn
CLOT Heart J. 2003;44:799–803.
25. Wintersperger BJ, Becker CR, Gulbins H, et al. Tumors of the cardiac valves: Imaging
• Variable signal on SE depending on age
findings in magnetic resonance imaging, electron beam computed tomography, and echo-
• No early enhancement after gadolinium chelate on T1 SE cardiography. Eur Radiol. 2003;10:443–449.
• Longer T1 nulling time than myocardium on delayed 26. Rienmuller R, Lloret JL, Tiling R, et al. MR imaging of pediatric cardiac tumors previously
gadolinium-enhanced sequence diagnosed by echocardiography. J Comput Assist Tomogr. 1989;13:621–626.
27. Abushaban L, Denham B, Duff D. 10 year review of cardiac tumors in childhood. Br Heart
• Lower signal than myocardium or cine GRE and cine SSFP
J. 1993;70:166–169.
LWBK1209-ch13_p176-195.indd 194 16/05/13 8:17 PM

28. Beghetti M, Gow RM, Haney I, et al. Pediatric primary benign cardiac tumors: a 15-year 41. Araoz PA, Eklund H, Welch TJ, et al. CT and MR imaging of primary cardiac malignan-
review. Am Heart J. 1997;134:1107–1114. cies. Radiographics. 1999;19:1421–1434.
29. Kiaffas MG, Powell AJ, Geva T. Magnetic resonance imaging evaluation of cardiac tumor 42. Ceresoli GL, Ferreri AJ, Bucci E, et al. Primary cardiac lymphoma in immunocompetent
characteristics in infants and children. Am J Cardiol. 2002;89:1229–1233. patients: diagnostic and therapeutic management. Cancer. 1997;80:1497–1506.
30. Brechtel K, Reddy GP, Higgins CB. Cardiac fibroma in an infant: magnetic resonance imag- 43. Anghel G, Zoli V, Petti N, et al. Primary cardiac lymphoma: report of two cases occurring
ing characteristics. J Cardiovasc Magn Reson. 1999;1:159–161. in immunocompetent subjects. Leuk Lymphoma. 2004;45:781–788.
31. Hamilton BH, Francis IR, Gross BH, et al. Intrapericardial paragangliomas (pheochromo- 44. Tada H, Asazuma K, Ohya E, et al. Images in cardiovascular medicine. Primary cardiac
cytomas): imaging features. Am J Roentgenol. 1997;168:109–113. B-cell lymphoma. Circulation. 1998;97:220–221.
32. Sahdev A, Sohaib A, Monson JP, et al. CT and MR imaging of unusual locations of extra- 45. Dorsay TA, Ho VB, Rovira MJ, et al. Primary cardiac lymphoma: CT and MR findings.
adrenal paragangliomas (pheochromocytomas). Eur Radiol. 2005;15:85–92. J Comput Assist Tomogr. 1993;17:978–981.
33. Brodwater B, Erasmus J, McAdams HP, et al. Case report. Pericardial hemangioma. 46. Hoffmann U, Globits S, Frank H. Cardiac and paracardiac masses. Current opinion on
J Comput Assist Tomogr. 1996;20:954–956. diagnostic evaluation by magnetic resonance imaging. Eur Heart J. 1998;19:553–563.
34. Oshima H, Hara M, Kono T, et al. Cardiac hemangioma of the left atrial appendage: CT 47. Kubo S, Tadamura E, Yamamuro M, et al. Primary cardiac lymphoma demonstrated
and MR findings. J Thorac Imaging. 2003;18:204–206. by delayed contrast-enhanced magnetic resonance imaging. J Comput Assist Tomogr.
35. Seline TH, Gross BH, Francis IR. CT and MR imaging of mediastinal hemangiomas. 2004;28:849–851.
J Comput Assist Tomogr. 1990;14:766–768. 48. Hanfling SM. Metastatic cancer to the heart. Review of the literature and report of 127
36. Kemp JL, Kessler RM, Raizada V, et al. Case report. MR and CT appearance of cardiac cases. Circulation. 1960;22:474–483.
hemangioma. J Comput Assist Tomogr. 1996;20:482–483. 49. Jungehulsing M, Sechtem U, Theissen P, et al. Left ventricular thrombi: evaluation with
37. Siripornpitak S, Higgins CB. MRI of primary malignant cardiovascular tumors. J Comput spin-echo and gradient-echo MR imaging. Radiology. 1992;182:225–229.
Assist Tomogr. 1997;21:462–466. 50. Seelos KC, Caputo GR, Carrol CL, et al. Cine gradient refocused echo (GRE) imaging of
38. Bruna J, Lockwood M. Primary heart angiosarcoma detected by computed tomography intravascular masses: differentiation between tumor and nontumor thrombus. J Comput
and magnetic resonance imaging. Eur Radiol. 1998;8:66–68. Assist Tomogr. 1992;16:169–175.
39. Villacampa VM, Villarreal M, Ros LH, et al. Cardiac rhabdomyosarcoma: diagnosis by 51. Barkhausen J, Hunold P, Eggebrecht H, et al. Detection and characterization of intracar-
MR imaging. Eur Radiol. 1999;9:634–637. diac thrombi on MR imaging. AJR Am J Roentgenol. 2002;179:1539–1544.
40. Watanabe AT, Teitelbaum GP, Henderson RW, et al. Magnetic resonance imaging of car- 52. Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging technique for the visual-
diac sarcomas. J Thorac Imaging. 1989;4:90–92. ization of myocardial infarction. Radiology. 2001;218:215–223.
This chapter is modified from Chapter 33 of the Second Edition of Webb, W.R. and Higgins,
C.B. Thoracic Imaging.
LWBK1209-ch13_p176-195.indd 195 16/05/13 8:17 PM

CHAPTER
14 Jan Bogaert
Andrew M. Taylor
Valvular Heart Disease
■■ Introduction progress and to help in determining the optimal timing for

surgical intervention. In contrast with other cardiac diseases,
■■ MRI Strategies to Evaluate Valvular Heart such as assessment of myocardial infarction and viability,
Disease where magnetic resonance imaging (MRI) can be consid-
Valvular Morphology ered as the reference technique, the role of MRI in assessing
Valvular Regurgitation patients with VHD is currently not well established. This is
Valvular Stenosis in part due to the availability of a good and well-accepted,
Associated Findings in VHD noninvasive first-line imaging modality, that is, echocardiog-
Assessment of Non-VHD Cardiac Diseases raphy. Moreover, clinicians are still unaware of the potential
clinical applications of MRI for assessing cardiac diseases
■■ Aortic Valve
(2,3). In this chapter, we focus on the different approaches
Aortic Stenosis to study VHD with MRI in the clinical setting.
Aortic Regurgitation
■■ Mitral Valve
Mitral Stenosis MRI Strategies to Evaluate
Mitral Regurgitation Valvular Heart Disease
Mitral Valve Prolapse
The heart has two atrioventricular and two ventriculoarterial
■■ Pulmonary Valve valves which guarantee a correct, unidirectional flow of blood
■■ Tricuspid Valve through the heart. This is achieved by alternated phases of valve
opening and valve closure. Normal valves are not flow-restrictive
■■ Mixed Valvular Heart Disease when opened and prevent backflow of blood when closed. A
■■ Prosthetic Valves small, physiologic regurgitation, however, may be present in
healthy subjects (usually the tricuspid and pulmonary valve,
■■ Future Directions
less frequently the mitral valve). In the diseased state, the valve
■■ Conclusion orifice may become narrowed, leading to an obstruction of the
blood flow (valvular stenosis), and/or the valve leaflets may not
completely coapt leading to a pathologic regurgitation of blood
Introduction into the proximal chamber (valvular regurgitation). Valvular
stenosis/regurgitation may be caused by diseases directly affect-
Although the incidence of valvular heart disease (VHD) in ing the valves (e.g., endocarditis, degenerative disease) or indi-
the developed countries is relatively low compared with the rectly by adjacent pathology or pathology at a distance (e.g.,
incidence of ischemic heart disease, it continues to cause con- dilated cardiomyopathy, papillary muscle infarction, Marfan’s
siderable morbidity and mortality, and in many patients sur- disease). Valve lesions lead to acute/chronic alterations in pre-
gical valve repair or replacement will be indicated (1). The load and/or afterload, with remodeling of the involved cardiac
diagnosis of VHD is usually based on the patients’ symp- chamber. When these initially compensatory responses, that
toms (e.g., dyspnea on exertion, exertional syncope, angina, are meant to preserve normal cardiac function, are severe and/
shortness of breath, palpitations, fatigue, heart failure) and or longstanding, the compensatory phenomena may fail and
abnormal cardiac auscultation (e.g., diastolic or systolic lead to a state of cardiac decompensation.
murmur). The role of additional imaging procedures, usu- The role of imaging in patients with suspected VHD is to
ally echocardiography with Doppler flow measurements and obtain information on the following issues: (a) Definition
less frequently cardiac catheterization, is to assess the sever- of valvular anatomy—number of valve leaflets, leaflet
ity of VHD and to evaluate its impact on cardiac function. thickness, presence of infective endocarditis; (b) assess-
Cardiac imaging is also indicated to monitor the disease ment of valvular function—degree of valvular stenosis or
196
LWBK1209-ch14_p196-218.indd 196 17/05/13 4:58 PM

Chapter 14 ■ Valvular Heart Disease 197
r egurgitation; (c) definition of the consequences of valvular Valvular Morphology

dysfunction on ventricular/atrial size, function and mass, and
Although fast spin-echo MRI techniques with dark blood
pulmonary artery pressure; (d) assessment of associated mor-
prepulses provide high-quality morphologic images of the
phologic abnormalities—great vessel anatomy or thrombus
heart, they are seldom of diagnostic quality to evaluate the
formation; and (e) exclusion of non-VHD–related cardiac
valvular anatomy (4). This is mainly related to the lack of
pathology—coronary artery disease and/or prior myocardial
contrast between the low-proton density of valve leaflets and
infarction. Much of the above required information can be
the surrounding dark blood (Fig. 14.1). With bright blood
obtained with a single investigation that is safe, noninvasive
MRI, using either the spoiled gradient-echo (GE) or balanced
and without exposure to x-rays: Cardiovascular MRI (2,3).
A B
C D
Figure 14.1. Valvular morphology on dark blood (A,B) and bright blood MRI (C,D). A: Coronal fast
spin-echo MRI of the aortic valve (arrows). B: Fast spin-echo MRI of the mitral (arrows) and tricuspid valve
(arrowheads) in the horizontal long-axis plane. The valve leaflets appear as low-signal intensity structures
surrounded by the dark blood. Vertical long-axis bright blood MRI using the balanced-SSFP technique at end-
systole (C) and during early ventricular filling (D). The leaflets of the mitral valve (arrows) and the tendinous
chords (arrowheads) are well visible. Also the mobility of the leaflets (black arrows) can be well appreciated.
LA, left atrium; LV, left ventricle.
LWBK1209-ch14_p196-218.indd 197 17/05/13 4:58 PM

Figure 14.2. Pseudoaneurysm

formation in a 56-year-old patient
with renal transplant and history of
endocarditis. Balanced-SSFP cine MRI
in axial plane (A) and vertical long-axis
through aortic root (B). The infectious
pseudoaneurysm is visible as a well-
defined hyperintense structure (asterisk)
behind the aortic root (Ao). MRI is
indicated to detect and follow up the
B formation of these pseudoaneurysms in
A
patients with endocarditis.
steady-state free precession (b-SSFP) techniques, valve leaf- Qualitative Assessment of Signal Loss with Cine MRI
lets are visible as low-signal intensity structures surrounded
by the bright blood (Fig. 14.1). Moreover, their depiction is The bright blood phenomenon on spoiled GE cine MRI tech-
greatly enhanced by using the cine-viewing mode, allowing niques is based on the laminar inflow of fully magnetized
for better visualization of the number of valve leaflets, leaf- spins. Intravoxel dephasing of the proton spins, due to flow
let morphology and integrity, as well as leaflet mobility and disturbance (turbulence, acceleration) leads to loss in signal.
motion. On the new b-SSFP techniques, combining a good This principle can be applied to assess valvular regurgita-
spatial with an excellent contrast resolution, the subvalvu- tion. Imaging over multiple frames throughout the cardiac
lar apparatus (e.g., tendinous chords) of the atrioventricular cycle enables visual assessment of flow disturbance caused
valves and papillary muscles are well depicted (Fig. 14.1). by the regurgitation. Depending on the severity a dark signal
Although echocardiography remains the investigation of is visible extending from the diseased valve into the receiv-
choice for imaging valve leaflet anatomy, MRI can be helping chamber, either pencil-beam–like or fan-like, during a
ful for unraveling valvular anatomy in patients with poor variable time period of the cardiac cycle (Fig. 14.3). Using
subjects for echocardiography, for example, due to heavily a combination of contiguous cine MRI slices, the direction
calcified mitral valve ring or aortic valvular calcifications. of the regurgitant jet can be well visualized. Because of the
The best imaging planes are a combination of longitudi- continuously changing loading conditions, the jet direction
nal and perpendicular views through the diseased valve(s). may vary over time (Fig. 14.4). Similar to cardiac catheter-
It should be emphasized that in patients with endocarditis, ization and cardiac ultrasound, the extent of signal loss in
echocardiography is superior to MRI in detecting valvular the receiving chamber can be used to grade the severity of
vegetations because of its real-time imaging features, while regurgitation. This semi-quantitative approach, however,
for endocarditis-related complications, such as abscess or suffers from similar shortcomings as the approach used by
pseudoaneurysm formation, MRI is often complementary to the other conventional imaging techniques. The signal loss
echocardiography (Fig. 14.2). is highly dependent on the imaging parameters (e.g., echo
time), and several studies have shown poor reproducibil-
Valvular Regurgitation ity of the technique between centers (5,6). The severity is
underestimated when the jet impinges on the cardiac cham-
Whereas conventional imaging techniques (i.e., echocardiog- ber wall, and this method is unable to separate turbulent
raphy and cardiac catheterization) use a semi-quantitative volumes when dual-valve disease exists. Nevertheless, cine
approach to grade the severity of valvular regurgitation, MRI remains an interesting approach to rapidly evaluate
MRI in contrast enables true quantitative assessment (2,3). the presence or absence of valvular regurgitation. In case of
This technique can image the regurgitant jet in any plane, valve regurgitation, other MRI techniques can be applied
and thus provide a three-dimensional (3D) appreciation of to quantify the valvular regurgitation severity. The new
the regurgitant jet. Besides, MRI can quantify blood flow and standard bright blood sequence that is now used to study
thus provide accurate information on the regurgitant vol- cardiac function, that is, b-SSFP technique, is designed to
ume, either as an absolute value or as a regurgitant fraction. be relatively flow-insensitive, yielding a better endocardial/
This information in combination with global ventricular blood pool definition than spoiled GE techniques. As a con-
function is of particular clinical relevance for the timing of sequence, however, this results in reduced visualization of
valve replacement. The following techniques can be applied flow disturbance secondary to valvular regurgitation. In par-
to evaluate the severity of valvular regurgitation: ticular, mild regurgitation may not be seen, or less appar-
•• Qualitative assessment of signal loss on cine MRI ent on b-SSFP cine imaging. Krombach et al. found a good
•• Quantitative assessment by velocity mapping correlation between b-SSFP cine imaging and color Doppler
•• Quantitative assessment of ventricular volumes, function, and cardiac catheterization in patients with known valvular
and mass dysfunction (7). Although the jet phenomenon was slightly
LWBK1209-ch14_p196-218.indd 198 17/05/13 4:58 PM

A B
C D
Flow
(mL/s)
Figure 14.3. Mitral regurgitation in a 15-year-old female patient

with severely dilated cardiomyopathy. Balanced-SSFP cine MRI in
horizontal long-axis (A), three-chamber view (B), basal short-axis
view (C), and velocity-mapping (phase images) (D) with flow volume
versus time curve (E). During cardiac systole presence of a regurgi-
tant jet through mitral valve (arrow, A,B,C). The direction of the jet
is excentrically located abutting the lateral left atrial wall. Velocity
mapping (systolic phase) shows white regurgitant jet (arrow, D).
The corresponding flow volume versus time curve shows regurgitant
volume of 14 mL (corresponding to a regurgitant fraction of 33%).
Quantitative assessment of LV and RV shows an end-diastolic vol-
ume of 236 mL and 83 mL and an ejection fraction of 24% and
E Time (ms) 37% for left and right ventricle, respectively.
LWBK1209-ch14_p196-218.indd 199 17/05/13 4:58 PM

Flow
(mL/s)
A B Time (ms)
Figure 14.4. Mixed tricuspid and bicuspid valve regurgitation in a 43-year-old man with surgically cor-
rected tetralogy of Fallot and recent history of inferior myocardial infarction, presenting with cardiac decom-
pensation. Balanced-SSFP cine MRI in horizontal long-axis, midsystolic time frame (A), and flow volume
versus time curve of mitral (red) and tricuspid (green) valve flow (B). Presence of excentrically (i.e., laterally)
located tricuspid (arrowheads, A) and mitral (arrows, A) regurgitation. Velocity mapping (B) shows important
(systolic) regurgitation through mitral and tricuspid valve. Quantitative assessment of LV and RV shows an
end-diastolic volume of 280 mL and 279 mL and an ejection fraction of 32% and 33% for left and right
ventricle, respectively.
more pronounced on spoiled GE cine MRI, the b-SSFP cine instantaneous flow volumes for all frames throughout the
technique resulted in a significantly better image quality. cardiac cycle enables calculation of the flow volume per
heartbeat. Velocity mapping has been well validated in vitro
and in vivo and can be considered the best available in vivo
Quantitative Assessment by Velocity Mapping
technique for flow measurements (8,9). In normal individu-
Noninvasive quantification of blood flow can be achieved als, the output of the left ventricle (LV) and the right ven-
with velocity mapping (also called velocity-encoded MRI or tricle (RV) obtained by quantification of the through-plane
phase-contrast velocity mapping). For an extensive descrip- flow in the proximal ascending aorta and pulmonary trunk
tion of the velocity mapping technique, we refer to Chapter 5 has a 1:1 ratio. Differences in ratio can be caused by left-to-
on MR blood flow measurements. In brief, spins moving right or right-to-left shunt or by valvular dysfunction.
along a magnetic field gradient are exposed to a different To obtain reliable measurements of flow velocities and
magnetic field than the stationary, nonmoving spins. As a volumes through valves with velocity mapping, several tech-
result, these moving spins experience a different degree nical issues need to be addressed. To improve the accuracy of
of dephasing than stationary spins, which is proportional measurements, it can be necessary to perform a baseline cor-
to the distance they move in time (and thus the velocity). rection using a stationary phantom (10,11). To keep phase
This phenomenon can be exploited to measure flow veloci- errors induced by eddy currents and Maxwell terms as low
ties and to extract flow volumes that pass through valves. as possible, the area of interest should be as close to the cen-
Velocity mapping can be applied in any direction, though ter of the main magnetic field (B0) as possible. The imaging
for flow quantification of valvular lesions, through-plane plane should be perpendicular to the valve. This ensures that
imaging is used. Stationary spins are depicted as midgray, the velocity vectors of most voxels are perpendicular to the
while increasing velocities in either direction are shown in imaging plane. This can be achieved by a thorough knowl-
increasing grades of white or black (Fig. 14.5). Precise deter- edge of cardiac anatomy and cardiac image-plane position-
mination of the velocity window as close as possible to the ing. The best position for flow measurements is not exactly
expected peak velocity is important to avoid aliasing while at the level of the valve, but just proximal or distal to the
maintaining the highest sensitivity for flow measurements. valve annulus. Correction for motion of the valve annulus
Measurement of the velocity of all pixels in a region of inter- through the imaging plane during the cardiac cycle can be
est enables calculation of the instantaneous flow volume at obtained with slice-tracking technique (12). This promising
any point of the cardiac cycle (Fig. 14.5). Integrating the technique is, however, not yet available for clinical use.
LWBK1209-ch14_p196-218.indd 200 17/05/13 4:58 PM

A B
Flow 600
(mL/s)
500
Figure 14.5. Quantitative through-plane flow assessment of the
400 aortic valve using the velocity mapping technique. Magnitude image (A),
phase image (B), and corresponding flow volume versus time map (C).
300 The three leaflets of the aortic valve in opened position are well visible
on the magnitude image (arrows, A). Forward flow through the aortic
200 valve is shown as dark signal on the phase image (arrows, B) while
stationary spins have a midgray signal intensity. Presence of regurgitant
100 or backward flow through the aortic valve would be visible on velocity
mapping as white signal. Delineation of the aortic valve orifice of all
0 images obtained during the entire cardiac cycle yields the flow volume
versus time map. Forward flow through the aortic valve is shown by
–100 positive values whilst retrograde flow has negative values. The surface
0 100 200 300 400 500 600 700 under the curve yields the flow volumes and enables to calculate ven-
C tricular stroke volumes (shaded surface).
Time (ms)
Quantitative Assessment of Ventricular Volumes, Valvular Stenosis

Function, and Mass
Exact determination of the stenotic valve area and transval-
Cardiac MRI is currently considered the reference technique vular gradient is crucial in assessing patients with valvular
for in vivo measurement of ventricular volumes and myo- stenosis. Although this is usually achieved by echo Doppler
cardial mass. For this task, breath-hold cine MRI, using the and cardiac catheterization, MRI is capable of assessing val-
b-SSFP sequence, is used. Based on the volumetric approach vular stenosis. Different strategies are currently available.
(Simpson’s rule), a set of cardiac short-axis cuts covering the The presence of valvular stenosis can be indirectly identi-
length of the ventricles enables to determine ventricular vol- fied by the presence of postvalvular signal loss on cine MR
umes and myocardial mass (see Chapter 6). In normal con- images. The orifice area can be measured on perpendicular
ditions, there is a 1:1 relationship between LV and RV stroke views through the stenotic valve. Velocity mapping may be
volumes. In the absence of a cardiac shunt, any discrepancy used to establish an accurate peak velocity across the valve
between stroke volumes in a patient with regurgitation will to quantify the severity of the stenosis, while b-SSFP cine
identify the regurgitant volume. Unfortunately, only patients MRI enables determination of the impact of valve stenosis
with a single regurgitant valve can be assessed. However, on ventricular volumes, function, and mass.
combination of velocity mapping with ventricular stroke
volume measurements allows for the assessment and quan-
Visual Assessment of Valve Stenosis on Cine MRI
tification of multivalvular disease. Moreover, wine MRI
allows for an excellent appreciation of the consequences of Use of signal loss extending from a stenotic valve into the
volume overload on ventricular/atrial remodeling. distal chamber or vessel is helpful for identification of a
LWBK1209-ch14_p196-218.indd 201 17/05/13 4:58 PM

A B
C D
Figure 14.6. Comprehensive MR imaging in aortic stenosis. A 71-year-old man known with stenotic
bicuspid aortic valve. Cine imaging (A) and velocity mapping (B) perpendicular through aortic valve. Cine
imaging in three-chamber view (C). Cine imaging in short-axis (D) and longitudinally through ascending aorta
(E). Contrast-enhanced 3D MR angiography of thoracic aorta (F). The presence of a bicuspid aortic valve can
be well appreciated on cine imaging and velocity mapping perpendicular through the aortic valve (arrows,
A,B). The aortic valve orifice measures 1.2 cm2 (direct planimetry) with a maximal flow velocity of 3.2 m/s
yielding a transvalvular gradient of 41 mm Hg. The flow acceleration with a postvalvular jet can be well seen
on the three-chamber view (arrow, C). Compensatory concentric LV hypertrophy with maximal septal thick-
ness 18 mm (D). Dilatation of the ascending aorta to 55 mm (E,F).
LWBK1209-ch14_p196-218.indd 202 17/05/13 4:58 PM

E F
Figure 14.6. (continued)
s tenotic valve (Fig. 14.6). The degree of signal loss, however, e stimate of the gradient across the valve, using the modified
is not only dependent on the degree of stenosis but also on Bernoulli equation:
the echo time used. For shorter echo times, less spin dephas-
∆P = 4V2
ing occurs. Kilner et al. have shown that in more severe ste-
noses shorter echo times need to be used to prevent signal where P is the pressure drop across the stenosis (in milli-
loss in the images (13). Similar as for the qualitative assess- meters of mercury) and V is velocity (m/s). The technique
ment of valvular regurgitation, signal loss is less marked on is comparable with Doppler echo measurements and has an
the b-SSFP cine MR images, and therefore it is more difficult in vitro accuracy of 4% (18). The main advantage of the
to get a feel for the peak velocity with b-SSFP sequences than technique over cardiac ultrasound is that the velocity jet can
with conventional spoiled GE cine imaging. be aligned easily in any direction without the limitation of
acoustic windows or the interference of (peri)valvular calci-
Valve Area Quantification fications.
The flow through a stenotic valve can be imaged in-
Valve area or valve orifice can be estimated indirectly using
plane (velocity jet parallel to the imaging plane) or through-
the peak systolic transvalvular gradient (cardiac catheteriza-
plane (velocity jet perpendicular to the imaging plane). The
tion) and the continuity equation (echo Doppler), or can be
in-plane method visualizes the entire jet and shows the point
calculated by direct planimetry (transesophageal echocar-
within the jet of peak velocity. However, peak velocity may
diography and MRI). MR images perpendicular through the
not always be accurately depicted, for instance, when the
stenotic valve, during maximal opening, allows a planimet-
jet is not aligned in a single two-dimensional (2D) plane,
ric quantification of the valve orifice (14,15). It is currently
or in case of tight narrow jets there may be partial volume
unclear whether b-SSFP sequences with short echo times are
averaging and motion within the imaging slice. For through-
better than the conventional spoiled GE techniques to obtain
plane motion, the jet will always pass through the imaging
reliable estimates of the valve stenosis (16). While spoiled GE
plane, but, as only part of the stenotic jet is sampled, the
cine MRI, especially when using longer echo times, is sensitive
peak velocity may not be measured. Since both strategies
to spin-dephasing artifacts, the b-SSFP technique may have
have advantages and disadvantages, it is the best to use a
different sources of artifacts, including sensitivity to areas of
combination of the two strategies with initial definition of
magnetic field disturbance (e.g., calcifications, tissue inter-
the jet in-plane and quantification with through-plane imag-
faces) and sometimes severe artifacts in areas of highly com-
ing at the site of maximum velocity on the in-plane image.
plex flow (16). Using data from velocity-encoded flow curves,
Use of an appropriate velocity window is essential to
the continuity equation (similar as used in echo Doppler) can
maintain sensitivity and accuracy of measurements, while
be used to calculate valve area (17) (see Aortic Valve).
avoiding aliasing. Thus, the velocity window should be
adapted to the expected peak velocity within the vessel or
Quantitative Evaluation with Velocity-encoded Cine MRI
the valve of interest. Practically, always check the phase im-
Information from the phase-contrast velocity map, that is, ages and repeat the sequence with an adapted velocity win-
peak velocity across the valve, can be used to obtain an dow if aliasing is present.
LWBK1209-ch14_p196-218.indd 203 17/05/13 4:58 PM

Associated Findings in VHD Aortic Valve

It is important to identify the presence of associated find-
ings in patients with VHD. The presence of VHD will Aortic Stenosis
have an immediate impact on the loading conditions of Severe AS is an important clinical entity because it is the
the ventricle, eventually leading to a geometric ventricu- most common valve lesion considered for valve replacement
lar remodeling with morphologic changes (cavity diameter, in the United States, especially in older people (1). AS may
wall thickness, and myocardial tissue characteristics) and be congenital (e.g., bicuspid aortic valve), or acquired (i.e.,
functional changes (systolic, diastolic (dys) functions). It is degenerative [fibrocalcific senile AS] and rheumatic heart
commonly accepted that MRI is more accurate than echo- disease) (Fig. 14.7). Besides stenosis of the aortic valve,
cardiography in describing this remodeling process (19). sub- and supravalvular stenosis may also lead to AS. The
Myocardial tagging can be applied noninvasively to study hemodynamic consequence of all forms is an increase in LV
abnormal myocardial deformation patterns in patients afterload leading to ventricular hypertrophy and myocardial
with VHD (20,21). replacment fibrosis (Fig. 14.6). It is essential in the diagnosis
A reversed remodeling is found early after valve replace- of AS to obtain information on the gradient across the aortic
ment (19,22). Assessment of the inflow patterns through valve and the absolute aortic valve area (AVA). The gradient
the atrioventricular valves or flow patterns in the pulmo- as such gives an incomplete assessment of the AS severity.
nary and systemic veins with velocity mapping can be used Since the gradient is dependent on the stroke volume, ejec-
to assess diastolic function, while contrast-enhanced inversion- tion time (both of which are load dependent and influenced
recovery MRI with late or delayed imaging (i.e., late gado- by myocardial contractility), and also the aortic pressure, it
linium enhancement [LGE] MRI) is able to depict abnormal is possible for a mean aortic valve gradient <50 mm Hg to
areas of focal (hyper)-enhancement reflecting areas of mac- be associated with severe, moderate, or even mild AS (24).
roscopic fibrosis or microinfarction in patients with long- Thus it is important to determine AVA. Severe AS is defined
standing aortic stenosis (AS). as an AVA ≤1 cm2 (indexed AVA ≤0.6 cm2/m2). It is cur-
In the presence of mitral stenosis or VHD with atrial rently believed that patients with severe AS can be divided
fibrillation, thrombus formation in the atria or atrial ap- into three groups: A first group of AS patients with high-
pendages needs to be ruled out. A combination of cine MRI gradient, normal-flow, and preserved LV ejection fraction;
sequences (preferably the b-SSFP technique) and LGE MRI a second group with low-gradient, low-flow, and decreased
is likely the best strategy to depict abnormal intracavity LV ejection fraction; while recently a third group has been
structures (23). The vertical long-axis view is well suited added—low-gradient, low-flow AS patients presenting with
to visualize the left atrial appendage. Abnormalities of the a preserved LV ejection fraction (24,25). The observed prev-
thoracic vessels are not uncommon in pulmonary and/or alence of the third group ranges between 14% and 24%.
aortic VHD. Aortic or pulmonary valve stenoses are not Whereas outcome in AS patients with “high-gradient” is
infrequently associated with aneurysmal dilatation of the good after aortic valve replacement, this is much less the
proximal ascending aorta or pulmonary trunk, respectively case in the other subgroups because these patients often
(Fig. 14.6). Moreover, there is an association between the have comorbid conditions (coronary artery disease, previ-
presence of bicuspid aortic valve and aortic coarctation. ous myocardial infarction, systemic hypertension, diabetes
MRI offers the possibility of an integrated approach com- mellitus, COPD).
bining valve imaging and great vessel imaging. The arse- On longitudinal cine MR images through the aortic valve,
nal of MRI techniques used to study the thoracic vessels the stenotic valve usually presents as a thickened, low-inten-
consists of fast spin-echo MRI, b-SSFP cine MRI sequences, sity, nearly immobile structure, with a systolic jet (visible as
and 3D contrast-enhanced MR angiography. a signal void) through the narrowed valve orifice. There may
be concomitant aortic regurgitation, visible as a diastolic sig-
nal void into the LV. Perpendicular cine views through the
Assessment of Non-VHD Cardiac Diseases aortic valve can clearly show the number of valve leaflets,
Concomitant non-VHD–related cardiac diseases are not the mobility of leaflets, and the aortic valve orifice. These
infrequent in patients with VHD. For example, LV dys- images enable direct measurement of the AVA (Fig. 14.7).
function with low cardiac output in the presence of AS For this measurement, the proximal part of the vena con-
may be related to the AS, but may also be caused by contracta is used, or in other words, the area of through-plane
comitant coronary artery disease or previous myocardial flow at the level of the origin of the flow jet (26). Similar
infarction. In order to determine the potential benefit of to echocardiography, AVA can be indirectly calculated using
an aortic valve replacement in such patients, other causes the continuity equation (also called “effective orifice area”).
of LV dysfunction need to be excluded. As highlighted in For this, the velocity time integral (VTI) is measured for sys-
other chapters in this book, MRI can be applied to image tolic forward flow in left ventricular outflow tract (LVOT)
the proximal segment and the midsegment of the coronary and aortic valve (Fig. 14.8). The AVA (cm2) is given by:
arteries (Chapter 20), to evaluate the myocardial perfusion
AVA = AreaLVOT(VTILVOT/VTIAo)
in rest and stress conditions (Chapter 15), and to determine
the presence and extent of myocardial necrosis and scar where AreaLVOT is LVOT area (cm2), VTILVOT is LVOT VTI,
tissue (Chapter 17). In this way, MRI is appealing to the and VTIAo is AoVTI.
clinicians as it offers a comprehensive approach to investi- Planimetric measurements of AVA have shown excellent
gate VHD patients. correlations with other techniques such as multidetector
LWBK1209-ch14_p196-218.indd 204 17/05/13 4:58 PM

A C E
B D F
Figure 14.7. Variation in the number of aortic valve leaflets. Tricuspid aortic valve (A,B), bicuspid aortic
valve (C,D), quadricuspid aortic valve (E,F). The diastolic images are shown in the upper row, the systolic
images in the lower row. Cine images obtained perpendicular through the aortic valve allow to visualize the
number of leaflets (arrows, arrowheads) and calculate the aortic valve orifice by direct planimetry.
computed tomography and transesophageal echocardiog-

raphy (17,27,28). Measurements of AVA based on b-SSFP
800 images correlate better with transesophageal echocar-
700
diography compared to spoiled GE images (29). Likewise,
continuity-equation–derived estimates of AVA using MRI
Velocity 600 match well with transthoracic echocardiography. However,
Aortic valve
(cm/s) planimetry of AVA tends to overestimate the effective orifice
500
LVOT area as measured by the continuity equation (30). Thus, the
400 anatomical maximum opening of a stenotic aortic valve is
300 larger than the size of the functional vena contracta (31).
Using an AVA of 1 cm2 on transthoracic echocardiography
200
as reference, the best threshold for detecting severe stenosis
100 on MRI was an AVA of 1.3 cm2, resulting in an accuracy of
0 96% (30).
Velocity mapping enables calculation of the gradient
–100 through the stenotic aortic valve using the peak velocity. As
1 4 7 10 13 16 19 22 25
mentioned above, a combination of in-plane and through-
Time (ms)
plane velocity measurements is the best approach to obtain
Figure 14.8. Calculation of the functional aortic valve area using the highest peak velocities through the stenotic valve (“vena
the velocity time integral (VTI) continuity equation for aortic steno-
contracta”). Depending on the severity of AS, velocity win-
sis. The graph shows the velocity time curves obtained perpendicular
through the aortic valve (dashed line) and left ventricular outflow
dows up to 5 m/s, or in extreme forms up to 10 m/s need
tract (LVOT) (full line). The peak velocity through the aortic valve is to be chosen (Fig. 14.8). Use of a short-echo time is rec-
7.15 m/s and through the LVOT 1.78 m/s. The LVOT area is 4.8 cm2. ommended to measure velocities in the vena contracta (32).
Summation of the area under the curve during systole gives the VTI This approach has shown a good in vivo agreement for a
enabling to calculate the functional aortic valve area. wide range of pressure gradients across the aortic valve
LWBK1209-ch14_p196-218.indd 205 17/05/13 4:58 PM

etween MR velocity mapping, and echo Doppler and car-

b cardiac hemodynamics is dependent on the speed of onset
diac catheterization (33). of regurgitation. In acute regurgitation, such as in bacte-
An important issue when imaging AS is the evaluation rial endocarditis or aortic dissection, the LV has no time for
of the impact on LV function and mass and on aortic root adaptation. LV filling pressures rapidly increase, cardiac out-
dimensions. Cine imaging can be used to measure the degree put decreases, and left atrial pressures increase. This cascade
of LV hypertrophy and to monitor the reduction in myo- of events leads to pulmonary edema and ultimately shock.
cardial mass after aortic valve replacement (19,22,34,35) The initial response in chronic aortic regurgitation is LV
(Fig. 14.6). Several studies, using myocardial tagging, have hypertrophy, as compensation for the increased wall stress
shown altered myocardial contractility and relaxation pat- secondary to volume overload. This is followed by a pro-
terns in AS patients. In pressure-overloaded, hypertrophied gressive LV dilatation and reduction in ventricular function.
ventricles, increased LV torsion with delayed and prolonged Regular follow-up is required to follow up disease pro-
diastolic untwisting were found (20–22). After aortic valve gression and to determine optimal timing of aortic valve
replacement, normalization of LV torsion is found. replacement. The timing of surgery is important, and is a
Caliber changes in the proximal thoracic aorta are not balance between operating too soon (operative risks, inser-
infrequent in patients with AS (Fig. 14.6). MRI can be con- tion of a valve that will not last forever, lifelong anticoagu-
sidered the reference technique to measure aortic diameters. lation), and operating too late (irreversible LV failure). At
It is important to mention that in patients with bicuspid aor- present, decision for surgical intervention usually depends
tic valve, changes in the aortic media are present indepen- on clinical symptoms, chest x-ray appearance, echocar-
dent of whether the valve is functionally normal, stenotic, or diographic findings, and the longitudinal changes in these
incompetent (36). As such, bicuspid aortic valve and associ- parameters (44).
ated aortic aneurysms are thought to be manifestations of Regurgitant aortic jets are best visualized in the coronal
a single gene defect. In patients with angina and AS, visu- or oblique coronal imaging plane. Though the extent of
alization of the coronary arteries is often requested to ex- signal void can be used to obtain a semi-quantitative esti-
clude ischemic heart disease as a contributing factor to the mate of the aortic regurgitation, the most accurate method
symptoms. To date, the role of coronary MR angiography for quantifying the regurgitant fraction is velocity mapping.
is still unclear, but other MRI techniques such as first-pass Care should be taken that the majority of flow is measured
perfusion MRI during vasodilatatory stress are appealing to through-plane (Figs. 14.9,14.10). This can be achieved by
accurately visualize myocardial ischemia in these patients. obtaining two perpendicular images through the ascending
In patients with AS, the LV hypertrophy compensates aorta and defining an oblique axial plane perpendicular to
for pressure overload. This may be accompanied by inter- both. The best position for imaging plane is placing the slice
stitial myocardial fibrosis affecting the systolic and diastolic between the aortic valve and the coronary ostia (45), though
function. MRI has the potential to visualize the presence of others recommend additional flow measurements more dis-
focal or diffuse myocardial fibrosis using LGE MRI and T1- tally in the thoracic aorta (46). A through-plane velocity
mapping techniques, respectively (37–39). Several recently window of 1.5 m/s can be recommended in isolated aortic
published studies have shown that the presence of myocar- regurgitation. In case of concomitant AS, a dual velocity
dial fibrosis in patients with moderate to severe AS predicts window, with a high systolic setting of ±5 m/s, changing to
lack of functional recovery after aortic valve replacement ±1.5 m/s for the diastolic frames, is necessary to obtain the
and is an independent predictor of mortality (37,40,41). In best sensitivities. Aortic regurgitation is expressed in abso-
patients with severe AS, a low gradient is associated with a lute terms, that is, millimeters per heart beat or liters per
higher degree of fibrosis despite preserved LV ejection frac- minute (milliliters per beat × heart rate) or as a regurgitant
tion (25). Moreover, LGE MRI is excellent for depicting fraction which is given by:
CAD-related myocardial fibrosis, and thus most helpful in
Regurgitant fraction (%) = [Aortic retrograde flow (mL/beat)/
determining the cause of LV dysfunction in AS patients pre-
Aortic forward flow (mL/beat)]
senting with low-gradient and low LV ejection fraction.
× 100
Finally, in AS patients scheduled for transcatheter aortic
valve implantation, MRI can measure accurately the aortic The severity of aortic regurgitation can be defined as
annulus diameters, and thus be considered a valuable al- follows (47):
ternative to other techniques such as catheter angiography,
•• Mild: Regurgitation fraction <15%
echocardiography, and multidetector computed tomography
•• Moderate: Regurgitation fraction 16% to 25%
(42,43).
•• Moderate to severe: Regurgitation fraction 26% to 48%
•• Severe: Regurgitation fraction >48%
Aortic Regurgitation
Besides quantitative assessment of the regurgitant jet with
Aortic regurgitation is caused by abnormalities of the valve velocity mapping, MRI can determine the anatomical regur-
leaflets (e.g., congenital bicuspid aortic valve, rheumatic gitant orifice using contiguous cross-sectional cine imaging
heart disease, bacterial endocarditis, myxomatous valve through the aortic valve, showing good agreement with sim-
associated with cystic medial necrosis, aortic valve prolapse), ilar measurements obtained with multidetector computed
or is secondary to dilatation of the aortic root (e.g., Marfan’s tomography and velocity mapping (48,49). Moreover, MRI
syndrome, syphilic aortitis, ankylosing spondylitis, Reiter findings regarding the underlying cause (i.e., aortic root dila-
disease, rheumatoid disease, Ehlers–Danlos syndrome), or tation, cusp prolapsed, cusp retraction) matched well with
to aortic dissection. The impact of aortic regurgitation on surgical findings.
LWBK1209-ch14_p196-218.indd 206 17/05/13 4:58 PM

A B
Figure 14.9. Typical appearance of aortic regurgitation on bright blood b-SSFP MRI. The regurgitant jet
is typically visible as a black fan-like (A) or a pencil-like jet (B) extending from the aortic valve (arrows) into
the left ventricle (LV). Ao, aorta; LA, left atrium.
Since the interstudy reproducibility of velocity mapping is of velocity using a short-axis plane just below the mitral
high (50), this technique may be an ideal technique in long- valve are recommended because of the complicated shape of
term patient follow-up. Moreover, a good correlation has the stenotic jet in most patients. This approach has shown
been shown between ventricular volumetric quantification good correlation with Doppler measurements at echocar-
and velocity mapping of flow volumes in the ascending aorta diography (53). The mitral valve area can also be measured
for the calculation of aortic regurgitant fractions (51). This on short-axis cine images through the mitral valve.
opens new avenues for cardiac MRI in the long-term follow- In mitral valve stenosis, velocity mapping shows high
up of patients with aortic incompetence. For instance, MRI velocities throughout cardiac diastole with loss of the nor-
is capable of demonstrating the beneficial effects of therapy mal dual-inflow peaks. Mitral pressure half-time (i.e., the
and is able to identify those patients with aortic regurgita- time for the pressure gradient to fall to half its peak value),
tion who respond favorably to angiotensin-converting en- a useful indicator of mitral stenosis severity can be also ex-
zyme (ACE) inhibition (52). Although the current role of tracted from the flow curves. Another parameter to assess
MRI in patient follow-up is not defined, MRI protocols, as mitral valve stenosis severity is assessment of the pulmonary
outlined above with accurate and reproducible measure- venous flow patterns, which is reversed in severe mitral ste-
ments of regurgitant volume and the impact on LV function, nosis (54). Left atrial size and atrial thrombus formation, as
may contribute to the development of more specific criteria well as the impact of pulmonary hypertension, secondary to
for the timing of surgery. mitral stenosis, on pulmonary valve and RV function, can
also be assessed using a comprehensive MRI approach.
Mitral Valve
Mitral Regurgitation
Mitral Stenosis
As with aortic regurgitation, mitral regurgitation may be
The most common cause of mitral stenosis is rheumatic heart acute or chronic. Causes of acute mitral regurgitation are bac-
disease. Less frequent causes are congenital mitral valve mal- terial endocarditis and myocardial infarction with involve-
formations or mitral valve obstruction caused by thrombus ment of the papillary muscle. Chronic mitral regurgitation is
or myxoma. When the mitral valve area is less than 2.5 cm2, caused by diseases of the mitral valve apparatus (e.g., rheu-
the left atrial outflow resistance will increase and lead to an matic heart disease, mitral valve prolapse, endomyocardial
abnormal diastolic transmitral gradient. The latter can be fibrosis), dilatation of the mitral annulus (e.g., ischemic or
assessed by measuring the peak velocity across the narrowed idiopathic dilated cardiomyopathy), or is secondary to other
mitral valve. Although this is usually performed with echo cardiac diseases (e.g., hypertrophic obstructive cardiomyopa-
Doppler, peak velocity measurements can be equally mea- thy) (55–57). To assess mitral regurgitation, similar strategies
sured with velocity mapping. Through-plane measurements as for aortic regurgitation can be used. Visually, the extent
LWBK1209-ch14_p196-218.indd 207 17/05/13 4:58 PM

A B
C D
Flow
(mL/s)
Figure 14.10. Mild aortic regurgitation in a 63-year-old man.

Cine imaging perpendicular through the aortic valve during systole
(A) and diastole (B). Cine imaging in three-chamber view (C) and
in LVOT view (D). Flow volume versus time curve of velocity map-
ping perpendicular through aortic valve (E). The aortic valve is
tricuspid with normal aortic orifice (6.2 cm2) (A). Nonclosure of the
aortic leaflets during systole (arrow, B), causing a pencil-like aortic
regurgitation (arrows, D) directed to the anterior mitral valve leaflet
(arrow, C). Velocity mapping (E) yields a regurgitant volume of
15 mL (i.e., regurgitant fraction of 11%). Mildly dilated aortic
E Time (ms)
root (41 × 38 mm).
LWBK1209-ch14_p196-218.indd 208 17/05/13 4:58 PM

of signal loss caused by mitral regurgitation on cine imaging cause of severe nonischemic mitral regurgitation requiring
shows good correlation with echo Doppler and cardiac cathe- surgery in the United States. Mitral valve prolapse is defined
terization. The best imaging plane to visualize the regurgitant by echocardiography as a condition in which a segment(s)
jet is the vertical or horizontal long-axis (Fig. 14.4) (58), while of one or both mitral leaflets extends beyond the plane of
short-axis imaging can be of interest to localize the exact posi- the mitral annulus in the parasternal long-axis view during
tion of valve leaflet noncoaptation. These views allow for ventricular systole by at least 2 mm (Fig. 14.13). Recently,
assessment of associated abnormalities such as MV prolapse MRI has gained a role for assessing MV prolapse, show-
or concomitant pathology of the papillary muscles. Different ing excellent agreement with echocardiography, and more-
MRI approaches are available to quantify either indirectly over, allowing the identification of myocardial fibrosis and
(A–C) or directly the regurgitant volume and fraction (D): dysfunction of the papillary muscles (63,64).
A. For isolated mitral regurgitation, in the absence of a car-
diac shunt, the difference in LV and RV stroke volumes
represents the regurgitant volume (mL per heart beat) or
Pulmonary Valve
the regurgitant fraction (%) when divided by LV end-
For pulmonary valve imaging, similar strategies as for the
diastolic volume (59).
aortic valve can be used. MRI has the advantage over echo-
B. In the presence of mixed valvular disease, the mitral regurgi-
cardiography of not being dependent on a good acoustic
tant volume can be calculated from the LV stroke (obtained
window, which is definitely advantageous for assessment of
by volumetric quantification) and the forward flow through
pulmonary valve pathology. Longitudinal views through the
the aortic valve (measured by velocity mapping). The differ-
RV outflow tract and pulmonary valve are useful to depict
ence between both values yields the regurgitant mitral flow.
the pulmonary valve and to visualize regurgitant or stenotic
In the presence of a concomitant aortic regurgitation, the
jets. Velocity mapping provides quantitative data on trans-
total forward flow through the aortic valve should be taken
valvular flow patterns. A cross-section plane just above the
into account to derive the mitral regurgitation.
pulmonary valve can be recommended (65). Velocity map-
C. The third approach uses the difference between LV
ping is usually combined with volumetric and functional
inflow and outflow by velocity-encoded cine MRI. LV
assessment of the RV or both ventricles (Fig. 14.14). These
inflow is measured at the mitral valve annulus during
studies can be performed at rest or during stress conditions,
diastole, while LV outflow is measured by the systolic
using a treadmill or pharmacology, in particular, in patients
forward aortic flow. In case of mitral regurgitation, LV
with tetralogy of Fallot postsurgically presenting with
inflow exceeds outflow, and the difference between both
chronic pulmonary regurgitation (66). Though the overall
values yields the regurgitant volume (Fig. 14.11) (60).
prognosis is excellent in these patients with a 30-year mor-
D. Finally, mitral regurgitation can be directly quantified
tality above 90%, chronic pulmonary regurgitation results
with velocity mapping (61). This can be achieved by first
in RV dilatation and dysfunction, decreased exercise toler-
visualizing the regurgitant jet on horizontal and verti-
ance, propensity for significant cardiac arrhythmia, and sud-
cal long-axis cine images. Next, through-plane velocity-
den cardiac death. There is irrefutable evidence that chronic
encoded cine MRI is performed perpendicular through
overload leads to irreversible RV damage with lack of func-
the jet, with the image plane positioned at the atrial side
tional improvement following pulmonary valve replacement
of the mitral valve (Fig. 14.4). During analysis, systolic
(67–69). MRI is excellent for longitudinal follow-up of the
forward flow in LVOT should be excluded from the (sys-
degree of valve dysfunction and chronic RV overload over
tolic) regurgitant mitral flow.
time, and to study the effects of pulmonary valve flow pat-
The severity of mitral regurgitation can be defined as terns, relief of RV overload, and improvement in ventricu-
follows (57): lar function in patients undergoing transcatheter pulmonary
•• Mild: Regurgitation fraction <30% valve placement (70–73). Wald et al. showed that pulmo-
•• Moderate: Regurgitation fraction 30% to 50% nary regurgitant volume and fraction are not interchange-
•• Severe: Regurgitation fraction >50% able, and that the regurgitant volume may more accurately
reflect RV preload and better represent physiologically sig-
A good correlation between the direct and indirect meth- nificant regurgitation (74). Moreover, use of a corrected RV
ods is reported, in particular in patients with low heart stroke volume, adjusting for the pulmonary regurgitation,
rate variability (61). Velocity mapping is increasingly used may be more appropriate to study recovery of RV function
to compare and validate other techniques such as real-time following reintervention (68,72).
3D echocardiography (62). Moreover, cardiac MRI is well
placed as imaging tool to accurately quantify and reproduc-
ibly monitor the severity of mitral regurgitation and the he- Tricuspid Valve
modynamic consequences on LV volumes and function. Also
changes in left atrial size and atrial thrombus formation can The MRI strategies to assess tricuspid valve disease are
be well depicted with MRI (Fig. 14.12). similar to those for the mitral valve. The regurgitant flow
of tricuspid regurgitation is best assessed along the longi-
tudinal cardiac image planes (Fig. 14.4). It should be noted
Mitral Valve Prolapse
that minor degrees of tricuspid regurgitation are a common
Mitral valve prolapse is a common disorder affecting 2% finding in normal subjects (75). Pathologic tricuspid regurgi-
to 3% of the general population and is the most common tation is most commonly seen secondary to dilatation of the
LWBK1209-ch14_p196-218.indd 209 17/05/13 4:58 PM

A B
600
1200
Flow (mL/s)
400 800
Flow (mL/s)
400
200
0
0 400 800
400 800
C Time (ms) D Time (ms)
Figure 14.11. Measurement of mitral regurgitation by using velocity mapping to measure the difference
between inflow across the mitral annulus (A) and outflow across the proximal ascending aorta (B). Phase
images acquired at the mitral annulus (arrow, A) and at the level of proximal part of ascending aorta (arrow,
B). The graphs show the flow volume versus time frame for mitral inflow (black) and aortic outflow (red) in
a normal subject (C) and in a patient with mitral regurgitation (D). The differences in the areas under the two
curves is the volume of mitral regurgitation.
RV, often due to mitral valve pathology and/or pulmonary with carcinoid heart disease present with signs of right heart
hypertension. Tricuspid regurgitation may also occur after failure as a consequence of combined tricuspid stenosis and
blunt chest trauma. Since the tricuspid valve orifice is larger regurgitation.
than the mitral valve orifice, tricuspid stenosis is rare. In a Ebstein’s anomaly is a congenital abnormality of the tri-
minority of patients with carcinoid syndrome the heart may cuspid valve, characterized by a more apical implantation
be involved called “carcinoid heart disease.” It consists of of the septal and mural leaflets (Fig. 14.15). This results in
a thickening of the tricuspid valve leaflets and subvalvular an atrialization of the RV and a malfunctioning, regurgitant
apparatus, leading to a thick, almost immobile valve (76). tricuspid valve. This anomaly is usually associated with an
Less frequently, the pulmonary valve may be involved as atrial septal defect with a right-to-left shunt at atrial level
well. The left heart valves are usually not affected. Patients (as a consequence of the raised right atrial pressure) and
LWBK1209-ch14_p196-218.indd 210 17/05/13 4:58 PM

A B
C D
Figure 14.12. Giant atria in a 72-year-old woman with restrictive cardiomyopathy of unknown origin
and secondary severe mitral and tricuspid regurgitation. Horizontal (A,B) and vertical (C,D) long-axis cine
imaging at end diastole (left panels) and midsystole (right panels). Volumetric assessment of ventricle shows
end-diastolic volumes of 287 mL and 230 mL and an ejection fraction of 60% and 61% for left and right ven-
tricle, respectively. Presence of a severe mitral regurgitation (80 mL regurgitant volume) (arrowheads B,D) and
moderate severe tricuspid regurgitation (47 mL regurgitant volume). Volumetric assessment of the atria yields
a volume of 1,700 mL and 1,300 mL for left and right atrium, respectively. LA, left atrium; RA, right atrium.
LWBK1209-ch14_p196-218.indd 211 17/05/13 4:58 PM

A B
C D
Flow
(mL/s)
Figure 14.13. Mitral valve prolapse in a 57-year-old man. Cine

imaging in horizontal long-axis (A,B) and vertical long-axis (C,D).
Flow volume versus time curve (E). Prolapse of both mitral valve
leaflets (arrows, A,C) with secondary mitral regurgitation (arrows,
B,D). The latter is reflected on the flow volume versus time curve by
the negative flow direction during systole. The difference between
the systolic forward flow in the ascending aorta (59 mL/heart beat)
and the diastolic forward flow through the mitral valve (81 mL/heart
beat) yielded a regurgitant volume of 22 mL. Volumetric quantifica-
tion of the ventricles yielded an end-diastolic volume of 178 mL and
128 mL and a stroke volume of 83 mL and 65 mL for the left and
right ventricle, respectively. The difference in stroke volume between
ventricles (22 mL) corresponded nicely with the regurgitant volume
E Time (ms) obtained by velocity mapping.
LWBK1209-ch14_p196-218.indd 212 17/05/13 4:59 PM

A B
C D
Flow
(mL/s)
Figure 14.14. Massive pulmonary regurgitation in a patient

with previous surgical repair for tetralogy of Fallot. Cine imag-
ing in horizontal long-axis (A) along RV outflow tract (B).
Velocity mapping (perpendicular through pulmonary valve) with
systolic (C) and diastolic (D) image. Flow volume versus time
curve (E). Volumetric quantification shows moderate RV dilata-
tion. Presence of moderate RV hypertrophy. Cine imaging along
RV outflow tract shows massive backward flow during diastole
E Time (ms) (arrows, B). The latter is also reflected by velocity mapping.
LWBK1209-ch14_p196-218.indd 213 17/05/13 4:59 PM

A B
Figure 14.15. Ebstein’s anomaly. Axial cine imaging at end diastole (A) and end systole (B). The abnormal
apical implantation of the mural leaflet of the tricuspid valve and the normal site anterosuperior leaflet (arrow-
heads). The atrialization of the right ventricle (RV) can be well appreciated on the axial images. RA, right atrium.
subsequent cyanosis. MRI can be used to assess valve morphol- mitral pre-6,000 series valves (77,78). At higher field strength
ogy and to quantify ventricular function, to size right atrial en- (i.e., 4.7 T), however, increased magnetism has been noted in
largement, and eventually to calculate the shunt volumes. the Carpentier–Edwards Physio Ring (79). For nonbiologic
As with echocardiography, the peak velocity within the prosthetic valves and stented biologic valves, the applied
tricuspid regurgitant flow jet can be used to estimate the RV magnetic field is distorted by differences in the local mag-
pressure and pulmonary artery pressure. This equation is netic field between the prosthesis and the biologic tissue, and
valid if there is no pulmonary valve or arterial stenosis. It is by eddy currents induced in the valve. These phenomena lead
estimated that the normal right atrial pressure is 5 to 10 mm to signal loss in the area of the prosthesis. Especially on GE
Hg. Adding the pressure gradient obtained by the modified images, these artifacts can be very severe and substantially
Bernoulli equation from the peak velocity of the regurgitant jet degrade image quality. Imaging of turbulent jets in the vicin-
gives an estimation of the RV and pulmonary artery pressure. ity of the prosthesis may be difficult. Homografts, autografts,
and stentless porcine valve replacements do not cause signal
artifact and can be imaged normally (Figs. 14.16,14.17). In
Mixed Valvular Heart Disease a recent study, von Knobelsdorff-Brenkenhoff et al. found
good correlation between transesophageal echocardiogra-
In patients with mixed VHD, a combination of biventricu- phy and cine imaging for assessment of orifice area of aortic
lar volumetric quantification with cine imaging, and veloc- bioprostheses (80).
ity mapping of the flow in ascending aorta and pulmonary Velocity mapping distal to the prosthetic valves, however,
trunk (optionally also the mitral and tricuspid valve flow) can still be accurately performed, and offers an ideal method
is recommended. Aortic and/or pulmonary regurgitation for the noninvasive follow-up of patients after valvular sur-
can be calculated from the backward flow during diastole gery (81,82). Assessment of transvalvular flow patterns can
in the velocity maps as discussed before. Subtraction of the be combined with morphologic evaluation of the LVOT,
systolic forward flow from the ventricular stroke volume aortic graft, and proximal coronary arteries (83). Flow pat-
yields the regurgitant volume over the atrioventricular valve. terns through aortic valve prostheses are different from flow
This approach, additionally, allows for the evaluation of the patterns through normal aortic valves. Peak velocities are
impact of concomitant cardiovascular shunt on intracardiac increased (1.9 ± 0.4 m/s vs. 1.2 ± 0.03 m/s, respectively),
flow volumes and volume overload. while a diastolic backward flow with a regurgitant fraction
of approximately 15% is typically present (83,84). The post-
valvular flow characteristics obtained with 2D or preferably
Prosthetic Valves 3D velocity mapping may be helpful in defining the best
prosthetic valve design and surgical positioning of the valve
All valvular prostheses can be safely imaged at current field (85). Volumetric quantification of ventricular volumes, func-
strengths (i.e., 1.5 T and 3 T), except for Starr–Edwards tion, and mass with MRI enables monitoring the ventricular
LWBK1209-ch14_p196-218.indd 214 17/05/13 4:59 PM

A B
C D
Figure 14.16. Severe aortic stenosis in a 73-year-old man. Cine imaging longitudinally (A,B) and perpen-
dicularly (C) through the aortic valve, velocity mapping (D). Tricuspid aortic valve with severely narrowed
aortic orifice (arrows, C). Cine imaging shows systolic jet (flow void) in the ascending aorta (arrows, A,B).
Aortic valve area measurement yields an area of 1 cm2 and velocity mapping shows a transvalvular gradient of
84 mm Hg.
response to valve replacement. For aortic valve replacement, diastolic RV function (increased effective stroke volume and
a reduction in LV mass is noticed at short-term (6 months to improved early filling) were observed (86).
1 year). This reduction is independent of the type of pros-
thesis used and is of the order of 19% to 26% reduction
(19,22,35). Improvement in LV ejection fraction is found Future Directions
after valve replacement for AS, but not for aortic regurgi-
tation (19). In patients with pulmonary valve replacement, From a technical point of view, further improvements in
subsequent RV remodeling and improvement in systolic and MR scanner hardware and sequence design will lead to
LWBK1209-ch14_p196-218.indd 215 17/05/13 4:59 PM

A B
C D
Figure 14.17. Follow-up CMR study postaortic valve replacement (similar patient as in Figure 14.16).
Cine imaging longitudinally (A,B) and perpendicularly (C) through the aortic valve, velocity mapping (D). The
systolic flow void has disappeared. The aortic orifice measures 2.8 cm2 and velocity mapping shows a transval-
vular gradient of 36 mm Hg.
faster acquisition enabling real-time 2D or 3D assessment of From a clinical point of view, there is need for large-scale
flow data. Moreover, slice-tracking techniques will correct long-term studies of the physiologic nature of VHD and the
for through-plane motion, while the development of more consequences of treatment, either medically or surgically
robust, automated contour detection programs will reduce with valve replacement. Such studies using the described ac-
the laborious postprocessing. Another domain of MR devel- curate MRI techniques should enable improved timing for
opment is the use of time-resolved 3D MR flow (4D) to non- optimal valve replacement to balance the risks of surgery
invasively study the hemodynamic consequences of VHD and prosthetic valve insertion and irreversible ventricu-
(87,88). Unfortunately, acquisition of 4D flow is time con- lar dysfunction. Several studies, as mentioned above, have
suming, limiting clinical use at present. shown the potential of contrast-enhanced MRI to depict
LWBK1209-ch14_p196-218.indd 216 17/05/13 4:59 PM

myocardial fibrosis. Whereas depiction of focal myocardial 20. Stuber M, Scheidegger MB, Fischer SE, et al. Alterations in the local myocardial motion
pattern in patients suffering from pressure overload due to aortic stenosis. Circulation.
fibrosis is straightforward, visualization of diffuse-intersti- 1999;100:361–368.
tial myocardial fibrosis remains more challenging and often 21. Nagel E, Stuber M, Burkland B, et al. Cardiac rotation and relaxation in patients with
aortic valve stenosis. Eur Heart J. 2000;21:582–589.
time consuming depending on the technique used (39). 22. Sandstede JJW, Beer M, Hofmann S, et al. Changes in left and right ventricular cardiac
function after valve replacement for aortic stenosis determined by cine MR imaging.
J Magn Reson Imaging. 2000;12:240–246.
Conclusion 23. Mollet NR, Dymarkowski S, Volders W, et al. Visualization of ventricular thrombi
with contrast-enhanced MRI in patients with ischemic heart disease. Circulation.
2002;106:2873–2876.
Although MRI is capable of providing anatomical infor- 24. Pibarot P, Dumesnil JG. Paradoxical low-flow, low-gradient aortic stenosis. Adding new
pieces to the puzzle. J Am Coll Cardiol. 2011;58:413–415.
mation with regard to leaflet number, valve thickness, and 25. Herrmann S, Störk S, Niemann M, et al. Low-gradient aortic valve stenosis. Myocardial
the presence of endocarditis, its strength lies mainly in the fibrosis and its influence on function and outcome. J Am Coll Cardiol. 2011;58:
402–412.
accurate assessment of valvular function. Besides visualiza- 26. Strohm O, Schulz-Menger J, Hänlein D, et al. Magnetic resonance planimetry of the vena
tion of valvular stenosis and regurgitation, the severity of contracta as a new approach to assessment of stenotic heart valves: An in vitro study.
both stenotic and regurgitant lesions can be quantified for J Magn Reson Imaging. 2001;14:31–34.
27. Bruder O, Jochims M, Hunold P, et al. Comparison of aortic valve area measured by
all four valves. Quantification can also be performed even magnetic resonance imaging and dual-source computed tomography. Acta Radiol.
when multiple valves are affected. 2009;50:645–651.
At present, MRI is the only noninvasive, accurate, and 28. Pouleur A, le Polain de Waroux J, Pasquet A, et al. Aortic valve area assessment:
Multidetector CT compared with cine MR imaging and transthoracic and transesophageal
reproducible method for quantifying valvular regurgitation. echocardiography. Radiology. 2007;244:745–754.
In combination with accurate ventricular function assessment, 29. Schlosser T, Malyar N, Jochims M, et al. Quantification of aortic valve stenosis in MRI—
comparison of steady-state free precession and fast low-angle shot sequences. Eur Radiol.
cardiovascular MRI is ideal for the long-term follow-up and 2007;17:1284–1290.
evaluation of patients awaiting valvular surgery. This makes 30. Westermann Y, Geigenmüller A, Elgeti T, et al. Planimetry of the aortic valve orifice:
MRI at least complementary among the existing imaging tech- Comparison of multislice spiral computed tomography and magnetic resonance imaging.
Eur J Radiol. 2011;77:426–435.
niques to assess VHD. 31. Pouleur A, le Polain de Waroux J, Pasquet A, et al. Planimetric and continuity equation
assessment of aortic valve area: Head to head comparison between cardiac magnetic reso-
nance and echocardiography. J Magn Reson Imaging. 2007;26:1436–1443.
32. O’Brien KR, Gabriel RS, Greiser A, et al. Aortic valve stenotic area calculation from
References phase contrast cardiovascular magnetic resonance: The importance of short echo time.
J Cardiovasc Magn Reson. 2009;11:49.
1. Rahimtoola SH. Severe aortic stenosis with low systolic gradient. The good and bad news. 33. Eichenberger AC, Jenni R, von Schulthess GK. Aortic valve pressure gradients in patients
Circulation. 2000;101:1892–1894. with aortic valve stenosis: Quantification with velocity-encoded cine MR imaging. Am J
2. Masci PG, Dymarkowski S, Bogaert J. Valvular heart disease: What does cardiovascular Roentgenol. 1993;160:971–977.
MRI add? Eur Radiol. 2008;18:197–208. 34. Rajappan K, Melina G, Bellinger NG, et al. Evaluation of left ventricular function and
3. Cawley PJ, Maki JH, Otto CM. Cardiovascular magnetic resonance imaging for valvular mass after Medtronic Freestyle versus homograft aortic root replacement using cardiovas-
heart disease. Technique and validation. Circulation. 2009;119:468–478. cular magnetic resonance. J Heart Valve Dis. 2002;11:60–65.
4. Glockner JF, Johnston DL, McGee KP. Evaluation of cardiac valvular disease with MR 35. Sensky PR, Loubani M, Keal RP, et al. Does the type of prosthesis influence early left
imaging: Qualitative and quantitative techniques. Radiographics. 2003;23:e9. ventricular mass regression after aortic valve replacement? Assessment with magnetic reso-
5. Sechtem U, Pfugfelder PW, Cassidy MM, et al. Mitral and aortic regurgitation: Quantification nance imaging. Am Heart J. 2003;146:e13.
of regurgitant volumes with cine MR imaging. Radiology. 1988;167:425–430. 36. Siu SC, Silversides CK. Bicuspid aortic valve disease. J Am Coll Cardiol. 2010;55:2789–
6. Wagner S, Aufferman W, Buser P, et al. Diagnostic accuracy and estimation of the severity 2800.
of valvular regurgitation from the signal void on cine magnetic resonance imaging. Am 37. Weidemann F, Herrmann S, Störk S, et al. Impact of myocardial fibrosis in patients with
Heart J. 1989;118:760–767. symptomatic severe aortic stenosis. Circulation. 2009;120:577–584.
7. Krombach GA, Kuhl H, Bucker A, et al. Cine MR imaging of heart valve dysfunction 38. Iles L, Pfluger H, Phrommintikul A, et al. Evaluation of diffuse myocardial fibrosis in
with segmented cine imaging with steady state free precession. J Magn Reson Imaging. heart failure with cardiac magnetic resonance contrast-enhanced T1 mapping. J Am Coll
2004;19:59–67. Cardiol. 2008;52:1574–1580.
8. Firmin DN, Nayler GL, Klipstein RH, et al. In vivo validation of MR velocity imaging. J 39. Mewton N, Liu CY, Croisille P, et al. Assessment of myocardial fibrosis with cardiovascular
Comput Assist Tomogr. 1987;11:751–756. magnetic resonance. J Am Coll Cardiol. 2011;57:891–903.
9. Meier D, Maier S, Boesiger P. Quantitative flow measurements on phantoms and blood 40. Dweck MR, Joshi S, Murigu T, et al. Midwall fibrosis is an independent predictor of mor-
vessels with MR. Magn Reson Med. 1988;8:25–34. tality in patients with aortic stenosis. J Am Coll Cardiol. 2011;58:1271–1279.
10. Chernobelsky A, Shubayev O, Comeau CR, et al. Baseline correction of phase contrast 41. Azvevedo CF, Nigri M, Higuchi ML, et al. Prognostic significance of myocardial fibrosis
images improves quantification of blood flow in the great vessels. J Cardiovasc Magn quantification by histopathology and magnetic resonance imaging in patients with severe
Reson. 2007;9:681–685. aortic valve disease. J Am Coll Cardiol. 2010;56:278–287.
11. Gatehouse PD, Rolf MP, Graves MJ, et al. Flow measurement by cardiovascular magnetic 42. Paelinck BP, Van Herck PL, Rodrigus I, et al. Comparison of magnetic resonance imaging
resonance: A multi-centre multi-vendor study of background phase offset errors that can of aortic valve stenosis and aortic root to multimodality imaging for selection of transcath-
compromise the accuracy of derived regurgitant or shunt flow measurements. J Cardiovasc eter aortic valve implantation candidates. Am J Cardiol. 2011;108:92–98.
Magn Reson. 2010;12:5. 43. Jabbour A, Ismail TF, Moat N, et al. Multimodality imaging in transcatheter aortic valve
12. Kozerke S, Schwitter J, Pedersen EM, et al. Aortic and mitral regurgitation: Quantification implantation and post-procedural aortic regurgitation. J Am Coll Cardiol. 2011;58:2165–
using moving slice velocity mapping. JMRI. 2001;14:106–112. 2173.
13. Kilner PJ, Firmin DN, Rees RSO, et al. Valve and great vessel stenosis: Assessment with 44. Bonow RO, Carabello B, Chatterjee K, et al. 2008 Focused update incorporated into the
MR jet velocity mapping. Radiology. 1991;178:229–235. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease:
14. Friedrich M, Schulz-Menger J, Poetsch T, et al. Quantification of valvular aortic stenosis by a report of the American College of Cardiology/American Heart Association Task Force
magnetic resonance imaging. Am Heart J. 2002;144:329–334. on practice guidelines (Writing Committee to revise the 1998 guidelines for the manage-
15. John AS, Dill T, Brandt RR, et al. Magnetic resonance to assess the aortic valve area in ment of patients with valvular heart disease). Endorsed by the Society of Cardiovascular
aortic stenosis. How does it compare to current diagnostic standards? J Am Coll Cardiol. Anesthesiologists, Society for Cardiovascular Angiography and Interventions, and Society
2003;42:519–526. of Thoracic Surgeons. J Am Coll Cardiol. 2008;52:e1–e142.
16. Friedrich M, Schulz-Menger J, Dietz R. Magnetic resonance to assess the aortic valve area 45. Chatzimavroudis GP, Walker PG, Oshinski JN, et al. Slice location dependence of aor-
in aortic stenosis. J Am Coll Cardiol. 2004;43:2148–2149. tic regurgitation measurements with MR phase velocity mapping. Magn Reson Med.
17. Caruthers SD, Jiuan Lin S, Brown P, et al. Practical value of cardiac magnetic resonance 1997;37:545–551.
imaging for clinical quantification of aortic valve stenosis. Comparison with echocardiog- 46. Bolen MA, Popovic ZB, Rajiah P, et al. Cardiac MR assessment of aortic regurgitation:
raphy. Circulation. 2003;108:2236–2243. Holodiastolic flow reversal in the descending aorta helps stratify severity. Radiology.
18. Simpson IA, Maciel BC, Moises V, et al. Cine magnetic resonance imaging and Doppler 2011;260:98–104.
flow mapping displays of flow velocity, spatial acceleration and jet formation: A compara- 47. Gelfand EV, Hughes S, Hauser TH, et al. Severity of mitral and aortic regurgitation as
tive in vitro study. Am Heart J. 1993;126:1165–1174. assessed by cardiovascular magnetic resonance: Optimizing correlation with Doppler echo-
19. Lamb HJ, Beyerbracht HP, de Roos A, et al. Left ventricular remodeling early after aortic cardiography. J Cardiovasc Magn Reson. 2006;8:503–507.
valve replacement: Differential effects on diastolic function in aortic valve stenosis and 48. Debl K, Djavidani B, Heinicke N, et al. Assessment of the anatomic regurgitant orifice in
aortic regurgitation. J Am Coll Cardiol. 2002;40:2182–2188. aortic regurgitation: A clinical magnetic resonance imaging study. Heart. 2008;94:e8.
LWBK1209-ch14_p196-218.indd 217 17/05/13 4:59 PM

49. Goffinet C, Kersten V, Pouleur A, et al. Comprehensive assessment of the severity 70. Roest AAW, Helbing WA, Kunz P, et al. Exercise MR imaging in the assessment of pulmo-
and mechanism of aortic regurgitation using multidetector CT and MR. Eur Radiol. nary regurgitation and biventricular function in patients after Tetralogy of Fallot repair.
2010;20:326–336. Radiology. 2002;223:204–211.
50. Dulce MC, Mostbeck GH, O’Sullivan M, et al. Severity of aortic regurgitation: 71. Tulevski II, Hirsch A, Dodge-Khatami A, et al. Effect of pulmonary valve regurgitation on
Interstudy reproducibility of measurements with velocity-encoded cine MRI. Radiology. right ventricular function in patients with chronic right ventricular pressure overload. Am
1992;185:235–240. J Cardiol. 2003;92:113–116.
51. Sondergaard L, Lindvig K, Hildebrandt P, et al. Quantification of aortic regurgitation by 72. Lurz P, Puranik R, Nordmeyer J, et al. Improvement in left ventricular filling properties
magnetic resonance velocity mapping. Am Heart J. 1993;125:1081–1090. after relief of right ventricle to pulmonary artery conduit obstruction: Contribution of sep-
52. Globits S, Blake L, Bourne M, et al. Assessment of haemodynamic effects of ACE inhibition tal motion and interventricular mechanical delay. Eur Heart J. 2009;30:2266–2274.
therapy in chronic aortic regurgitation by using velocity-encoded cine magnetic resonance 73. McElhinney DF, Hellenbrand WE, Zahn EM, et al. Short and medium-term outcomes after
imaging. Am Heart J. 1996;131:289–293. transcatheter pulmonary valve placement in the expanded multicenter US Melody Valve
53. Heidenreich PA, Steffens JC, Fuijta N, et al. The evaluation of mitral stenosis with velocity- Trial. Circulation. 2010;122:507–516.
encoded cine MRI. Am J Cardiol. 1992;75:365–369. 74. Wald RM, Redington AN, Pereira A, et al. Refining the assessment of pulmonary regurgita-
54. Mohiaddin RH, Amanuma M, Kilner PJ, et al. MR phase-shift velocity mapping of mitral tion in adults after tetralogy of Fallot repair: Should we be measuring regurgitant fraction
and pulmonary venous flow. J Comput Assit Tomogr. 1991;15:227–243. or regurgitant volume? Eur Heart J. 2009;30:356–361.
55. Enriquez-Sarano M, Akins CW, Vahanian A. Mitral regurgitation. Lancet. 2009;373:1382– 75. Waggoner AD, Quinones MA, Young JB, et al. Pulsed Doppler echocardiography detec-
1394. tion of right-sided valve regurgitation. Experimental results and clinical significance. Am
56. Woo A, Jedrzkiewicz S. The mitral valve in hypertrophic cardiomyopathy. It’s a long story. J Cardiol. 1981;47:279–283.
Circulation. 2011;12:9–12. 76. Mollet NR, Dymarkowski S, Bogaert J. MRI and CT revealing carcinoid heart disease. Eur
57. Gaasch WH, Meyer TE. Left ventricular response to mitral regurgitation: Implications for Rad. 2003;13:L14–L18.
management. Circulation. 2008;118:2298–2303. 77. Randall PA, Kohman LJ, Scalzetti EM, et al. Magnetic resonance imaging of prosthetic
58. Aurigemma G, Reichek N, Schiebler M, et al. Evaluation of mitral regurgitation by cine cardiac valves in vitro and in vivo. Am J Cardiol. 1988;62:973–976.
MRI. Am J Cardiol. 1990;66:621–625. 78. Shellock FG, Crues JV. MR procedures: Biologic effects, safety, and patient care. Radiology.
59. Hundley WG, Li HF, Willard JE, et al. Magnetic resonance imaging assessment of the 2004;232:635–652.
severity of mitral regurgitation. Circulation. 1995;92:1151–1158. 79. Edwards MB, Ordidge RJ, Hand JW, et al. Assessment of magnetic field (4.7 T) induced
60. Fujita N, Chazouillers AF, Hartiala JJ. Quantification of mitral regurgitation by veloc- forces on prosthetic heart valves and annuloplasty rings. J Magn Reson Imaging.
ity encoded cine nuclear magnetic resonance imaging. J Am Coll Cardiol. 1994;23: 2005;22:311–317.
951–958. 80. von Knobelsdorff-Brenkenhoff F, Rudolph A, Wassmuth R, et al. Feasibility of cardiovas-
61. Myerson S, Francis JM, Neubauer S. Direct and indirect quantification of mitral regurgita- cular magnetic resonance to assess the orifice area of aortic bioprostheses. Circ Cardiovasc
tion with cardiovascular magnetic resonance, and the effect of heart rate variability. Magn Imaging. 2009;2:397–404.
Reson Mater Phy. 2010;23:243–259. 81. Deutsch HJ, Bachmann R, Sechtem U, et al. Regurgitant flow in cardiac valve prosthe-
62. Shanks M, Siebelink HJ, Delgado V, et al. Quantitative assessment of mitral regurgitation. sis: Diagnostic value of gradient echo nuclear magnetic resonance imaging in reference
Comparison between three-dimensional transesophageal echocardiography and magnetic to transesophageal two-dimensional color Doppler echocardiography. J Am Coll Cardiol.
resonance imaging. Circ Cardiovasc Imaging. 2010;3:694–700. 1992;19:1500–1507.
63. Han Y, Peters DC, Salton CJ, et al. Cardiovascular magnetic resonance characterization of 82. Fontaine AA, Heinrich RS, Walker PG, et al. Comparison of magnetic resonance imaging
mitral valve prolapse. J Am Coll Cardiol Img. 2008;1:294–303. and laser Doppler anemometry velocity measurements downstream of replacement heart
64. Delling FN, Kang LL, Yeon SB, et al. CMR predictors of mitral regurgitation in mitral valves: Implications for in vivo assessment of prosthetic valve function. J Heart Valve Dis.
valve prolapse. J Am Coll Cardiol Img. 2010;3:1037–1045. 1996;5:66–73.
65. Kivelitz DE, Dohmen PM, Lembcke A, et al. Visualization of the pulmonary valve using 83. Botnar R, Nagel E, Scheidegger MB, et al. Assessment of prosthetic aortic valve perfor-
cine MR imaging. Acta Radiologica. 2003;44:172–176. mance by magnetic resonance velocity imaging. MAGMA. 2000;10:18–26.
66. Valverde I, Parish V, Tzifa A, et al. Cardiovascular MR dobutamine stress in adult tetralogy 84. Kozerke S, Hasenkam JM, Nygaard H, et al. Heart motion-adapted MR velocity map-
of Fallot: Disparity between CMR volumetry and flow for cardiovascular function. J Magn ping of blood velocity distribution downstream aortic valve prostheses: Initial experience.
Reson Imaging. 2011;33:1341–1350. Radiology. 2001;218:548–555.
67. Therrien J, Siu SC, McLaughin PR, et al. Pulmonary valve replacement in adult late after 85. Kozerke S, Hasenkam JM, Pedersen EM, et al. Visualization of flow patterns distal to aor-
repair of Tetralogy of Fallot: Are we operating too late? J Am Coll Cardiol. 2000;36: tic valve prostheses in humans using a fast approach for cine 3D velocity mapping. J Magn
1670–1675. Reson Imaging. 2001;13:690–698.
68. Vliegen HW, Van Straten A, de Roos A, et al. Magnetic resonance imaging to assess the 86. Roest AA, de Roos A. Imaging of patients with congenital heart disease. Nat Rev Cardiol.
hemodynamic effects of pulmonary valve replacement in adults late after repair of tetralogy 2011;9:101–115.
of Fallot. Circulation. 2002;106:1703–1707. 87. Markl M, Chan FP, Alley MT, et al. Time-resolved three-dimensional phase-contrast MRI.
69. Razavi R, Muthurangu V, Taylor AM, et al. Magnetic resonance assessment of per- J Magn Reson Imaging. 2003;17:499–506.
cutaneous pulmonary valve-stent implantation (abstract). J Cardiovasc Magn Reson. 88. Hope MD, Hople TA, Crook SES, et al. 4D Flow CMR in assessment of valve-related
2004;6:429. ascending aortic disease. J Am Coll Cardiol Img. 2011;4:781–787.
LWBK1209-ch14_p196-218.indd 218 17/05/13 4:59 PM

PART
3
Ischemic Heart Disease
LWBK1209-ch15_p219-238.indd 219 17/05/13 5:22 PM

LWBK1209-ch15_p219-238.indd 220 17/05/13 5:22 PM
C h apt e r
Juerg Schwitter 15
Myocardial Perfusion in
Ischemic Heart Disease
■■ Introduction 82 million in 2020 (3). The economic burden is also impor-

tant. For instance, the total direct and indirect costs of CAD
■■ Pathophysiology of Coronary Artery Disease and stroke in the United States for 2008 was estimated at
Development of Coronary Artery Lesions $156 billion (4).
■■ Detection of Hemodynamically Significant Treatment of CAD showed intriguing progress in the
Coronary Lesions past years based on therapeutic strategies aimed at risk
factor modifications, for example, through lipid-lowering
■■ Technical Aspects of MR Perfusion Imaging interventions, and through new options for revasculariza-
Noncontrast Medium-based Approaches tions, for example, by the introduction of drug-eluting
Magnetic Resonance Pulse Sequences for stents. These improvements were somehow contrasted by a
Contrast Media First-pass Studies lack of efficient diagnostics. Invasive coronary angiography
Contrast Media detects disease reliably, but is costly and bears a small risk
T1-enhancing Extravascular CM for complications. Noninvasive techniques such as stress
Intravascular CM echocardiography is safe, but its robustness is suboptimal
Hyperpolarized CM resulting in a compromised reproducibility and wrong di-
agnoses (5). Scintigraphic techniques demonstrate adequate
■■ Analysis of Perfusion Data reproducibility, but some limitations in spatial resolution
Visual Assessment reduce sensitivity, while attenuation artifacts compromise
Quantitative Approach—Parameters Linked to specificity (6,7). As a consequence of these limitations, in
Tissue Perfusion a randomized controlled prospective trial an exercise ECG
Quantitative Approach—Absolute Tissue was shown to be as effective as SPECT to guide the man-
Perfusion agement of women with suspected CAD and consequently,
exercise ECG reduced costs of CAD management substan-
■■ Performance of MR Perfusion Imaging tially (8). Finally, SPECT exposes patients to radiation,
Animal Studies which becomes important if the technique should screen
Human Studies patients for presence of disease and/or should monitor pro-
■■ Outlook gression of disease in patients with known CAD. Similarly,
this limitation of radiation also applies for CT coronary
angiography. MR perfusion imaging is not affected by any
of these limitations. Instead, it shows an unbeatable ver-
Introduction satility. Therefore, in the first place, the pathophysiologic
events during CAD development will be summarized fol-
Coronary artery disease (CAD) is the leading cause of death lowed by some technical considerations concerning pulse
in the industrialized world causing more than 36% of all sequences, imaging parameters, contrast media (CM), and
deaths in the United States (1). More than 50,000 individu- data analysis. Taking its technical complexity into account,
als per million are estimated at risk for a heart attack at the performance of MR perfusion imaging in multicenter
some time during the course of their CAD (2). In addition to trials is of particular importance and there is an increasing
causing more than seven million deaths worldwide each year number of positive trials available (9–12). These and future
(3), the total human burden of CAD was estimated at 47 studies will define the position of MR perfusion imaging in
millions of disability-adjusted life years in 1990 and rising to the work-up of cardiac patients.
221
LWBK1209-ch15_p219-238.indd 221 17/05/13 5:22 PM

222 Part 3 ■ Ischemic Heart Disease
Pathophysiology of Coronary 2 × 2 mm2 for perfusion imaging, slice thickness of 8 mm).

Artery Disease It seems obvious that a change in stenosis severity would
be more reliably detected by interrogating 1,875 pixels in
Development of Coronary Artery Lesions the myocardium than 27 pixels covering the cross-sectional
area of the coronary vessel. Besides these technical consid-
A large number of clinical and experimental studies could erations, there is overwhelming evidence from the literature
relate endothelial dysfunction to cardiovascular risk factors that ischemia burden is one of the most important predic-
(RFs), thereby supporting the concept of endothelial dys- tors of future complications in patients with CAD (28–30).
function being a crucial initial step in the development of Thus, ischemia is identifying the patient population at risk
atherosclerosis (13–15). In dysfunctional endothelium, the which should benefit from treatment/revascularization. This
production of NO is reduced which in turn enhances the concept is supported by a large number of studies demon-
expression of endothelial adhesion molecules, and conse- strating an excellent prognosis in patients without ischemia
quently, promotes the recruitment of leucocytes to the vessel (this aspect is discussed in detail in Chapter 19). Also, the
intima (16). While mononuclear leucocytes typically become assessment of ischemia by fractional flow reserve (FFR), an
foam cells by accumulating lipids and form reversible fatty invasive tool, proved the importance of ischemia as a prog-
streaks, accumulation of smooth muscle cells and their nostic factor and its usefulness to guide treatment decisions.
production of extracellular matrix may induce the forma- In the FAME trial in >1,000 patients, treatment decisions,
tion of fibrous lesions during this phase of atherogenesis. when based on ischemia assessment by FFR, improved patient
Inflammatory mediators can impede the ability of smooth outcome (31) and FFR-based decisions were cost-effective
muscle cells to maintain the collagen content in atheroscle- over coronary angiography alone (32). In this light, the high
rotic lesions (17). Further, activated macrophages are also diagnostic agreement between ischemia assessment by per-
known to produce matrix-degrading enzymes (18), which fusion CMR and FFR is of particular importance (see also
can cause destabilization of the fibrous cap of atheromas, Table 15.2).
finally resulting in plaque rupture (19). Repetitive plaque
ruptures, often clinically silent, and subsequent healing may
account for an intermittent progression of atherosclerotic Detection of Hemodynamically
lesions (20–22). With progressive disease, ruptures of larger Significant Coronary Lesions
plaques containing considerable amounts of lipids and tissue
factor, a powerful procoagulant can cause ultimate throm- Once the decision is made to assess myocardial perfusion,
bus formation and acute vessel occlusion (23). In line with in order to assess coronary artery stenoses, two major
this concept, invasive studies demonstrated an increasing strategies are currently proposed to achieve this goal. One
risk for rupture for lesions with increasing severity of ste- is based on the concept that a hemodynamically signifi-
nosis (22,24,25). Accordingly, vulnerable plaques are char- cant stenosis limits flow during hyperemia, which can be
acterized by four major criteria (26): (1) Thin cap with large induced by IV infusion of a pharmacologic vasodilator
lipid core, (2) stenosis >90%, (3) endothelial denudation such as adenosine (or dipyridamole). In order to detect a
with superficial platelet aggregation, and (4) fissured plaque. reduction in hyperemic flow in a given subject, the mea-
A major goal of noninvasive imaging is to identify vulner- sured hyperemic flow must be compared with a normal
able plaques in the various perfusion beds, particularly the value for hyperemic flow. Conversely, the concept propos-
heart, the carotids, renals, and peripheral arteries of the ing coronary flow reserve (CFR) in order to detect hemo-
lower limbs. dynamically significant stenoses, suggests to “normal-
Ideally, identification of vulnerable plaques would in- ize” hyperemic flow by dividing it through resting flow,
clude determination of location, mass (i.e., stenosis severity), yielding a ratio (flowhyperemic/flowrest). Several PET studies
and composition of lesions. Direct visualization of plaques demonstrated a close correlation of CFR versus percent
in the coronary circulation has been performed in several area stenosis with a tendency toward better correlations
studies and is discussed in detail in Chapters 27 and 28. In of hyperemic flow versus percent area stenosis (47–49).
this context, it might be useful to consider the following: In While the first approach, that is, a stress-only protocol,
a 3-mm coronary vessel imaged with a currently available requires a normal database to provide threshold values for
spatial resolution of 0.5 × 0.5 mm2 for coronary MRA, sig- the hyperemic flow state, the second approach, that is, the
nal distribution in as little as 27 pixels would determine the CFR approach, incorporates measurements of resting flow.
degree of stenosis. With respect to plaque composition, as Division of hyperemic flow by resting flow may become
little as 6 pixels per component would characterize plaque problematic taking into account that many factors influ-
tissue of a 70% stenosis (considering three components in encing resting flow (e.g., heart rate, contractility, loading
the plaque: Fibrosis, lipid necrotic core, and inflammatory conditions) (50) cannot be controlled for during a clini-
cell-rich component, and each occupying a third of plaque cal application of the technique. Further, matching myo-
area) in a 3-mm vessel. Alternatively, one could aim at de- cardial regions, for example, the subendocardial layer, for
termining the severity of stenosis by assessing myocardial both rest and hyperemic condition may be difficult (since
perfusion disturbances, whereby one would neglect the geometry of the heart is changing with changing heart rate,
additional information contained in the plaque composi- loading, etc.). And finally, to obtain accurate results for the
tion. In this case, a 3-mm vessel is assumed to supply ap- CFR calculation, one must guarantee a linear relationship
proximately 60 g of muscle tissue (27), which corresponds between MR parameters representing flow over a wide
to about 1,875 pixels (considering a spatial resolution of range of flow conditions (i.e., resting and hyperemic flow).
LWBK1209-ch15_p219-238.indd 222 17/05/13 5:22 PM

Chapter 15 ■ Myocardial Perfusion in Ischemic Heart Disease 223
Technical Aspects of Hyperpolarized

13C-CM
Gd Fe
USPIO
Gd Gd Gd Dy Gd Gd
Mr Perfusion Imaging
IR SR SR IR/SR DE SR IR
Parallel
Noncontrast Medium-based Approaches
An elegant approach for perfusion measurements aims at SSFP Turbo- Turbo- Hybrid- Turbo- SSFP
exploiting the decrease in T2 relaxation time of deoxygenated FLASH SE EPI FLASH
and thus, paramagnetic hemoglobin (while oxygenated hemo-
globin is slightly diamagnetic causing less of T2 shortening).
With this blood oxygen level dependent (BOLD) method, Intra- Extra-
a T2- or T2*-weighted pulse sequence allows for an estima- vascular CM vascular CM
tion of an increased content of oxygenated blood, which

is expected to occur in the presence of unrestricted hyper-
Spin labeling CM first pass Flow BOLD
emic flow (which is associated with a lower O2 extraction).
However, this approach is sensitive to magnetic field inhomo-
geneities and the robustness of the method is not yet demon-
Human — Multicenter
Human — Single Center
strated (51,52). Also, signal differences in normally perfused Animal Studies
regions versus hyperemic regions (with a fourfold increase in Figure 15.1. Overview on various strategies and pulse sequences
flow) is as low as 32% in an experimental setting (53). This aimed at perfusion imaging.The most frequently used contrast
low signal difference might be problematic, since studies with media combined with these pulse sequences are also mentioned. All
CM first-pass approaches performed in humans showed dif- approaches given in the figure have been reported for cardiac imag-
ferences in the order of 80% to 100% to be clinically inad- ing except for hyperpolarized 13C-CM, which have only been used for
equate while signal differences in the order of 250% to 300% cerebral perfusion imaging so far. BOLD, blood oxygen level depen-
dent (imaging); CM, contrast medium; DE, driven equilibrium (yield-
were required for reliable detection of ischemic regions in
ing T2 weighting); Gd-chelate, gadolinium-chelate; EPI, echo-planar
patients (54). Another technique which also does not require imaging; Fe, iron-based CM; FLASH, fast low-angle shot (imaging);
the administration of exogenous CM is arterial spin labeling IR, inversion recovery; SE, spin-echo; SR, saturation recovery; SSFP,
(55–57). This approach is based on T1 measurements after steady-state free precession; USPIO, ultra-small paramagnetic iron
global and slice-selective spin preparation (using appropriate oxide particle.
ECG-triggered pulse sequences). Due to the inflow of unsatu-
rated proton spins, T1 in tissue is shortened after slice-selective
preparation compared with global preparation. Assuming a conditions only last 5 to 10 seconds, breathing motion is
two-compartment model absolute tissue perfusion is calcu- typically minimized by a breath-hold maneuver while car-
lated. However, this concept is not yet well established for diac motion is eliminated by ECG triggering. In contrast to
an application in humans. Magnetization transfer techniques scintigraphic techniques, this control of cardiac and breath-
are also under investigation to assess myocardial perfusion ing motion preserves the high spatial resolution of the MR
(58). A schematic of the various pulse sequences for perfusion perfusion data in the order of 1 to 3 × 1 to 3 mm. Most
assessment is given in Figure 15.1. commonly used are the extravascular gadolinium chelates
in combination with heavily T1-weighted pulse sequences.
Signal increase in the myocardium during first-pass of these
Magnetic Resonance Pulse Sequences for
CM is dependent on both, perfusion of the myocardium
Contrast Media First-pass Studies
and diffusion of CM into the interstitial space. Since these
Most experience exists for the so-called CM first-pass CM are excluded from the intracellular compartment, that
approaches used for the detection of perfusion abnormali- is, of viable cells with intact cell membranes, a perfusion
ties. With this technique, a bolus of CM is injected into a deficit during first pass can reflect either hypoperfused viable
peripheral vein and its effect on myocardial signal during myocardium or scar tissue. Conversely, for scintigraphic
hyperemia (stress-only protocol) and during resting condi- techniques, the flow tracers are taken up by viable myocytes
tions (stress–rest protocol) is monitored with fast MR pulse causing tracer accumulation, and consequently, an increase
sequences. Common for all CM first-pass techniques either in signal (a schematic is given in Figure 15.2). In order to dif-
under development or in clinical routine are the requirements ferentiate by MR imaging viable hypoperfused myocardium
to (1) provide high spatial resolution to permit detection of from scar, both demonstrating slow CM wash-in during MR
small subendocardial perfusion deficits, (2) provide adequate first-pass studies, it is recommended to wait for establish-
cardiac coverage to allow for assessment of the extent of per- ment of equilibrium distribution of CM in the various com-
fusion deficits, and (3) feature high CM sensitivity to gener- partments, that is, blood and extracellular, interstitial space
ate optimum contrast between normally and abnormally (of viable myocardium and scar), which occurs within 10 to
perfused myocardium during CM first pass. Finally, acquisi- 20 minutes post-CM injection. During this condition, late
tion of perfusion data should occur every 1 to 2 heartbeats enhancement imaging (with the inversion time set to null
to yield signal intensity—time curves of adequate temporal normal myocardium) displays hypoperfused but viable myo-
resolution which allow for extraction of various perfusion cardium as dark tissue, while scar tissue appears bright (due
parameters (see below). To fulfill these requirements, speed to the high distribution volume of CM in the enlarged inter-
of data acquisition and time-efficient magnetization prepa- stitial space of fibrous tissue). For comparison, during equi-
ration are crucial. Since passage of CM during hyperemic librium distribution of scintigraphic tracers (occurring after
LWBK1209-ch15_p219-238.indd 223 17/05/13 5:22 PM

Nuclear MR Mt /Mo TPR = 30 ms, n = 1 Mt /Mo TPR = 80 ms, n = 1

Normal Ischemic Scar Normal Ischemic Scar 85°
0.8 85° 0.8 70°
Preinjection Preinjection
70° 55°
0.6 55° 0.6
40°
CM: CM: 40°
Radiolabeled 0.4 0.4 25°
Extracellular 25°
tracer Gd-chelate 0.2 0.2
15° 15°
c /mM c /mM
Perfusion: Perfusion: A 2 4 6 8 10 12 14 B 2 4 6 8 10 12 14
Stress injection Stress injection
Acquisition:
Mt /Mo TPR = 130 ms, n = 1 Mt /Mo Steady state, n = ∞
First-Pass Acquisition:
-Minutes -Seconds 85°
0.8 70° 0.8
-(ECG triggered) -ECG-triggered
55°
-Breath-hold 0.6 0.6
Delay: Several hours Delay: 15–20 minutes 40°
Viability: Viability: 0.4 25°
0.4
Rest injection } 25°–85°
Rest injection 0.2 15°
0.2
Re-distribution 15°
LE-Acquisition:
Acquisition: c /mM
-Seconds
C 2 4 6 8 10 12 14 D 2 4 6 8 10 12 14 c /mM
-Minutes -ECG-triggered
-ECG-triggered -Breath-hold
Figure 15.3. The simulations show the influence of the delay time
Figure 15.2. Mechanism for MR and scintigraphic assessment of between a 90-degree saturation recovery preparation pulse and read-
perfusion and viability. Upper row: The three main conditions of myo- out (= time of preparation recovery: TPR) on the signal response (fast
cardium are depicted before administration of contrast media (CM), gradient-echo pulse sequence with TR/TE of 5/1 milliseconds) for vari-
that is, normal (nonischemic) myocardium, ischemic myocardium, and ous contrast medium (CM) concentrations c with r1 and r2 of 4 and 6 /
chronic scar tissue. Large rectangles represent intact myocytes and mmol/s, respectively (native T1 and T2: 1 second and 35 milliseconds).
small rectangles represent fibrocytes (within large interstitial compart- The signal–response curves are shown for the first TR period (n = 1;
ment). Middle row demonstrates perfusion assessment: Radiolabeled A–C) and for the steady state (n = ∞,D). The read-out flip angles are
tracer is trapped within viable myocytes, less tracer accumulation 15, 25, 40, 55, 70, and 85 degrees (with dash lengths increasing with
occurs in hypoperfused, ischemic tissue (middle column), which flip angle). The maximum signal and the sensitivity to changes of the
appears as cold spot. For MR, in normally perfused myocardium, CM concentration c at low concentrations increase with TPR (Figure
MR-CM accumulates in the extracellular compartment during first 15.3A–C). At higher concentrations, however, c cannot be unambigu-
pass. In hypoperfused myocardium (middle column), wash-in kinetics ously inferred from the signal with increasing TPR (clipping of the
are delayed. In hypoperfused scar tissue, negligible radiotracer and signal, Figure 15.3B,C). The steady-state signal does no longer depend
MR-CM accumulation occur. For viability assessment (bottom row), on details of the magnetization preparation (Figure 15.3D).
equilibrium distribution is required for both radiotracer and MR-CM.
Fibrocytes do not accumulate radiotracer and scar appears as a cold
spot (= fixed defect, see middle row). Ischemic myocardium with delay times (TPR) (68,69). However, this approach limited
viable myocytes accumulates radiotracer yielding a hot spot (reversible the dynamic range of signal response (54). Therefore, nowa-
defect, see above). For MR, distribution volume for MR-CM in scar days a saturation recovery preparation by a 90-degree pulse
tissue is large with consequent high concentration of MR-CM: Scar is typically applied, since it shortens the delay time TPR to
tissue appears bright on late enhancement MR imaging (with inver- 100 to 150 milliseconds (54), and additionally it renders the
sion time nulling viable myocardium).For MR imaging, the combined
sequence heart rate independent. Simulations in Figure 15.3
perfusion–viability study is short since equilibrium distribution is typi-
cally achieved within 15 to 20 minutes. Data acquisition by MR last a
demonstrate the influence of increasing TPR following a
few seconds only, allowing to use breath-holding and ECG triggering, 90-degree preparation on signal response for a fast gradient-
which preserves a high spatial resolution of perfusion data. echo pulse sequence (TR/TE: 5/1 milliseconds, r1 and r2* of
CM is 4 and 6 /mmol/sec, respectively). With increasing TPR
the sensitivity of the pulse sequence for low CM concentra-
tions increases, however, at the cost of reduced sensitivity of
rest injection or rest re-injection) tracer is not taken up by higher CM concentrations. In order to select the optimum
scar tissue, which consequently appears as a cold spot, while TPR, not only the signal intensity—CM concentration re-
viable tissue appears as a hot spot (see also Figure 15.2). lationship is important—but also the number of excitations
required achieving steady state of magnetization. Further,
magnetization preparation. In the beginning of first- TPR should result in acquisition windows which are placed
pass perfusion-CMR imaging, a high contrast between ab- within phases of minimal motion during the cardiac cycle. A
normally and normally perfused myocardium during peak TPR of 120 milliseconds in combination with a hybrid echo-
effect of bolus was achieved by preparing the myocardium planar readout scheme can result in an acquisition window
by a 180-degree inversion pulse (IR technique). To obtain as short as 240 milliseconds per slice (33). This combination
maximum contrast, the signal of precontrast myocardium avoids data acquisitions during rapid cardiac motion, that is,
was nulled by applying inversion recovery times in the order during early systole, early diastole, and late diastole (atrial
of 300 to 400 milliseconds (time from preparation to re- contraction) over a wide range of heart rates which may
covery, i.e., readout of central k-space lines: TPR) (59–67). occur during hyperemia. This very fast approach allows for
However, combining this IR-preparation scheme with a multislice data acquisition, while preserving a high sampling
turboFLASH (fast low-angle shot), readout resulted in long rate. To speed up the acquisition, it was suggested to play
acquisition windows of 650 to 750 milliseconds per slice, out one single nonslice-selective 90-degree saturation pulse,
which precludes the acquisition of multiple slices during 1 and to acquire all slices in series thereafter (36). With this
to 2 heartbeats. Partial preparation flip angles, for example, scheme, however, the delay time varies from slice to slice,
of 45 to 60 degrees were suggested which allow for shorter and consequently, CM sensitivity becomes dependent upon
LWBK1209-ch15_p219-238.indd 224 17/05/13 5:22 PM

slice position, which potentially complicates data analysis. While a 3D acquisition might be advantageous with respect
In a modified version of a saturation preparation, the entire to SNR and registration during postprocessing, larger stud-
myocardium is experiencing a 90-degree saturation pulse ex- ies, ideally, multicenter trials would be needed to show an
cept that slice of tissue which is immediately imaged after incremental benefit of these highly accelerated strategies ver-
preparation (70). With this scheme the time period for read- sus conventional perfusion sequences.
out of slicen is utilized to prepare slicen+1 and so on, that is,
the acquisition window equals the saturation recovery time. imaging parameters and myocardial signal re-
This approach allows acquisitions of up to seven slices every sponse. The concept of CM first-pass perfusion imaging
2 heartbeats at heart rates below 115 beats per minute. is based on the assumption that the CM administered in-
travenously is delivered to the myocardium in relation to
mr data readout. Conventional TurboFlash pulse se- tissue perfusion, and consequently, the relationship between
quences were often used in the past for perfusion imaging. signal intensity and CM concentration should be known.
With this technique each k-line is preceded by a radiofre- With fast gradient-echo pulse sequences the extravascular
quency excitation, which results in a readout duration in the Gd-chelates typically show an increase in myocardial signal
order of 350 to 450 milliseconds depending on the number intensity up to a maximum (T1-dominated behavior), and
of phase-encoding steps and duration of TR (59–66). Since signal decreases at higher CM concentrations (T2-dominated
this acquisition window is too long to allow for multislice behavior). It is desirable that the pulse sequence applied
imaging, faster echo-planar pulse sequences became popu- shows a linear CM concentration-–signal intensity relation-
lar (71–73), particularly when broken up in fewer lines per ship. This relationship is dependent on the type of pulse se-
radiofrequency excitation, known as hybrid echo-planar quence and the imaging parameters such as magnetization
pulse sequences (33,54,68,69). This technique reduces TE, preparation, delay time between preparation and readout
and consequently, renders the sequence more robust with re- (TPR), readout flip angle, k-space trajectories during read-
spect to potential susceptibility artifacts. With these acceler- out, and others. The simulations in Figure 15.3 demonstrate
ated pulse sequences several k-lines are acquired following that signal response is not only influenced by TPR, but
one single radiofrequency excitation reducing the TR per also be the readout flip angle. During the readout period,
k-line down to <2 milliseconds and consequently, reducing all the curves shown in Figures 15.3A–C approach steady-
total acquisition window/slice substantially. Since fast imag- state magnetization which yields the curves shown in Figure
ing is inherently conflicting with lower signal-to-noise ratio, 15.3D. Varying signal responses for different TR periods,
readout strategies during steady-state conditions of magneti- however, cause signal inhomogeneities in k-space, and thus,
zation appear promising, since they preserve magnetization image artifacts. The simulations in Figure 15.4 show the
and thus, a high signal-to-noise ratio. This technique has course of signal response during successive TR periods in
been employed successfully in animal models for monitor- relation to TPR and readout flip angle. At higher flip angles
ing CM first pass through the myocardium (74). In a human (e.g., 55 degrees), the steady state is approached earlier (= 6
volunteer study, however, ECG triggering was problematic TR periods at a TPR of 130 milliseconds), while a shorter
as well as the presence of artifacts (banding artifact proba- TPR of 11.6 milliseconds achieves steady state already after
bly due to off-resonance) and a low signal increase (approxi- first TR. These considerations may illustrate that imaging
mately 40% of baseline signal) (75). A high image quality, parameters strongly affect signal response and should be op-
reflected by a substantial signal increase in the myocardium timized with respect to dynamic range of signal response,
during first pass, appears crucial for reliable CAD detec- signal intensity–CM concentration relationship, homogene-
tion, and even offsets some compromise in cardiac cover- ity of signal in k-space, and timing of acquisition during the
age (54). Nevertheless, improving cardiac coverage without cardiac cycle. In a study in phantoms, it could be shown that
the need to reduce spatial and temporal resolution would optimization of a hybrid echo-planar pulse sequence (68)
be beneficial, since extent of CAD correlates with outcome. with respect to TPR, readout flip angle and preparation flip
Parallel imaging approaches can increase coverage without angle, improved signal response of Gd-DTPA-doped phan-
compromise in spatial and temporal resolution. In a recent toms from 80% (of nondoped phantom) to 250%, which
study in human volunteers, the loss in signal-to-noise ratio ultimately increased diagnostic performance in patients sig-
given by g × vR (with g being the so-called geometry fac- nificantly (54). However, changes in these parameters typi-
tor and R the accelerating factor) was compensated for by a cally go along with changes in the acquisition window/slice
longer TR (due to implementation of TSENSE, a modifica- which can affect cardiac coverage.
tion of SENSE (76)) and increasing the readout flip angle Ideally, perfusion data would be evaluated by dedicated
from 20 to 30 degrees. This accelerated hybrid echo-planar algorithms rather than by subjective reading (see below).
saturation recovery technique yielded twice as many slices This approach would allow comparing signal response
as obtained with the nonparallel approach, while signal-to- in an individual patient with a normal database, which
noise ratio improved by approximately 20% (77). Finally, could render the technique less observer dependent. Such
even faster imaging is possible by exploiting correlations in observer-independent comparisons with a normal data-
k-space and time (k-t) combined with sensitivity encoding base would require adaptation of normal values in case,
(78) yielding spatial resolutions of 1.4 × 1.4 mm2 with high the pulse sequence and/or imaging parameters had been
diagnostic yield (79,80). This approach was recently used changed. These technical considerations illustrate that
to acquire three-dimensional (3D) myocardial perfusion myocardial signal response is strongly dependent on im-
data at 3 T with a spatial resolution of 2.3 × 2.3 × 10 mm3 aging parameters. Therefore, two important multicenter
(16 slices) acquired within a 200-millisecond window (40). prospective trials (MR-IMPACT I and II) proved the high
LWBK1209-ch15_p219-238.indd 225 17/05/13 5:22 PM

Mt /Mo α = 25°, TPR = 130 ms Mt /Mo α = 55°, TPR = 130 ms

0.6 0.6
0.4 0.4
0.2 0.2
c /mM c /mM
A 2 4 6 8 10 12 14 B 2 4 6 8 10 12 14
Figure 15.4. The simulations show the change in signal response for different TR periods, n. Signal–
response curves for every second of the first 64 TR periods are shown for two different flip angles 25
degrees and 55 degrees in A and B, respectively (other assumptions are as for simulations shown in Figure
15.3). Assuming ideal spoiling of transverse magnetization, the steady state is approached in earlier TR
periods for higher flip angles. At 55 degrees, all signals acquired after approximately six dummy scans well
match the steady-state signal, which results in a signal response that is homogenous in k-space. This could
possibly reduce artifacts along the phase-encoding direction and keep their magnitude at a constant level for
all contrast medium concentrations. In contrast, for a flip angle of 25 degrees and a delay time TPR of
130 milliseconds, the signal for different TR periods varies considerably.
diagnostic yield of pulse sequences being based on compa-

rable sequence parameters implemented on various vendor
platforms (12,44).
Contrast Media
In MR first-pass perfusion imaging, an exogenous CM is
administered into the circulation, in general into a periph-
eral vein. Myocardial signal changes induced by CM wash-
in during first pass then reflect changes in the concentration
of the CM and thus, are related to perfusion. However, the
relation between signal change and change in CM concen-
tration is complex. While conventional T1-shortening CM,
such as extravascular Gd-chelates, induce a signal increase
in normally perfused myocardium during first pass, other
classes of extravascular CM, for example, T2*-shortening
CM such as dysprosium chelates, induce a signal loss (Figure
15.5). Even Gd-chelates at higher concentrations can induce
a signal decrease during first pass when combined with a
T2-weighted sequence (81,82). Thus, the signal intensity–
CM concentration relationship is typically nonlinear and
can even be inverse depending on CM type, dose, and pulse
sequences applied. This is fundamentally different from Figure 15.5. Effect of type of contrast medium and pulse sequence
on signal response in blood and myocardium. Top row: A conven-
ischemia detection based on the assessment of wall motion,
tional extravascular gadolinium-chelate (Gd-DTPA-BMA) is admin-
where a decrease in motion during inotropic stress unam- istered in a dog model with critical stenosis of the left circumflex
biguously indicates the presence of ischemia. coronary artery during vasodilation (induced by dipyridamole). At the
peak effect of the bolus (left), hypoperfused myocardium appears dark
(arrows), normally perfused myocardium appears bright. The gray
T1-Enhancing Extravascular
signal intensity–time curve on the right shows normal signal increase,
Contrast Media
the black curve shows delayed signal increase of hypoperfused myo-
These CM are most frequently used and represent the con- cardium. Bottom row: An extravascular dysprosium-chelate is applied
ventional extravascular Gd-chelates. Administration of these to the same animal model. In combination with a T2-weighted (driven
equilibrium) fast gradient-echo sequence, the CM induces a signal loss.
CM typically induce an increase in myocardial signal during
Consequently, at the peak effect of the CM (left), the hypoperfused
first pass, which allows detection of hypoperfused myocar-
myocardium appears bright (arrows), normally perfused myocardium
dium as territories of lower signal by visual assessment, or appears dark. Accordingly, the signal intensity–time curve of normal
by a quantitative approach extracting various parameters myocardium shows a signal drop (gray curve). (Images with courtesy
from the signal intensity–time curves (see below). Studies of C.B. Higgins, MD, M. Saeed, PhD, and M. Wendland, PhD, from
demonstrated a strong impact of maximum signal increase the University of California, San Francisco.)
LWBK1209-ch15_p219-238.indd 226 17/05/13 5:22 PM

in the myocardium achievable during first pass and diag- Alternatively, however, the polarization level of the spin
nostic performance (54). Therefore, best test performance population of specific nuclei, such as liquid 13C in various
is expected for high CM doses (in combination with pulse compounds, can be increased by a factor of up to 100,000
sequences with strong T1 weighting), which provide a large compared with polarization of water protons at the ther-
dynamic range of signal response in the myocardium, that mal equilibrium by two techniques, that is, dynamic nuclear
is, a large signal difference between normally perfused and polarization (88) and parahydrogen-induced polarization
hypoperfused territories. However, at higher CM doses, (89). A major advantage of hyperpolarized 13C-CM is the
susceptibility artifacts may occur at the subendocardium, fact that signal is only received from the 13C-nuclei, thus, no
where differences in CM concentrations between blood signal from background tissue is obtained. This theoretically
and myocardium are particularly high during first pass. To allows for absolute quantification of perfusion, since signal
optimize the CM dose (providing high signal increase in the is directly proportional to the amount of 13C-molecules (90).
myocardium without causing susceptibility artifacts in the However, since the spin population of 13C-molecules is far
subendocardium), several multicenter MR perfusion trials away from thermal equilibrium, application of RF pulsing
were performed. For strongly T1-weighted sequences, they irreversibly destroys longitudinal magnetization. This means
showed absence of a susceptibility-induced signal drop in that CM depolarization must be taken into account similarly
the subendocardial layer up to doses of 0.15 mmol/kg of a to radiotracer modeling which corrects for tracer decay. In
Gd-chelate (9). addition, depolarization due to RF pulsing has to be con-
The safety of Gd-based CM when applied for stress per- sidered, which is depending on the pulse sequence and the
fusion CMR was recently reported from the European CMR imaging parameters. Johansson et al. (91) showed that depo-
registry (83). In 17,767 patients, no severe adverse event larization can be approximated by a monoexponential func-
occurred. Thirty acute adverse reactions were observed tion with a time constant TD and successfully applied this
(0.17%), all classified as mild according to the American concept for cerebral perfusion quantification.
College of Radiology definition (83).
Intravascular Contrast Media

Analysis of Perfusion Data
T1-enhancing intravascular CM such as albumin-targeted A variety of studies demonstrated visual assessment of myo-
MS-325 (69) or poly-lysine-Gd (84,85) have been applied cardial CM wash-in kinetics to allow for the detection of
in animal models. In these studies, differences (for group hypoperfused myocardium supplied by stenosed coronary
means) between normally perfused and stenosis-dependent, arteries (9,10,33,66,67,92). A visual approach exploits the
hypoperfused myocardium were reported. However, as to experience of the observer to differentiate, for example,
our knowledge, these intravascular Gd-based CM have not between reduced signals in the myocardium during first pass
yet been tested in humans. Intravascular superparamagnetic due to hypoperfusion versus artifacts. Thus, a given data
iron oxide nanoparticles (USPIO) with a starch coating were set of perfusion studies may yield a high portion of correct
used for perfusion studies in humans (86). In combina- diagnosis when interpreted by an experienced reader, but at
tion with a T2-weighted turbo spin-echo sequence, a signal the cost of suboptimal inter-reader reproducibility, unless
drop in normal myocardium of 59% (in eight volunteers) reading criteria are strictly defined and easy to recognize by
was observed (this CM is no longer under investigation the human eye. Therefore, a computer-assisted or fully auto-
due to iron accumulation in the liver). Since susceptibility matic analysis of perfusion data would potentially render
influences signal drop of T2*-enhancing CM, not only the MR perfusion studies fully reproducible, that is, observer
concentration of CM in the voxel but also intravoxel dis- independent. Generation of such observer-independent
tribution (homogeneous vs. inhomogeneous) determines its results would provide a firm basis for cost-effective calcula-
T2*-shortening effect, thus, rendering such an approach sus- tions and would save human resources otherwise spent for
ceptible to vessel architecture (orientation, inter-vessel dis- data analysis.
tance) within the myocardium. This principle was demon- Thus, perfusion data analyses can be based on visual in-
strated with a dysprosium chelate in an elegant experimental terpretation, or may involve a manual, semi-automatic, or
model of reperfusion injury (87). Since T2* approaches are automatic quantification of perfusion parameters derived
affected by these geometry factors (which are not known from the signal intensity–time curves. These parameters are
a priori), measurements of CM concentrations by T1 tech- either linked to perfusion or represent perfusion in absolute
niques are considered superior. units of mL/min/g tissue (Figure 15.6). In this context, it is
important to note that data quality is a pivotal prerequisite
for an appropriate analysis whatever approach will be used
Hyperpolarized Contrast Media
(visual or quantitative). Thus, any analysis should start with
Conventional Gd-chelates act by modulating the signal from a quality control, which can combine subjective criteria (such
surrounding water molecules by accelerating their relax- as ghosting, wrap-around artifacts, ECG mis-triggering, and
ation rates. The low polarization level of the spin popula- breathing motion during first pass) with quantitative criteria
tion of the NMR-active nuclei at the thermal equilibrium (maximal and relative signal increase in the myocardium,
at a field strength of 1.5 T is compensated by the vast homogeneity of the saturation preparation pulse, spatial res-
abundance of water molecules. One way to increase sig- olution, and duration of acquisition window). Such criteria
nal could be achieved by increasing the field strength, since are given by Schwitter et al. (2) and are currently applied in
the polarization level increases with magnetic flux density. the Quality Core Lab of the European CMR registry (93).
LWBK1209-ch15_p219-238.indd 227 17/05/13 5:22 PM

Perfusion data assessment approach involves registration of the first-pass data over
time, that is, motion in the data caused by breathing and/
Quality control or diaphragmatic drift during first pass should be elimi-
Qualitative measures nated either by a manual procedure or by (semi)-automatic
Quality score algorithms (97–99). From the resulting signal intensity–time
Quantitative measures
curves of various transmural or subendocardial segments
covering ideally the entire left ventricular myocardium, a
Visual analysis Quantitative analysis variety of parameters can be extracted such as peak signal
intensity (59,60,69,100), signal change over time (slope)
(33,34,54,63–65,100–102), arrival time, time to peak signal,
Perfusion-related parameters Absolute perfusion mean transit time (100,101,103,104), area under the signal
e.g., upslope, MTT, etc. unit: mL/min/g intensity–time curve (74), and others. A candidate parame-
ter for a broader clinical application should demonstrate on
Manual Semi- Automatic Manual Semi- Automatic one hand a close correlation with tissue perfusion, and on
automatic automatic the other hand should prove a low interobserver and intertest
Figure 15.6. Perfusion data can be assessed in many ways and variability. For the upslope of the signal intensity–time curve,
some inconsistency is noted when analysis procedures are reported. a relatively close correlation with perfusion data is reported
A scheme is proposed for better definition of possible analysis strate- in both, animal (101,105) and human studies (33), while its
gies. In this scheme, “quantitative” means, that results are obtained in robustness could be demonstrated in multicenter trials (9).
numbers, which allows for comparison of studies in both, cross-section This upslope parameter is relatively insensitive to CM recir-
and longitudinally. Such quantitative results are obtained either manu- culation; since it uses the initial portion of the STC only, and
ally (and thus, are associated with some observer dependence), in a it also reduces the sensitivity of this parameter for motion
semi-automatic fashion, that is, some observer interaction with the
(most patients are able to hold their breath for this short time
data set is present (again associated with some minor observer depen-
dence), and finally, such quantitative data are obtained automatically,
period). Taking into account, that cardiac perfusion studies
thus eliminating any observer interaction with the data, which com- are performed with phased-array coils, analyses of signal
pletely eliminates analysis variability. Of importance, any type of per- intensity–time curves have to correct for inhomogeneous coil
fusion analysis should be preceded by a grading of image data quality sensitivities. As a correction, it has been suggested to subtract
to render results comparable between different studies and/or different precontrast signal from the first-pass signal intensities (36).
patient populations. However, signal reception by a phased-array coil does not
cause a constant off-set of signal over the field of view, but a
Visual Assessment signal decreases with increasing distance from the coil, which
consequently requires a division of first-pass signal by precon-
In order to minimize observer variability, it would be desir- trast signal for correction (9,33). Once, this correction is per-
able to define criteria which are sensitive for the detection of formed, signal response in the myocardium is dependent on
delayed CM wash-in, while they are specific and hence, are the arterial input. To obtain an estimate of an arterial input,
unlikely to be mixed up with image artifacts. Relatively long myocardial upslope data are often divided by the upslope of
persistence of zones with low signal during first pass are sug- signal in the left ventricular blood pool (33,54,64,65,102).
gestive for the presence of hypoperfusion, particularly, when This approach is suboptimal for input correction and may fail
located in the subendocardial layer (33,94). It is conceivable in situations with substantial variations in hemodynamics. In
that criteria improving specificity will decrease sensitivity an experimental study, it was shown that for this attempt for
and vice versa. Since the threshold defined to detect disease input correction, linearity between the upslope and perfusion
is negatively correlated with test specificity, it is suggested measured by microspheres was given for low perfusion values
to determine test performance based on receiver–operator only. At perfusion higher than approximately 1.5 mL/min/g,
curve characteristics (ROC) (95,96). Such ROC analysis the upslope underestimated true perfusion (105). A similar
demonstrated that various readers tend to read at differ- relationship was described for humans using PET perfusion
ent thresholds, which results in high observer variabilities measurements as the standard of reference (33). This non-
with kappa values of 0.30 to 0.74 (10). Using a quantitative linearity in the high-flow range would be particularly prob-
approach would allow to select a given threshold, which is lematic, if measurements would involve low- and high-flow
then maintained for all analyses, which reduces variability. states as it is the case for calculations of perfusion reserve,
Further, the threshold can be selected in a way to optimize that is, the ratio of hyperemic flow divided by resting flow.
sensitivity/specificity for a given test. Assuming relatively Absolute quantification of perfusion would be helpful for this
low costs for further evaluation of a false-positive case, approach. It is worth mentioning in this context, that a stress-
while missing a true-positive case would represent a high only protocol not only is less demanding with respect to per-
risk for complications, would force one to use a rather low fusion quantification, but also avoids the need for a spatial
threshold for best cost-effectiveness. co-registration of hyperemic and resting perfusion data sets.
Quantitative Approach—Parameters Quantitative Approach—Absolute

Linked to Tissue Perfusion Tissue Perfusion
Since first-pass perfusion studies are typically performed dur- Several models exist for perfusion measurements. In the fol-
ing a breath-hold, a first step in any type of a quantitative lowing, the first-pass techniques with residue detection will
LWBK1209-ch15_p219-238.indd 228 17/05/13 5:22 PM

be addressed. It is assumed that CM diluted in blood enters where a small CM bolus is injected first for determination
the tissue via the artery, passes through the capillaries, and of an arterial input function, followed by a larger CM bolus
leaves the tissue through the venous system. The CM con- to achieve an adequate signal response in the myocardium.
centration in the tissue is then solely determined by the arte- In a canine model, this approach yielded absolute values of
rial input and venous outflow through the conservation of tissue perfusion in close agreement with microsphere mea-
mass (Fick principle, 1870). The relationship between arte- surements. Alternatively, Gatehouse et al. (115), proposed
rial CM concentration (arterial input function), tissue CM a dual-T1–sensitivity method, which uses a single high-
concentration, and tissue perfusion is described by a convo- dose CM bolus injection, which provides adequate signal
lution integral. This model introduced by Kety takes both, response in the myocardium while preventing clipping of
CM inflow and outflow into account, and assumes a freely blood signal at peak effect of the bolus. This is achieved by
diffusible CM which is homogeneously mixed within a sin- applying a short saturation recovery time to measure blood
gle-tissue compartment (106). It is described by the formula signal (low T1 sensitivity for very short T1) while saturation
recovery time is longer for measurement of myocardial sig-
t nal (high T1 sensitivity for longer T1). Furthermore, to speed
C T (t) = F ⋅ ∫ C A ⋅ (τ ) ⋅ e−F / λ(t−τ )dτ up the acquisition for the low T1-sensitive blood pool mea-
0 (1) surements, a low spatial resolution is implemented. With
= C A (t) ⊗ F ⋅ e −F /λ ⋅ t this dual-T1–sensitivity single-bolus approach, perfusion re-
serve estimates closely matched the dual bolus results in a
where CT is the tissue CM concentration at time t, CA is the set of volunteers. Once a reliable AIF is obtained, different
CM concentration in blood, F is tissue perfusion (mL/min/g), models for calculation of perfusion may be applied. Since
λ is the tissue–blood partition coefficient, and ⊗ denotes con- conventional CM exert their effect indirectly through water
volution. This approach is not requiring assumptions on the proton relaxation, diffusion of water molecules between
shape of the arterial input function and after measurements of the intravascular and extravascular compartments modifies
CT and CA, F and λ can be determined by a two-parameter fit the MR signal during first pass (116–119). If the effect of
(by application of a three-parameter fit, blood volume VB can an intravascular CM would be confined to the intravascu-
be incorporated in the model as well). Several deconvolution lar compartment, maximum achievable signal during first
routines have been used to obtain F (105,107,108). In case of pass would approximate 10% of fully relaxed magnetiza-
nonfreely diffusible CM, the extraction fraction E must also tion (assuming a blood volume in tissue of 10% and absence
be considered (109–111), where E is related to F by of any water exchange between compartments), whereas an
extravascular CM mixed homogeneously in the extravascu-
Ki = E × F = F(1 − e-PS/F) (2)
lar space by diffusion would yield a maximum signal during
with PS being the permeability surface area product first pass of 20% (assuming an extracellular compartment
(mL/g/sec) and Ki is the unidirectional influx constant Ki of 20% in myocardium and the absence of any water ex-
(mL/g/sec). change between compartments). As shown in Figure 15.7
For blood pool CM, which are restricted to the intra- intravascular and extravascular CM yielded approximately
vascular compartment, and therefore, do not mix homo- 50% and 70% of fully relaxed magnetization during first
geneously in the tissue compartment, the so-called bolus pass, respectively, indicating that water exchange across
tracking approach is typically applied, which is based on the both, capillary vessel wall and cell membranes, strongly af-
central volume principle (Stewart, 1894) described by fects the signal generation (116). Generally, water exchange
conditions are categorized into fast, intermediate, and slow.
VB = F × MTT (3)
For fast water exchange, myocardial signal response in the
with MTT being the mean transit time (= expected distribu- presence of a CM would be unaffected by water exchange,
tion of transit times for the blood through the tissue volume). that is, tissue 1/T1 (= r1) increases linearly with intravascu-
Since the extravascular Gd-chelates do not cross the intact lar r1 (with the slope of this relationship representing blood
blood–brain barrier, this model is the preferred one for cere- volume), thus, changes in R1 in tissue reflect changes in r1
bral perfusion. However, hyperpolarized 13Carbon CM can of blood. Fast water exchange exists, if the rate of water
act as intravascular tracers (91) and hence, this model could exchange between compartments is considerably higher
become attractive for cardiac studies using these types of CM. than the difference in r1 between the compartments in the
For all models mentioned above, measurement of the ar- presence of CM. Larsson et al. (120) performed simulations
terial CM concentration over time, that is, the arterial input based on the tracer kinetic model proposed by Kety (106) to
function, is crucial. As a prerequisite, an accurate math- calculate the unidirectional influx constant Ki for diffusion
ematic description of the signal intensity–CM concentra- of CM over the capillary membrane (mL/min/g), where Ki
tion relationship over the full range of CM concentrations = E × F. For low doses of extravascular CM (= 0.1 mmol/
occurring in the blood pool during first-pass conditions is kg body weight to limit differences in r1 between compart-
required (84,112–114). To achieve an appropriate contrast- ments and assuming E > 0.3), water exchange only mini-
to-noise level for the signal response in the myocardium, mally affected Ki, while water exchange strongly affected the
relatively high CM doses would be desirable. However, such signal generated by intravascular CM (120). Similar findings
high doses are likely to cause a clipping of the signal inten- were reported by other groups (117–119) demonstrating
sity–time curve from which a conversion of signal intensities that intra–extravascular water exchange is the rate-limiting
into CM concentration is not possible. To solve this prob- step for signal generation in tissue (slow exchange regime).
lem, Christian et al. (105) presented a dual-bolus approach, Accordingly, neglecting water exchange for intravascular
LWBK1209-ch15_p219-238.indd 229 17/05/13 5:22 PM

Sl Intravascular Sl
(% fully relaxed) GdDOTA-PL: Intra (% fully relaxed) GdDTPA-BMA: Extra
0.00313 mmol Gd/kg 0.00625 mmol Gd/kg
100 100
80 80
60 60
LV blood t LV blood
myocardium Fas myocardium
40 40
20 20
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (sec) Myocyte Interstitium Time (sec)
A C
Capillary
Sl
Sl
GdDOTA-PL: Intra (% fully relaxed) GdDTPA-BMA: Extra
(% fully relaxed) Extravascular
0.050 mmol Gd/kg 100 0.1 mmol Gd/kg
100
80
80
60
60
t 40
40 Fas
LV blood
LV blood 20 myocardium
20
myocardium
0
0
0 10 20 30 40 50 60
0 10 20 30 40 50 60 Time (sec)
B Time (sec) D
Figure 15.7. Figures A/B show signal intensity–time curves (STC) after administration of an intravascular
gadolinium-chelate (bound to polylysine: PL). Note the different shape of the STC for blood (open circles)
with marked clipping of the peak signal for the higher dose (B) compared with the lower dose (A). For a
series of increasing CM doses, myocardial signal for the intravascular CM saturated at approximately 50%
of fully relaxed signal (B). This signal behavior is explained by a restricted water exchange between intra- and
extravascular compartment—small arrow in the corresponding scheme—and the fast water exchange between
extravascular/extracellular and intracellular compartment—large arrow in the scheme (top). Figure C/D show
the STC after administration of an extravascular gadolinium-chelate (Gd-DTPA-BMA). Again, the STC for
blood (open circles) is clipped for the higher dose. For a series of increasing CM doses, myocardial signal for
the extravascular CM saturated at approximately 70% of fully relaxed signal (D). This signal behavior is
explained by perfusion and additional diffusion of CM into the extravascular compartment and the fast water
exchange between extravascular/extracellular and intracellular compartment—large arrow in the scheme
(bottom). (STC are modified from Wendland MF, Saeed M, Yu KK, et al. Inversion recovery EPI of bolus
transit in rat myocardium using intravascular and extravascular gadolinium-based MR contrast media: Dose
effects on peak signal enhancement. Magn Reson Med. 1994;32:319–329, reproduced with permission by
MRM.)
CM could result in considerable errors in perfusion mea- volume determinations (125,126). Despite a large body of
surements. In addition to water exchange, in ischemic tis- simulations (109,125) and experimental data (84,114), the
sue leakage of intravascular CM across capillary membranes value of quantitative measures for the detection of CAD still
may occur (121,122). For extravascular CM, Tong et al. remains unclear.
(111) demonstrated substantial changes of E for different For hyperpolarized 13C-CM water exchange does no lon-
levels of myocardial perfusion similar to radioactive trac- ger affect the MR signal and consequently, perfusion quan-
ers. Nevertheless, other investigators assumed no relevant tification becomes feasible without assumptions on water
changes in E for resting and hyperemic conditions allowing exchange. From the signal received from the 13C molecules,
them to determine Ki at rest and during hyperemia to calcu- their concentration can be readily calculated, if the level of
late a perfusion index or CFR (123,124). Besides precision, polarization is known. Since depolarization of a hyperpolar-
the robustness of parameter estimates is another important ized CM in an imaged slice is dependent on relaxation times
aspect of different models in combination with extra- and (which are in the order of 30 to 60 seconds depending on
intravascular CM. Based on simulations, regional blood vol- the specific type of 13C molecules) and the pulse sequence,
ume and E were most sensitive for noise when determined the time constant TD of the polarization function can be cal-
by extravascular CM (109), while other studies showed culated to enter the appropriate models (e.g., Eq. 1 or 3) to
intravascular CM being more reliable for regional blood take depolarization into account (91).
LWBK1209-ch15_p219-238.indd 230 17/05/13 5:22 PM

Performance of Mr Perfusion ment yielded a sensitivity and specificity of 90% and 85%,
Imaging respectively, for the detection of stenoses of >70% diameter
reduction on QCA. In comparison with SPECT (n = 69),
Animal Studies MR performed better with AUC of 0.89 to 0.91 for observer
one and two, respectively, versus AUC of 0.71 to 0.75 for
In animal models, the feasibility of myocardial first-pass SPECT (37). Recently, a large single-center study confirmed
MR perfusion imaging to detect differences in perfusion these results in 628 patients using a visual assessment yield-
has been demonstrated repeatedly (59,60,127,128). In ing sensitivities and specificities for the detection of ≥50%
addition, experimental studies offer the possibility to relate diameter stenoses on QCA of 82% and 86%, respectively
estimates of myocardial perfusion derived from MR data (AUC of 0.86) (39). In this study, perfusion CMR also per-
sets to microsphere measurements, which are generally formed significantly better than SPECT in detecting ≥50%
accepted as the standard of reference. The Table 15.1 lists coronary stenoses (AUC of SPECT: 0.74, p < 0.0001 vs.
several contrast- and noncontrast-based MR techniques CMR). High-resolution perfusion-CMR pulse sequences
and provides the correlations achieved versus microsphere with temporospatial acceleration strategies yielded similarly
measurements. For BOLD imaging, a good correlation is high diagnostic performance with an AUC of 0.85 to detect
obtained between microsphere measurements (ratio between ≥50% diameter stenoses (38). Such acceleration strategies
normally perfused and hyperemic regions) and the same allowed to acquire 3D perfusion data sets on a 3-T system
ratio for differences in percent signal enhancement on T2- yielding similarly high diagnostic performance with a sen-
weighted images; however, the slope is far away from unity sitivity and specificity of 92% and 74%, respectively (40).
(53). In this regard, Δr2 measurements yielded a slope of
0.94 (129). However, even when precision seems high, the
sensitivity to detect differences in perfusion by this tech- Single Center Studies—(Semi) Automatic Analysis
nique appears rather low, since a 100% increase in flow Calculation of rate of myocardial signal change (upslope) dur-
yields a signal increase of 5% only (= change in myocardial ing first-pass conditions is probably the most widely used type
r2 of 0.94/sec); thus, differences in flow of 300% would be of a (semi)-automatic approach yielding a parameter linked
required to yield statistically significant changes in R2 (129). to tissue perfusion. Based on a single-slice IR acquisition,
For another non-CM–based technique, that is, arterial spin Al-Saadi et al. calculated a CFR index (= slopehyperemia/sloperest)
labeling, an excellent agreement with microsphere measure- yielding a sensitivity and specificity of 90% and 83%, respec-
ments was reported, however, these results were obtained in tively, at a CFR threshold of 1.5 for the detection of stenoses
nonbeating rat hearts at 4.7 T (55). While several investiga- with ≥75% area reduction on QCA (65). In this study, as a
tors calculated slope or 1/MTT from γ-variate fits of signal rough estimate of the arterial input, the upslope of signal in
intensity–time curves (85,101,104), others demonstrated the the LV blood pool was used and the dose of the extravascu-
possibility to even quantify perfusion in absolute measures lar CM Gd-DTPA was kept at 0.025 mmol/kg body weight
of mL/min/g when applying sophisticated kinetic models to to minimize clipping of the blood pool signal intensity–time
the MR data (84,107,114). However, 95% confidence inter- curve. Peripheral administration of the same CM dose and
vals of −45% to 82% to +45% to 82% (84,107) indicate acquisition in a multislice mode yielded a similar performance
that the sensitivity of these techniques to detect changes in for the CFR index (calculated as described above) with a sen-
perfusion might be limited. Furthermore, to our knowledge, sitivity and specificity of 88% and 90%, respectively (AUC
the utility of these absolute quantitative approaches for the 0.93), for the detection of stenoses with ≥75% area reduc-
detection of CAD in unselected patient populations has not tion on QCA (36). A stress-only protocol was proposed by
yet been demonstrated. our group (33) and a comparable performance of the hyper-
emic upslope parameter was reported when thresholds were
derived from a normal data base yielding a sensitivity and
Human Studies specificity of 87% and 85%, respectively, for the detection
of stenoses of ≥50% diameter reduction on QCA (Figure
Single Center Studies—Visual Assessment
15.8). A stress-only protocol avoids the need for matching
Visual assessment of MR perfusion data is popular since myocardial regions for rest and stress condition, which is dif-
it avoids time-consuming data postprocessing and artifacts ficult, if for example, the heart rate, and thus, loading varies.
within the data can be “overread” by the observer, at least, Consequently, this protocol offers easy analysis of the sub-
if he is experienced in interpreting MR perfusion studies. endocardial layer, where perfusion abnormalities are most
However, this approach also introduces variability depend- severe (33,94). While coronary anatomy and diameter steno-
ing upon the observer performance resulting in diagnostic sis is not an ideal approach to detect relevant coronary artery
accuracy with 95% confidence intervals for AUC rang- disease, FFR provides a measure of ischemia which correlates
ing from 0.54 to 0.90 in multicenter data (10). In order to with outcome. Recently, several studies demonstrated a high
reduce these differences, it would be desirable to use semi- agreement between noninvasively determined ischemia bur-
automatic or even fully automatic analysis algorithms, den by perfusion CMR versus FFR (41,42).
thereby eliminating any observer variability.
MR first-pass perfusion imaging was compared with Multicenter Studies
SPECT in a recent study in 104 patients (37) (see also Table
15.2 (35)). At dose of 0.075 mmol/kg of Gd-DTPA injected As discussed above, MR perfusion imaging provides excel-
into a peripheral vein during vasodilatation, visual assess- lent spatial and temporal resolution for signal monitoring
LWBK1209-ch15_p219-238.indd 231 17/05/13 5:22 PM

232
Table 15.1 Experimental MR Perfusion Imaging: Comparison Versus Microsphere Measurements
LWBK1209-ch15_p219-238.indd 232
Authors MR approach CM Analysis Corr. r Slope Intercept Animal No. Model
BOLD
Fieno et al. (28) Bold T2-weighted no % SI-enhancement ratio 0.80 0.08 0.94 dogs 13 occlusions adeno
Wright et al. (102) Bold 3D T2-weighted no ΔR2 ratio 0.80 0.94 0.02 dogs 9 adeno
Spin-labeling
Reeder et al. (30) spin-labeling no 1/T1app 0.91 to 0.62 to 0.67 0.45 to rabbits 3 occlusion non-beating
(at 4.7 T) 0.97 0.46 heart (Langendorff)
Poncelet et al. (31) spin-labeling no 1/T1app 0.83 0.62 0.64 pigs 6 adeno
Extravascular CM
Wilke et al. (73) IR turbo-FLASH extra γ-variate fit 1/MTT 0.89 0.02 0.03 dogs 3 occlusions stenosis/dip
upslope of STC 0.66 — —
Wilke et al. (82) SR turbo-FLASH extra fermi function ratio: RF (0) 0.88 0.92 0.10 pigs 8 occlusions adeno
Klocke et al. (52) IR true-Fisp extra ratio: AUC 0.93 0.96 0.07 dogs 12 occlusions stenosis/
adeno
Christian et al. (77) SR hybrid EPI extra dual bolus fermi function 0.95 0.95 0.10 dogs 16 occlusions adeno
upslope of STC 0.69 — —
Intravascular CM
Wilke et al. (58) IR turbo-FLASH intra multiple pathway axially -0.1a -0.92 to 0.72b dogs 7 stenosis/adeno
distributed model
(ml/min/g)
Kraitchman et al. (59) IR turbo-FLASH intra γ-variate fit ratio: upslope 0.77 — — dogs 7 stenosis/dobut
Lombardi et al. (76) IR turbo-FLASH intra γ-variate fit 1/MTT 0.70 — — pigs 5 stenosis/adeno
(at 0.5 T)
Kraitchman et al. (45) Partial presaturation intra % SI-enhancement ratio 0.80 0.58 0.33 pigs 12 stenosis/dip
Hybrid echo-planar intra
Jerosch-Herold et al. SR trubo-FLASH model-indep. decon. blood −0.16a −45% to 45%b pigs 4 occlusions adeno
(79) flow (ml/min/g) 0.0a
perfusion reserve −57% to 57%b
All studies were performed on 1.5 T systems unless noted otherwise.

Adeno, adenosine; AUC, area under the signal intensity time curve; BOLD, blood oxygen level dependent (imaging); CM, contrast medium; dip, dipyridamole; dobut, dobutamine; EPI, echo-
planar imaging; extra, extravascular CM; Fisp, fast imaging with steady-state precession; FLASH, fast low-angle shot acquisition; intra, intravascular CM; IR, inversion recovery; micro, radio-
labeled microspheres; mod-indep. decon, model-independent deconvolution; MTT, mean transit time; RF (0) = F, the impulse response amplitude represents relative measure of perfusion; Ratio,
ischemic/normal region (for this correlation the same ratio is calculated for microsphere measurements); SI, signal intensity; SR, saturation recovery; STC, signal intensity time curve; T1app,
apparent T1 of myocardium in presence of flow.
a
Mean difference in mL/min/g (minus = understimation by MR).
b
Confidence interval of 95%.
17/05/13 5:22 PM

LWBK1209-ch15_p219-238.indd 233
Table 15.2 Diagnostic Performance of First-pass Perfusion CMR in Patients
Dose of
Field GdB-CM
Study (Reference) n (n) Sites Strength Comparator (mmol/kg) Stress Analysis Sens. (%) Spec.(%) AUC
Single (33) 57 — 1.5 T CXA (≥50%) 0.1 dip (stress only) upslopesubendo 87 85 0.91
Single (33) 43 — 1.5 T PET (CFR) 0.1 dip (stress only) upslopesubendo 91 94 0.93
Single (34) 92 — 1.5 T CXA (≥70%) 0.05 adeno (rest/stress) upslope MPR 88 82 0.91
Single (35) 79 — 1.5 T CXA (≥50%) 0.05 adeno (stress/rest) visual 91 62 —
Single (36) 84 — 1.5 T CXA (≥75%)a 0.025 adeno (rest/stress) upslope MRR 86 87 0.92
Single (37) 104 — 1.5 T CXA (≥70%) 0.075 dip (stress/rest) visual 90 85 0.90
Single (38) 51 — 1.5 T CXA (≥50%) 0.1 adeno (stress only) visual — — 0.85
Single (39) CE-MARC 628 — 1.5 T CXA (≥50%) 0.05 adeno (rest/stress)b visual 82 86 0.86
Single (40)c 146 — 3T CXA (≥50%) 0.1 adeno (rest/stress) visual 92 74 —
Single (41) 43 — 1.5 T FFR <0.75 0.05 adeno (rest/stress) Upslope, MPR <1.5 88 90 0.93
Single (42) 38 — 3T FFR <0.75 0.05 adeno (rest/stress) Fermi, MPR <1.58 80 89 0.89
Multicenter (43)d 50 7 1.5 T CXA (≥50%) 0.05 ATP (stress/rest) visual 86 75 0.88
Multicenter (9) 80 (24) 3 1.5 T CXA (≥50%) 0.1 adeno (stress only) upslopesubendo 91 78 0.91
Low dose (9) 80 (29) 3 1.5 T CXA (≥50%) 0.05 adeno (stress only) upslopesubendo — — 0.53
MR-IMPACT I (44) 212 (45) 18 1.5 T CXA (≥50%) 0.075 adeno (stress only) visual 85 67 0.86
MR-IMPACT II (46) 425 33 1.5 T CXA (≥75%)a 0.075 adeno (stress/rest) visual 75 59 —
n: In CM dose finding studies (n) indicates participants in a specific dose group.

a
Area stenosis.
b
CMR positive: Any evidence of regional functional abnormality and/or any infarct (scar) on late gadolinium-enhancement images.
c
3D acquisition.
d
Study results are given as mean of two observers.
PRI, perfusion reserve index; GdB-CM, Gadolinium-based contrast medium; AUC, area under the receiver–operator characteristics curve; CXA, invasive coronary angiography; PET, positron
emission computed tomography; FFR, Fractional flow reserve; dip, dipyridamole; adeno, adenosine; MPR, myocardial perfusion reserve (= perfusion during stress/perfusion at rest).
233
17/05/13 5:22 PM
during CM first pass and is potentially the most accurate study at the dose of 0.15 mmol/kg is shown in Figure 15.9.
method currently available for the noninvasive assessment At this level of data quality, the upslope approach proved to
of myocardial perfusion in humans (see also Table 15.2). be sensitive, specific, and robust with a kappa value of 0.73
The optimum CM dose was recently determined by a multi- for interobserver agreement (9). For comparison, Figure
center trial (9). In a stress-only protocol, the upslope param- 15.10 shows kappa values for interobserver agreement for
eter performed best with relatively high CM doses ranging a multicenter scintigraphy study (130) as well as for multi-
from 0.10 to 0.15 mmol/kg Gd-DTPA yielding an AUC of center MR perfusion studies (9–11). This figure clearly dem-
0.91 ± 0.07 and 0.86 ± 0.08, respectively, with correspond- onstrates a potential disadvantage of a pure visual assess-
ing sensitivity/specificity of 91/78% and 94/71% for doses ment of perfusion studies.
of 0.10 and 0.15 mmol/kg, respectively (9). Thus, this mul- As a next step, a large multicenter, multivendor trial was
ticenter trial confirmed the high diagnostic performance of performed to compare MR perfusion imaging not only with
the stress-only approach in combination with an upslope QCA, but also with SPECT imaging (11). In 18 centers
analysis proposed in an earlier single-center study (33) worldwide, 241 patients were studied in MR-IMPACT (MR-
(Figure 15.8). In this multicenter, single-vendor trial, the Imaging for Myocardial Assessment of Coronary artery dis-
high diagnostic performance was achieved in data with ade- ease Trial). This study again confirmed the high diagnostic
quate quality only, that is, a quality reading was performed performance of the MR first-pass perfusion approach and
first, which eliminated 14% of the studies. An example for a also showed superiority over SPECT imaging (11). As shown
1
Single center: Sens/spec AUC
CMR 87%/85% 0.91
PET: CFR 91%/81% 0.93
0.8
CE-MARC
CMR 82%/86% 0.86
SPECT 0.69
0.6
Multicenter:
Sensitivity
Single vendor –3 sites

MR: Visual 85%/91% 0.92
MR: Upslope 91%/78% 0.91
0.4 MR-IMPACT: Multivendor –18 sites
MR 78%/82% 0.75
SPECT 0.65
CE-MARC
0.2 CMR vs. SPECT: p < 0.0001
MR-IMPACT
CMR vs. SPECT: p < 0.013
0
0 0.2 0.4 0.6 0.8 1
1-specificity
Figure 15.8. Comparison of diagnostic performance of MR perfusion imaging in single-center and multicenter trials and with scintigraphic
studies.A single-center study evaluated a first-pass MR stress-only protocol (saturation recovery hybrid echo-planar pulse sequence) with analysis
of the upslope parameter (quantitative semi-automatic analysis of parameter linked to perfusion). A high sensitivity and specificity for the detection
of coronary artery stenoses (≥50% diameter reduction in quantitative coronary angiography) was obtained (black thin line). Numbers represent
sensitivity/specificity and the area under the receiver–operator characteristics curve, respectively. A similar performance was achieved for posi-
tron emission tomography in that study (thin gray dotted line) (From Schwitter J, Nanz D, Kneifel S, et al. Assessment of myocardial perfusion
in coronary artery disease by magnetic resonance: A comparison with positron emission tomography and coronary angiography. Circulation.
2001;103:2230–2235). These MR results could be confirmed in a multicenter, single-vendor trial, when using a visual assessment (best of three-
blinded core lab readers). This approach (blue line), however, is susceptible for inter-reader variability, which is shown in Figure 15.10. Analysis of
the upslope parameter (thick red line) in a multicenter data set, as performed in the single-center trial, again yielded high sensitivity and specificity,
confirming the results of the single-center study. In addition, semi-automatic analysis of this upslope parameter linked to perfusion is associated
with fair to good interobserver agreement (see Figure 15.10). The MR-IMPACT I trial (very thick yellow line) yielded excellent results confirming
the other trials. SPECT performance was significantly inferior in both comparative trials. (With permission from: Giang T, Nanz D, Coulden R, et
al. Detection of coronary artery disease by magnetic resonance myocardial perfusion imaging with various contrast medium doses: First European
multicenter experienc. Eur Heart J. 2004;25:1657–1665; Wolff S, Schwitter J, Coulden R, et al. Myocardial first-pass perfusion magnetic reso-
nance imaging: A multicenter dose-ranging study. Circulation. 2004;110:732–737; Schwitter J, Wacker CM, van Rossum AC, et al. MR-IMPACT:
Magnetic Resonance imaging for myocardial perfusion assessment in coronary artery disease trial: Comparison of perfusion CMR with single
photon emission computed tomography for the detection of coronary artery disease in a multicenter, multivendor, randomized trial. Eur Heart
J. 2008;29:480–489; and Greenwood J, Maredia N, Younger J, et al. Cardiovascular magnetic resonance and single-photon emission computed
tomography for diagnosis of coronary heart disease (CE-MARC): A prospective trial. Lancet. 2012;379:453–460.)
LWBK1209-ch15_p219-238.indd 234 17/05/13 5:22 PM

A B C D E F
G H I J K L
Figure 15.9 Example of an MR perfusion study for a dose of 0.15 mmol/kg body weight of Gd-DTPA
(saturation recovery hybrid echo-planar pulse sequence). Distal stenosis in the large left circumflex coronary
artery (LCX) is shown (arrow head, A); another stenosis is present in the small, first diagonal branch of the
LAD (arrow, A) on x-ray coronary angiography. Five short-axis MR perfusion images at peak bolus effect in B
to F demonstrate low signal areas in the inferior wall (B–F) and a small hypointense area in the anterior wall
(E,F). Corresponding upslope maps are shown in H to J and polar maps for the subendocardial layer are given
with thresholds applied pixel-wise (K) and for the entire segments (L) with blue/red encoding for upslope val-
ues below/above the thresholds, respectively. This study further demonstrates the limitation of using coronary
anatomy as a reference for perfusions studies: The perfusion defects in segment five and six (typically assigned
to the right coronary artery) represent a mismatch with the right coronary artery perfusion territory which
is supplied by a large LCX in this patient. (Reproduced with permission from Giang T, Nanz D, Coulden R,
et al. Detection of coronary artery disease by magnetic resonance myocardial perfusion imaging with various
contrast medium doses: First European multicenter experienc. Eur Heart J. 2004;25:1657–1665.)
in Figure 15.10, the low kappa value for visual reading of these Outlook
multicenter, multivendor MR perfusion data underscores the
need for standardization of data acquisition, data quality, and Several single-center and large multicenter MR perfu-
data interpretation. To demonstrate superiority of perfusion sion trials could demonstrate a high diagnostic value
CMR versus SPECT, and in particular, versus gated-SPECT, of this technique, which can perform better than estab-
MR-IMPACT II was performed including 465 patients in 33 lished SPECT techniques. Preliminary cost-effectiveness
centers in Europe and the United States (12). Reading of the estimations are encouraging (93) and larger prospective
MR-IMPACT II data at a single threshold yielded superior- evaluations are planned to run within the European CMR
ity for sensitivity and inferiority for specificity (12), while the registry. Dissemination of the technique should also be
ROC analysis yielded superiority for CMR versus SPECT accompanied by quality assurance mechanisms and for
over all thresholds (46) in agreement with MR-IMPACT I and this reason, the European CMR registry was expanded by
other multicenter perfusion-CMR trials (9,44,43). a Reference Center for Quality Assessment with the aim
to assess quality in all the prospective sub-studies running
The European Registry of Cardiac Magnetic Resonance within the CMR registry and to assess quality in repre-
sentative samples of the remaining CMR examinations
A registry collecting data on CMR examinations throughout (93). One of these prospective sub-studies in the registry,
Europe was started in 2007 and yielded first results in 2008. called “suspected-CAD”, is currently ongoing and will
In more than 11,000 patients included in 2008, approxi- assess the impact of ischemia burden on patient outcome
mately one-third of all examinations were performed to and thus, will yield the important information on how
assess myocardial ischemia (131). much of ischemic myocardium can be treated medically
In this registry, 7,285 patients were studied for suspected and at what amount of ischemia revascularization would
CAD until March 2011. In these patients first-pass perfu- be required.
sion CMR proved to be very safe with only seven patients In combination with well-trained personnel, which will
experiencing an adverse reaction all of which being mild to perform standardized MR perfusion studies, the data qual-
moderate (none severe). In the pilot population of 2,717 ity will most likely allow for (semi)-automatic analysis. Such
patients examined for suspected CAD, cost-effectiveness a perfusion assessment, based on a method, which can be
estimations were performed and showed costs savings of repeated without affecting patient safety, could well shift the
23% and 30% for the Swiss and German health care sys- paradigm of “re-action,” that is, testing in case of symp-
tems, respectively, versus outpatient coronary angiography toms and/or complications to that of “action,” that is, moni-
examinations (132). toring disease in a preclinical (asymptomatic) state before
LWBK1209-ch15_p219-238.indd 235 17/05/13 5:22 PM

1 there might be the possibility to even perform revasculariza-

MR single vendor tions in the interventional MR laboratory as the ultimate
Excellent
MR multivendor extension of the one-stop shop.
0.8 Scintigraphy multivendor
References
0.6 Fair 1. Roger VL, Go AS, Lloyd-Jones DM, et al. Executive summary: Heart disease and stroke
statistics—2012 update: A report from the American Heart Association. Circulation.
Kappa
2012;125:188–197.
2. Schwitter J, Lombardi M, Plein S. Chapter 4: Coronary artery disease: Ischemia–
perfusion. In: Schwitter J, ed. CMR Update. 2nd ed. Lausanne, Switzerland: J. Schwitter;
0.4 2012:54–73. ISBN: 978-2-8399-1030-9. www.herz-mri.ch.
3. WHO. The atlas of heart disease and stroke. 2002.
Poor 4. Roger V, Go A, Lloyd-Jones D, et al. Heart disease and stroke statistics—2011 update: A
report from the American Heart Association. Circulation. 2011;123:e18–e209.
0.2 5. Hoffmann R, Marwick TH, Poldermans D, et al. Refinements in stress echocardiographic
techniques improve inter-institutional agreement in interpretation of dobutamine stress
echocardiograms. Eur. Heart J. 2002;23:821–829.
6. Hendel RC, Berman DS, Cullom SJ, et al. Multicenter clinical trial to evaluate the efficacy
Chance of correction for photon attenuation and scatter in SPECT myocardial perfusion imaging.
0 Circulation. 1999;99:2742–2749.
No Uniform Uniform 7. Taillefer R, DeDuey EG, Udelson JE, et al. Comparative diagnostic accuracy of Tl-201
standard display quantification and Tc-99m sestamibi SPECT imaging (perfusion and ECG-gated SPECT) in detecting
visual visual coronary artery disease in women. J Am Coll Cardiol. 1997;29:69–77.
8. Shaw LJ, Mieres JH, Hendel RH, et al. Comparative effectiveness of exercise electro-
Figure 15.10. A high reproducibility of perfusion data analysis cardiography with or without myocardial perfusion single photon emission computed
tomography in women with suspected coronary artery disease: Results from the What
is crucial for a broader application of a diagnostic technique in both,
is the Optimal Method for Ischemia Evaluation in Women (WOMEN) trial. Circulation.
clinical routine and research. Kappa values for inter-reader agreement 2011;124:1239–1249.
are given for three approaches, involving no standardization of visual 9. Giang T, Nanz D, Coulden R, et al. Detection of coronary artery disease by magnetic reso-
readings, some standardization for visual reading procedures, and uni- nance myocardial perfusion imaging with various contrast medium doses: First European
multi-center experience. Eur Heart J. 2004;25:1657–1665.
form quantitative algorithms (semi-automatic quantitative approach
10. Wolff S, Schwitter J, Coulden R, et al. Myocardial first-pass perfusion magnetic resonance
yielding upslope values). In this comparison visual readings of mul- imaging: A multicenter dose-ranging study. Circulation. 2004;110:732–737.
ticenter, multivendor data with some standardizations performed 11. Schwitter J, Bauer W, van Rossum AC, et al. MR perfusion imaging for detection coronary
poorly (red bar) (11), while single-vendor data yielded poor to fair artery disease: A multicenter dose finding study in comparison with X-ray coronary angi-
ography and single photon emission computed tomography. Circulation. 2004;110:643.
results (green bar, middle column) (10). For a semi-automatic analysis
12. Schwitter J, Wacker C, Wilke N, et al. MR-IMPACT II. Magnetic Resonance Imaging
of multicenter, single-vendor data (green bar, right), however, fair to for Myocardial Perfusion Assessment in Coronary Artery Disease Trial Perfusion-
good results were achieved (9). Similar observations were reported for CMR versus Single Photon Emission Computed Tomography for the Detection of
multicenter, multivendor scintigraphic data (blue bars). (Reproduced Coronary Artery Disease: A Comparative Multicenter, Multivendor Trial. Eur Heart
J. 2012;34:775–781.
with permission from Giang T, Nanz D, Coulden R, et al. Detection
13. Celermajer DS, Sorensen KE, Gooch VM, et al. Non-invasive detection of endothelial dys-
of coronary artery disease by magnetic resonance myocardial perfu- function in children and adults at risk of atherosclerosis. Lancet. 1992;340:1111–1115.
sion imaging with various contrast medium doses: First European 14. Suwaidi JA, Hamasaki S, Higano ST, et al. Long-term follow-up of patients with mild
multicenter experienc. Eur Heart J. 2004;25:1657–1665; Wolff S, coronary artery disease and endothelial dysfunction. Circulation. 2000;101:948–954.
15. Clarkson P, Celermajer DS, Donald AE, et al. Impaired vascular reactivity in insulin-
Schwitter J, Coulden R, et al. Myocardial first-pass perfusion magnetic
dependent diabetes mellitus is related to disease duration and low density lipoprotein
resonance imaging: A multicenter dose-ranging study. Circulation. cholesterol levels. J Am Coll Cardiol. 1996;28:573–579.
2004;110:732–737; Schwitter J, Bauer W, van Rossum AC, et al. MR 16. Lefer AM, Ma XL. Decreased basal nitric oxide release in hypercholesterolemia increases
perfusion imaging for detection coronary artery disease: A multicenter neutrophil adherence to rabbit coronary artery endothelium. Arterioscler Thromb.
1993;13:771–776.
dose finding study in comparison with x-ray coronary angiography
17. Libby P. Molecular bases of the acute coronary syndromes. Circulation. 1995;91:2844–2850.
and single photon emission computed tomography. Circulation. 18. Galis ZS, Sukhova GK, Lark MW, et al. Increased expression of matrix metalloprotein-
2004;110:643; and Wackers FJT, Bodenheimer M, Fleiss JL, et al. ases and matrix degrading activity in vulnerable regions of human atherosclerotic plaques.
Factors affecting uniformity in interpretation of planar Thallium-201 J Clin Invest. 1994;94:2493–2503.
19. Finn AV, Nakano M, Polavarapu R, et al. Hemoglobin directs macrophage differentiation
imaging in a multicenter trial. J Am Coll Cardiol. 1993;21:1064–1074.)
and prevents foam cell formation in human atherosclerotic plaques. J Am Coll Cardiol.
2012;59:166–177.
20. Finn AV, Nakano M, Narula J, et al. Concept of vulnerable/unstable plaque. Arterioscl
Thromb Vasc Biol. 2010;30:1282–1292.
c omplications occur. This could reduce the number of acute 21. Davies MJ. A macro and micro view of coronary vascular insult in ischemic heart disease.
Circulation. 1990;82:II38–II46.
myocardial infarctions as a first manifestation of CAD, 22. Bruschke AV, Kramer JR, Jr., Bal ET, et al. The dynamics of progression of coronary
which still show a high mortality despite novel treatment atherosclerosis studied in 168 medically treated patients who underwent coronary arteri-
strategies, and which are costly as well. Monitoring CAD ography three times. Am Heart J. 1989;117:296–305.
23. Wilcox JN, Smith KM, Schwartz SM, et al. Localization of tissue factor in the normal ves-
would also allow to assess the effect of rigorous modifica- sel wall and in the atherosclerotic plaque. Proc Natl Acad Sci U S A. 1989;86:2839–2843.
tions of risk factors and to detect recurrent and/or progres- 24. Waters D, Lesperance J, Francetich M, et al. A controlled clinical trial to assess the effect
sive CAD in patients after revascularizations. of a calcium channel blocker on the progression of coronary atherosclerosis. Circulation.
1990;82:1940–1953.
Today, a robust perfusion-CMR approach is the back- 25. Ojio S, Takatsu H, Tanaka T, et al. Considerable time from the onset of plaque rupture
bone for the workup of patients with known or suspected and/or thrombi until the onset of acute myocardial infarction in humans: Coronary angio-
graphic findings within 1 week before the onset of infarction. Circulation. 2000;102:2063–
CAD, and it is combined with functional and viability tests 2069.
to yield a comprehensive evaluation of patients with known 26. Naghavi M, Libby P, Falk E, et al. From vulnerable plaque to vulnerable patient: A call for
or suspected CAD. new definitions and risk assessment strategies: Part I. Circulation. 2003;108:1664–1672.
27. Seiler C, Kirkeeide RL, Gould KL. Measurement from arteriograms of regional myocar-
Further studies will evaluate, to what extent perfusion dial bed size distal to any point in the coronary vascular tree for assessing anatomic area
imaging at 3 T would increase test performance. Finally, at risk. J Am Coll Cardiol. 1993;21:783–797.
LWBK1209-ch15_p219-238.indd 236 17/05/13 5:22 PM

28. Ladenheim ML, Pollock BH, Rozanski A, et al. Extent and severity of myocardial hypo- 56. Wendland MF, Saeed M, Lauerma K, et al. Endogenous susceptibility contrast in myocar-
perfusion as predictors of prognosis in patients with suspected coronary artery disease. J dium during apnea measured using gradient recalled echo planar imaging. Magn Reson
Am Coll Cardiol. 1986;7:464–471. Med. 1993;29:273–276.
29. Iskander S, Iskandrian AE. Risk assessment using single-photon emission computed tomo- 57. Belle V, Kahler E, Waller C, et al. In vivo quantitative mapping of cardiac perfusion in
graphic technetium-99m sestamibi imaging. J Am Coll Cardiol. 1998;32:57–62. rats using a noninvasive MR spin-labeling method. J Magn Reson Imaging. 1998;8:
30. Hachamovitch R, Rozanski A, Shaw LJ, et al. Impact of ischaemia and scar on the 1240–1245.
therapeutic benefit derived from myocardial revascularization vs. medical therapy 58. Prasad PV, Burstein D, Edelman RR. MRI evaluation of myocardial perfusion without a
among patients undergoing stress-rest myocardial perfusion scintigraphy. Eur Heart J. contrast agent using magnetization transfer. Magn Reson Med. 1993;30:267–270.
2011;32:1012–1024. 59. Saeed M, Wendland MF, Sakuma H, et al. Coronary artery stenosis: Detection with con-
31. Pijls NHJ, van Schaardenburgh P, Manoharan G, et al. Percutaneous coronary interven- trast-enhanced MR imaging in dogs. Radiology. 1995;196:79–84.
tion of functionally nonsignificant stenosis: 5-year follow-up of the DEFER study. J Am 60. Schwitter J, Saeed M, Wendland MF, et al. Assessment of myocardial function and per-
Coll Cardiol. 2007;49:2105–2111. fusion in a canine model of non-occlusive coronary artery stenosis using fast magnetic
32. Fearon WF, Bornschein B, Tonino PAL, et al. Economic evaluation of fractional flow resonance imaging. JMRI. 1998;9:101–110.
reserve-guided percutaneous coronary intervention in patients with multivessel disease. 61. Manning WJ, Atkinson DJ, Grossman W, et al. First-pass nuclear magnetic resonance
Circulation. 2010;122:2545–2550. imaging studies using gadolinium-DTPA in patients with coronary artery disease. J Am
33. Schwitter J, Nanz D, Kneifel S, et al. Assessment of myocardial perfusion in coronary Coll Cardiol. 1991;18:959–965.
artery disease by magnetic resonance: A comparison with positron emission tomography 62. Lauerma K, Virtanen KS, Sipila LM, et al. Multislice MRI in assessment of myocardial
and coronary angiography. Circulation. 2001;103:2230–2235. perfusion in patients with single-vessel proximal left anterior descending coronary artery
34. Plein S, Radjenovic A, Ridgway JP, et al. Coronary artery disease: Myocardial perfu- disease before and after revascularization. Circulation. 1997;96:2859–2867.
sion MR imaging with sensitivity encoding versus conventional angiography. Radiology. 63. Matheijssen NA, Louwerenburg HW, van Rugge F, et al. Comparison of ultrafast dipyri-
2005;235:423–430. damole magnetic resonance imaging with dipyridamole SestaMIBI SPECT for detection
35. Paetsch I, Jahnke C, Wahl A, et al. Comparison of dobutamine stress magnetic resonance, of perfusion abnormalities in patients with one-vessel coronary artery disease: Assessment
adenosine stress magnetic resonance, and adenosine stress magnetic resonance perfusion. by quantitative model fitting. Magn Reson Med. 1996;35:221–228.
Circulation. 2004;110:835–842. 64. Eichenberger AC, Schuiki E, Kochli VD, et al. Ischemic heart disease: Assessment with
36. Nagel E, Klein C, Paetsch I, et al. Magnetic resonance perfusion measurements for the gadolinium-enhanced ultrafast MR imaging and dipyridamole stress. J Magn Reson
noninvasive detection of coronary artery disease. Circulation. 2003;108:432–437. Imaging. 1994;4:425–431.
37. Ishida N, Sakuma H, Motoyasu M, et al. Noninfarcted myocardium: Correlation between 65. Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of myocardial ischemia
dynamic first-pass contrast-enhanced myocardial MR imaging and quantitative coronary from perfusion reserve based on cardiovascular magnetic resonance. Circulation.
angiography. Radiology. 2003;229:209–216. 2000;101:1379–1383.
38. Plein S, Kozerke S, Suerder D, et al. High spatial resolution myocardial perfusion car- 66. Sensky PR, Jivan A, Hudson NM, et al. Coronary artery disease: Combined stress MR
diac magnetic resonance for the detection of coronary artery disease. Eur Heart J. imaging protocol-one-stop evaluation of myocardial perfusion and function. Radiology.
2008;29:2148–2155. 2000;215:608–614.
39. Greenwood J, Maredia N, Younger J, et al. Cardiovascular magnetic resonance and 67. Walsh EG, Doyle M, Lawson MA, et al. Multislice first-pass myocardial perfusion imag-
single-photon emission computed tomography for diagnosis of coronary heart disease ing on a conventional clinical scanner. Magn Reson Med. 1995;34:39–47.
(CE-MARC): A prospective trial. Lancet. 2012;379:453–460. 68. Ding S, Wolff SD, Epstein FH. Improved coverage in dynamic contrast-enhanced cardiac
40. Manka R, Jahnke C, Kozerke S, et al. Dynamic 3-dimensional stress cardiac magnetic MRI using interleaved gradient-echo EPI. Magn Reson Med. 1998;39:514–519.
resonance perfusion imaging: Detection of coronary artery disease and volumetry of 69. Kraitchman DL, Chin BB, Heldman AW, et al. MRI detection of myocardial perfusion
myocardial hypoenhancement before and after coronary stenting. J Am Coll Cardiol. defects due to coronary artery stenosis with MS-325. J Magn Reson Imaging. 2002;15:
2011;57:437–444. 149–158.
41. Rieber J, Huber A, Erhard I, et al. Cardiac magnetic resonance perfusion imaging for the 70. Slavin GS, Wolff SD, Gupta SN, et al. First-pass myocardial perfusion MR imaging with
functional assessment of coronary artery disease: A comparison with coronary angiogra- interleaved notched saturation: Feasibility study. Radiology. 2001;219:258–263.
phy and fractional flow reserve. Eur Heart Jl 2006;27:1465–1471. 71. Edelman RR, Li W. Contrast-enhanced echo-planar MR imaging of myocardial perfusion:
42. Lockie T, Ishida M, Perera D, et al. High-resolution magnetic resonance myocardial per- Preliminary study in humans. Radiology. 1994;190:771–777.
fusion imaging at 3.0-Tesla to detect hemodynamically significant coronary stenoses as 72. Schwitter J, Debatin JF, von Schulthess GK, et al. Normal myocardial perfusion assessed
determined by fractional flow reserve. J Am Coll Cardiol. 2011;57:70–75. with multishot echo-planar imaging. Magn Reson Med. 1997;37:140–147.
43. Kitagawa K, Sakuma H, Nagata M, et al. Diagnostic accuracy of stress myocardial perfu- 73. Schwitter J, Sakuma H, Saeed M, et al. Very fast cardiac imaging. Magn Reson Imaging
sion MRI and late gadolinium-enhanced MRI for detecting flow-limiting coronary artery Clin N Am. 1996;4:419–432.
disease: A multicenter study. Eur Radiol. 2008;18:2808–2816. 74. Klocke FJ, Simonetti OP, Judd RM, et al. Limits of detection of regional differences in
44. Schwitter J, Wacker CM, van Rossum AC, et al. MR-IMPACT: Magnetic resonance imaging vasodilated flow in viable myocardium by first-pass magnetic resonance perfusion imag-
for myocardial perfusion assessment in coronary artery disease trial: Comparison of perfusion ing. Circulation. 2001;104:2412–2416.
CMR with single photon emission computed tomography for the detection of coronary artery 75. Schreiber WG, Schmitt M, Kalden P, et al. Dynamic contrast-enhanced myocardial per-
disease in a multicenter, multivendor, randomized trial. Eur Heart J. 2008;29:480–489. fusion imaging using saturation-prepared TrueFISP. J Magn Reson Imaging. 2002;16:
45. Poncelet BP, Koelling TM, Schmidt CJ, et al. Measurement of human myocardial per- 641–652.
fusion by double-gated flow alternating inversion recovery EPI. Magn Reson Med. 76. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: Sensitivity encoding for fast
1999;41:510–519. MRI. Magn Reson Med. 1999;42:952–962.
46. Schwitter J, Wacker CM, Wilke N, et al. Superior diagnostic performance of perfusion-CMR 77. Kellman P, Derbyshire JA, Agyeman KO, et al. Extended coverage of first-pass perfusion
versus SPECT to detect coronary artery disease: The secondary endpoints of the multicenter imaging using slice-interleaved TSENSE. Magn Reson Med. 2004;51:200–204.
multivendor MR-IMPACT II. J Cardiovasc Magn Reson. 2012;14:61–71. 78. Vitanis V, Manka R, Giese D, et al. High resolution three-dimensional cardiac perfusion
47. Uren NG, Melin JA, De Bruyne B, et al. Relation between myocardial blood flow and the imaging using compartment-based k-t principal component analysis. Magn Reson Med.
severity of coronary artery stenosis. N Engl J Med. 1994;330:1782–1788. 2011;65:575–587.
48. Di Carli M, Czernin J, Hoh CK, et al. Relation among stenosis severity, myocardial 79. Plein S, Kozerke S, Suerder D, et al. High spatial resolution myocardial perfusion car-
blood flow, and flow reserve in patients with coronary artery disease. Circulation. diac magnetic resonance for the detection of coronary artery disease. Eur Heart J.
1995;91:1944–1951. 2008;29:2148–2155.
49. Sambuceti G, Parodi O, Marcassa C, et al. Alterations in regulation of myocardial blood 80. Plein S, Schwitter J, Suerder D, et al. k-t SENSE-accelerated myocardial perfusion MR
flow in one-vessel coronary artery disease determined by positron emission tomography. imaging at 3.0 Tesla—comparison with 1.5 Tesla. Radiology. 2008;249:493–500.
Am J Cardiol. 1993;72:538–543. 81. Saeed M, Wendland MF, Yu KK, et al. Dual effects of gadodiamide injection in depiction
50. Schwitter J, DeMarco T, Kneifel S, et al. Magnetic resonance-based assessment of global of the region of myocardial ischemia. J Magn Reson Imaging. 1993;3:21–29.
coronary flow and flow reserve and its relation to left ventricular functional parameters: 82. Sakuma H, O’Sullivan M, Lucas J, et al. Effect of magnetic susceptibility contrast medium
A comparison with positron emission tomography. Circulation. 2000;101:2696–2702. on myocardial signal intensity with fast gradient-recalled echo and spin-echo MR imag-
51. Li D, Dhawale P, Rubin PJ, et al. Myocardial signal response to dipyridamole and dobu- ing: initial experience in humans. Radiology. 1994;190:161–166.
tamine: Demonstration of the BOLD effect using a double-echo gradient-echo sequence. 83. Bruder O, Schneider S, Nothnagel D, et al. Acute adverse reactions to gadolinium-based
Magn Reson Med. 1996;36:16–20. contrast agents in CMR: Multicenter experience with 17,767 patients from the EuroCMR
52. Atalay MK, Reeder SB, Zerhouni EA, et al. Blood oxygenation dependence of T1 and T2 Registry. JACC Cardiovasc Imaging. 2011;4:1171–1176.
in the isolated, perfused rabbit heart at 4.7T. Magn Reson Med. 1995;34:623–627. 84. Wilke N, Kroll K, Merkle H, et al. Regional myocardial blood volume and flow: First-pass
53. Fieno DS, Shea SM, Li Y, et al. Myocardial perfusion imaging based on the blood oxygen MR imaging with polylysine-Gd-DTPA. J Magn Reson Imaging. 1995;5:227–237.
level-dependent effect using T2-prepared steady-state free-precession magnetic resonance 85. Kraitchman DL, Wilke N, Hexeberg E, et al. Myocardial perfusion and function in dogs
imaging. Circulation. 2004;110:1284–1290. with moderate coronary stenosis. Magn Reson Med. 1996;35:771–780.
54. Bertschinger KM, Nanz D, Buechi M, et al. Magnetic resonance myocardial first-pass 86. Bjerner T, Johansson L, Ericsson A, et al. First-pass myocardial perfusion MR imaging
perfusion imaging: Parameter optimization for signal response and cardiac coverage. J with outer-volume suppression and the intravascular contrast agent NC100150 injection:
Magn Reson Imaging. 2001;14:556–562. Preliminary results in eight patients. Radiology. 2001;221:822–826.
55. Reeder SB, Atalay MK, McVeigh ER, et al. Quantitative cardiac perfusion: A non- 87. Saeed M, Wendland MF, Masui T, et al. Reperfused myocardial infarctions on T1- and
invasive spin-labeling method that exploits coronary vessel geometry. Radiology. susceptibility-enhanced MRI: evidence for loss of compartmentalization of contrast
1996;200:177–184. media. Magn Reson Med. 1994;31:31–39.
LWBK1209-ch15_p219-238.indd 237 17/05/13 5:22 PM

88. Hall DA, Maus DC, Gerfen GJ, et al. Polarization-enhanced NMR spectroscopy of bio- 112. Fritz-Hansen T, Rostrup E, Larsson HB, et al. Measurement of the arterial concentra-
molecules in frozen solution. Science. 1997;276:930–932. tion of Gd-DTPA using MRI: A step toward quantitative perfusion imaging. Magn Reson
89. Golman K, Axelsson O, Johannesson H, et al. Parahydrogen-induced polarization in Med. 1996;36:225–231.
imaging: Subsecond 13C angiography. Magn Reson Med. 2001;46:1–5. 113. Larsson HB, Stubgaard M, Sondergaard L, et al. In vivo quantification of the unidirec-
90. Schwitter J. Extending the frontiers of cardiac magnetic resonance. Circulation. tional influx constant for Gd-DTPA diffusion across the myocardial capillaries with MR
2008;118:109–112. imaging. J Magn Reson Imaging. 1994;4:433–440.
91. Johansson E, Mansson S, Wirestam R, et al. Cerebral perfusion assessment by bolus 114. Kroll K, Wilke N, Jerosch Herold M, et al. Modeling regional myocardial flows from
tracking using hyperpolarized 13C. Magn Reson Med. 2004;51:464–472. residue functions of an intravascular indicator. Am J Physiol. 1996;271:H1643–H1655.
92. Hartnell G, Cerel A, Kamalesh M, et al. Detection of myocardial ischemia: value of com- 115. Gatehouse PD, Elkington AG, Ablitt NA, et al. Accurate assessment of the arterial input
bined myocardial perfusion and cineangiographic MR imaging. AJR Am J Roentgenol. function during high-dose myocardial perfusion cardiovascular magnetic resonance. J
1994;163:1061–1067. Magn Reson Imaging. 2004;20:39–45.
93. Muzarelli S, Locca D, Lauriers N, et al. Quality Assessment of CMR Images in the 116. Wendland MF, Saeed M, Yu KK, et al. Inversion recovery EPI of bolus transit in rat
Setting of the European Cardiovascular Magnetic Resonance Registry: Description and myocardium using intravascular and extravascular gadolinium-based MR contrast media:
Validation of Standardized Criteria. J Cardiovasc Magn Reson. 2013; in press. Dose effects on peak signal enhancement. Magn Reson Med. 1994;32:319–329.
94. Keijer JT, van Rossum AC, Wilke N, et al. Magnetic resonance imaging of myocardial 117. Judd RM, Atalay MK, Rottman GA, et al. Effects of myocardial water exchange on T1
perfusion in single-vessel coronary artery disease: Implications for transmural assessment enhancement during bolus administration of MR contrast agents. Magn Reson Med.
of myocardial perfusion. J Cardiovasc Magn Reson. 2000;2:189–200. 1995;33:215–223.
95. Swets JA. Measuring the accuracy of diagnostic systems. Science. 1988;240:1285–1288. 118. Judd RM, Reeder SB, May Newman K. Effects of water exchange on the measure-
96. Metz CE. ROC methodology in radiologic imaging. Invest Radiol. 1986;21:720–733. ment of myocardial perfusion using paramagnetic contrast agents. Magn Reson Med.
97. Holland AE, Goldfarb JW, Edelman RR. Diaphragmatic and cardiac motion during sus- 1999;41:334–342.
pended breathing: preliminary experience and implications for breath-hold MR imaging. 119. Donahue KM, Burstein D, Manning WJ, et al. Studies of Gd-DTPA relaxivity and proton
Radiology. 1998;209:483–489. exchange rates in tissue. Magn Reson Med. 1994;32:66–76.
98. McConnell MV, Khasgiwala VC, Savord BJ, et al. Prospective adaptive navigator correc- 120. Larsson HB, Rosenbaum S, Fritz Hansen T. Quantification of the effect of water exchange
tion for breath-hold MR coronary angiography. Magn Reson Med. 1997;37:148–152. in dynamic contrast MRI perfusion measurements in the brain and heart. Magn Reson
99. Chuang ML, Chen MH, Khasgiwala VC, et al. Adaptive correction of imaging plane Med. 2001;46:272–281.
position in segmented k-space cine cardiac MRI. J Magn Reson Imaging. 1997;7: 121. Dauber IM, VanBenthuysen KM, McMurtry IF, et al. Functional coronary microvascular
811–814. injury evident as increased permeability due to brief ischemia and reperfusion. Circ Res.
100. Keijer JT, van Rossum AC, van Eenige MJ, et al. Magnetic resonance imaging of regional 1990;66:986–998.
myocardial perfusion in patients with single-vessel coronary artery disease: Quantitative 122. Svendsen JH, Bjerrum PJ, Haunso S. Myocardial capillary permeability after regional
comparison with (201)Thallium-SPECT and coronary angiography. J Magn Reson ischemia and reperfusion in the in vivo canine heart. Effect of superoxide dismutase. Circ
Imaging. 2000;11:607–615. Res. 1991;68:174–184.
101. Wilke N, Simm C, Zhang J, et al. Contrast-enhanced first pass myocardial perfusion imag- 123. Fritz-Hansen T, Rostrup E, Sondergaard L, et al. Capillary transfer constant of Gd-DTPA
ing: Correlation between myocardial blood flow in dogs at rest and during hyperemia. in the myocardium at rest and during vasodilation assessed by MRI. Magn Reson Med.
Magn Reson Med. 1993;29:485–497. 1998;40:922–929.
102. Panting JR, Gatehouse PD, Yang GZ, et al. Abnormal subendocardial perfusion in car- 124. Cullen JH, Horsfield MA, Reek CR, et al. A myocardial perfusion reserve index in
diac syndrome X detected by cardiovascular magnetic resonance imaging. N Engl J Med. humans using first-pass contrast-enhanced magnetic resonance imaging. J Am Coll
2002;346:1948–1953. Cardiol. 1999;33:1386–1394.
103. Keijer JT, van Rossum A, van Eenige M, et al. Semiquantitation of regional myocardial 125. Neyran B, Janier MF, Casali C, et al. Mapping myocardial perfusion with an intravascular
blood flow in normal human subjects by first-pass magnetic resonance imaging. Am Heart MR contrast agent: Robustness of deconvolution methods at various blood flows. Magn
J. 1995;130:893–901. Reson Med. 2002;48:166–179.
104. Lombardi M, Jones RA, Westby J, et al. Use of the mean transit time of an intravascular 126. Bjornerud A, Bjerner T, Johansson LO, et al. Assessment of myocardial blood volume
contrast agent as an exchange-insensitive index of myocardial perfusion. J Magn Reson and water exchange: Theoretical considerations and in vivo results. Magn Reson Med.
Imaging. 1999;9:402–408. 2003;49:828–837.
105. Christian TF, Rettmann DW, Aletras AH, et al. Absolute myocardial perfusion in 127. Wendland MF, Saeed M, Masui T, et al. Echo-planar MR imaging of normal and ischemic
canines measured by using dual-bolus first-pass MR imaging. Radiology. 2004;232: myocardium with gadodiamide injection. Radiology. 1993;186:535–542.
677–684. 128. Wendland MF, Saeed M, Masui T, et al. First pass of an MR susceptibility contrast agent
106. Kety S. The theory and application of the exchange of inert gas in lung and tissues. through normal and ischemic heart: Gradient-recalled echo-planar imaging. J Magn
Pharmacol Rev. 1951;3:1–41. Reson Imaging. 1993;3:755–760.
107. Jerosch Herold M, Hu X, Murthy NS, et al. Magnetic resonance imaging of myocardial 129. Wright KB, Klocke FJ, Deshpande VS, et al. Assessment of regional differences in myocardial
contrast enhancement with MS-325 and its relation to myocardial blood flow and the blood flow using T2-weighted 3D BOLD imaging. Magn Reson Med. 2001;46:573–578.
perfusion reserve. J Magn Reson Imaging. 2003;18:544–554. 130. Wackers FJT, Bodenheimer M, Fleiss JL, et al. Factors affecting uniformity in inter-
108. Jerosch-Herold M, Swingen C, Seethamraju RV. Myocardial blood flow quantification pretation of planar Thallium-201 imaging in a multicenter trial. J Am Coll Cardiol.
with MRI by model-independent deconvolution. Med Phys. 2002;29:886–897. 1993;21:1064–1074.
109. Larsson HB, Fritz Hansen T, Rostrup E, et al. Myocardial perfusion modeling using MRI. 131. Bruder O, Schneider S, Nothnagel D, et al. EuroCMR (European Cardiovascular
Magn Reson Med. 1996;35:716–726. Magnetic Resonance) Registry: Results of the German Pilot Phase. J Am Coll Cardiol.
110. Diesbourg L, Prato F, Wisenberg G, et al. Quantification of myocardial blood flow and 2009;54:1457–1466.
extracellular volumes using a bolus injection of Gd-DTPA: Kinetic modelling in canine 132. Moschetti K, Muzzarelli S, Pinget C, et al. Cost evaluation of cardiac magnetic resonance
ischemic disease. Mag Reson Med. 1992;23:239–253. imaging versus coronary angiography for the diagnostic work-up of coronary artery dis-
111. Tong CY, Prato FS, Wisenberg G, et al. Measurement of the extraction efficiency and disease: Application of the European Cardiovascular Magnetic Resonance registry data to
tribution volume for Gd-DTPA in normal and diseased canine myocardium. Magn Reson the German, United Kingdom, Swiss, and United States health care systems. J Cardiovasc
Med. 1993;30:337–346. Magn Reson. 2012:35–44.
LWBK1209-ch15_p219-238.indd 238 17/05/13 5:22 PM

Chapter
16
Rolf Gebker
Andreas Schuster
Eike Nagel
Assessment of Ischemic Heart Disease

Utilizing Dobutamine
■■ Introduction such as bypass surgery or angioplasty. Even more impor-

tantly, it has been shown that revascularization procedures in
■■ Pathophysiology
patients without flow-limiting coronary artery stenoses result
■■ How to Perform Dobutamine Stress Testing in a higher event rate than optimal medical therapy alone
Ergometric Stress (1,2). As a consequence, current guidelines have stressed the
Pharmacologic Stress importance of documenting ischemia, preferably using non-
Image Acquisition invasive functional testing before elective invasive procedures
(3). Thus, a noninvasive test with a high rate of diagnostic
Safety
accuracy is required to optimally guide patient management
Image Interpretation and avoid unnecessary invasive tests or potentially harmful
■■ Accuracy of Stress-Induced Wall-Motion procedures.
Abnormalities in the Diagnosis of Ischemia CVMR has evolved into a new technique for the nonin-
vasive detection of obstructive CAD. The ability of CVMR
■■ Additional Value of Perfusion Imaging
to visualize global and regional wall motion and systolic
During Dobutamine Stress Testing thickening of the left ventricle (LV) with a high degree of
■■ Prognostic Value of Dobutamine Stress spatial and temporal resolutions makes it possible to detect
Magnetic Resonance Imaging even subtle abnormalities of wall motion. In addition, perfu-
sion defects and reductions in coronary flow reserve can be
■■ Quantitative Methods for Evaluating Wall
assessed. Except for high-grade coronary artery stenoses, ab-
Motion normalities can, for the most part, be identified only under
■■ Viability Imaging with Dobutamine Stress stress conditions. These can be induced by physical exercise
Magnetic Resonance or by means of standardized stress protocols with infusions
of pharmacologic agents such as dobutamine/atropine,
■■ Summary dipyridamole, or adenosine. To date, the most reliable clini-
cal data are either based on the analysis of LV wall motion
and thickening during dobutamine stress or on the analysis
Introduction of first pass perfusion during adenosine stress. In this chap-
ter, the evidence and a detailed description of how to utilize
Ischemic heart disease is one of the most common health dobutamine as a pharmacologic agent to assess both viabil-
problems of the Western world. A variety of tests are avail- ity and ischemia are presented.
able in routine clinical practice for the noninvasive diagno-
sis of coronary artery disease (CAD) such as exercise elec-
trocardiography (ECG), echocardiography, single photon Pathophysiology
emission computed tomography (SPECT), positron emission
tomography (PET), and cardiovascular magnetic resonance Ischemic heart disease can be caused by a variety of patho-
(CVMR) imaging. The advantages and disadvantages of each physiologic conditions. Stenosis of the coronary arteries is
technique are summarized in Table 16.1. Many noninvasive the most common of these, but LV hypertrophy, alterations
diagnostic tools are suboptimal and both patients and physi- of the microcirculation, and a reduction of energy uptake are
cians want a reliable diagnosis. Consequently, 40% to 60% other less frequent causes. Patients with CAD usually have
of all patients who undergo invasive cardiac catheterization sufficient blood flow at rest; however, during stress, which
procedures do not require a revascularization procedure induces a four- to fivefold increase in blood flow in healthy
239
LWBK1209-ch16_p239-250.indd 239 16/05/13 9:38 PM

ifferent Techniques for the Evaluation of

D
Table 16.1
Myocardial Ischemia
Advantage Disadvantage
Exercise ECG Low cost Low diagnostic accuracy

Wide availability Many equivocal results
Target heart rate not reached in many
patients
Stress echocardiography Low cost Nondiagnostic examinations in 10–15%
Wide availability
SPECT Low cost Radiation
Low sensitivity
Attenuation artifacts, especially in women
PET Quantitative evaluation Limited availability
No attenuation artifacts High cost
Radiation
CMR High diagnostic accuracy Limited availability
Combination of wall motion, High cost
perfusion, and late gadolinium
enhancement (LGE)
No radiation
ECG, electrocardiography; SPECT, single photon emission computed tomography; PET, positron emission
tomography; CMR, cardiovascular magnetic resonance imaging.
persons, the myocardium supplied by the stenotic coronary such as peripheral vascular or degenerous musculoskeletal
arteries does not receive enough blood because blood flow disease or due to chronic deconditioning. As a result, a large
is impeded through the narrowed coronary artery lumen. number of tests may be nondiagnostic because of an inad-
Thus, except in very severe cases in which the patients have equate level of stress or image quality. Thus, the most robust
ischemia at rest, stress testing is required to induce ischemia. approach for stress testing within the CVMR environment is
pharmacologic stress testing.
How to Perform Dobutamine
Stress Testing Pharmacologic Stress
Pharmacologic stress is preferred by many clinicians because
Stress testing can be performed with either physical or phar-
the results are highly reproducible and diagnostic in most
macologic stimulation. In general, to detect ischemic heart
patients. Careful monitoring is required for patient safety.
disease with a high level of sensitivity and reproducibility, a
Two different approaches are used: Oxygen consumption
defined endpoint (submaximal stress) must be reached. This
can be increased by increasing the heart rate and contrac-
endpoint is defined by the heart rate as follows:
tility (dobutamine), or vasodilation with the induction of
Target heart rate = 0.85(220 – Age) regional flow heterogeneities (steal effect) can be induced
(dipyridamole/adenosine). The different pharmacologic
Ergometric Stress stress protocols are summarized in Table 16.2.
Dobutamine is a sympathomimetic drug with β1-, β2-,
Ergometric stress is the most physiologic stress test and thus and slight α1-receptor-stimulating properties. Infusion of
generally the preferred method for testing. Patients exercise the drug increases the cardiac contractility and rate and
on a bicycle or treadmill ergometer with incremental work- decreases the systolic vascular resistance. During low-dose
loads. infusion (≤10 μg/kg/min), an increase in contractility is the
Despite the fact that ergometric stress in combination major effect, whereas at higher doses, the augmented con-
with CVMR has been shown to be feasible both for using sumption of oxygen causes contraction abnormalities in
a magnetic resonance (MR)-compatible bicycle while lying myocardial segments supplied by stenotic coronary arteries.
on the examination table or a treadmill placed outside the A low dose of dobutamine is defined as an infusion of
5-Gauss line inside the MRI room, there are several limita- up to 10 μg/min/kg of body weight. Such a dose is suffi-
tions associated with these techniques (4,5). Patients have cient to stimulate myocardium that does not contract at rest
to stop exercising and must be moved back into the bore but may benefit from revascularization (viable or hibernat-
for scanning, which may cause the heart rate to fall below ing myocardium); however, this dose is insufficient to induce
the target heart rate thus reducing sensitivity. Due to physi- ischemia.
cal exercise, breathing is generally more labored and breath- A high dose of dobutamine is defined as an infusion
holding more challenging, which may introduce motion ar- aimed at inducing ischemia. As previously explained, it is es-
tifacts. Most importantly; however, patients are often not sential that patients reach their target heart rate. To achieve
able to reach the anticipated target because of comorbidities this goal, atropine, which increases the heart rate by an
LWBK1209-ch16_p239-250.indd 240 16/05/13 9:38 PM

Chapter 16 ■ Assessment of Ischemic Heart Disease Utilizing Dobutamine 241
Table 16.2 Stress Protocols
Stress Test Patient Instructions Protocol Antidote
Dobutamine for the 5, 10 μg/kg BW per minute for

assessment of viability >3 min
Dobutamine/atropine No β-blockers and nitrates 24 h 5, 10, 20, 30, 40 μg/kg BW per β-blocker (esmolol) 0.5 mg/kg as
for the detection of before the examination minute for 3 min each, up to slowly injected bolus, additional
myocardial ischemia (wall 2 mg atropine (8 × 0.25 mg) bolus of 0.2 mg/kg as needed,
motion and perfusion) until submaximal heart sublingual nitroglycerin
rate [(220 − age) × 0.85] is
reached (half-life 2 min)
Dipyridamole (perfusion) No caffeine (tea, coffee, 0.56 mg/kg BW per minute for Aminophylline 250 mg IV slowly
aminophylline 250 mg chocolate, 4 min, maximal effect after injected with ECG monitoring,
etc.) or medications such as approximately 3–4 min (half- sublingual nitroglycerin
aminophylline or nitrates 24 h life 30 min)
before the examination
Adenosine (perfusion) Same as for dipyridamole 140–210 μg/kg BW per minute Stop infusion (in occasional cases
for 3–6 min (half-life 4–10 s) aminophylline 250 mg IV slowly
injected with ECG monitoring)
BW, body weight; ECG, electrocardiogram.
a nticholinergic mechanism, is commonly added. The stress eation of the endocardial border, especially in long-axis
protocol most widely used is described in the guidelines views (7). The in-plane spatial resolution is in the range of
of the Society of Cardiovascular Magnetic Resonance (6). 1.6 × 1.6 mm with a slice thickness of 6 to 8 mm. The heart
Stress is induced by increasing doses of dobutamine, started can be visualized with either contiguous short-axis slices or
at 10 μg/kg of body weight per minute for 3 minutes and in- a combination of several short-axis (typically basal, mid,
creased in increments of 10 μg/kg of body weight per minute and apical slice) and long-axis (typically two-, three-, and
every 3 minutes, until a maximal dose of 40 μg/kg of body four-chamber) views (Fig. 16.1).
weight per minute is reached. (Some investigators use 50 μg/ Imaging is performed at rest and then immediately after
kg of body weight per minute as a maximum.) If the target the dobutamine dose is increased for each stress level. Images
heart rate is not reached, up to 2 mg of atropine is added in are acquired and reviewed immediately at rest and during
0.25-mg fractions. The test must be stopped if certain crite- each stress level to detect new wall-motion abnormalities
ria are fulfilled (Table 16.3). during stress. The stress protocol, details of monitoring,
contraindications, and termination criteria are summarized
in Tables 16.2–16.5. Dobutamine stress magnetic resonance
Image Acquisition
(DSMR) is generally performed on 1.5-T scanners. With the
For the assessment of wall motion, cine loops of the heart advent of 3 T for cardiac imaging, there has been a growing
are acquired with steady state free precession (SSFP) usu- interest to perform high-dose dobutamine stress protocols
ally in combination with parallel image acquisition and under these high field conditions. However, this approach
retrospective ECG gating, resulting in 25 to 50 phases per has been hampered by an increase in artifacts (banding,
cardiac cycle up to heart rates of 200 beats per minute dur- flow-related artifacts), especially applying SSFP imaging
ing a breath-hold of 6 to 10 seconds. In contrast to turbo during higher heart rates. Turbo gradient-echo cine imaging
gradient-echo techniques that have been used in older stud- with administration of contrast agent offers more robust im-
ies, imaging contrast in SSFP results in a much better delin- aging under these conditions (8). More recently, the develop-
ment of dual-source RF transmission with patient-adaptive
local RF shimming has led to an improvement in SSFP image
obutamine Termination
D quality during DSMR at 3 T (9).
Table 16.3
Criteria
Safety
Submaximal heart rate reached [(220 − age) × 0.85]
Systolic pressure decrease >20 mm Hg below baseline systolic During stress examinations in which low or high doses of
blood pressure or decrease >40 mm Hg from a previous level dobutamine are administered, monitoring the patient within
Blood pressure increase >240/120 mm Hg the magnet is mandatory. In general, monitoring during a
Intractable symptoms CVMR examination requires that the same precautions be
New or worsening wall-motion abnormalities in at least 2 of 17 taken and the same emergency equipment be available as
adjacent left ventricular segments
in any other stress examination. A physician trained in car-
Complex cardiac arrhythmias
diovascular emergencies and resuscitation must be at the
LWBK1209-ch16_p239-250.indd 241 16/05/13 9:38 PM

SAX 4ch 2ch 3ch
Rest
ED
Rest
ES
Peak
stress
ES
Figure 16.1. DSMR in a patient with new onset angina after coronary bypass grafting and resting wall-
motion abnormalities of the inferolateral basal wall. During high-dose dobutamine stress, the patient develops
an extensive stress-inducible wall-motion abnormality of the anterior/anterolateral wall from the base to the
apex (white arrows). Invasive coronary angiography demonstrated extensive coronary disease with a high-
grade stenosis of the LAD distal to the anastomosis of the LIMA-graft (black arrow). SAX, short axis views;
ED, end diastolic frame; ES, end systolic frame.
scanner. Apart from specific contraindications to CVMR high-dose DSMR in 1,000 consecutive patients and showed
such as the presence of retro-orbital metal, cerebral clips, or a safety profile virtually identical to that previously reported
non-MRI compatible pacemakers, the contraindications are for dobutamine stress echocardiography. One patient suffered
identical to those for stress echocardiography and are listed sustained ventricular tachycardia and emergent defibrillation
in Table 16.4. was carried out, while no cases of death or myocardial in-
In a routine clinical setting, DSMR proved to be feasible farction occurred over the years (Table 16.6). Approximately
and safe, and it resulted in a high number of diagnostic ex- 50% of the patients had inducible ischemia during stress
aminations (89.5%) in patients without contraindications to CVMR, thereby closely reflecting the clinical practice in a
MRI (10). Wahl et al. (10) reported a 5-year experience of tertiary center. Data from the large Euro CVMR registry has
LWBK1209-ch16_p239-250.indd 242 16/05/13 9:38 PM

ontraindications to
C oderate and Severe Side
M
Table 16.4 Magnetic Resonance Stress Table 16.6 Effects in 1,000 Patients,
Tests Including Positive Tests
MR examination Incompatible metallic implants (e.g., Sustained VT 1 (0.1%)
pacemakers, retro-orbital metal, Nonsustained VT 4 (0.4%)
cerebral artery clips) Paroxysmal atrial fibrillation 16 (1.6%)
Claustrophobia Temporary AV block II 2:1 2 (0.2%)
Dobutamine Severe arterial hypertension RR-increase (>240/120) 5 (0.5%)
(>220/120 mm Hg) RR-decrease >40 mm Hg 5 (0.5%)
Unstable angina pectoris Nausea 31 (3.1%)
Significant aortic stenosis (aortic valve Total 64 (6.4%)
gradient >50 mm Hg or aortic valve
area >1 cm2) With permission from: Wahl A, Paetsch I, Gollesch A, et al. Safety
Complex cardiac arrhythmias and feasibility of high-dose dobutamine–atropine stress cardiovas-
Significant hypertrophic obstructive cular magnetic resonance for diagnosis of myocardial ischaemia:
cardiomyopathy Experience in 1000 consecutive cases. Eur Heart J 2004;25(14):
Myocarditis, endocarditis, pericarditis 1230–1236.
Other major disease
Dipyridamole/adenosine Myocardial infarction <3 d
Unstable angina pectoris
Severe arterial hypertension the radiofrequency cage or special CVMR-compatible equip-
Asthma or severe obstructive ment, available at many CVMR sites, can be used. A defi-
pulmonary disease
brillator and all medications needed for emergency treatment
Atrioventricular block >Iia
must be available at the CVMR site. A specific problem asso-
ciated with monitoring within the magnet is that of assessing
ST-segment changes from the ECG. However, because wall-
motion abnormalities precede ST changes and because such
confirmed the low periprocedural complication rate of high-
abnormalities can be readily detected with fast CVMR, moni-
dose DSMR (11). Thus, although adverse events are rare,
toring is effective without a diagnostic ECG and can also be
the staff must be prepared to remove the patient from the
performed in patients with left-bundle branch block, who are
magnet rapidly if necessary and must comply closely with
routinely evaluated with dobutamine stress echocardiography
the test termination criteria (Table 16.3). Whereas in most
despite nondiagnostic ST segments. Such wall-motion abnor-
other modalities, “eye-to-eye” contact between patient and
malities can be detected immediately after image reconstruc-
examiner takes place, communication during CVMR is usu-
tion, which is completed 5 to 10 seconds after image acqui-
ally established via a microphone system and video cameras,
sition. In addition, real-time imaging permits the immediate
although it can also be conducted personally. This does not
detection of wall-motion abnormalities and can be used for
hinder the safety process if the patient is monitored care-
monitoring (12). Such real-time acquisition has been found
fully for symptoms, blood pressure changes, and wall-mo-
to yield a similar diagnostic accuracy for the detection of in-
tion abnormalities (Table 16.2). Either standard equipment
ducible wall-motion abnormalities (sensitivity and specificity
can be placed outside the scanner room and connected to
for real-time MR imaging 81% and 83%) when compared to
the patient with special extensions through a waveguide in
conventional MR cine imaging (13). However, spatial resolu-
tion is less, and thus we recommend ECG-triggered breath-
hold imaging as the basis for the final diagnosis.
onitoring Requirements
M Image Interpretation
Table 16.5 for Stress Magnetic
Resonance Imaging For image interpretation, a multiple cine loop display that
allows different stress levels to be assessed at the same time
Dobutamine Adenosine is essential. The ventricle is typically analyzed for 17 LV seg-
ments per stress level according to the standards suggested
Heart rate and Continuously Continuously by the Society of Cardiovascular Magnetic Resonance (6)
rhythm (single-lead
and the American Heart Association (Fig. 16.2). Each seg-
ECG)
Blood pressure Every minute Every minute
ment is assigned to a specific coronary artery; however,
Pulse oximetrya Continuously Continuously depending on the coronary artery anatomy or degree of
Symptoms Continuously Continuously collateralization, some segments may be supplied by differ-
Wall-motion Every dose increment Not necessary ent arteries. Thus, it is sometimes not possible to define a
abnormalities stenotic coronary artery from a wall-motion study. Image
quality is graded as excellent, good, moderate, or non-
a
Only required for additional rhythm control if no vector ECG is diagnostic, and the number of diagnostic segments is
available. reported. Segmental wall motion is classified as normoki-
ECG, electrocardiogram. netic, hypokinetic, akinetic, or dyskinetic and assigned one
LWBK1209-ch16_p239-250.indd 243 16/05/13 9:38 PM

Vertical minute intravenously) (Table 16.7) (16,17,36). However,

Short axis Long axis to ensure a high sensitivity, it is necessary to achieve sub-
Apical Mid Basal Mid maximal heart rates, which normally is not attained at these
1 doses. The first prospective study applying a high-dose dobu-
7 tamine/atropine protocol for CVMR compared the results of
13 2 6
8 12
DSMR versus dobutamine stress echocardiography with in-
14 16 17 vasive angiography as the reference method (18). This study
9 11 3 5 found significantly better values for the sensitivity (86% vs.
15
10
4
74%), specificity (86% vs. 70%), and diagnostic accuracy
(86% vs. 73%) of DSMR versus transthoracic echocar-
diography. The main reason for the superiority of DSMR
LAD RCA LCX
was based on its improved image quality (37). A study by
Figure 16.2. Seventeen-segment model suggested by the American Hundley et al. (19) demonstrated that 94% of patients with
Heart Association. The coronary artery territories are shown in the nondiagnostic echocardiographic image quality could be ad-
graph.
equately examined with DSMR at a sensitivity and speci-
ficity of 83%. The high diagnostic accuracy of DSMR was
to four points (Fig. 16.3). The sum of the points is divided confirmed in patients with pre-existing wall-motion abnor-
by the number of analyzed segments to yield a wall-motion malities at rest or after revascularization, which are chal-
score. Normal contraction results in a wall-motion score of lenging to assess with echocardiography (22). Because high-
one, and a higher score indicates wall-motion abnormality. dose dobutamine stress CVMR is highly accurate and can be
During stress with increasing doses of dobutamine, either a performed within 30 minutes, it has replaced dobutamine
lack of increase in wall motion or systolic wall thickening, stress echocardiography for the detection of CAD in patients
or a reduction in wall motion or thickening is regarded as a with nondiagnostic or suboptimal echocardiography image
pathologic finding (Fig. 16.1). quality in many institutions.
Overall, women are underreported in studies assessing the
diagnostic utility of cardiovascular imaging methods, despite
Accuracy of Stress-Induced Wall- the fact that women die more often than men from cardiac
Motion Abnormalities in the events related to CAD (38). Furthermore, both transthoracic
Diagnosis of Ischemia echocardiography and radionuclide scintigraphy suffer from
gender-related differences in diagnostic accuracy, which is
The echocardiographic detection of wall-motion abnor- related to smaller chamber size, poor acoustic windows, and
malities during high-dose dobutamine or exercise stress has more attenuation artifacts, respectively. DSMR has been re-
been shown to be an accurate diagnostic tool for screening ported to provide a gender-independent high diagnostic per-
patients with suspected CAD. Sensitivities of 54% to 96% formance using DSMR for the detection of CAD (28). The
and specificities of 60% to 100% have been reported (14), diagnostic values (sensitivity/specificity/accuracy) were simi-
depending on the pretest likelihood of disease and the expe- lar for men (86%/83%/85%) and women (85%/86%/85%).
rience of the stress centers. However, the value of stress Importantly, DSMR is associated with low interobserver
echocardiography is limited by a 10% to 15% rate of non- variability and a high reproducibility (23,39).
diagnostic results (14) and low specificities for the basal–lat- In a study by Paetsch et al. (21), a direct comparison of
eral and basal–inferior segments of the LV (15). DSMR, adenosine stress MR, and adenosine stress MR per-
The first studies applying dobutamine as a stress agent in fusion was performed during a single, combined examina-
the CVMR environment have yielded good results in the de- tion using coronary angiography as the reference standard.
tection of wall-motion abnormalities at intermediate doses It was found that dobutamine is superior to adenosine stress
of dobutamine (maximum of 20 μg/kg of body weight per for the detection of inducible wall-motion abnormalities re-
lated to the presence of epicardial coronary stenoses >50%,
with DSMR and adenosine stress MR yielding an overall
Normal diagnostic accuracy of 86% and 58%, respectively. Only for
the detection of coronary stenosis >75%, a reasonably good
diagnostic accuracy of adenosine stress MR was found.
Thus, adenosine stress MR cannot be recommended for the
Ischemic detection of inducible wall-motion abnormalities resulting
from epicardial coronary stenoses.
Hibernating
+ ischemic Additional Value of Perfusion
Imaging During Dobutamine
Rest Low dose Max. stress Stress Testing
stress
Figure 16.3. Schematic display of normal and pathologic The traditional indicator of ischemia during DSMR is a
response of regional left ventricular wall motion to low- and high- stress-inducible wall-motion abnormality. With the develop-
dose dobutamine. ment of improved perfusion techniques allowing more robust
LWBK1209-ch16_p239-250.indd 244 16/05/13 9:38 PM

Table 16.7 Diagnostic Accuracy of DSMR
Mean Dobutamine Stenosis

Patients Men Age Dose (μg/ Sensitivity Specificity MRI Field Definition
Authors Year (n) (%) (years) kg/min) (%) (%) Technique Strength (%)
Pennell et al. 1992 25 74 52 20 91 100 Wall motion 1, 5 ≥50

(16)
van Rugge 1993 45 82 61 20 81 100 Wall motion 1, 5 ≥50
et al. (17)
Nagel et al. 1999 208 71 60 40 + atropine 86 86 Wall motion 1, 5 ≥50
(18)
Hundley 1999 163 56 NS 40 + atropine 83 83 Wall motion 1, 5 ≥50
et al. (19)
Kuijpers 2003 194 67 62 40 96 95 Myocardial 1, 5 ≥50
et al. (20) tagging
Paetsch et al. 2004 79 66 61 40 + atropine 89 80 Wall motion 1, 5 ≥50
(21)
Wahl et al. 2004 160 82 59 40 + atropine 89 84 Wall motion 1, 5 ≥50
(22)
Paetsch et al. 2006 150 83 61 40 + atropine 78 87 Wall motion 1, 5 ≥50
(23)
Jahnke et al. 2006 40 75 63 40 + atropine 89 83 Wall motion 1, 5 ≥50
(24)
Gebker et al. 2008 414 65 64 40 + atropine 91 70 Wall 1, 5 ≥70
(25) motion +
perfusion
Kelle et al. 2008 30 24 66 40 + atropine 80 86 Wall motion 3 ≥50
(8)
Korosoglou 2009 101 71 61 40 + atropine 98 86 SENC 1, 5 ≥50
et al. (26)
Korosoglou 2009 65 78 64 40 + atropine 81 96 Myocardial 1, 5 ≥50
et al. (27) tagging
et al. (27)
Gebker et al. 2010 745 73 64 40 + atropine 86 84 Wall motion 1, 5 ≥70
(28)
Manka et al. 2010 41 66 64 40 + atropine 92 75 Wall 1, 5 ≥50
(29) motion +
perfusion
(30) motion +
perfusion
Korosoglou 2010 1493 74 65 40 + atropine 89 94 Wall 1, 5 ≥50
et al. (31) motion +
perfusion
Korosoglou 2010 80 72 62 20 76 88 SENC 1, 5 ≥50
et al. (32)
(33) motion +
perfusion
et al.
(34)
Gebker 2012 78 76 65 40 + atropine 92 83 Wall 1, 5 ≥70
et al. (33) motion +
perfusion
Gebker 2011 1532 67 63 40 + atropine 86 93 Wall motion 1, 5 ≥50
et al. (35)
MRI, magnetic resonance imaging; SENC, strain-encoding.
LWBK1209-ch16_p239-250.indd 245 16/05/13 9:38 PM

imaging at higher heart rates, combining the assessment of further corroborating the pathophysiologic concept of the
both wall motion and perfusion during the same examina- ischemic cascade, where disturbances of perfusion precede
tion seemed as a natural progression in further advancing wall-motion abnormalities (41). A concomitant decrease in
DSMR. To this end, several studies have demonstrated the specificity was observed during these earlier trials, which
feasibility and clinical value of additional perfusion imaging may have been caused by dark rim artifacts resulting from
during high-dose DSMR perfusion (25,40). Its major impact insufficient spatial resolution and motion artifacts. These
on diagnostic accuracy is related to an increase in sensitivity issues could in part be overcome by introducing high-spatial
(Fig. 16.4) (25). More recently, it could be shown that DSMR resolution imaging (1.5 × 1.5 mm) accomplished by accel-
perfusion improved sensitivity in patients with intermediate erating imaging with k-space and time sensitivity encoding
coronary stenosis, but not in those with severe stenosis, thus (k–t SENSE) (33). At the same time the true extent of CAD
Rest ED
Rest ES
Stress ES
Stress
perfusion
Rest
perfusion
LGE
Figure 16.4. Clinical example for the additional value of perfusion imaging during DSMR. The stress-
inducible wall-motion abnormality is limited to the inferior apical segment, whereas the dobutamine
stress-induced perfusion deficit involves the inferior apical, inferior/inferoseptal medial, and inferoseptal/
inferior basal segments. A small perfusion deficit is seen inferior apical and medial at rest perfusion due to a
subtotal stenosis of the RCA on invasive coronary angiography. There is no scar on late gadolinium enhance-
ment (LGE) imaging.
LWBK1209-ch16_p239-250.indd 246 16/05/13 9:38 PM

as expressed by the number of diseased segments was better as an independent risk factor for cardiac events, regardless
defined using DSMR perfusion than wall motion alone (33). of the induction of ischemia. In a second study, LV mass
The identification of ischemia using cine wall motion was was measured among individuals with a range of LVEF who
shown to be negatively affected by concentric hypertrophy underwent DSMR (47). The authors showed that LV hyper-
or remodeling of the LV (30). At the same time, these pa- trophy was an independent predictor of future myocardial
tients particularly benefit from additional DSMR perfusion infarction and cardiac death in patients with or without in-
(30). The decrease in sensitivity observed in these geometric ducible ischemia during dobutamine cardiac stress testing.
subtypes using DSMR is most likely caused by obliteration Hence, these results suggest that increased LV wall thick-
of the LV cavity during dobutamine stress, thus impeding ness and left ventricular hypertrophy should be reported in
the recognition of stress-inducible wall-motion abnormali- patients referred for DSMR, particularly in those without
ties. On the contrary, the detection of perfusion deficits inducible ischemia, in whom, otherwise, one would assume
under these morphologic conditions is rather improved. In a favorable cardiac prognosis.
a study, directly comparing first-pass CVMR perfusion im- Many patients presenting for cardiovascular care have re-
aging during adenosine and high-dose dobutamine/atropine duced LV function at rest due to pre-existing coronary arte-
stress resulted in an equally high sensitivity and specificity riosclerosis. Recently, Dall’Armellina et al. (48) performed a
for the stenosis detection on a per patient basis (92% and study to identify the prognostic utility of DSMR in patients
75% for both stressors) (29). with impaired baseline LV function (LVEF < 40%) who were
poorly suited for dobutamine stress echocardiography. They
demonstrated that stress-inducible wall-motion abnormali-
Prognostic Value of Dobutamine ties did not add further prognostic information to the al-
Stress Magnetic Resonance ready impaired prognosis in this population.
Imaging As described above, the versatility of CVMR allows
the assessment of both wall motion and perfusion during
Over the past 10 years, several studies have highlighted the a single examination with high-dose dobutamine stress. In
value of DSMR to predict cardiac prognosis. Hundley et al. order to investigate the prognostic impact of this combined
(42) were the first to provide data in this context and showed protocol, Korosoglou et al. (31) investigated 1,493 patients
that the presence of inducible ischemia or an LV ejection who were followed for 2 years. In this large, single center
fraction (LVEF) <40% was associated with hard clinical study, dobutamine-induced wall-motion abnormalities were
events such as myocardial infarction or cardiac death inde- of added value for the risk stratification of patients with and
pendent of the presence of other risk factors for coronary without inducible perfusion deficits, whereas perfusion defi-
arteriosclerosis. Jahnke et al. performed a large study where cits forecasted a poorer prognosis only in the absence of in-
patients received both adenosine stress CVMR and DSMR ducible wall-motion abnormalities. Furthermore, the prog-
during the same examination in order to determine their nostic benefit of additional perfusion deficits was limited to
respective prognostic accuracy. During a 2-year follow-up patients with resting wall-motion abnormalities, those with
period, no significant difference regarding the prognostic known CAD and left ventricular hypertrophy.
value was found between the two tests. More importantly,
a negative result in either test was associated with a 2-year
relatively event-free period of cardiac events. A large study Quantitative Methods for
including 1,463 patients on the long-term prognosis of Evaluating Wall Motion
DSMR further demonstrated the independent prognostic
value of DSMR in patients with suspected and known CAD Despite the superior image quality and diagnostic accuracy
over a mean follow-up time of 44 ± 24 months (43). A nega- of DSMR compared to dobutamine stress echocardiography,
tive DSMR carried an excellent prognosis with an annual assessment of CVMR cine studies relies mostly on the sub-
cardiac event rate of 1.1% over 6 years (0.8% in the first jective interpretation of wall motion in clinical routine. Over
3 years and 1.4% between the fourth and sixth year). Thus, the past years, several interesting methods allowing objec-
DSMR may not only distinguish high-risk patients in whom tive assessment of regional wall motion have been devel-
further interventions are necessary but also identifies low- oped. Visual interpretation may be facilitated with the use of
risk patients in whom additional procedures and intensive myocardial tagging. Kuijpers et al. (20) demonstrated that
medical follow-up may not be required. additional correct positive findings could be made by the use
Recently, several gender-based studies have clarified that of tagging without decreasing specificity. The main limita-
DSMR is an excellent tool for prognostication of cardiac tion of routine clinical application of myocardial tagging has
events in both men and women (44,45). Jahnke et al. (44) been the time-consuming process of its quantitative evalu-
demonstrated that CVMR adenosine perfusion imaging and ation. Harmonic phase (HARP) imaging, which provides
DSMR wall-motion testing exhibited equally high utility for automatic and near real-time analysis of tagged cardiac MR
cardiac risk stratification in both genders with an incremen- images, was a considerable advancement in this context
tal value over conventional cardiovascular risk factors. (49). HARP, however, suffers from relatively low spatial res-
Two studies on DSMR have demonstrated an associa- olution, not allowing differentiation of strain across the wall
tion between the presence of left ventricular hypertrophy or of the heart. More recently, strain-encoding (SENC) (50) has
increased wall thickness and an adverse cardiac prognosis. been introduced for the investigation of circumferential and
Walsh et al. (46) identified an end-diastolic wall thickness longitudinal myocardial strain. The advantage of this tech-
≥12 mm measured at the base of the septum or lateral wall nique compared to conventional CVMR tagging sequences
LWBK1209-ch16_p239-250.indd 247 16/05/13 9:38 PM

is based on its superior temporal resolution, total scan dura- LDDSMR yielded the highest overall accuracy of 84% (95%
tion, and time required for postprocessing of the acquired CI 82% to 86%) with a sensitivity of 81% (95% CI 73% to
data (27). The same group also identified a superior prog- 86%), specificity of 91% (95% CI 84% to 95%), positive pre-
nostic utility of SENC beyond the assessment of atherogenic dictive value (PPV) of 93% (95% CI 87% to 97%), and nega-
risk factors and conventional reading of CVMR cine images tive predictive value (NPV) of 75% (95% CI 65% to 83%).
(34). Recently, CVMR feature tracking, a novel method to LGE imaging had an overall accuracy of 70% (95% CI 69%
assess myocardial strain off-line on routine cine images, was to 71%) with a sensitivity of 95% (95% CI 93% to 97%),
presented (51). First data on intermediate-dose dobutamine specificity of 51% (95% CI 40% to 62%), PPV of 69% (95%
stress using feature tracking demonstrated its reliability to CI 56% to 80%), and NPV of 90% (95% CI 85% to 93%).
determine quantitative wall motion and strain derived from Looking at these results, a combination of LGE with excellent
SSFP cine imaging during inotropic stimulation (52). In sum- sensitivity (95%) and NPV (90%) and LDDSMR with excel-
mary, there is great potential that quantification of abnor- lent specificity (91%) and PPV (93%) seems highly attractive
malities of deformation within the myocardium improves (58–60). This is particularly true for segments with intermedi-
overall accuracy of dobutamine stress CVMR as compared ate scar transmurality (1% to 50%), which can be appreciated
with visual assessment of wall motion. from the landmark paper by Kim et al. (61) published in the
New England Journal of Medicine. In this study the likelihood
of functional recovery was ∼78% in the absence of scar and
Viability Imaging with ∼8% if the scar transmurality exceeded 50% (61). In these
Dobutamine Stress Magnetic cases LGE imaging alone is sufficient to predict functional re-
Resonance covery. However in case of segments with a scar transmurality
of 1% to 50%, the likelihood of functional recovery is ∼53%
CVMR offers comprehensive assessment of viable myocar- (61). LDDSMR should be added to guide further management
dium or better hibernating myocardium (53). It accurately (62). Wellnhofer et al. (58) demonstrated that in these seg-
determines global left ventricular volumes and function as ments, the prediction of functional recovery based on assess-
well as regional wall thickness and motion (54). Low-dose ment of contractile reserve with LDDSMR is superior to scar
dobutamine stress magnetic resonance (LDDSMR) can be quantification. A combination of LDDSMR and LGE imaging
performed as an adjunct to high-dose dobutamine stress or as thus provides complementary information and seems most ef-
the first step in patients with low ejection fraction. Improved ficient reaching the highest accuracy (58–60). Low-dose dobu-
wall motion at low stress levels (5 to 10 μg/kg/min) in seg- tamine stimulation and scar imaging can be easily combined
ments with wall-motion abnormalities at rest is suggestive within a single CVMR examination and Figure 16.5 provides
of hibernating myocardium, whereas the occurrence of new a practical algorithm to decide which technique to choose [73].
wall-motion abnormalities or a biphasic response during
high-dose dobutamine stress (20 to 40 μg/kg/min + atropine)
is indicative of myocardial ischemia (53). A combination of
high- and LDDSMR has the advantage of delivering infor-
mation on ischemia in addition to accurate assessment of
hibernating myocardium. Furthermore, dobutamine stress
adds a functional component to late gadolinium enhance-
ment (LGE) imaging, which detects irreversibly damaged
myocardium (necrosis, scar) with unprecedented accuracy
(55). With LGE, hibernating tissue is defined as dysfunc-
tional myocardium without scar or significant remaining
viable tissue (<50% transmurality of scarring).
It is important to remember how hibernating myocar-
dium is defined in the clinical scenario. The most accepted
definition is the improvement or recovery of regional func-
tion after revascularization. In clinical practice we need to
base the decision whether or not to revascularize our pa-
tients on the presence or absence of hibernating myocardium
and therefore all diagnostic tests must use a surrogate defi-
nition of hibernating myocardium (56). However, different
noninvasive tests provide different surrogate definitions of
hibernating myocardium (e.g., metabolism, scar, or contrac-
tile reserve) and consequently the result of a given investiga-
tion may depend on the chosen imaging test (56).
A recent meta-analysis reports the value of CVMR and its
different approaches to predict functional recovery after re-
vascularization (57). Romero et al. pooled the data from 24
studies and 698 patients. Among currently available CVMR Figure 16.5. Algorithm to assess hibernating myocardium with
parameters, essentially LGE imaging and LDDSMR were the CMR. Adapted from: Nagel E, Schuster A. Myocardial viability: Dead or
most accurate techniques to assess hibernating myocardium. alive is not the question! JACC Cardiovasc Imaging. 2012;5:509–512.
LWBK1209-ch16_p239-250.indd 248 16/05/13 9:38 PM

There is evidence to suggest that assessment of hibernating 2. Tonino PA, De Bruyne B, Pijls NH, et al. Fractional flow reserve versus angiography for
guiding percutaneous coronary intervention. N Engl J Med. 2009;360:213–224.
myocardium with CVMR has also prognostic implications. 3. Wijns W, Kolh P, Danchin N, et al. Guidelines on myocardial revascularization: The Task
Scar detected by LGE-CVMR is a stronger predictor of ad- Force on Myocardial Revascularization of the European Society of Cardiology (ESC) and the
European Association for Cardio-Thoracic Surgery (EACTS). Eur Heart J. 2010;31:2501–2555.
verse clinical outcome than LVEF and volumes (63). In pa- 4. Jeneson JA, Schmitz JP, Hilbers PA, et al. An MR-compatible bicycle ergometer for in-magnet
tients with fewer than six scarred segments, LGE predicts out- whole-body human exercise testing. Magn Reson Med. 2010;63:257–261.
come more accurately than hibernating myocardium assessed 5. Raman SV, Dickerson JA, Jekic M, et al. Real-time cine and myocardial perfusion with
treadmill exercise stress cardiovascular magnetic resonance in patients referred for stress
by LDDSMR. Conversely in patients with scar in more than SPECT. J Cardiovasc Magn Reson. 2010;12:41.
six segments, hibernating myocardium assessed by LDDSMR 6. Kramer CM, Barkhausen J, Flamm SD, et al. Standardized cardiovascular magnetic reso-
is a better predictor (64). In addition, CVMR has been suc- nance imaging (CMR) protocols, society for cardiovascular magnetic resonance: Board of
trustees task force on standardized protocols. J Cardiovasc Magn Reson. 2008;10:35.
cessfully used to guide patient management and improve 7. Thiele H, Nagel E, Paetsch I, et al. Functional cardiac MR imaging with steady-state free
outcome if the relevant information are incorporated into precession (SSFP) significantly improves endocardial border delineation without contrast
agents. J Magn Reson Imaging. 2001;14:362–367.
the management of the patients. Gerber et al. prospectively 8. Kelle S, Hamdan A, Schnackenburg B, et al. Dobutamine stress cardiovascular magnetic
evaluated 144 consecutive patients that either underwent cor- resonance at 3 Tesla. J Cardiovasc Magn Reson. 2008;10:44.
onary revascularisation and medical therapy or medical ther- 9. Berger A, Schnackenburg B, Schneeweis C, et al. High-dose dobutamine stress SSFP cine
MRI at 3 Tesla with patient adaptive local RF shimming using dual-source RF transmis-
apy alone and followed them up for approximately 3 years. sion. J Cardiovasc Magn Reson. 2012;14(suppl 1):P252.
They were able to demonstrate that patients with hibernating 10. Wahl A, Paetsch I, Gollesch A, et al. Safety and feasibility of high-dose dobutamine–atro-
myocardium defined by LGE that underwent revascularisa- pine stress cardiovascular magnetic resonance for diagnosis of myocardial ischaemia:
Experience in 1000 consecutive cases. Eur Heart J. 2004;25:1230–1236.
tion had a mortality benefit over those that received medical 11. Bruder O, Schneider S, Nothnagel D, et al. EuroCMR (European Cardiovascular Magnetic
therapy only (65). This evidence shows that the management Resonance) registry: Results of the German pilot phase. J Am Coll Cardiol. 2009;54:1457–
1466.
of patients with hibernating myocardium defined by LGE- 12. Schalla S, Nagel E, Lehmkuhl H, et al. Comparison of magnetic resonance real-time imag-
CVMR (65) confers a prognostic benefit as opposed to its ing of left ventricular function with conventional magnetic resonance imaging and echocar-
definition by SPECT or dobutamine stress echocardiography diography. Am J Cardiol. 2001;87:95–99.
13. Schalla S, Klein C, Paetsch I, et al. Real-time MR image acquisition during high-dose
that at present does not yield similar value (66). dobutamine hydrochloride stress for detecting left ventricular wall-motion abnormalities
Potentially, the sensitivity of LDDSMR to detect areas in patients with coronary arterial disease. Radiology. 2002;224:845–851.
likely to improve after revascularization could be improved 14. Geleijnse ML, Fioretti PM, Roelandt JR. Methodology, feasibility, safety and diagnostic
accuracy of dobutamine stress echocardiography. J Am Coll Cardiol. 1997;30:595–606.
using quantitative analysis such as strain; for example, using 15. Bach DS, Muller DW, Gros BJ, et al. False positive dobutamine stress echocardiograms:
myocardial tagging (67) or feature tracking (68). Promising Characterization of clinical, echocardiographic and angiographic findings. J Am Coll
Cardiol. 1994;24:928–933.
preliminary results; however, need to be confirmed in con- 16. Pennell DJ, Underwood SR, Manzara CC, et al. Magnetic resonance imaging during dobu-
trolled prospective investigations. Similarly there are novel tamine stress in coronary artery disease. Am J Cardiol. 1992;70:34–40.
sequences for scar imaging available such as the dual inver- 17. van Rugge FP, van der Wall EE, de Roos A, et al. Dobutamine stress magnetic resonance
imaging for detection of coronary artery disease. J Am Coll Cardiol. 1993;22:431–439.
sion recovery prepulse black blood sequence with enhanced 18. Nagel E, Lehmkuhl HB, Bocksch W, et al. Noninvasive diagnosis of ischemia-induced wall
contrast between blood and scar that may particularly in- motion abnormalities with the use of high-dose dobutamine stress MRI: Comparison with
crease the limited specificity of LGE (69). Future studies need dobutamine stress echocardiography. Circulation. 1999;99:763–770.
19. Hundley WG, Hamilton CA, Thomas MS, et al. Utility of fast cine magnetic resonance
to investigate whether the survival benefit associated with a imaging and display for the detection of myocardial ischemia in patients not well suited for
CVMR-based patient management can be further increased second harmonic stress echocardiography. Circulation. 1999;100:1697–1702.
20. Kuijpers D, Ho KY, van Dijkman PR, et al. Dobutamine cardiovascular magnetic resonance
using an optimal combination of the different CVMR meth- for the detection of myocardial ischemia with the use of myocardial tagging. Circulation.
odologies for the detection of hibernating myocardium. 2003;10:1592–1597.
21. Paetsch I, Jahnke C, Wahl A, et al. Comparison of dobutamine stress magnetic resonance,
adenosine stress magnetic resonance, and adenosine stress magnetic resonance perfusion.
Circulation. 2004;110:835–842.
Summary 22. Wahl A, Paetsch I, Roethemeyer S, et al. High-dose dobutamine–atropine stress cardiovas-
cular MR imaging after coronary revascularization in patients with wall motion abnor-
malities at rest. Radiology. 2004;233:210–216.
Over the past decade there has been a growing body of evi- 23. Paetsch I, Jahnke C, Ferrari VA, et al. Determination of interobserver variability for iden-
dence proving the usefulness of CVMR techniques for the tifying inducible left ventricular wall motion abnormalities during dobutamine stress mag-
netic resonance imaging. Eur Heart J. 2006;27:1459–1464.
assessment of ischemic heart disease during dobutamine 24. Jahnke C, Paetsch I, Gebker R, et al. Accelerated 4D dobutamine stress MR imaging with
stress. When performed by trained staff using appropriate k-t BLAST: Feasibility and diagnostic performance. Radiology. 2006;241:718–728.
monitoring, these studies are safe and have a high rate of 25. Gebker R, Jahnke C, Manka R, et al. Additional value of myocardial perfusion imaging
during dobutamine stress magnetic resonance for the assessment of coronary artery disease.
diagnostic accuracy, superior to that of dobutamine stress Circ Cardiovasc Imaging. 2008;1:122–130.
echocardiography. Thus, DSMR should be used in patients 26. Korosoglou G, Lossnitzer D, Schellberg D, et al. Strain-encoded cardiac MRI as an adjunct
for dobutamine stress testing: Incremental value to conventional wall motion analysis. Circ
with nondiagnostic or moderate echocardiographic image Cardiovasc Imaging. 2009;2:132–140.
quality to detect stress-induced wall-motion abnormalities. 27. Korosoglou G, Futterer S, Humpert PM, et al. Strain-encoded cardiac MR during high-
Since DSMR offers the unique capability to combine the dose dobutamine stress testing: Comparison to cine imaging and to myocardial tagging. J
Magn Reson Imaging. 2009;29:1053–1061.
assessment of wall motion, perfusion, and late gadolinium 28. Gebker R, Jahnke C, Hucko T, et al. Dobutamine stress magnetic resonance imaging for
enhancement during a single session with a single stress the detection of coronary artery disease in women. Heart. 2010;96:616–620.
agent, this technique may lead to improved clinical decision 29. Manka R, Jahnke C, Gebker R, et al. Head-to-head comparison of first-pass MR perfu-
sion imaging during adenosine and high-dose dobutamine/atropine stress. Int J Cardiovasc
making for patients with known and suspected CAD. Imaging. 2010.
30. Gebker R, Mirelis JG, Jahnke C, et al. Influence of left ventricular hypertrophy and geom-
etry on diagnostic accuracy of wall motion and perfusion magnetic resonance during dobu-
tamine stress. Circ Cardiovasc Imaging. 2010;3:507–514.
References 31. Korosoglou G, Elhmidi Y, Steen H, et al. Prognostic value of high-dose dobutamine stress
magnetic resonance imaging in 1493 consecutive patients: Assessment of myocardial wall
1. Pijls NH, van Schaardenburgh P, Manoharan G, et al. Percutaneous coronary intervention motion and perfusion. J Am Coll Cardiol. 2010;56:1225–1234.
of functionally nonsignificant stenosis: 5-year follow-up of the DEFER Study. J Am Coll 32. Korosoglou G, Lehrke S, Wochele A, et al. Strain-encoded CMR for the detection of induc-
Cardiol. 2007;49:2105–2111. ible ischemia during intermediate stress. JACC Cardiovasc Imaging. 2010;3:361–371.
LWBK1209-ch16_p239-250.indd 249 16/05/13 9:38 PM

33. Gebker R, Jahnke C, Manka R, et al. High spatial resolution myocardial perfusion imaging 51. Hor KN, Gottliebson WM, Carson C, et al. Comparison of magnetic resonance feature
during high dose dobutamine/atropine stress magnetic resonance using k-t SENSE. Int J tracking for strain calculation with harmonic phase imaging analysis. JACC Cardiovasc
Cardiol. 2012;158:411–416. Imaging. 2010;3:144–151.
34. Korosoglou G, Gitsioudis G, Voss A, et al. Strain-encoded cardiac magnetic resonance 52. Schuster A, Kutty S, Padiyath A, et al. Cardiovascular magnetic resonance myocardial
during high-dose dobutamine stress testing for the estimation of cardiac outcomes com- feature tracking detects quantitative wall motion during dobutamine stress. J Cardiovasc
parison to clinical parameters and conventional wall motion readings. J Am Coll Cardiol. Magn Reson. 2011;13:58.
2011;58:1140–1149. 53. Schuster A, Morton G, Chiribiri A, et al. Imaging in the management of ischemic cardio-
35. Gebker R, Jahnke C, Manka R, et al. The role of dobutamine stress cardiovascular mag- myopathy: Special focus on magnetic resonance. J Am Coll Cardiol. 2012;59:359–370.
netic resonance in the clinical management of patients with suspected and known coronary 54. Attili AK, Schuster A, Nagel E, et al. Quantification in cardiac MRI: Advances in image
artery disease. J Cardiovasc Magn Reson. 2011;13:46. acquisition and processing. Int J Cardiovasc Imaging. 2010;26(suppl 1):27–40.
36. Baer FM, Voth E, Theissen P, et al. Gradient-echo magnetic resonance imaging during 55. Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed contrast enhancement to
incremental dobutamine infusion for the localization of coronary artery stenoses. Eur irreversible injury, infarct age, and contractile function. Circulation. 1999;100:1992–2002.
Heart J. 1994;15:218–225. 56. Nagel E, Schuster A. Shortening without contraction: New insights into hibernating myo-
37. Nagel E, Lehmkuhl HB, Klein C, et al. [Influence of image quality on the diagnostic accu- cardium. JACC Cardiovasc Imaging. 2010;3:731–733.
racy of dobutamine stress magnetic resonance imaging in comparison with dobutamine 57. Romero J, Xue X, Gonzalez W, et al. CMR imaging assessing viability in patients with
stress echocardiography for the noninvasive detection of myocardial ischemia]. Z Kardiol. chronic ventricular dysfunction due to coronary artery disease: A meta-analysis of prospec-
1999;88:622–630. tive trials. JACC Cardiovasc Imaging. 2012;5:494–508.
38. Mieres JH, Shaw LJ, Arai A, et al. Role of noninvasive testing in the clinical evaluation 58. Wellnhofer E, Olariu A, Klein C, et al. Magnetic resonance low-dose dobutamine test is
of women with suspected coronary artery disease: Consensus statement from the Cardiac superior to SCAR quantification for the prediction of functional recovery. Circulation.
Imaging Committee, Council on Clinical Cardiology, and the Cardiovascular Imaging 2004;109:2172–2174.
and Intervention Committee, Council on Cardiovascular Radiology and Intervention, 59. Kirschbaum SW, Rossi A, Boersma E, et al. Combining magnetic resonance viability vari-
American Heart Association. Circulation. 2005;111:682–696. ables better predicts improvement of myocardial function prior to percutaneous coronary
39. Syed MA, Paterson DI, Ingkanisorn WP, et al. Reproducibility and inter-observer variabil- intervention. Int J Cardiol. 2012;159:192–197.
ity of dobutamine stress CMR in patients with severe coronary disease: Implications for 60. Glaveckaite S, Valeviciene N, Palionis D, et al. Value of scar imaging and inotropic reserve
clinical research. J Cardiovasc Magn Reson. 2005;7:763–768. combination for the prediction of segmental and global left ventricular functional recovery
40. Lubbers DD, Janssen CH, Kuijpers D, et al. The additional value of first pass myocardial after revascularisation. J Cardiovasc Magn Reson. 2011;13:35.
perfusion imaging during peak dose of dobutamine stress cardiac MRI for the detection of 61. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance imaging
myocardial ischemia. Int J Cardiovasc Imaging. 2008;24:69–76. to identify reversible myocardial dysfunction. N Engl J Med. 2000;343:1445–1453.
41. Gebker R, Frick M, Jahnke C, et al. Value of additional myocardial perfusion imaging 62. Morton G, Schuster A, Perera D, et al. Cardiac magnetic resonance imaging to guide com-
during dobutamine stress magnetic resonance for the assessment of intermediate coronary plex revascularization in stable coronary artery disease. Eur Heart J. 2010;31:2209–2215.
artery disease. Int J Cardiovasc Imaging. 2012;28:89–97. 63. Roes SD, Kelle S, Kaandorp TAM, et al. Comparison of myocardial infarct size assessed
42. Hundley WG, Morgan TM, Neagle CM, et al. Magnetic resonance imaging determination with contrast-enhanced magnetic resonance imaging and left ventricular function and vol-
of cardiac prognosis. Circulation. 2002;106:2328–2333. umes to predict mortality in patients with healed myocardial infarction. Am J Cardiol.
43. Kelle S, Chiribiri A, Vierecke J, et al. Long-term prognostic value of dobutamine stress 2007;100:930–936.
CMR. JACC Cardiovasc Imaging. 2011;4:161–172. 64. Kelle S, Roes S, Klein C, et al. Prognostic value of myocardial infarct size and contractile
44. Jahnke C, Furundzija V, Gebker R, et al. Gender-based prognostic value of pharmaco- reserve using magnetic resonance imaging. J Am Coll Cardiol. 2009;54:1770–1777.
logical cardiac magnetic resonance stress testing: Head-to-head comparison of adenosine 65. Gerber BL, Rousseau MF, Ahn SA, et al. Prognostic value of myocardial viability by
perfusion and dobutamine wall motion imaging. Int J Cardiovasc Imaging. 2012;28:1087– delayed-enhanced magnetic resonance in patients with coronary artery disease and low
1098. ejection fraction: Impact of revascularization therapy. J Am Coll Cardiol. 2012;59:825–
45. Wallace EL, Morgan TM, Walsh TF, et al. Dobutamine cardiac magnetic resonance results 835.
predict cardiac prognosis in women with known or suspected ischemic heart disease. JACC 66. Bonow RO, Maurer G, Lee KL, et al. Myocardial viability and survival in ischemic left
Cardiovasc Imaging. 2009;2:299–307. ventricular dysfunction. N Engl J Med. 2011;364:1617–1625.
46. Walsh TF, Dall’Armellina E, Chughtai H, et al. Adverse effect of increased left ventricu- 67. Bree D, Wollmuth JR, Cupps BP, et al. Low-dose dobutamine tissue-tagged magnetic reso-
lar wall thickness on five year outcomes of patients with negative dobutamine stress. J nance imaging with 3-dimensional strain analysis allows assessment of myocardial viability
Cardiovasc Magn Reson. 2009;11:25. in patients with ischemic cardiomyopathy. Circulation. 2006;114:I33–I36.
47. Charoenpanichkit C, Morgan TM, Hamilton CA, et al. Left ventricular hypertrophy influ- 68. Schuster A, Paul M, Bettencourt N, et al. Cardiovascular magnetic resonance myocardial
ences cardiac prognosis in patients undergoing dobutamine cardiac stress testing. Circ feature tracking for quantitative viability assessment in ischemic cardiomyopathy. Int J
Cardiovasc Imaging. 2010;3:392–397. Cardiol. 2011;[E-pub ahead of print], doi:10.1016/j.ijcard.2011.10.137.
48. Dall’Armellina E, Morgan TM, Mandapaka S, et al. Prediction of cardiac events in patients 69. Peel SA, Morton G, Chiribiri A, et al. Dual inversion-recovery MR imaging sequence for
with reduced left ventricular ejection fraction with dobutamine cardiovascular magnetic reduced blood signal on late gadolinium-enhanced images of myocardial scar. Radiology.
resonance assessment of wall motion score index. J Am Coll Cardiol. 2008;52:279–286. 2012;264:242–249.
49. Kraitchman DL, Sampath S, Castillo E, et al. Quantitative ischemia detection during car- 70. Garot J, Bluemke DA, Osman NF, et al. Fast determination of regional myocardial strain
diac magnetic resonance stress testing by use of FastHARP. Circulation. 2003;107:2025– fields from tagged cardiac images using harmonic phase MRI. Circulation. 2000;101:981–
2030. 988.
50. Pan L, Stuber M, Kraitchman DL, et al. Real-time imaging of regional myocardial function 71. Nagel E, Schuster A. Myocardial viability: Dead or alive is not the question! JACC
using fast-SENC. Magn Reson Med. 2006;55:386–395. Cardiovasc Imaging. 2012;5:509–512.
LWBK1209-ch16_p239-250.indd 250 16/05/13 9:38 PM

Chapter
17
Lowie M.R. Van Assche
Han W. Kim
Raymond J. Kim
Assessment of Myocardial Viability

by Contrast Enhancement
■■ Introduction stunning and hibernation. “Myocardial stunning” describes

Background the state of post-ischemic myocardial dysfunction in the pres-
Clinical Importance of Myocardial Viability ence of relatively normal blood flow (1). On light micros-
copy, stunned myocardium appears normal. However, there
Scope of the Chapter
is decreased adenosine triphosphate and subtle ultrastructural
■■ Definition of Myocardial Viability abnormalities (2). The mechanism underlying stunning is not
■■ Delayed Enhancement Magnetic Resonance completely understood but has been proposed to involve myo-
filament desensitization to calcium, possibly by initial calcium
Imaging
overload during reperfusion, leading to troponin-I proteolysis
Background and/or oxygen radical formation during ischemia–reperfusion
Current Technique damaging the contractile cell membranes. In general, stunned
Overall Procedure myocardium has been shown to recover function within hours
Pulse Sequence Timing to days, and recovery may take longer in cases of longer isch-
Imaging Parameters emic episodes prior to revascularization (3).
Inversion Time The term “myocardial hibernation” describes the state
Imaging Time After Contrast Administration of chronic myocardial dysfunction at rest that can be par-
tially or completely restored to normal either by improving
■■ Pathophysiologic Validation
blood flow and/or by reducing demand (4). This physiology
Background was first noted when ventricular wall motion abnormali-
Animal Studies ties improved after coronary artery bypass grafting (CABG)
Human Studies (5–7). Biopsy specimens from regions of dysfunctional myo-
■■ The Ideal Technique cardium that recovered function after CABG showed that
Knowing How Much is Alive is Not Enough ∼90% of the volume consisted of viable cells (8,9). However,
increased glycogen is seen in areas previously occupied by
Is Thinned Myocardium Always Dead?
myofilaments and small mitochondria are present (10). The
Additional Characterization of Dead Tissue mechanism underlying hibernating myocardium is not com-
■■ Common Assumptions and Clinical pletely understood but many investigators believe it is due to
Applications a chronic reduction in resting coronary blood flow. Others
Intermediate Levels of Viability suggest that blood flow is normal in hibernating myocar-
Functional Recovery as the Standard of Truth dium, and that the dysfunction results from repetitive epi-
Overview of Clinical Interpretation sodes of stunning (10). In contrast to the relatively early re-
covery of function seen with myocardial stunning, recovery
■■ Summary of function has been shown to take up to 1 year after revas-
cularization in myocardial hibernation (3). Thus, myocar-
dial stunning and hibernation both represent myocardium
Introduction that is alive (or viable), rather than dead (or nonviable).
Background Clinical Importance of

Myocardial Viability
The concept that dysfunctional myocardium may be viable
has been known for decades. Specific physiologic states In the acute setting following an episode of myocardial
that can result in dysfunctional, yet viable myocardium are ischemia and reperfusion, there are several reasons why it
251
LWBK1209-ch17_p251-282.indd 251 17/05/13 5:09 PM

is important to distinguish between stunned and infarcted rates were similar for revascularization and medical therapy
myocardium. First, the patient prognosis is changed. Several (7.7% vs. 6.2%, p = NS).
studies have shown that patients with acute ventricular dys- Based on data such as these, the ACC/AHA guidelines on
function primarily resulting from myocardial necrosis have the diagnosis and management of heart failure from 2009
a worse prognosis than patients with ventricular dysfunction and CABG surgery from 2011 state that CABG, in addition
that is primarily reversible (11,12). Second, patient manage- to those with the coronary anatomy described above, may
ment during the acute setting may be changed. Viable but be performed in patients with LV dysfunction (35% to 50%)
injured myocardium, such as stunned myocardium, is poten- and significant CAD if the target vessels supply a large area
tially at risk for future infarction if the reperfusion therapy of viable myocardium (Class IIa recommendation [weight of
was not complete and significant stenosis remains (11,13). evidence/opinion is in favor of usefulness/efficacy]) (18,21).
In addition, a determination of the extent of nonviable and In contrast, CABG is not recommended in patients with only
viable myocardium across the ventricular wall in a dysfunc- mild ischemia or only a small area of viable myocardium
tional region may be valuable in selecting patients most (Class III recommendation [evidence/agreement that treat-
likely to benefit from therapy that can modulate ventricular ment is not useful/efficacious or may be harmful]). However,
remodeling after acute myocardial infarction (MI), such as the evidence for these recommendations has limitations.
angiotensin-converting enzyme inhibitors and beta-blockers First, studies were small, not randomized, observational,
(14). Third, infarct size determined accurately in the acute and retrospective leading to potential patient selection bias.
setting may prove to be an adequate surrogate endpoint for Second, the methodology and criteria for defining viability,
the assessment of new therapies (15,16). For example, the as well as the treatment regimens were not standardized
efficacy of experimental therapies for acute coronary syn- among the different studies. Finally, the success and com-
dromes or acute MI could be evaluated without the need pleteness of myocardial revascularization was not investi-
for “mega” trials with large sample sizes that use mortality gated by follow-up angiography or stress imaging data in
as an endpoint. It is expected that the number of drug and most studies (19). Given these flaws, meta-analyses incorpo-
device trials that employ cardiac magnetic resonance imag- rating these data are subject to the same limitations. In part,
ing parameters as a primary endpoint will increase substan- due to these limitations, the Surgical Treatment for Ischemic
tially in the future (17). Heart Failure trial (STICH) was undertaken (22).
In the setting of chronic coronary artery disease (CAD), de- STICH was a randomized, multicenter, nonblinded trial
termining which patients will benefit from revascularization is funded by the National Heart, Lung, and Blood Institute
of obvious clinical importance. Guidelines for revasculariza- (22). Patients with angiographic documentation of CAD
tion are provided by the American College of Cardiology and amenable to surgical revascularization and LV systolic dys-
American Heart Association (ACC/AHA), last updated in function (ejection fraction [EF] ≤35%) were eligible for en-
2011 (18). In patients with chronic CAD, surgical revascu- rollment. Exclusion criteria were left main stenosis >50%,
larization is given a Class I recommendation (evidence/agree- cardiogenic shock, MI within 3 months, and need for aortic
ment that treatment is useful/efficacious) to improve survival valve surgery. In hypothesis 1, enrolled patients were ran-
in patients with the following coronary anatomy: Significant domly assigned to receive medical therapy alone or medical
(≥50% diameter stenosis) left main coronary artery stenosis therapy plus CABG. In hypotheses 2, enrolled patients were
(1); significant (≥70% diameter) stenosis in three major coro- randomly assigned to receive medical therapy plus CABG or
nary arteries or in the proximal LAD plus one other major medical therapy plus CABG and surgical ventricular recon-
coronary artery (18). These recommendations are based struction. Between 2002 and 2007, 1,212 patients from 99
largely upon data from randomized trials that focused pri- centers in 22 countries were enrolled in hypothesis 1. The
marily on the angiographic appearance of CAD. While these primary endpoint was death from any cause. In the initial
studies showed benefit in the overall population with severe design of the STICH trial, viability testing with single-pho-
angiographic disease, in patients with left ventricular (LV) ton emission computed tomography (SPECT) was required.
dysfunction, from a mechanistic point of view, one might ex- However, this requirement proved to be an impediment to
pect that additional information such as the presence or ab- study enrollment. Therefore, the protocol was revised in
sence of extensive viable myocardium could improve patient 2004 to make viability testing optional and the decision to
risk stratification for revascularization. perform the test was left up to the recruiting investigators. In
The potential importance of myocardial viability testing in addition, the viability test options were expanded to include
reference to prognosis after revascularization has been evalu- SPECT (four separate protocols), dobutamine stress echo-
ated in many small, single-center studies including a meta- cardiography (DSE) (23), or both (24).
analysis demonstrating decreased annual mortality in patients Overall, there was no significant difference between pa-
with viability undergoing revascularization compared to those tients with medical therapy alone and medical therapy plus
not undergoing revascularization (19). Allman et al. (20), in CABG with respect to the primary endpoint of death from
a meta-analysis of 24 studies, which included 3,088 patients any cause at a median of 5.1 years of follow-up (22). In
with a mean left ventricular ejection fraction (LVEF) of 32%, the 601 patients who received a viability study, there was
concluded that in patients with significant viability (as deter- a significant association between viability and outcome
mined by thallium perfusion imaging, F-18 fluorodeoxyglucose on univariate analysis, but not on multivariate analysis.
metabolic imaging, or dobutamine echocardiography), revas- Surprisingly, the assessment of myocardial viability did not
cularization was associated with 79.6% reduction in annual identify patients with a differential survival benefit from
mortality compared with medical treatment (3.2% vs. 16%, CABG, as compared with medical therapy alone, in contrast
p < 0.0001). Conversely, in patients without viability, mortality with the prior literature (24).
LWBK1209-ch17_p251-282.indd 252 17/05/13 5:09 PM

Chapter 17 ■ Assessment of Myocardial Viability by Contrast Enhancement 253
The authors point out that conclusions drawn from Markers of infarct size
STICH are limited by a number of factors (24). First, slightly
less than half of the 1,212 patients enrolled in the hypothesis Less precise Wall motion abnormality
1 comparison underwent viability testing. Second, patients
were not randomized to viability testing, and third, testing Q-waves
could have been performed prior to, on the day of, or after
study enrollment. These factors may have led to patient se- Total enzyme leak
lection/enrollment bias and influenced subsequent clinical
decision making. In addition, only a small percentage of pa-
No reflow or low reflow
tients were deemed not to have viable myocardium (19%),
which limited the power of the analysis to detect a differ-
ential effect of CABG, as compared with medical therapy Change in tissue composition
alone, in patients with myocardial viability as compared to
those without viability. Lastly, the viability analyses were Myocyte integrity
limited to SPECT and DSE imaging. While the results were More precise
similar for SPECT and DSE, caution should be taken to not Figure 17.1. Clinical and physiologic markers to determine the
extrapolate these results to other imaging modalities that size of infarction. (Adapted from Kaul S. Assessing the myocar-
were not tested in STICH, such as positron emission tomog- dium after attempted reperfusion: Should we bother? Circulation.
raphy (PET) and delayed enhancement magnetic resonance 1998;98(7):625–627, with permission.)
imaging (DE-MRI). Despite these limitations, the STICH
trial is a landmark investigation that raises an important
question—is viability assessment important? From a patho- because infarction results in the loss of viability. A number
physiologic viewpoint, it would be difficult to interpret the of techniques are available in the clinical setting to deter-
STICH results as concluding that alive myocardium is the mine whether or not infarction has occurred and if so, how
same as dead myocardium. On the other hand, from a clini- much of the injured territory is not yet infarcted and can be
cal viewpoint, the STICH results show that using “status salvaged. In a recent review article, Kaul (25) summarized
quo” viability methods to determine who will benefit from clinical markers of infarct size and ranked them from least
CABG is highly problematic. As a result, there clearly is a to most precise (Fig. 17.1). As discussed previously, obser-
need to reassess the underlying mechanisms regarding how vation of a wall motion abnormality alone does not pro-
viability tests are assumed to work and are interpreted. vide information regarding viability because both stunned
and hibernating myocardium are dysfunctional. The elec-
trocardiogram (ECG), although useful, is recognized as
Scope of the Chapter being insensitive to infarction because patients with smaller
In this chapter we focus on a contrast MRI technique—delayed infarcts may demonstrate minimal ECG changes during the
enhancement magnetic resonance imaging (DE-MRI)—to acute event and often will not have Q-waves chronically.
assess myocardial viability. We will start, however, with the Serum markers such as creatine kinase (CK) and troponin-I
basic definition of myocardial viability. Although the defi- or T can be extremely useful but even these are associated
nition may appear to be self-evident, we will find that dis- with several limitations. For example, CK and troponin lev-
crepancies between the results of DE-MRI and other clinical els may exhibit differing time courses depending on whether
indexes of viability may arise because of assumptions con- or not reperfusion has occurred (26), and neither can be
cerning the definition of viability. We proceed by providing a used to localize the infarction to a specific coronary artery
comprehensive overview of the technical aspects of DE-MRI, territory. Perhaps, most importantly, serum levels of CK and
followed by a review of the original pathophysiologic vali- troponin are not elevated beyond the first few days (27), pre-
dation studies. We then demonstrate the clinical application cluding the detection of older infarcts.
of DE-MRI and compare this technique to other clinically According to Kaul (25), the most precise way to define
available modalities used to assess myocardial viability. Next, infarction, and therefore the loss of viability, is to deter-
we explore what might constitute the “ideal” technique for mine whether or not myocyte death has occurred. All
assessing viability in order to get insight into physiologic as ischemic events prior to cell death are at least in principle
well as technical limitations that exist. Finally, we examine reversible; therefore, the further we deviate from a direct
common assumptions in the metrics used to evaluate viability assessment of cell death, the more imprecise we become
techniques, and finish with a general summary of the clinical in defining infarction. Likewise, the most precise definition
interpretation process. of myocardial viability is the presence of living myocytes.
The presence or absence of living myocytes can be readily
established in tissue specimens by light microscopy, elec-
tron microscopy, or by the use of histologic stains such
Definition of Myocardial as triphenyl tetrazolium chloride (TTC) (28). Testing for
Viability viability by microscopy or histologic staining, however, is
obviously not practical in a clinical setting. Accordingly, a
Before any technique used to identify myocardial viabil- number of less precise definitions of viability that are based
ity can be evaluated, a definition of viability is required. on parameters more easily measured in patients have been
The definition of viability is directly related to that of MI developed (Table 17.1).
LWBK1209-ch17_p251-282.indd 253 17/05/13 5:09 PM

Ideal imaging method

ommon Clinical
C
Table 17.1 Definitions of Myocardial
Viability Question:
• Improvement in contraction after revascularization Is the anterior wall
• Improvement in contraction with low-dose dobutamine viable or not viable?
• Preserved perfusion
• Preserved radionuclide tracer uptake
• Preserved glucose uptake
• Preserved wall thickness and/or thickening
In the literature, viability is often defined as improvement

in contractile function after coronary revascularization. This
definition is frequently the clinical “gold standard” to which Figure 17.2. Cartoon showing infarction of the subendocardial
imaging techniques are compared. Although convenient for half of the anterior wall (white area). See text for details.
clinical purposes, this definition can be inaccurate. If con-
tractile function improves after revascularization, it is safe to
assume that there is a significant amount of viability; how- inversion recovery sequence have been previously published
ever, the converse is not true. In fact, analysis of transmural (30,31). The primary action of most MRI contrast agents
needle biopsy specimens taken during CABG demonstrated currently approved for use in humans is to shorten the
that some regions that did not improve after revasculariza- longitudinal relaxation time (T1). Accordingly, the goal of
tion did have a significant amount of viability. For example, most MRI pulse sequences used for the purpose of exam-
Dakik et al. (8) reported that the extent of viability was ining contrast enhancement patterns is to make image
nearly 70% of total myocardium in their samples. intensities a strong function of T1 (T1-weighted images).
Since the correct definition of viability is the presence of Early approaches to acquiring T1-weighted images of the
living myocytes, the ideal imaging method for assessing vi- heart often used ECG-gated spin-echo imaging in which
ability should be able to delineate infarcted tissue from vi- one k-space line was acquired during each cardiac cycle.
able tissue with high spatial resolution. Unfortunately, cur- Since the duration of the cardiac cycle is comparable
rently available techniques, such as SPECT, PET, and DSE, with the myocardial T1 (∼800 milliseconds), the resulting
have various limitations. First, what is measured is not the images were T1-weighted. Following the administration
direct presence and exact quantity of viable myocytes, but of gadolinium contrast, myocardial T1 was shortened and
rather a physiologic parameter, such as contractile reserve or image intensities increased. Using this approach, a number
perfusion that has only an indirect relationship to viability. of investigators reported that as image intensities increased
Second, there are technique-specific limitations, including throughout the heart, regions associated with acute MI
partial volume effects due to poor spatial resolution (SPECT, became particularly bright (hyperenhanced) on a timescale
PET), attenuation and scatter artifacts (SPECT), errors in of minutes to tens of minutes after contrast administration
registration between comparison images, and the occasional (32–38). The use of ECG-gated spin-echo imaging, how-
inability to visualize all parts of the LV myocardium. Third, ever, has several intrinsic limitations that adversely affect
all of these techniques interpret viability as an all-or-none image quality. One such limitation is the need for relatively
phenomenon within a myocardial region since none can as- long acquisition times (minutes), which introduces artifacts
sess the transmural extent of viability across the ventricu- caused by respiratory motion.
lar wall. Figure 17.2 demonstrates the discrepancy that can
arise due to this limitation. In this particular example, as- Current Technique
sessment of the anterior wall using current clinical methods
and definitions will be incorrect regardless of whether the Since the early use of ECG-gated spin-echo imaging, a num-
anterior wall is determined to be viable or not viable since ber of improvements have been made. One of the most
the correct assessment is that the subendocardial half of the important among these is the use of segmented k-space
wall is not viable and the epicardial half is viable. This will (39), in which multiple k-space lines are acquired after each
be discussed in more detail later in the chapter. cardiac cycle. As a result, imaging times are reduced to the
point at which an entire image can be acquired during a sin-
gle breath-hold (∼8 seconds), thereby eliminating image arti-
Delayed Enhancement Magnetic facts caused by respiration. In addition, preparation of the
Resonance Imaging magnetization before image acquisition with an inversion
pulse significantly increases the degree of T1-weighting in the
Background images. The segmented inversion recovery pulse sequence
was compared with nine other pulse sequences in a dog
The phenomenon of delayed hyperenhancement was first model of acute MI (40). Table 17.2 summarizes the litera-
described over 20 years ago (29). Several excellent reviews ture in humans and in in vivo large animal models regarding
of the literature concerning the interpretation of myocardial the depiction of infarcted regions by gadolinium-enhanced
hyperenhancement before the development of the segmented MRI prior to the development of the segmented inversion
LWBK1209-ch17_p251-282.indd 254 17/05/13 5:09 PM

ercentage Elevations in MR Signal Intensity of Infarcted Versus Normal Myocardium

P
Table 17.2
and Voxel Sizes: Previous Studies Compared With New Technique
Canine Human
Breath- ΔI/R Voxel size Voxel Size

Year Reference Techniquea hold (%)b (mm3) ΔI/R (%)b (mm3)
1986 Rehr et al. (37) Spin-echo No 80 NS

1986 Tscholakoff et al.g Spin-echo No 70 NS
1986 Eichstaedt et al. (32) Spin-echo No 42c 29.3d
1988 de Roos et al.h Spin-echo No 60 29.3d
1989 de Roos et al.i Spin-echo No 36 29.3d
1990 van der Wall et al. (41) Spin-echo No 32c 31.3d
1991 van Dijkman et al. (42) Spin-echo No 31c 27.5
1991 Matheijssen et al. (33) Spin-echo No 42 29.3d
1994 Fedele et al. (43) Spin-echo No 41c 29.3d
1995 Lima et al. (44) MD-SPGRE Yes 103c 33.3
1995 Judd et al. (45) MD-SPGRE Yes 123e 39.6
1998 Ramani et al. (46) MD-SPGRE Yes 58e 30.8
1999 Pereira et al.j MD-SPGRE Yes 19e 14.7
1999 Rogers et al. (47) Single-shot Yes 39 54.9f
inversion
recovery GRE
Mean of previous 86 27.2 48 32.4
studies
New Technique Simonetti et al (40) Segmented Yes 1,080 6.2 485 16.8
inversion
recovery GRE
a
All images were acquired in vivo at least 5 minutes after the administration of an U.S. Food and Drug Administration-approved MR imaging
contrast agent.
b
AI/R = percent elevation in MR signal intensity of infarcted myocardium compared with normal myocardium.
c
Published data were reported as precontrast versus postcontrast values; values in Table 17.2 were calculated as follows: (Postcontrast
value − precontrast value)/precontrast value.
d
Assuming a field of view of 320 mm.
e
Estimated from data reported in graphical format.
f
Assuming a rectangular (6/8) field of view.
g
Tscholakoff D, Higgins CB, Sechtem U, et al. Occlusive and reperfused myocardial infarcts: Effect of Gd-DTPA on ECG-gated MR imaging.
Radiology. 1986;160:515–519.
h
de Roos A, Doornbos J, van der Wall EE, et al. MR imaging of acute myocardial infarction: Value of Gd-DTPA. AJR Am J Roentgenol.
1988;150:531–534.
i
de Roos A, van Rossum AC, van der Wall E, et al. Reperfused and nonreperfused myocardial infarction: Diagnostic potential of Gd-DTPA-
enhanced MR imaging. Radiology. 1989;172:717–720.
j
Pereira RS, Prato FS, Sykes J, et al. Assessment of myocardial viability using MRI during a constant infusion of Gd-DTPA: Further studies at early
and late periods of reperfusion. Magn Reson Med. 1999;42:60–68.
MD-SPGRE, magnetization-driven spoiled gradient-echo; NS, not stated; GRE, gradient-recalled echo.
Adapted from Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging technique for the visualization of myocardial infarction.
Radiology. 2001;218:215–223.
recovery sequence. From 1986 to 1999, image intensities patient enters the MRI scanner, and does not require phar-
in “hyperenhanced” regions were generally 50% to 100% macologic or physiologic stress. After obtaining scout images
higher than normal regions. The use of a segmented inver- to delineate the short- and long-axis views of the heart, we
sion recovery pulse sequence with the inversion time (TI) set obtain cine images to provide a matched assessment of LV
to null signal from normal myocardium increased this dif- morphology and contractile function, and to aid in the
ferential approximately 10-fold to 1,080% in animals and detection of small subendocardial infarcts (to be discussed).
485% in humans (labeled “New Technique” in Table 17.2). Short-axis views (6-mm slice thickness with 4-mm gap to
match contrast enhancement images) are taken every 10 mm
from mitral valve insertion to LV apex along with two to
Overall Procedure
three long-axis views in order to encompass the entire LV. The
The procedure for DE-MRI is relatively simple. It can be patient is then given a bolus of 0.10 to 0.15 mmol/kg intra-
performed in a single, brief examination, requires only a venous gadolinium by hand injection. After a 10 to 15 min-
peripheral intravenous catheter that is placed before the utes delay (timing issues will be discussed further), high
LWBK1209-ch17_p251-282.indd 255 17/05/13 5:09 PM

Typical patient scan Pulse Sequence Timing

The timing diagram for the seg IR-FGE pulse sequence is
shown in Figure 17.4. Immediately after the onset of the
R-wave trigger, there is a delay or wait period (which is
referred to as the trigger delay or TD) before a nonselec-
tive 180-degree hyperbolic secant adiabatic inversion pulse
is applied. Following this inversion pulse, a second variable
wait period (which is referred to as the inversion time or
TI) occurs, corresponding to the time between the inversion
pulse and the center of the data acquisition window (for lin-
early ordered k-space acquisition). The data acquisition win-
dow is generally 140 to 200 milliseconds long, depending
on the patient heart rate, and is placed during mid-diastole,
when the heart is relatively motionless. A group of k-space
lines are acquired during this acquisition window, where the
flip angle used for radiofrequency excitation of each k-space
line is shallow (20 to 30 degrees) to retain regional differ-
ences in magnetization that result from the inversion pulse
and TI delay. The number of k-space lines in the group is
Figure 17.3. Images from a typical patient scan. Cine and delayed limited by the repetition time (TR) between each k-space
contrast-enhanced images are acquired at 6 to 8 short-axis locations line (5 to 10 milliseconds) and the duration of mid-diastole.
and 2 to 3 long-axis locations during repeated breath-holds. Images In the implementation shown on Figure 17.4, 23 lines of
are interpreted with cine images immediately adjacent to contrast k-space are acquired during each data acquisition window,
images. This particular patient had a myocardial infarction caused which occurs every other heartbeat. With this implemen-
by occlusion of the right coronary artery. Note hyperenhancement tation, typically breath-hold duration of 10 to 12 cardiac
of the inferior wall. (From Kim RJ, Shah DJ, Judd RM. How we cycles is required to obtain all the k-space lines for the image
perform delayed enhancement imaging. J Cardiovasc Magn Reson. matrix.
2003;5(3):505–514, with permission.)
Imaging Parameters
spatial resolution delayed enhancement images of the heart The typical settings that we use for the seg IR-FGE sequence
are obtained at the same slice locations as the cine images are shown in Table 17.3. The dose of gadolinium given is
using a segmented inversion recovery fast gradient-echo (seg usually 0.1 to 0.2 mmol/kg. Recently, the performance of
IR-FGE) pulse sequence. Figure 17.3 demonstrates cine and DE-MRI for the detection of MI was tested in an interna-
DE-MRI images from a typical patient scan. Each delayed tional multicenter trial (48). In total, 282 patients with acute
enhancement image is acquired during an 8 to 10 seconds and 284 with chronic first AMI were scanned in 26 centers
breath-hold, and the imaging time for the entire examina- throughout the United States, Europe, and South America.
tion is generally 25 to 35 minutes. The sensitivity of DE-MRI increased with increasing
R R R R
Diastole Diastole Diastole

ECG
Figure 17.4. Timing dia-
gram of two-dimensional seg-
Nonselective Nonselective mented inversion recovery fast
180° inversion 180° inversion gradient-echo pulse sequence.
ECG, electrocardiogram; α,
α1 α12 α23 α1 α12 α23 shallow flip angle excitation;
RF pulses RF, radiofrequency; TD, trigger
delay; TI, inversion time delay.
Note that in this implementa-
1 12 23 1 12 23 tion, 23 lines of k-space are
Data acquired every other heartbeat.
acquisition See text for further details.
Acquisition Acquisition
(From Kim RJ, Shah DJ,
window window
Judd RM. How we perform
delayed enhancement imag-
ing. J Cardiovasc Magn Reson.
Timing 2003;5(3):505–514, with
delays TD TI TD TI permission.)
LWBK1209-ch17_p251-282.indd 256 17/05/13 5:09 PM

0.10 to 0.15 mmol/kg still provides excellent image contrast

Table 17.3 Typical Parameters
between injured and normal myocardium, and reduces
Parameter Typical Values the time required to wait between intravenous contrast
administration and delayed enhancement imaging (49–52).
Gadolinium dose 0.10–0.2 mmol/kg Sufficient time is required in order to allow the blood pool
Field of view 300–380 mm signal in the LV cavity to decline and provide discernment
In-plane voxel size 1.2–1.8 × 1.2–1.8 mm between LV cavity and hyperenhanced myocardium.
Slice thickness 6–8 mma The field of view (FOV) in both read- and phase-encoded
Flip angle 20–30 degrees
directions is minimized to improve spatial resolution and
Segmentsb 13–31
minimize breath-hold time without resulting in phase alias-
TI (Inversion time) Variable
Bandwidth 90–250 Hz/pixel ing or “wrap” artifact in the area of interest. For patients
TE (Echo time) 3–4 ms with heart rates less than 90 beats per minute, we typically
TR (Repetition time)c 5–10 ms acquire 23 lines of k-space data during the middiastolic por-
Gating factord 2 tion of the cardiac cycle. For a TR of 8 milliseconds, the data
K-space ordering Linear acquisition window is 184 milliseconds in duration (8 × 23 =
Fat saturation No 184). Since the middiastolic period of relative cardiac stand-
Asymmetric echo Yes still is reduced in patients with faster heart rates, we decrease
Gradient moment refocusinge Yes the number of segments (k-space lines) acquired per cardiac
a cycle in order to reduce the length of the imaging window.
Short-axis slices acquired every 10 mm to achieve identical posi-
tions as the cine images.
This eliminates blurring from cardiac motion during the
b
Fewer segments are used for higher heart rates. See text for details. k-space collection. In order to allow for adequate longitu-
c
TR is generally defined as the time between RF pulses; however, dinal relaxation between successive 180-degree inversion
scanner manufacturers occasionally redefine the TR to represent the pulses, we typically image every other heartbeat (gating fac-
time from the ECG trigger (R-wave) to the center or end of the data tor of 2). In our experience, an in-plane resolution of 1.2 to
acquisition window (i.e., ∼100 ms less than the ECG R–R interval). 1.8 mm by 1.2 to 1.8 mm with a slice thickness of 6 mm pro-
d
Image every other heartbeat. See text for details. vides an ideal balance between adequate signal-to-noise ratio
e
Also known as gradient moment nulling, gradient-moment rephras- while avoiding significant partial volume effects. As stated
ing, or flow compensation. See text for details. TI, inversion time; previously, the flip angle is kept shallow to retain the effects
TE, echo time; TR, repetition time.
of the inversion prepulse, but it can be relatively greater (30
Adapte From Kim RJ, Shah DJ, Judd RM. How we perform delayed
enhancement imaging. J Cardiovasc Magn Reson. 2003;5(3):505–514,
degrees) if larger doses of gadolinium are given (0.2 mmol/
with permission. kg) and the T1 of myocardium is correspondingly shorter.
In patients who are unable to hold their breath for the
duration required for the standard seg IR-FGE sequence, a
g adolinium dose, reaching 99% and 94% in acute and number of options are available to reduce the breath-hold
chronic MI, respectively, with a 0.3 mmol/kg dose (Fig. duration. Some simple strategies include (1) minimizing the
17.5). Furthermore, with a dose of 0.2 mmol/kg, if MI was FOV in the phase-encode direction (FOV phase); (2) imag-
identified, it was in the correct location in more than 97% ing with only the anterior coil elements (keeping the poste-
of patients. Clinically, we found that using doses as low as rior or spine coil elements turned off), allowing a smaller
Sensitivity (%)
Acute MI (n = 282) Chronic MI (n = 284)
99 99
95 (97–100) 94
92 (98–100) 92
(90–100) 87 (89–100)
84 (86–98) 84 (86–99)
83 (79–95)
(76–93) 80 (76–93)
(74–91)
(70–89) 73
(62–83)
52
48 45 (40–63)
(36–59) 44
(34–56) (33–52)
17
14 13
(8–26) 8 10
(6–22) (5–21) 6
3 (2–15) (3–17)
(1–12)
(0–7)
Precontrast Post-10 min Post-30 min Precontrast Post-10 min Post-30 min
Figure 17.5. Sensitivity of delayed enhancement magnetic resonance imaging (DE-MRI) for acute and
chronic myocardial infarction (MI) is summarized according to gadoversetamide dose group and imaging time
point. Numbers in parentheses are 95% confidence intervals. See text for further details. (From Kim RJ, Albert
TS, Wible JH, et al. Performance of delayed-enhancement imaging with gadoversetamide contrast for the
detection and assessment of myocardial infarction: An international, multicenter, double-blinded, randomized
trial. Circulation. 2008:117:629–637, with permission.)
LWBK1209-ch17_p251-282.indd 257 17/05/13 5:09 PM

FOV phase than expected without resulting in wrap-around 1.00

artifact over the heart; and (3) increasing the number of k-
space lines acquired per cardiac cycle (i.e., segments) with
the trade-off of worse temporal resolution, or with the same
0.50 Infarct
temporal resolution using a steady-state free precession
(SSFP) instead of FGE readout (resultant shorter echo time
Magnetization
Normal
[TE] and TR) at the expense of a reduction in pure T1 con-
trast effects. In some individuals, even with the approaches 0.00
outlined here, there is still inadequate breath-holding. In this 0 500 1000 1500 2000
situation, the use of respiratory navigators may be helpful “null” point
although imaging time is prolonged. Recently a fast, “sin-
gle-shot” version of DE-MRI has been developed that can –0.50
acquire snap-shot images during free breathing (53). This
technique uses an SSFP readout with parallel imaging accel-
eration and provides complete LV coverage in less than 30 A –1.00
seconds. This technique could be considered the preferred
approach in patients more acutely ill, unable to breath-hold, 1.00
or with irregular heart rhythm. However, compared with
standard, segmented DE-MRI, sensitivity for detecting MI
is mildly reduced, and the transmural extent of infarction
0.75
(TEI) may be underestimated (53,54).
Inversion Time Image intensity

0.50 Infarct
Selecting the appropriate TI is extremely important for Normal
obtaining accurate imaging results. The TI is chosen to
“null” normal myocardium, the time at which the magneti-
0.25
zation of normal myocardium reaches the zero crossing (Fig.
17.6A). It is at this point (or immediately just after) that
the image intensity difference between infarcted and normal
myocardium is maximized (Fig. 17.6C). If the TI is set too 0.00
short, normal myocardium will be below the zero crossing B 0 500 1000 1500 2000
and will have a negative magnetization vector at the time
1.00
of k-space data acquisition. Since the image intensity cor-
responds to the magnitude of the magnetization vector, the
image intensity of normal myocardium will increase as the
TI becomes shorter and shorter, whereas the image inten- 0.75 Image intensity difference
sity of infarcted myocardium will decrease until it reaches between infarct and
Image intensity
its own zero crossing (Fig 17.6B). At this point, infarcted normal myocardium
myocardium will be nulled and normal myocardium will 0.50
be hyperenhanced. On the opposite extreme, if the TI is set
too long, the magnetization of normal myocardium will be
above zero and will appear gray (not “nulled”). Although
areas of infarction will have high image intensity, the rela- 0.25 Max intensity
tive contrast between infarcted and normal myocardium will difference
be reduced. In principle, the optimal TI at which normal
myocardium is “nulled” must be determined by imaging 0.00
0 500 1000 1500 2000
iteratively with different TIs. In practice, however, only one
C Time (ms)
or two “test” images need to be acquired; with experience,
one can estimate the optimal TI on the basis of the amount Figure 17.6. A: Inversion recovery curves of normal and infarcted
of contrast agent that is administered and the time after con- myocardium assuming T1 of normal myocardium is 450 milliseconds
trast agent administration. Figure 17.7 shows images of a and infarcted myocardium is 250 milliseconds. The time at which
patient with an anterior wall MI in whom the TI has been the magnetization of normal myocardium reaches the zero crossing is
varied from too short to too long. Note that with the TI defined as the inversion time to “null” normal myocardium (312 milli-
set moderately too short (Fig. 17.7B), the anterior wall has seconds in this example). B: Image intensities resulting from an inversion
prepulse with various inversion delay times. Note that image intensities
some regions that are hyperenhanced; however, the total
correspond to the magnitude of the magnetization vector and cannot be
extent of hyperenhancement is less than that seen when the negative. C: Difference in image intensities between infarcted and nor-
TI is set correctly (Fig. 17.7C). This is due to the periph- mal myocardium as a function of inversion time. The optimal inversion
ery of the infarcted region passing through a zero cross- time is when the maximum intensity difference occurs. (From Kim RJ,
ing, thereby affecting its apparent size (55). If the TI is set Shah DJ, Judd RM. How we perform delayed enhancement imaging. J
too long (Fig. 17.7D), although the contrast is reduced as Cardiovasc Magn Reson. 2003;5(3):505–514, with permission.)
LWBK1209-ch17_p251-282.indd 258 17/05/13 5:09 PM

Figure 17.7. Delayed enhancement images in a subject with an anterior wall myocardial infarction in which
the TI has been varied from too short to too long. See text for details. (From Kim RJ, Shah DJ, Judd RM. How
we perform delayed enhancement imaging. J Cardiovasc Magn Reson. 2003;5(3):505–514, with permission.)
reviously stated, the total extent of hyperenhancement does

p null time for normal myocardium (56). These techniques, by
not change. Lastly, if the TI is set far too short (Fig. 17.7A), restoring signal polarity, can provide consistent contrast be-
infarct will be “nulled” and normal myocardium will hyper- tween infarcted and normal myocardium over a wide range
enhance. Thus, it is far better to err on the side of setting the of TIs and can eliminate the apparent reduction in infarct
TI too long rather than too short. size that is seen on images acquired with TIs that are too
As stated previously, data are acquired every other heart- short (56). TI “scout” sequences also have been developed to
beat in order to allow for adequate longitudinal relaxation aid in selecting the correct TI (57). In our experience, these
between successive inversion pulses. The time for recovery of sequences only provide a reasonable “first guess” at the cor-
96% of the bulk magnetization is approximately four times rect TI since they can be off by as much as 40 to 50 mil-
the T1 [M(t) = M0(1 – 2e−t/T1) or M(4T1) = M0(1 − 2e−4)]. liseconds. This discrepancy may be due to several reasons.
For example, if the T1 of normal myocardium after admin- For example, TI scout sequences generally do not account
istration of gadolinium is 400 milliseconds, then in order for the gating factor that is used in the seg IR-FGE sequence,
to achieve adequate longitudinal relaxation, there should which can lead to changes in the correct TI as described pre-
be approximately 1,600 milliseconds between successive in- viously. In addition, recovery of longitudinal magnetization
version pulses or every other heartbeat imaging in patients is not identical for inversion recovery prepared SSFP (often
with heart rates of 75 beats per minute (R–R interval = 800 used for TI scouting) (57) and inversion recovery prepared
milliseconds). In this example, the optimal TI to null nor- FGE sequences (58). In our laboratory, TI scout sequences
mal myocardium would be approximately 280 milliseconds are used infrequently.
[TI(null) = ln(2) × T1 = 0.69 × T1]. Occasionally, imaging is
performed every third heartbeat (gating factor of 3) if the
Imaging Time After Contrast
patient is tachycardic. Conversely, owing to limitations in
Administration
breath-hold duration and/or bradycardia, every heartbeat
imaging may be performed. In general, once the optimal TI has been determined, it does
Fortunately, for those that are unfamiliar with inversion not require adjustment if all delayed enhancement images
recovery relaxation curves, newer pulse sequences that allow can be acquired within approximately 5 minutes. However,
phase-sensitive reconstruction of inversion recovery data it is important to keep in mind that the gadolinium con-
may allow a nominal TI to be used rather than a precise centration within normal myocardium gradually washes out
LWBK1209-ch17_p251-282.indd 259 17/05/13 5:09 PM

1 postcontrast. Again, with appropriate adjustment of TI to

null normal myocardium, a recent multicenter investigation
0.9 420 of DE-MRI demonstrated nearly identical performance for
10 minutes postcontrast imaging compared with 30 minutes
Plasma [Gd-DTPA]
0.8 postcontrast imaging (48).
Inversion time (ms)

379 ms
(mmol/L)
0.7 370
Pathophysiologic Validation
0.6
Background
0.5 320
The clinical utility of DE-MRI is determined both by the
304 ms
0.4 quality of the images and their relationship to the underlying
pathophysiology. While clinical studies clearly help to define
0.3 270 the information available, direct comparison of the MR
0 5 10 15 20 25 30 35 40 45 50 55 60 images to histologic tissue samples obtained from animals is
Time after injection of Gd-DTPA a unique and important source of information. In principle,
(min) several species of animals can be employed for the study of
Figure 17.8. Solid line: Monoexponential fits to the plasma gado- myocardial contrast enhancement patterns. However, large
linium concentration (left-hand y axis) as a function of time postcon- animals such as the dog provide a practical advantage in
trast administration extrapolated from data from Weinmann et al. (59) that they can be studied on clinical scanners with the identi-
in humans. Dashed green line: MRI inversion time (right-hand y axis) cal pulse sequences used for humans, thereby ensuring the
to null normal myocardium calculated from the data of the solid blue clinical relevance of the findings.
line. Open circles: The calculated TI to null normal myocardium at In humans, the type and age of an ischemic injury are
5 and 30 minutes postcontrast administration. (From Kim RJ, Shah
complex and often are only partially documented. Using ani-
DJ, Judd RM. How we perform delayed enhancement imaging.
J Cardiovasc Magn Reson. 2003;5(3):505–514, with permission.)
mal models, however, several well-defined states of ischemic
injury can be studied in a controlled setting. These states
include acute infarction, chronic infarction, and severe but
with time, and the TI will need to be adjusted upward if reversible ischemic injury. In this section, we review the data
delayed enhancement imaging is performed over a long time concerning MRI contrast enhancement patterns and their
span (>5 minutes). For example, Figure 17.8 demonstrates relationship to the underlying pathophysiology in animal
that the plasma concentration of gadolinium decreases expo- models of these three pathologic states. Then we turn to
nentially with time following contrast administration (solid studies performed in humans and determine if the concepts
blue line). Since interstitial concentrations of Gd-DTPA in gleaned from the animal models can be applied successfully
the myocardium depend primarily upon plasma concentra- in the clinical setting. Since histopathologic correlations are
tions, the correct TI needed to null normal myocardium can rare in human studies, particular attention should be paid
be estimated from basic physical principles (59). The dashed to the “gold standard” used to define MI and viability when
green line in Figure 17.8 depicts the correct TI that is needed reviewing the clinical literature.
to account for the pharmacokinetics of the contrast agent. In
this example (gadolinium dose of 0.125 mmol/kg), the cor- Animal Studies
rect TI to null normal myocardium at 30 minutes postcon-
Acute Myocardial Infarction
trast will be 379 milliseconds as compared to 304 millisec-
onds at 5 minutes. The basic premise here is not that one can Patterns of contrast enhancement have been described in
calculate the correct TI for a given time point after contrast the setting of acute infarction with and without reperfusion.
administration, but that the TI needs to be adjusted, some- Most, if not all, studies report that regions of acute infarc-
times significantly, if the imaging time is prolonged. tion appear hyperenhanced in T1-weighted images acquired
Oshinski et al. (60) suggested that the “accurate determi- more than a few minutes postcontrast (32–38,41,44,45,47,
nation of infarct size by delayed enhancement MRI requires 61–64). The exact relationship of the observed hyperen-
imaging at specific times after gadolinium-DTPA injection.” hanced regions to the underlying pathophysiology, however,
This conclusion was based on the observation that the size has been a subject of debate in the past and deserves some
of the hyperenhanced regions decreased with increasing additional consideration. Understanding the issues requires
time after contrast administration. However, in this study, some background regarding the development of ischemic
the TI was not adjusted to “null” normal myocardium but myocardial injury.
held constant throughout the 40 minutes of imaging post- In the traditional view of ischemic myocardial injury,
contrast. In this situation, one is in effect choosing a TI an “area at risk” (65) is defined as the territory with re-
that is further and further from the correct TI (too short) as duced blood flow following occlusion of a coronary artery.
the time after contrast administration increases. Mahrholdt Following occlusion, myocardial contractile function falls al-
et al. (59) showed, in the setting of chronic infarction, that most immediately throughout the “area at risk” (66). Little
when the TI is adjusted appropriately to “null” normal or no cellular necrosis, however, is found until about 15
myocardium, the size of hyperenhanced regions does not minutes after occlusion (67,68). After 15 minutes, a “wave-
change if imaging is performed between 10 and 30 minutes front” of necrosis begins in the subendocardium and grows
LWBK1209-ch17_p251-282.indd 260 17/05/13 5:09 PM

toward the epicardium over the next few hours (65,69). in a series of slices, usually short-axis slices from base to apex.
During this period, the size of the “area at risk” remains the This “one point per animal” approach, however, effectively
same, but the size of the infarcted region within the “area discards the much larger amount of information that could be
at risk” increases. The interpretation of MRI hyperenhance- obtained by careful image registration.
ment in the setting of acute infarction, therefore, requires a To address this issue in our laboratory we introduced
direct comparison of the MR images with the “area at risk” the intermediate step of high-resolution ex vivo imaging.
and the “area of infarction.” Immediately after the in vivo DE-MRI images were acquired,
Both the “area at risk” and the “area of infarction” we removed the hearts, cooled them to 4°C, attached three
can only be precisely defined in histologic tissue sections. MRI-visible registration markers, placed a balloon in the LV
The “area at risk” can be identified experimentally with cavity (filled with deuterium to cause a proton signal void),
techniques involving microspheres (70) or blue dye (65), and suspended the hearts in an extremity radiofrequency
whereas the “area of infarction” can be defined by staining coil for high-resolution imaging. The absence of cardiac mo-
with TTC (28). Understanding the relationship of MRI con- tion allowed us to acquire T1-weighted images with spatial
trast enhancement patterns to the underlying pathophysiol- resolutions of 500 × 500 × 500 μm. After imaging, the hearts
ogy depends importantly, therefore, on registration of the in were further cooled and made partially stiff by repeated im-
vivo MR images to the postmortem tissue samples. mersion in −80°C ethanol and then sliced every 2 mm in a
In practice, the accurate registration of in vivo MR im- commercial rotating meat slicer along the same planes used
ages to histologic tissue sections is difficult for several reasons. for imaging (defined by the three MRI-visible markers).
First, the three-dimensional (3D) shapes of both the “area at After this, each tissue slice was stained for myocyte necrosis
risk” and the “area of infarction” are very complex, and the with TTC (28) and compared with the MR images.
details of their shapes are generally beyond the resolution of Figure 17.9 shows a comparison of MRI to histology in
in vivo MRI. Second, after removal from the chest, the heart an animal with acute infarction using this approach. The
itself is easily deformed and almost impossible to cut along the “match” between TTC and MRI was extremely close, and
exact plane used for imaging. As a result, tissue slices aligned even minute details such as “fingers” of necrosis defined by
by eye and cut by hand almost certainly will not be registered TTC were readily identified in the T1-weighted MR images.
precisely enough to allow slice-by-slice comparisons of the de- This “match” held in a series of animals that were studied
tailed shapes of hyperenhanced regions observed in vivo by with acute infarction both with and without reperfusion
MRI with those defined histologically. In this setting, many (Fig. 17.10). On the basis of these findings, we concluded
investigators have avoided the image registration problem en- that in the setting of acute infarction, the spatial extent of
tirely by computing the percentage of the entire LV that is at hyperenhancement by MRI is identical to the spatial extent
risk or infarcted on the basis of the sum of regions identified of myocardial necrosis (71).
Figure 17.9. Comparison of ex vivo,

high-resolution, delayed enhanced MR
images with acute myocardial necrosis
defined histologically by TTC staining.
See text for details. TTC, triphenyl tetro-
zolium. (From Kim RJ, Fieno DS, Parrish
TB, et al. Relationship of MRI delayed
contrast enhancement to irreversible injury,
infarct age, and contractile function.
Circulation. 1999;100(19):1992–2002,
with permission.)
LWBK1209-ch17_p251-282.indd 261 17/05/13 5:09 PM

1 Day 3 Days 8 Weeks

45 90 12
R = 0.99 R = 0.99 R = 0.97
40 80
y = 1.02x + 0.28 y = 1.01x + 0.59 10 y = 1.03x − 0.14
MRI hyperenhancement
35 70
30 60 8
25 50
6
20 40
15 30 4
10 Identity 20 Identity
Not reperfused Not reperfused 2 Identity
5 10 Not reperfused
Reperfused Reperfused
0 0 0
0 5 10 15 20 25 30 35 40 45 0 10 20 30 40 50 60 70 80 90 0 2 4 6 8 10 12
Infarct size (% of slice)
Figure 17.10. Comparison of hyperenhanced regions by MRI with infarct size measured by TTC at 1 day,
3 days, and 8 weeks with and without reperfusion. See text for details. TTC, triphenyl tetrozolium.
(From Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed contrast enhancement to irreversible
injury, infarct age, and contractile function. Circulation. 1999;100(19):1992–2002, with permission.)
The mechanisms at the cellular level responsible for hy- Chronic Myocardial Infarction
perenhancement have not been fully elucidated. There is evi-
dence that gadolinium concentrations are elevated in regions Unlike acute infarcts, which are characterized by necrotic
of acute infarction (64,72), and this observation would ex- myocytes, chronic infarcts are characterized by a dense col-
plain the shortened T1 in these regions. Figure 17.11 describes lagenous scar. Owing to these underlying structural differ-
one possible mechanism for hyperenhancement of acute in- ences, there is no reason a priori to believe that acute and
farction. Whole-body data strongly suggest that in normal chronic infarcts will appear similar in contrast-enhanced
myocardial regions, gadolinium is excluded from the myocyte MR images. To address this issue, we scanned dogs 8 weeks
intracellular space by intact sarcolemmal membranes (73,74). after MI, when infarct healing had clearly progressed.
The hypothesis demonstrated in Figure 17.11 states that in Figure 17.12 shows an example of a dog with an 8-week-old
acutely infarcted regions, the myocyte membranes are rup- infarct in which the hyperenhanced region observed in vivo
tured, allowing gadolinium to passively diffuse into the in- is clearly associated with a dense collagenous scar (verified
tracellular space. The result is an increased concentration of postmortem). Using the same technique of high-resolution
gadolinium at the tissue level and therefore hyperenhance- ex vivo imaging described previously in the setting of acute
ment. Loss of sarcolemmal membrane integrity is thought to infarction, the regions of hyperenhancement observed in the
be very tightly related to cell death (66–68), and the idea that setting of chronic infarction also appeared to be identical
an event specific to cell death is related to hyperenhancement in shape and size to the infarcted regions defined histologi-
would explain the nearly one-to-one relationship of hyperen- cally (Fig. 17.10) (71). Furthermore, these data indicate that
hancement to necrosis shown in Figure 17.9. chronic infarcts systematically hyperenhance.
Figure 17.11. Potential

mechanisms of hyperenhance-
ment in acute and chronic myo-
cardial infarction.See text for
details.
LWBK1209-ch17_p251-282.indd 262 17/05/13 5:09 PM

In this case, the concentration of gadolinium in scar would be

greater than in normal myocardium because of the expanded
volume of distribution of gadolinium. Higher concentrations
of gadolinium would lower T1, as in acute infarction, and the
regions of scar would appear hyperenhanced by DE-MRI.
Reversible Ischemic Injury

Given that both acute and chronic myocardial infarcts
exhibit hyperenhancement, the question of whether severe
but reversible ischemic injury exhibits hyperenhancement
takes on added importance. In our laboratory, we used two
separate experimental approaches to examine this question.
Figure 17.12. In vivo delayed contrast-enhanced images of a dog In the first approach, severe but reversible ischemic injury
with an 8-week-old myocardial infarction (right). Despite replace- was induced in the magnet and then delayed enhanced
ment of necrotic myocytes with dense collagenous scar (revealed by images were acquired. In the second approach, high-reso-
trichrome staining of infarcted region) (inset in left panel), hyper- lution ex vivo images were compared with the “area at risk
enhancement is still observed. (From Kim RJ, Fieno DS, Parrish but not infarcted” defined histologically.
TB, et al. Relationship of MRI delayed contrast enhancement to Figure 17.13 shows examples of the first approach (71).
irreversible injury, infarct age, and contractile function. Circulation. Under sterile conditions, two coronary arteries were ma-
1999;100(19):1992–2002, with permission.) nipulated. The first coronary artery was permanently ligated
to cause MI. The second coronary artery was instrumented
with a reversible hydraulic occluder and a Doppler flow
The mechanism of hyperenhancement in chronic infarcts meter. The animals were then allowed to recover and were
remains to be elucidated. One potential mechanism is pro- studied 3 days later. While in the magnet, regional wall
posed in Figure 17.11. Although myocardial scar is charac- motion was examined with cine MRI. The reversible oc-
terized by a dense collagenous matrix, at a cellular level, the cluder was then inflated for 15 minutes. During this period
interstitial space between collagen fibers may be significantly of occlusion, cine MRI was repeated to identify new
greater than the interstitial space between densely packed liv- wall motion abnormalities to ensure that the region distal
ing myocytes that are characteristic of normal myocardium. to the reversible occluder was within the imaging plane.
CIRC
flowmeter
Reversible −
Figure 17.13. Stunned myocardium does not exhibit hyperenhancement. See text for details. (From Kim
RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed contrast enhancement to irreversible injury,
infarct age, and contractile function. Circulation. 1999;100(19):1992–2002, with permission.)
LWBK1209-ch17_p251-282.indd 263 17/05/13 5:09 PM

After 15 minutes, the occluder was released, and flow was intensities in order to provide data that were independent of
restored (verified by Doppler flow). The purpose of the the MRI technique. To measure gadolinium concentrations
15-minutes occlusion was to induce severe but reversible over a range of myocardial injuries, electron probe x-ray mi-
ischemic injury (67,68). Cine MRI was then repeated a third croanalysis (EPXMA) was performed in animal models with
time to document myocardial stunning. The colored bull’s- reversible and irreversible injury. Infarcted regions were
eye plots in Figure 17.13 correspond to wall thickening at defined by antimyoglobin antibody or TTC staining. Regions
this time. As can be seen in Figure 17.13, both the region at risk were defined by fluorescent microparticles adminis-
associated with infarction (yellow arrows) and the region as- tered during coronary occlusion. The results demonstrated
sociated with severe but reversible ischemic injury (green ar- that compared with normal regions, gadolinium concentra-
rows) show abnormal wall thickening. Gadolinium was then tion was increased in both acute and chronic infarction but
injected, and the images were inspected for hyperenhance- not increased in “at risk but not infarcted” regions (Fig.
ment. The region of infarction exhibited hyperenhancement 17.14). The conclusion was that regional elevations in myo-
while the region of severe but reversible ischemic injury cardial gadolinium concentrations are exclusively associated
(“stunned” myocardium) did not. When these same animals with irreversible ischemic injury.
were scanned 8 weeks later, wall thickening had returned The data of Figure 17.13 underscore that dissociation be-
to normal in the region of severe but reversible ischemic injury tween wall thickening and delayed enhancement can occur
(71). Histologic examination (also at 8 weeks) revealed infarc- in which wall thickening is impaired but hyperenhancement
tion in the territory subtended by the permanently occluded is not observed. In the setting of acute ischemic injury, this
artery, but no evidence of infarction was found distal to the condition is related to the phenomenon of myocardial “stun-
reversible occluder. These in vivo data support the view that ning” (1,77), in which cell death has not occurred but con-
severe but reversible ischemic injury does not result in hyper- tractile dysfunction persists for days or even weeks after a
enhancement (71). severe ischemic event. The data suggest that in the setting of
In the second approach, fluorescent microparticles in- an acute ischemic event, regions exhibiting contractile dys-
jected into the left atrium during coronary artery occlusion function can be subdivided into irreversibly and reversibly
were used to define the “area at risk” and TTC was used injured regions defined as those with and without hyperen-
to define the usually smaller area of infarction. Using both hancement, respectively. This concept suggests that the pres-
datasets, we could identify the region that was “at risk but ence or absence of hyperenhancement in regions of contrac-
not infarcted.” High-resolution ex vivo MR images clearly tile dysfunction may be useful in the detection of myocardial
demonstrated that the “at risk but not infarcted” myocar- salvage early after acute infarction. If correct, this approach
dium did not exhibit hyperenhancement. Light microscopy would represent a new technique to define myocardial sal-
of this region verified that normal myocyte architecture was vage in patients following acute infarction.
present. These data also strongly support the view that se- To test the hypothesis that DE-MRI can be used to index
vere but reversible ischemic injury does not result in hyper- myocardial salvage, we initiated another study in which
enhancement (75). animals were imaged 3 days, 10 days, and 4 weeks after
Further evidence that hyperenhancement is exclusively transient coronary artery occlusion (78). The hypothesis
associated with irreversible injury is provided by Rehwald et al. was that recovery of contractile dysfunction at 4 weeks
(76). Since MR image intensities are highly dependent on the could be predicted by DE-MRI at 3 days. Specific examples
pulse sequence that is used, Rehwald et al. directly exam- from that study are shown in Figure 17.15, in which each
ined regional gadolinium concentrations rather than image row represents a different animal. The first three columns
Figure 17.14. Reversible ischemic injury.

Representative results in remote (R) and reversibly
injured but not infarcted (RNI) myocardium.
A: Anticardiac myoglobin stain shows that all
regions are viable. B: Hematoxylin and eosin stain
also shows all regions are viable. C: Lack of micro-
spheres identifies RNI region. D: Photograph of
cardiac slice on stub. Spectra from R and RNI are
very similar, showing similar elemental concentra-
tions in both regions. SEM, scanning electron micro-
scope. (From Rehwald WG, Fieno DS, Chen EL, et al.
Myocardial magnetic resonance imaging contrast
agent concentrations after reversible and irreversible
ischemic injury. Circulation. 2002;105:224–229.)
LWBK1209-ch17_p251-282.indd 264 17/05/13 5:09 PM

Figure 17.15. Contrast enhancement and wall thickening at 3 days (columns 1 to 3) compared to wall
thickening at 28 days (columns 5 to 6) in three different animals (rows). See text for details. (From Hillenbrand
HB, Kim RJ, Parker MA, et al. Early assessment of myocardial salvage by contrast-enhanced magnetic resonance
imaging. Circulation. 2000;102(14):1678–1683, with permission.)
show MRI data acquired at 3 days, whereas the data in myocardium, and no hyperenhancement of these regions
columns four and five were acquired at 4 weeks. At 3 days, would be observed.
all three animals showed contractile dysfunction in the left
anterior descending (LAD) coronary artery perfusion terri-
tory by cine MRI (columns 2 and 3). The first animal (row Human Studies
1), however, showed nearly transmural hyperenhancement,
Acute Myocardial Infarction
the second animal (row 2) about 50% transmural hyper-
enhancement, and the third animal (row 3) almost no hy- Simonetti et al. (40) performed the first clinical study to
perenhancement. Four weeks later, contractile function did evaluate segmented inversion recovery DE-MRI in the set-
not improve in the first animal, improved partially in the ting of acute MI. In this study, 18 consecutive patients were
second animal, and completely recovered in the third ani- imaged 19 ± 7 days after acute MI documented by an appro-
mal. A summary of the results from this study is shown in priate rise (> two times the upper limit of normal) and fall
Figure 17.16. These data underscore that as predicted by in CK-myocardial band (CK-MB) isoenzyme levels. Figure
the results of Figure 17.13, contrast enhancement patterns 17.17 shows representative images in three patients from
observed early after MI can be used to index the extent of this study. Starting from the left panel, infarction was due
myocardial salvage (78). to occlusion of the LAD, left circumflex artery, and right
The mechanism for the lack of hyperenhancement in regions coronary artery (RCA), respectively. Myocardial hyperen-
of severe but reversible ischemic injury may be directly related hancement is clearly visible in these patients in the appropri-
to the mechanism for hyperenhancement in acute infarcts ate infarct-related artery (IRA) perfusion territory. Similar
(i.e., it may relate to the integrity of the myocyte mem- results were observed in the other 15 patients. On average,
brane). Although severe but reversible ischemic injury has hyperenhanced regions had image intensities that were 485%
many effects on the myocyte, the sarcolemmal membrane ± 43% higher than those in normal myocardial regions. As
remains intact (67,68), and therefore, presumably continues discussed previously, the degree of hyperenhancement was
to exclude the MRI contrast agent from the intracellular approximately 10-fold greater than that in previous reports
space. In this setting, the volume of distribution of the (Table 17.2).
contrast agent in regions of severe but reversible ischemic In a later study, we evaluated a series of 24 patients
injury would be expected to remain similar to that in normal who presented with their first MI (documented by cardiac
LWBK1209-ch17_p251-282.indd 265 17/05/13 5:09 PM

100 cardiographic infarct location. Ingkanisorn et al. (81) stud-

P < 0.0001 for trend
improved wall motion–Day 28
ied 33 patients within 5 days after first-time acute MI. All

80
57 8
Day 10 patients had hyperenhancement despite the fact that many
368 47
had small infarcts (troponin-I ≤ 9 ng/mL). There was a good
% Segments with
Day 28
21 correlation between infarct size measured by DE-MRI and

60 88 peak troponin-I levels in the 23 patients who were reper-
fused acutely (r = 0.83, p < 0.001), whereas there was a poor
40 correlation in the remaining 10 patients who were not (r =
5
47
0.28, p = NS). An international multicenter trial confirmed
20 that DE-MRI is a highly sensitive and accurate method for
264 33 37 11 2 the detection of acute MI (48).
368 47 88 47 17
0
0 1–25 26–50 51–75 76–100
Chronic Myocardial Infarction
40 n = 200 n = 30 The studies of chronic infarction, perhaps, best demon-
strate the importance of image quality in delayed contrast
Day 3 enhancement imaging. For example, using older tech-
Wall thickening (%)
30
Day 10 niques, Eichstaedt et al. (32), Nishimura et al. (82), and
n = 72
Day 28 van Dijkman et al. (42), all observed gadolinium hyperen-
20 hancement in patients with acute MI but found no hyperen-
hancement in patients with chronic infarction. These reports
n = 39 formed the basis for the initial widespread conclusion that
n = 16
10 chronic infarcts do not hyperenhance. Later, Fedele et al.
(43) and Ramani et al. (46) suggested that this conclusion
is erroneous. They described hyperenhancement in patients
0
0 1–25 26–50 51–75 76–100 with chronic CAD and a high clinical likelihood of chronic
% Hyperenhancement on day 3 infarction. Unfortunately, biochemical evidence for infarc-
tion was not provided, the age of infarction was unknown,
Figure 17.16. The likelihood of wall thickening improvement
(upper panel) and absolute wall thickening (lower panel) as a function
and differences in image intensity were modest, with hyper-
of the transmural extent of hyperenhancement. (From Hillenbrand enhanced regions having on average less than 60% increase
HB, Kim RJ, Parker MA, et al. Early assessment of myocardial sal- in image intensity over nonhyperenhanced regions.
vage by contrast-enhanced magnetic resonance imaging. Circulation. Despite these conflicting results, we postulated that
2000;102(14):1678–1683, with permission.) human chronic MI hyperenhances. This hypothesis was
based on our experimental animal data demonstrating that
collagenous scar hyperenhances in both ex vivo (Fig. 17.10)
e nzymes) and were successfully reperfused (thrombolytics and in vivo imaging protocols (Fig. 17.12). To test this hy-
or angioplasty) (79). All patients underwent DE-MRI using pothesis, we enrolled patients at the time of acute infarction
the seg IR-FGE sequence within 7 days of their MI. There on the basis of abnormal CK release and then performed
was a strong correlation between infarct size defined by DE-MRI several months later after infarct healing (52). To
DE-MRI and peak CK-MB values (Fig. 17.18). In a similar assess the specificity of the findings, DE-MRI was also per-
study, Beek et al. (80) studied 30 patients 7 ± 3 days after a formed in patients with nonischemic cardiomyopathy and in
first-time reperfused acute MI. They found that 29 patients healthy volunteers.
(97%) showed regional hyperenhancement, and in all 29, In the patients with chronic MI, we observed hyperen-
the area of hyperenhancement corresponded to the electro- hancement in a variety of sizes, ranging from large, fully
Figure 17.17. Short-axis

DE-MRI images in three
patients with acute myocardial
infarction.The arrows point
to the hyperenhanced region,
which was in the appropriate
infarct-related artery perfu-
sion territory.(From Simonetti
OP, Kim RJ, Fieno DS, et al.
An improved MR imaging
technique for the visualiza-
tion of myocardial infarction.
Radiology. 2001;218(1):215–
223, with permission.)
LWBK1209-ch17_p251-282.indd 266 17/05/13 5:09 PM

60 territories. Despite the small volume of hyperenhancement,

the hyperenhancement zone was visually distinct and clearly
50 nontransmural. In all patients with hyperenhancement, the
difference in image intensity between hyperenhanced and
Infarct size by MRI (% LV)
r = 0.9
40
p < 0.001 remote myocardium was more than six standard deviations
of remote region intensity (mean difference = 17 standard
deviations).
30
Twenty-nine of 32 patients with 3-month-old infarcts
(91%) and all 19 with 14-month-old infarcts exhibited hy-
20 perenhancement. For the patients with hyperenhancement in
whom the IRA territory could be determined by coronary
10 angiography, 24 of 25 (96%) with 3-month-old infarcts and
all 14 with 14-month-old infarcts had hyperenhancement in
0 the correct IRA perfusion territory. Regardless of the pres-
0 200 400 600 800 1000 ence or absence of Q-waves, the majority of patients with
Peak CK-MB hyperenhancement had only nontransmural involvement.
Figure 17.18. Relationship of infarct size by MRI to peak Transmural infarction was present in 12 of the 29 (41%)
CK-MB. CK-MB, creatine kinase-myocardial band. (From Choi KM, Q-wave infarctions and in 3 of the 19 (16%) non-Q–wave
Kim RJ, Gubernikoff G, et al. Transmural extent of acute myocardial infarctions. None of the 20 patients with nonischemic di-
infarction predicts long-term improvement in contractile function. lated cardiomyopathy exhibited hyperenhancement despite
Circulation. 2001;104(10):1101–1107, with permission.) significant LV systolic dysfunction. Likewise, none of the
11 normal volunteers exhibited hyperenhancement. The
sensitivity of DE-MRI for the detection of healed infarction
transmural hyperenhancement that extended over several was 91% and 100% in the 3- and 14-month-old groups,
short-axis slices to small, subendocardial hyperenhance- respectively. The specificity was 100% when patients with
ment that was visible only in a single sector of a single view. nonischemic dilated cardiomyopathy and normal volunteers
Figure 17.19 shows typical images in three patients with were considered.
large transmural hyperenhancement in different coronary To establish the clinical reproducibility of DE-MRI for
artery territories. Figure 17.20 demonstrates typical images the measurement of chronic infarct size, Mahrholdt et al.
in three patients with minor CK-MB elevations and small (59) performed two scans (by different operators) on the
regions of hyperenhancement in different coronary artery same day in 20 patients with chronic MI. All patients had
Figure 17.19. Typical short-

axis and long-axis views of three
patients with large transmural
hyperenhancement in different
coronary artery territories.
(From Wu E, Judd RM, Vargas
JD, et al. Visualisation of presence,
location, and transmural extent of
healed Q-wave and non-Q–wave
myocardial infarction. Lancet
2001;357(9249):21–28, with per-
mission.)
LWBK1209-ch17_p251-282.indd 267 17/05/13 5:09 PM

Figure 17.20. Typical short-

and long-axis views of three
patients with minor CK-MB
elevations and small regions of
hyperenhancement in different
coronary artery territories.
CK-MB, creatine kinase-
myocardial band. (From Wu
E, Judd RM, Vargas JD, et al.
Visualisation of presence, loca-
tion, and transmural extent of
healed Q-wave and non-Q–wave
myocardial infarction. Lancet.
2001;357(9249):21–28, with
permission.)
cardiac enzyme documented MI that was at least 1 year old. In is due to myocardial necrosis or to myocardial stunning.
15 patients, two SPECT scans (Tc-MIBI) were also performed Differentiation between these two conditions can have sig-
on the same day of the MRI scans. Hyperenhancement was nificant clinical implications.
observed on all of the MRI scans, and the mean difference in The experimental studies described in the earlier sections
infarct size between the paired scans was negligible (bias = indicate that DE-MRI can be used to distinguish necrotic
−0.1% LV). The coefficient of repeatability for DE-MRI was from stunned myocardium since only the former exhibits
±2.4% LV, which was smaller than that for SPECT (±4% hyperenhancement. Moreover, the extent of hyperenhance-
LV). Likewise, Ingkanisorn et al. (81) performed DE-MRI ment early post-MI was found to be useful in predicting
in 30 patients with documented chronic MI. Infarct size was the magnitude of functional improvement that could be
highly reproducible if imaging was repeated at a chronic expected weeks to months later. To test whether these re-
time point, whereas infarct size was smaller compared with lationships would be equally valid in humans, we studied
images that were obtained at an acute time point, most likely 24 consecutive patients who presented with their first acute
owing to infarct shrinkage during the healing process. MI and were successfully revascularized (79). All patients
The results of these studies, along with the gadoverset- underwent cine and delayed enhancement imaging within
amide multicenter trial (48) are consistent with the animal 7 days of their MI, and then cine imaging was repeated
studies discussed earlier and confirm that chronic MI in hu- 8 to 12 weeks later. Figure 17.21 shows the percentage of
mans hyperenhances and that DE-MRI can provide an accurate improved segments at the follow-up time point as a function
and permanent record of prior MI. of the TEI determined by DE-MRI early after acute MI. The
percentage of improved segments decreased with increasing
TEI (p < 0.001). For example, 213 of 275 segments (77%)
Reversible Ischemic Injury
without any infarction showed improvement, whereas only
In the setting of acute ischemic disease, restoration of blood 3 of 64 segments (5%) with over 75% infarction showed
flow with primary angioplasty or thrombolytic therapy improvement. When segmental analysis was performed
has been shown to result in salvage of ischemic but viable using a mixed-effects model, which takes into account the
myocardium, improvement in EF, and long-term improve- multiple measurements from the same patient and a vari-
ment in survival (46,82–85). Although the magnitude of the able number of segments from each patient, the TEI was
benefit appears to be related to the duration of occlusion, the best predictor of segmental improvement (p < 0.001).
other factors such as intermittency of occlusion, mainte- In addition to predicting improvement in regional func-
nance of patency, myocardial oxygen demands, and extent tion, DE-MRI also provided the best predictor of improve-
of collateral circulation also affect on the rate of necrosis ment in global function. Figure 17.22 demonstrates that
and, therefore, on the amount of myocardial salvage (86). the percentage of the LV that was dysfunctional but viable
In the immediate postinfarction setting, even after success- (i.e., ≤25% TEI) was directly related to the change in mean
ful reperfusion has been achieved, myocardial dysfunction LV wall thickening score (r = 0.87, p < 0.0001) and LVEF
can persist, and it can be difficult to determine whether this (r = 0.65, p = 0.002).
LWBK1209-ch17_p251-282.indd 268 17/05/13 5:09 PM

A B
Figure 17.21. Results of segmental analysis. For segments that were dysfunctional on the early post-MI
scan, the likelihood that contractile function improved on the follow-up scan decreased as the transmural
extent of infarction (hyperenhancement on DE-MRI) increased. Numbers above each column refer to the
number of segments. (From Choi KM, Kim RJ, Gubernikoff G, et al. Transmural extent of acute myocardial
infarction predicts long-term improvement in contractile function. Circulation. 2001;104(10):1101–1107,
with permission.)
40
2.0
r = 0.65
r = 0.87 30
p = 0.002
Change in mean wall thickening score
1.5 p < 0.0001

Change in ejection fraction (%)
20
1.0
10
0.5
0
0 20 40 60 80 100
0.0
0 20 40 60 80 100 −10
A −0.5 Dysfunctional but viable (% LV) −20 Dysfunctional but viable (% LV) B
Figure 17.22. Results of global analysis. The size of the dysfunctional but viable region was the best
predictor of both ejection fraction and mean wall thickening score. (From Choi KM, Kim RJ, Gubernikoff
G, et al. Transmural extent of acute myocardial infarction predicts long-term improvement in contractile func-
tion. Circulation. 2001;104(10):1101–1107, with permission.)
LWBK1209-ch17_p251-282.indd 269 17/05/13 5:09 PM

The ability of DE-MRI to predict LV functional improve- for those without hyperenhancement to 0% improving for
ment in humans after reperfused acute MI has now been those with more than 75% hyperenhancement). When the
confirmed by several groups. Gerber et al. (87) studied 20 volume of dysfunctional but viable myocardium was cal-
patients in whom the transmural extent of hyperenhance- culated on a patient-by-patient basis from the DE-MRI im-
ment early (4 ± 2 days) after acute MI was inversely related ages, this parameter predicted the magnitude of improve-
to functional improvement 7 months later. Beek et al. (80) ment in global function after revascularization as measured
studied 30 patients in whom DE-MRI 7 ± 3 days after acute by mean LV wall motion score and EF (p < 0.001 for both)
MI predicted functional improvement at 13 ± 3 weeks. (Fig. 17.24B). Schvartzman et al. (90) demonstrated similar
Kitagawa et al. (88) observed similar findings in 22 patients findings for a patient cohort with more severe cardiomy-
with acute MI. opathy (LVEF 28% ± 10%). In their study of 29 patients,
In the setting of chronic ischemic disease, viability testing 82% of dysfunctional regions with no hyperenhancement
is often performed to differentiate hibernating myocardium improved compared to only 18% with ≥50% hyperen-
from scar tissue and to predict contractile response to coro- hancement. Selvanayagam et al. (91) demonstrated nearly
nary revascularization. In the initial article that reported that identical results regarding the predictive accuracy of TEI
DE-MRI could be used to predict functional improvement for the recovery of regional function in a cohort of 52 pa-
after coronary revascularization, cine and delayed enhance- tients with relatively well-preserved systolic function (LVEF
ment imaging was performed in 50 consecutive patients with 62% ± 11%) undergoing CABG surgery.
LV dysfunction before they underwent surgical or percuta- These data demonstrate that there is a smooth progres-
neous revascularization (89). Cine MRI was repeated in 41 sive relationship between the likelihood of functional im-
patients approximately 11 weeks after revascularization in provement and the transmural extent of hyperenhancement
order to document changes, if any, in regional and global by DE-MRI. This indicates that the use of a single cutoff
wall motion. Figure 17.23 shows representative cine and value of hyperenhancement on which to base predictions of
contrast images from two patients in the study. Regional functional improvement would not have a physiologic basis
function recovered in the patient without hyperenhance- and would be suboptimal. In fact, the ability to approach
ment of the dysfunctional region, whereas it did not recover viability as a continuum rather than in a binary manner is one
in the patient with nearly transmural hyperenhancement of of the greatest strengths of DE-MRI. For example, accord-
the dysfunctional region. Similar to the findings in the acute ing to the data from Kim et al. (89), if a cutoff value of 25%
setting, regional wall motion analysis demonstrated that the hyperenhancement were chosen, the positive and negative
likelihood of functional improvement was inversely related predictive accuracy rates for functional improvement would
in a progressive stepwise fashion to the transmural extent of be 71% and 79%, respectively. Although these predictive
hyperenhancement (Fig. 17.24A). For instance, 78% of seg- accuracy rates compare favorably with those reported previ-
ments with no hyperenhancement improved, whereas only ously for other imaging modalities (92), the full diagnostic
2% of segments with more than 75% hyperenhancement information portrayed by DE-MRI is not utilized. Consider
improved. The relationship was even steeper for segments the situation in which a region has more than 75% hyper-
with at least severe hypokinesia at baseline (86% improving enhancement. The data from Kim et al. demonstrate that 57
Figure 17.23. Representative cine and contrast enhancement images from two different patients. See text
for details. (From Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance imaging to
identify reversible myocardial dysfunction. N Engl J Med. 2000;343(20):1445–1453, with permission.)
LWBK1209-ch17_p251-282.indd 270 17/05/13 5:09 PM

Regional Global
A (%) B
Figure 17.24. Prediction of improvement in regional and global systolic function by delayed enhancement
MRI. A: Relation between the transmural extent of hyperenhancement (infarction) before revascularization
and the likelihood of improved contractility after revascularization. B: Relation between amount of dysfunc-
tional but viable left ventricle before revascularization and improvement in global function after revasculariza-
tion. Decreases in wall motion scores indicate increases in contractility (From Kim RJ, Wu E, Rafael A, et al.
The use of contrast enhanced magnetic resonance imaging to identify reversible myocardial dysfunction.
N Engl J Med. 2000;343(20):1445–1453, with permission.)
of 58 such segments did not improve, a negative predictive The Ideal Technique
accuracy of 98%. Thus, in this situation, rather than simply
assuming that the negative predictive accuracy remains at Knowing How Much is Alive is Not Enough
79% (using the single cutoff of 25% hyperenhancement),
approaching viability as a continuum would allow us to sig- A useful exercise is to consider what might constitute an
nificantly raise the negative predictive accuracy to 98%. “ideal” technique for assessing viability. Certainly, it would
This example highlights an important advantage of DE- be important for the technique to be fast, safe, and provide
MRI over the other imaging modalities used to assess vi- reproducible results. However, what might not be so apparent
ability. Myocardial regions are not interpreted in a binary is that even if a technique were available that could offer pre-
fashion as either viable or nonviable; rather, the transmural cise quantification of viability without technical limitations
extent of viable and infarcted myocardium is directly visual- (infinite spatial resolution, no artifacts, etc.), this would still
ized. Knowledge of the transmural extent of viability can be insufficient to provide a complete characterization of via-
then be used to predict functional improvement more ac- bility and thus insufficient to provide the highest accuracy in
curately, as in the previous example, and can also be used predicting wall motion improvement or clinical benefit after
to understand the underlying physiology of functional im- coronary revascularization (93). Certainly, there are addi-
provement. For instance, in the previous study, the average tional factors that are not related to limitations in noninvasive
extent of hyperenhancement across the ventricular wall was testing that could reduce accuracy in predicting functional
10% ± 7% for all dysfunctional segments that improved and improvement. These have been well described and include
41% ± 14% for those that did not improve (p < 0.001). This peri- or postoperative occult MI (91), incomplete revascu-
result, which is consistent with previous studies that ana- larization as a result of diffuse disease (94), or tethering of
lyzed needle biopsy specimens taken during bypass surgery regions with extensive scarring adjacent to viable regions
(8,9), indicates that significant degrees of myocardial viabil- (95). For the purposes of this section we will not consider
ity can be present without eventual functional improvement. these clinical factors, but instead focus on issues related to the
These data underscore the importance of differentiating noninvasive assessment of viability and examine the concept
between the current clinical “gold standard” definition of that “knowing how much is alive is not enough.”
myocardial viability, which is improvement in wall motion A central tenet of this concept is that a technique that
after revascularization, and the correct definition, which is offers even a perfect assessment of what is alive (viable)
the presence of living myocytes. has substantial and practical limitations compared with a
LWBK1209-ch17_p251-282.indd 271 17/05/13 5:09 PM

Figure 17.25. A: Left panels show delayed

enhancement images of a normal subject
(Patient A) and a patient with a subendocar-
dial, inferior wall MI (Patient B). The center
panels show tracing of the borders of viable
myocardium. The right panels plot the extent
of viability (thickness) as a function of LV
location. Note that there is significant regional
variation in the extent of viable myocardium.
See text for details. (Adapted from Kim RJ,
Shah DJ. Fundamental concepts in myo-
cardial viability assessment revisited: When
knowing how much is “alive” is not enough.
Heart. 2004:90;137–140, with permission).
A B: Cartoon highlighting differences between
an indirect and direct method of quantifying
regional viability. The indirect method is based
on assessment of viable regions only. The
direct method is based on assessment of viable
and nonviable regions. Viable myocardium
is black, and infarct is white. “Remote” zone
represents segment with maximum amount of
viability. (From Fuster V, Kim RJ. Frontiers
in cardiovascular magnetic resonance.
Circulation. 2005;112(1):135–144, with
B permission.)
t echnique that offers both an assessment of what is alive and myocardium at different locations of the normal LV, as was
what is dead (nonviable). This tenet may be counterintui- found in Patient A.
tive, since it would seem that by knowing exactly how much The principle that normal hearts can have significant het-
is alive, by deductive logic, one should know how much is erogeneity in the extent of viable myocardium has direct clini-
dead. However, this is an incorrect assumption, and the pa- cal implications. For example, a region with 70% the viability
tient examples in Figure 17.25A may be illustrative. of the region with the maximum amount of viability, may
The left panel in the top row (Patient A) shows a mid- represent either a normal region with 70% the wall thick-
ventricular, short-axis DE-MRI image of a normal heart in ness of the thickest region or a region with a subendocardial
diastole. With DE-MRI, nonviable regions such as infarction MI. The images from Patient B (Fig. 17.25A) underscore this
appear bright (hyperenhanced), whereas viable regions ap- concept. This particular patient had a clinically documented
pear black. First, consider only a “nonviability” or bright MI due to occlusion of the RCA that was reopened during
region assessment. A quick visual inspection shows no evi- primary angioplasty. Again, consider first only a “nonviabil-
dence of myocardial bright regions, thus one rapidly con- ity” assessment. From a quick visual perusal (left panel), one
cludes that no infarction is present. In comparison, now can determine that there is a subendocardial bright region in
consider only a “viability” or black region assessment. One the inferoseptal wall consistent with the patient’s prior MI in
can trace the endocardial and epicardial borders of viable the RCA perfusion territory. Next, consider only a “viability”
myocardium (center panel), and then plot the transmural assessment, similar to that performed in Patient A. Again, en-
extent of viability as a function of position starting from docardial and epicardial contours are traced (center panel),
the anterior wall at 12-o’clock and moving counterclock- however, one difference should be noted. The endocardial
wise around the LV (right panel). From the plot, it is obvious contour is along the border of viable (97) myocardium, and
that there is marked heterogeneity in the transmural extent the infarcted region is not included. The plot showing the
of viability, with as much as 12 mm of viable myocardium extent of viable myocardium (right panel) once more shows
in the portion of the inferior wall adjacent to the posterior marked regional variability, and from this assessment the
papillary muscle and as little as 7 mm in the anteroseptal presence and/or location of infarction is not evident. Thus,
wall at the RV insertion site. Concerning this finding, it is in this patient, the intrinsic variation in the extent of viable
important to note that it has been long known that there can myocardium for noninfarcted regions is greater than the
be significant variation in diastolic wall thickness at different reduction in viable myocardium for the region with subendo-
points around the LV, even if papillary muscles are excluded cardial infarction. This renders the subendocardial infarction
and healthy volunteers are studied (96). This observation “invisible” for any technique that can assess only viable myo-
then naturally proceeds to the conclusion that there can cardium, even if the technique has perfect accuracy and has
be significant variability in the transmural extent of viable infinite spatial resolution.
LWBK1209-ch17_p251-282.indd 272 17/05/13 5:09 PM

These two patient examples highlight the differences be- is the reason, it should be evident that even infinite spatial
tween techniques that can visualize only viable myocardium resolution would not improve the detection of subendocar-
as opposed to techniques that can visualize viable and nonvi- dial infarction by these or other methodologies based on as-
able myocardium. In addition, it is important to recognize sessing only viable myocardium. Conversely, a methodology
that the use of different techniques often lead to differences in with poor spatial resolution (e.g., a single voxel across the
the way in which viability is quantified although the nomen- LV wall) would still be able to detect subendocardial infarc-
clature used may be the same. For instance, when only viable tion if it had the ability to assess both alive and dead tissues.
myocardium can be visualized, the percentage of viability in
any given segment is assessed indirectly and generally refers
Is thinned myocardium always dead?
to the amount of viability in the segment normalized to the
segment with the maximum amount of viability or to data Prior studies report that in patients with coronary disease
from a gender-specific database of controls. Conversely, when and LV dysfunction, regions with thinned myocardium rep-
both viable and nonviable myocardium can be visualized, the resent scar tissue and cannot improve in contractile func-
percentage of viability can be assessed directly and expressed tion after revascularization (100). However, the ability of
as the amount of viability in the segment normalized to the delayed enhancement imaging to directly visualize both alive
amount of viability plus infarction in the same segment (Fig. and dead myocardium has led to some recent observations
17.25B). These differences in the way in which viability is that appear to refute this concept (101,102).
measured can alter clinical interpretation. For the normal Figure 17.27 shows MRI images of two patients (C and
heart in Patient A, the indirect method would show that there D) who have severe CAD and chronic LV dysfunction. The
is significant regional variability in the extent of viable myo- left panels represent long-axis views acquired before coro-
cardium: 60% to 100% of maximum viability; whereas, the nary revascularization, and two points should be noted on
direct method would show essentially no variability since all the cine images. First, both patients have severe contractile
segments would be classified as 100% viable. Likewise for dysfunction of the anterior wall. Second, Patient D has asso-
Patient B, the indirect method would not be able to identify ciated thinning of the anterior wall (diastolic wall thickness,
the region of subendocardial infarction since the extent of 5 mm) while Patient C does not (diastolic wall thickness,
viable myocardium in this sector is within the normal varia- 8 mm). Based on these cine images and the existing litera-
tion; whereas, the direct method would clearly identify the ture, one might expect that there is more viable myocardium
region with subendocardial infarction since this region would in the anterior wall of Patient C than in Patient D, and in
be the only region with less than 100% viable myocardium. fact question the need for viability testing at all in Patient D
These patient examples also clarify the mechanism by since the thinned, akinetic anterior wall must undoubtedly be
which subendocardial infarcts are routinely missed by tech- scar tissue and thus nonviable. The DE-MRI images acquired
niques such as SPECT and PET (51,98,99) (Fig. 17.26). before revascularization, however, indicate a different clinical
Although it is often reported that limited spatial resolution interpretation. In Patient C, there is a bright endocardial rim
Figure 17.26. Typical comparison

of SPECT and DE-MRI with histology
in three canines with subendocardial
infarcts (arrows). Note that the infarcts
are not evident on SPECT. (From
Wagner A, Mahrholdt H, Holly TA,
et al. Contrast-enhanced MRI and rou-
tine single photon emission computed
tomography (SPECT) perfusion imag-
ing for detection of subendocardial
myocardial infarcts: An imaging study.
Lancet. 2003;361(9355):374–379, with
permission.)
LWBK1209-ch17_p251-282.indd 273 17/05/13 5:09 PM

Figure 17.27. Cine and DE-MRI

images of two patients (C,D) before
coronary revascularization and cine
images 2 months after revascularization.
See text for details. (From Kim RJ, Shah
DJ. Fundamental concepts in myocardial
viability assessment revisited: When
knowing how much is “alive” is not
enough. Heart. 2004:90;137–140, with
permission.)
of hyperenhancement (infarction) that measures on average of viable myocardium in a given region is dynamic and can
4.5 mm in thickness. The remaining epicardial rim of tissue increase or decrease as a result of ventricular remodeling.
which is black (viable) measures 3.5 mm in thickness (total Whereas it is common knowledge that myocardial viability
thickness, 8 mm). In Patient D, there is also an endocardial can decrease, for example, due to MI with associated wall
rim of hyperenhancement; however, it measures on average thinning, the reverse process in which regions of thin myo-
only 1.5 mm in thickness. The epicardial rim which is viable cardium become thick with an absolute increase in the trans-
measures 3.5 mm in thickness (total thickness, 5 mm). Note mural extent of viability (as in patient D) has only recently
that in both patients the absolute amount of viable myo- been described (103). Third, it appears that quantification of
cardium is the same (3.5 mm); however, when the extent of nonviable in addition to viable myocardium is important in
viability is expressed as a percentage of the total amount of predicting contractile improvement following revasculariza-
myocardium in the segment, Patient C has less than 50% tion. For instance, incorporating information regarding non-
viable myocardium (3.5/8 = 44%) whereas Patient D has viable myocardium into a ratio of viable to total myocar-
greater than 50% viable myocardium (3.5/5 = 70%). The dium (viable plus nonviable) within the same region would
right-most panels represent cine images acquired 2 months lead to the conclusion that the anterior wall of Patient D
after coronary revascularization. Patient C exhibits no has a higher percentage of viable myocardium (70%) than
improvement in contractile function in the anterior wall Patient C (44%) and thus is more likely to improve in con-
and in fact develops diastolic wall thinning in this region. tractile function. The follow-up cine images in Figure 17.27
Conversely, Patient D exhibits not only significant improve- demonstrate that this prediction is correct. An investigation
ment in anterior wall contractile function, but also recovery systematically evaluating these concepts has recently been re-
of diastolic wall thickness in this region (from 5 to 9 mm). ported. Shah et al. studied 1055 patients with CAD referred
Three fundamental points are raised by these patient for DE-MRI and found that 201 (19%) had regional wall
examples. First, it is apparent that a method that can quan- thinning. Limited scar burden was present in 18% of thinned
tify only viable myocardium, even if technically flawless (infi- regions and was associated with improved contractility and
nite spatial resolution, no artifacts, etc.), provides insufficient resolution of wall thinning after revascularization (103).
information to allow a comprehensive assessment of myocar-
dial viability. Since both patients had the same, reduced
Additional Characterization
amount of viable myocardium (3.5-mm thick) we would
of Dead Tissue
have predicted, incorrectly as it turns out, that both patients
would not improve in contractile function following revas- It may be important clinically to provide a characteriza-
cularization. Second, it is evident that the absolute amount tion of dead tissue in addition to measuring its extent. For
LWBK1209-ch17_p251-282.indd 274 17/05/13 5:10 PM

Figure 17.28. Microvascular obstruction revealed by DE-MRI. Labels refer to time after administration
of gadolinium contrast. The subendocardial black zone surrounded by hyperenhancement corresponds to the
region of microvascular obstruction (arrow) within the acute infarction. This region can be distinguished from
normal myocardium since it is encompassed in 3D space by hyperenhanced myocardium or the LV cavity, and
by the fact that it slowly becomes hyperenhanced over time. (From Kim RJ, Choi KM, Judd RM. Assessment
of myocardial viability by contrast enhancement. In: Higgins CB, deRoos A, eds. Cardiovascular MRI & MRA.
Philadelphia, PA: Lippincott Williams & Wilkins; 2003:209–237, with permission.)
example, rather than simply identifying a region of acute cavity blood flow (Fig. 17.29). If the thrombus protrudes into
infarction as nonviable, DE-MRI (45,104) has the ability to the LV cavity it is easily identified by cine MRI or echocar-
differentiate between acute infarcts with necrotic myocytes diography, however, layered mural thrombus may often be
and acute infarcts with necrotic myocytes and damaged mistaken as akinetic, nonviable myocardium. Importantly, the
microvasculature. The latter, termed microvascular obstruc- diagnosis of LV thrombus will likely affect patient manage-
tion, indicates compromised tissue perfusion despite restora- ment and prognosis. Recent data indicate that DE-MRI has
tion of epicardial artery patency. The presence and extent improved sensitivity for detecting LV thrombus compared with
of microvascular obstruction appears to be associated with transthoracic or transesophageal echocardiography (109,110).
worse LV remodeling and outcome (105,106). Although the Mollet et al. (109) reported that DE-MRI identified LV
initial MR studies of microvascular obstruction used single- thrombus in substantially more patients than cine MRI or
shot perfusion sequences 1 to 2 minutes after contrast injec- transthoracic echocardiography; however, a reference stan-
tion (44), DE-MRI performed 5 to 10 minutes after contrast dard was not available. Srichai et al. (110) evaluated a pro-
will provide higher image quality and will be more specific tocol combining cine MRI and DE-MRI for the diagnosis of
for microvascular obstruction (107). On DE-MRI, microvas- LV thrombus in patients with advanced ischemic cardiomy-
cular obstruction can be identified as a hypoenhanced region opathy undergoing surgical LV reconstruction. Among 160
at the core of acute, usually transmural infarction (45,108). patients (using a reference standard that consisted of direct
Thus, the dark area of microvascular obstruction is distin- visualization of the LV cavity during surgery and/or patho-
guishable from viable myocardium (which is also dark) since logic confirmation), MRI showed higher sensitivity and
it is always surrounded by hyperenhanced myocardium and/ specificity (88% and 99%, respectively) than transthoracic
or the bright LV blood pool. In addition, if repeated imaging (23% and 96%, respectively) and transesophageal echocar-
over time is performed after contrast administration, micro- diography (40% and 96%, respectively) for the diagnosis of
vascular obstruction can be distinguished from viable myo- LV thrombus. Weinsaft et al. (111) assessed the prevalence
cardium since it slowly becomes hyperenhanced (Fig. 17.28). of LV thrombus by cine and DE-MRI in 784 consecutive pa-
There is increased propensity for thrombus formation near tients with systolic dysfunction (LVEF < 50%). Among this
infarcted regions with impaired contractility and stagnant LV broad population, DE-MRI detected thrombus in 7% (55
patients) and cine MRI in 4.7% (37 patients), and clinical
follow-up for embolic events or pathologic confirmation was
consistent with DE-MRI as a better reference standard than
cine MRI. Interestingly, patients with ischemic cardiomyop-
athy were more than five times more likely to have thrombus
than those with nonischemic cardiomyopathy despite similar
mean LVEF. In addition, myocardial scarring (also detected
by DE-MRI) was identified as a novel risk factor for throm-
bus. In a second study, Weinsaft et al. (112) showed that
even when sonographic contrast is routinely utilized, echo-
cardiography may fail to identify up to 39% of thrombi de-
tected by DE-MRI. Based on these data, DE-MRI may repre-
sent the emerging gold standard for the clinical diagnosis of
LV thrombus. The ability of DE-MRI to identify thrombus
based on tissue characteristics rather than just anatomical
Figure 17.29. DE-MRI images in a patient with an LV thrombus
adjacent to the interventricular septum. Note that increasing the inver-
appearance likely explains its improved performance com-
sion time (TI) from that needed to “null” normal myocardium (∼300 pared with cine MRI or noncontrast echocardiography. The
milliseconds) to “null” thrombus (∼600 milliseconds at 1.5T) results in basic underlying principle is that thrombi are avascular and
increased image intensity for all myocardial regions while thrombus is have essentially no gadolinium uptake. Thus, thrombus can
now homogeneously black. be identified as a nonenhancing defect surrounded by bright
LWBK1209-ch17_p251-282.indd 275 17/05/13 5:10 PM

ventricular blood and contrast-enhanced myocardium on DE- requires not only the ability to directly visualize both viable and
MRI images. Image intensity differences between normal myo- nonviable myocardium but also high spatial resolution.
cardium and thrombus can be accentuated by using a DE-MRI
sequence in which the TI is increased to null avascular tissue
such as thrombus (500 to 600 milliseconds at 1.5T) (112). Common Assumptions and
With long-inversion time (long-IT) imaging, regions with con-
Clinical Applications
trast uptake such as viable myocardium increase in image in-
tensity (i.e., appear grey rather than black), whereas thrombus
Intermediate Levels of Viability
appears homogeneously black, and there is improved delinea-
tion, particularly of mural thrombus (Fig. 17.29). The reports of predicting functional recovery with DE-MRI
Finally, the specific location and shape of hyperenhancement have shown that for any given myocardial region there is
within the ventricular wall may shed light on its etiology. Figure a smooth progressive relationship between the transmural
17.30 and 17.31 show different hyperenhancement patterns extent of viability and the likelihood of functional recov-
that may be encountered in clinical practice. Myocardial cell ery (89–91,114). In other words, with each increase in the
death from ischemic injury progresses as a “wavefront” from amount of viability (or decrease in the amount of nonvi-
the subendocardium to the epicardium (65). Correspondingly, ability) there is a corresponding increment in the likelihood
hyperenhancement patterns that spare the subendocardium for functional recovery (see Fig. 17.24A). Two questions are
and are limited to the middle or epicardial portion of the LV often raised regarding this relationship. The first is in regards
wall are clearly in a non-CAD pattern (123). Determination to intermediate levels of viability (transmural extent, 50%
of the pattern of dead myocardium within the ventricular wall to 75%), a situation where DE-MRI appears to have limited
Figure 17.30. Hyperenhancement

(HE) patterns that may be encountered
in clinical practice. Since myocardial
necrosis due to coronary artery disease
(CAD) progresses as a wavefront from
the subendocardium to the epicardium,
HE (if present) should always involve the
subendocardium in patients with ischemic
disease. Isolated midwall or epicardial HE
strongly suggests a nonischemic etiology.
In addition, endocardial HE that occurs
globally (i.e., throughout the left ventricle)
is uncommon, even with diffuse CAD,
and therefore, a nonischemic etiology
should be considered. HTN, hyperten-
sion. (Reproduced from Shah DJ, Kim RJ.
Magnetic resonance of myocardial viabil-
ity. In: Edelman RR, eds. Clinical Magnetic
Resonance Imaging. 3rd ed. New York,
NY: Elsevier; 2005, with permission.)
LWBK1209-ch17_p251-282.indd 276 17/05/13 5:10 PM

Figure 17.31. Representative delayed enhancement images in patients with various nonischemic car-
diomyopathies. The hyperenhancement patterns in all patients are distinctly noncoronary artery disease
(CAD) type. Dilated cardiomyopathy (DCM): Arrows point to a linear stripe of hyperenhancement that
is limited to the midwall of the interventricular septum. Hypertrophic cardiomyopathy (HCM): Arrows
point to multiple foci of hyperenhancement, which are predominantly midmyocardial in location and
occur in the hypertrophied septum and not in the lateral left ventricle (LV) free wall. The junctions of the
right ventricle free wall and interventricular septum are commonly involved. Myocarditis: Arrows point
to two separate regions of hyperenhancement: A linear midwall stripe in the interventricular septum and
a large confluent region affecting the epicardial half of the LV lateral wall. Amyloidosis: Arrows point to
hyperenhancement affecting the subendocardial half of the myocardial wall diffusely throughout the entire
LV. (Reproduced from Kim. Magnetic resonance imaging of the heart. In: Fuster, eds. Hurst’s The Heart.
13th ed. Columbus, OH: McGraw Hill; 2011, with permission.)
predictive value, in that only ∼50% of these regions will consider most regions with intermediate levels of viability as
have functional recovery. In this case, would viability assess- having a high chance for an intermediate amount of improve-
ment be improved by considering additional information ment rather than 50% likely to have complete recovery and
such as provided by dobutamine contractile reserve testing? the other 50%, no improvement whatsoever.
The second question is related. Since other techniques such Complicating this issue is the belief that a threshold phe-
as radionuclide imaging can also demonstrate a smooth pro- nomenon exists between the TEI and systolic thickening.
gressive relationship between the amount of viability (tracer This assumption is based primarily on results by Lieberman
activity) and the likelihood for functional recovery (115), et al. (116) who demonstrated in a canine model of acute in-
in what way exactly does DE-MRI with all of its technical farction that akinesia is expected if infarction involves ≥20%
advantages provide an improved diagnosis for the clinician? of the wall thickness. The animal model, however, involved
The questions, in part, are a consequence of some com- permanent coronary ligation and likely included effects from
mon assumptions regarding functional recovery. First, it myocardial stunning and/or ongoing ischemia. Evaluation
is important to distinguish between what would be desir- by DE-MRI in humans with chronic infarction suggests that
able clinically from what can actually occur physiologi- a threshold phenomenon does not exist (117,118). These data
cally. While we may ask for an improved technique with indicate that it is problematic to extrapolate the results of
better predictive value, perhaps in reality we are asking for Lieberman et al. who did not consider the effects of stunning
a different physiologic relationship—an abrupt threshold and/or ongoing ischemia, to humans following revasculariza-
of viability, below which there is virtually no chance for tion in whom substantial stunning, ischemia, or hibernation
functional recovery, and above which nearly all have func- may no longer be present.
tional recovery. Figure 17.32(A) demonstrates that with an The finding that two techniques can demonstrate a similar
abrupt threshold the problem of predicting an intermedi- relationship between the amount of viability and the likeli-
ate likelihood for functional recovery is greatly minimized. hood of functional recovery does not mean that the techniques
Unfortunately, this physiologic relationship may not exist. have the same diagnostic capability. For example, both an ac-
Second, it is essential to remember that functional recovery curate and a less accurate technique may show that for a given
is a continuum (as is viability itself), and not simply a bi- intermediate level of viability, on average 50% of segments
nary—yes or no—function. Thus, part of the problem with recover function. However, part of the problem again is de-
intermediate levels of viability is expecting that these regions fining functional recovery as a binary variable. Figure 17.32B
will or will not have functional recovery in this binary fash- demonstrates the utility of expressing functional recovery as
ion. Since it more closely reflects reality, it may be better to a continuous variable (e.g., millimeters of improvement in
LWBK1209-ch17_p251-282.indd 277 17/05/13 5:10 PM

Improved technique? CABG. The biopsies from regions that were dysfunctional
prior to revascularization but recovered afterward showed
89% ± 6% of the volume consisted of viable cells. Similarly,
Dakik et al. (8) reported that among 21 biopsied dyssynergic
myocardial segments, the 11 which recovered function after
CABG had 93% ± 4% viability by volume.
However, a common misconception is that if a region does
not have functional improvement, then this region is nonvi-
able. In fact, Maes et al. (9) and Dakik et al. (8) observed
that biopsy specimens from regions that did not improve
after revascularization still had 65% ± 25% and 69% ± 21%
viability by volume, respectively. There are several reasons
A
why regions that are predominantly viable may not recover
Techniques with different accuracy function following revascularization. First, revascularization
may often be incomplete in patients with extensive athero-
sclerosis (94). Therefore, there may be persistent areas of
hibernating myocardium. Second, viable myocardium may be
juxtaposed to regions with extensive scarring and unable to
respond to revascularization because of tethering (95). Third,
there might be new perioperative myocardial necrosis in re-
gions that were viable prior to revascularization (91). Finally,
the use of a single evaluation of ventricular function soon
after revascularization may lead to an underestimation of
the true rate of functional recovery. Bax et al. (3) evaluated
B functional recovery in 26 patients at 3 and 14 months fol-
Figure 17.32. Assumptions in evaluating viability techniques. lowing CABG. At 3 months, only 31% of hibernating seg-
A: A technique that demonstrates a smooth linear relationship between ments recovered, whereas at 14 months an additional 61%
viability and functional recovery may be considered limited in that recovered.
intermediate levels of viability predict intermediate likelihood for Even with these limitations, one could argue that func-
functional recovery. With an abrupt threshold relationship between tional recovery is an appropriate truth standard for viabil-
viability and functional recovery, the problem of predicting intermedi-
ity since without recovery there may be no benefit to the
ate likelihood for functional recovery is greatly minimized. See text
for details. B: Two techniques with different diagnostic capability may
patient for undergoing revascularization. On this point,
show a similar overall relationship between viability and functional the results by Samady et al. (119) are instructive. Of 104
improvement. However, if functional improvement is expressed as a consecutive patients that underwent pre- and post-CABG
continuous rather than binomial variable, the better technique may LVEF assessment, 68 had improvement in LVEF (>5% in-
reduce the variability (smaller error bars) in predicting the absolute crease) and 36 had no significant change. Surprisingly, the
amount of functional improvement. See text for details. two groups, however, had similar postoperative improvement
in angina and heart failure scores, and there was no differ-
systolic wall thickening). Although the overall relationship ence in cardiovascular mortality with a mean follow-up of
is the same, for any given level of viability, the better tech- 32 months. The authors concluded that a lack of improve-
nique will provide a smaller variability (smaller error bars) ment in LVEF after CABG is not associated with poorer
in the absolute amount of functional improvement than the outcome, and speculated that many patients without im-
worse technique. provement in LVEF may nonetheless have substantial viable
myocardium that can respond to effective revascularization
Functional Recovery as the Standard with beneficial effects on prognosis.
of Truth Thus, functional recovery is a flawed truth standard for
evaluating viability imaging techniques. The primary prob-
The ideal reference standard for evaluating viability imag- lem is that regions (or patients) without functional recov-
ing techniques would be histopathologic examination. ery may have substantial viability that may be important
However, for the vast majority of clinical studies this is not to detect clinically. This has implications for the published
feasible, and functional recovery following revascularization studies that have used functional recovery as the standard for
has become the de facto standard. This clinical endpoint evaluating tests of viability. Poor specificity (for functional re-
makes sense for a variety of reasons. Since LV dysfunction covery) should be considered less important than poor sensi-
portends poor prognosis, one would expect that functional tivity, since for the former, a number of “false positives” (test
recovery would be an important and beneficial clinical showed viability but there was no functional recovery) may
outcome of revascularization. In addition, if a myocardial represent a problem with the truth standard rather than the
region recovers function then one might confidently assume technique. Further work needs to be completed to determine
that the region has a substantial amount of viable myocytes. how various assessments of viability predict other clinically
Indeed, this has been confirmed by histopathologic examina- relevant endpoints such as improvement in symptoms of an-
tion. In 33 patients with LAD disease, Maes et al. (9) took gina and heart failure, improvement in exercise tolerance,
transmural needle biopsies from the LV anterior wall during and reduction in future MI, arrhythmias, and death.
LWBK1209-ch17_p251-282.indd 278 17/05/13 5:10 PM

Overview of Clinical Interpretation v iability assessment to determine whether the dysfunction in

viable regions is clearly due to ischemic heart disease. As dis-
For clinical purposes, we interpret DE-MRI scans for via-
cussed earlier, it is important to remember that the pattern
bility assessment using a standard 17-segment model rec-
of nonviable myocardium on DE-MRI may also shed light
ommended by the American Heart Association (120). The
on the state of viable myocardium. For stress testing, we
extent of hyperenhanced or nonviable tissue within each seg-
prefer adenosine perfusion MRI over high-dose dobutamine
ment is graded using a 5-point scale: 0, 0%; 1, 1% to 25%
cine MRI because of logistic issues, including scan time, ease
volume; 2, 26% to 50% volume; 3, 51% to 75% volume;
of patient monitoring, and staffing (122).
and 4, 76% to 100% volume. Although the interpretation
is visual and thus rapid, note that the LV myocardium can
have a total of 85 (17 × 5) different levels of nonviable tis-
sue, and one can approach viability as a continuum rather Summary
than in a binary manner.
We do not use low-dose dobutamine cine MRI to im- A diagnostic test that can distinguish between viable and
prove the assessment of viability. For dysfunctional regions nonviable myocardium is essential in the clinical assessment
with ≥50% scar by DE-MRI the negative predictive value and management of patients with ischemic heart disease.
for functional improvement is quite high (92% in the ini- Although viability is often thought to be synonymous with
tial report by Kim et al. (89), and likely will be higher than clinical findings such as recovery of wall motion following
that found by dobutamine MRI (121). For regions with revascularization, the true definition is the presence of liv-
<25% scar, it is possible that dobutamine MRI may provide ing myocytes. This issue is not simply an academic exercise;
a higher positive predictive value than DE-MRI. However, discordant interpretations of viability may be generated by
in this group there is such a large amount of viability and the use of different definitions so that patient management is
the potential for benefit is so great, it is sensible to be con- adversely affected.
cerned about the possibility of a false-negative dobutamine Advances in DE-MRI have led to significant improve-
MRI result. Thus in this group, with all other clinical is- ments in image quality. Techniques such as the breath-hold,
sues being equal, it is usually preferable to err on the side segmented k-space, and IR-FGE sequence yield image inten-
of overestimating rather than underestimating the clinical sities in “hyperenhanced” regions that are typically 500%
benefit of revascularization. For regions with 25% to 50% higher than those in “nonhyperenhanced” regions. Better
scar, one should recognize that this group represents only a image quality reduces observer subjectivity in detecting hy-
small portion of dysfunctional segments: 14% in the initial perenhanced regions and, importantly, allows a clear delin-
report by Kim et al. Nonetheless, it is likely that if one di- eation of the transmural extent of hyperenhancement across
chotomizes this group into those that are contractile reserve a ventricular wall.
positive and those that are negative, these two groups will Experimental studies using animal models have shown
have differences in the rate of functional improvement after that DE-MRI can distinguish between viable and nonviable
revascularization (114). But in this case, recall from the pre- myocardium independent of the level of ventricular function
vious section that functional improvement is a flawed truth and the age of infarction. Studies in patients indicate that
standard for viability. The question whether one can predict DE-MRI can detect both acute and chronic infarcts with a
a subgroup with high likelihood of functional improvement sensitivity approaching that of serum assays for cardiac en-
is only part of the issue. Perhaps, a more important question zymes. However, unlike cardiac enzymes, which are cleared
is whether a substantial amount of viability (e.g., 50% to from the blood in a few days, DE-MRI provides a perma-
75%) may be sufficient to provide clinical benefit even with- nent record of infarction, localizes the infarct to a specific
out functional improvement after revascularization. The re- coronary artery territory, and can be combined with cine
sults of Samady et al. (119) strongly suggest that there may MRI to allow differentiation between several distinct forms
be an intermediate level of viability that increases the likeli- of ischemic injury. The combination of cine- and DE-MRI
hood of a good clinical outcome but is, meanwhile, insuf- can be used in patients before revascularization procedures
ficient to provide improved resting function. It is likely that to predict the likelihood of wall motion recovery following
improved clinical outcome is related to a number of factors revascularization.
including prevention or even reversal of adverse remodel- Importantly, delayed enhancement imaging provides vi-
ing, improved contractile response under stress, prevention sualization of both alive and dead myocardium. Although
of recurrent myocardial ischemia/infarction, or prevention perhaps counterintuitive, a technique that offers a perfect
of arrhythmias. assessment of what is alive (infinite spatial resolution, no
The decision for revascularization is dependent on sev- artifacts, etc.) has substantial and practical limitations com-
eral factors besides the amount of viability, including specif- pared with a technique that offers both an assessment of
ics regarding the coronary anatomy, the presence of angina what is alive and dead. In particular, the settings of suben-
or ischemia, and patient comorbidities. In addition, viable docardial infarction and myocardial “thinning” highlight
myocardium can be in different states (i.e., normal, isch- this concept. In addition, the assessment of the “pattern” of
emic, hibernating, cardiomyopathic from a nonischemic pro- dead myocardium may shed light regarding the etiology
cess, etc.) and it is likely that differentiating between these of the cardiomyopathic process. There are many assump-
states may be important for revascularization decisions in tions regarding viability testing. Several arise, in part, from
some patients. Thus, although we do not perform low-dose considering viability and/or functional recovery as a simple
dobutamine MRI, we often incorporate stress testing with binary—yes or no—function, rather than as a continuum,
LWBK1209-ch17_p251-282.indd 279 17/05/13 5:10 PM

the later of which better reflects reality. Although an impor- 25. Kaul S. Assessing the myocardium after attempted reperfusion: Should we bother?
Circulation. 1998;98(7):625–627.
tant clinical endpoint, functional recovery after revascular- 26. Zabel M, Hohnloser SH, Koster W, et al. Analysis of creatine kinase, CK-MB, myoglobin,
ization is a flawed truth standard for evaluating viability and troponin T time-activity curves for early assessment of coronary artery reperfusion
after intravenous thrombolysis. Circulation. 1993;87(5):1542–1550.
techniques, since myocardium that does not recover may be 27. Adams JED, Abendschein DR, Jaffe AS. Biochemical markers of myocardial injury. Is MB
predominantly viable rather than nonviable. The decision for creatine kinase the choice for the 1990s? Circulation. 1993;88(2):750–763.
revascularization involves many factors besides the amount 28. Fishbein MC, Meerbaum S, Rit J, et al. Early phase acute myocardial infarct size quanti-
fication: Validation of the triphenyl tetrazolium chloride tissue enzyme staining technique.
of viability (and nonviability), and incorporating adenosine Am Heart J. 1981;101(5):593–600.
stress perfusion MRI may be helpful in selected patients. 29. Wesbey G, Higgins CB, Lanzer P, et al. Imaging and characterization of acute myocardial
infarction in vivo by gated nuclear magnetic resonance. Circulation. 1984;69(1):125–130.
30. Manning WJ, Edelman RR. Magnetic resonance imaging assessment of ischemic heart
disease. In: Skorton DJ, Schelbert HR, Wolf GL, Brundage BH, eds. Cardiac Imaging:
References A Companion to Braunwalds’s Heart Disease. 2nd ed. Philadelphia, PA:W.B. Saunders;
1996:719–744.
1. Braunwald E, Kloner RA. The stunned myocardium: Prolonged, postischemic ventricular 31. van der Wall EE, Vliegen HW, de Roos A, et al. Magnetic resonance imaging in coronary
dysfunction. Circulation. 1982;66(6):1146–1149. artery disease. Circulation. 1995;92(9):2723–2739.
2. Kloner RA, Bolli R, Marban E, et al. Medical and cellular implications of stunning, 32. Eichstaedt HW, Felix R, Dougherty FC, et al. Magnetic resonance imaging (MRI) in dif-
hibernation, and preconditioning: An NHLBI Workshop. Circulation. [Congresses]. ferent stages of myocardial infarction using the contrast agent gadolinium-DTPA. Clin
1998;97(18):1848–1867. Cardiol. 1986;9(11):527–535.
3. Bax JJ, Visser FC, Poldermans D, et al. Time course of functional recovery of stunned 33. Matheijssen NA, de Roos A, van der Wall EE, et al. Acute myocardial infarction:
and hibernating segments after surgical revascularization. Circulation. 2001;104(12 suppl Comparison of T2-weighted and T1-weighted gadolinium-DTPA enhanced MR imaging.
1):I314–I318. Magn Reson Med. 1991;17(2):460–469.
4. Rahimtoola SH. The hibernating myocardium. Am Heart J. 1989;117(1):211–221. 34. McNamara MT, Tscholakoff D, Revel D, et al. Differentiation of reversible and irrevers-
5. Alderman EL, Fisher LD, Litwin P, et al. Results of coronary artery surgery in patients ible myocardial injury by MR imaging with and without gadolinium-DTPA. Radiology.
with poor left ventricular function (CASS). Circulation. 1983;68(4):785–795. 1986;158(3):765–769.
6. Chatterjee K, Swan HJ, Parmley WW, et al. Influence of direct myocardial revasculariza- 35. Nishimura T, Yamada Y, Hayashi M, et al. Determination of infarct size of acute myocar-
tion on left ventricular asynergy and function in patients with coronary heart disease. dial infarction in dogs by magnetic resonance imaging and gadolinium-DTPA: Comparison
With and without previous myocardial infarction. Circulation. 1973;47(2):276–286. with indium-111 antimyosin imaging. Am J Physiol Imaging. 1989;4(3):83–88.
7. Rees G, Bristow JD, Kremkau EL, et al. Influence of aortocoronary bypass surgery on left 36. Peshock RM, Malloy CR, Buja LM, et al. Magnetic resonance imaging of acute myocar-
ventricular performance. N Engl J Med. 1971;284(20):1116–1120. dial infarction: Gadolinium diethylenetriamine pentaacetic acid as a marker of reperfu-
8. Dakik HA, Howell JF, Lawrie GM, et al. Assessment of myocardial viability with 99mTc- sion. Circulation. 1986;74(6):1434–1440.
sestamibi tomography before coronary bypass graft surgery: Correlation with histopathol- 37. Rehr RB, Peshock RM, Malloy CR, et al. Improved in vivo magnetic resonance imaging
ogy and postoperative improvement in cardiac function. Circulation. 1997;96(9):2892– of acute myocardial infarction after intravenous paramagnetic contrast agent administra-
2898. tion. Am J Cardiol. 1986;57(10):864–868.
9. Maes A, Flameng W, Nuyts J, et al. Histological alterations in chronically hypoperfused 38. Van Rossum AC, Visser FC, Van Eenige MJ, et al. Value of gadolinium-diethylene-
myocardium. Correlation with PET findings. Circulation. 1994;90(2):735–745. triamine pentaacetic acid dynamics in magnetic resonance imaging of acute myocardial
10. Wijns W, Vatner SF, Camici PG. Hibernating myocardium. N Engl J Med. [Research Support, infarction with occluded and reperfused coronary arteries after thrombolysis. Am J
Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S. Review]. 1998;339(3):173–181. Cardiol. 1990;65(13):845–851.
11. Anselmi M, Golia G, Cicoira M, et al. Prognostic value of detection of myocardial viabil- 39. Edelman RR, Wallner B, Singer A, et al. Segmented turboFLASH: Method for breath-hold
ity using low-dose dobutamine echocardiography in infarcted patients. Am J Cardiol. MR imaging of the liver with flexible contrast. Radiology. 1990;177(2):515–521.
1998;81(12A):21G–28G. 40. Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging technique for the visu-
12. Picano E, Sicari R, Landi P, et al. Prognostic value of myocardial viability in medically alization of myocardial infarction. Radiology. 2001;218(1):215–223.
treated patients with global left ventricular dysfunction early after an acute uncompli- 41. van der Wall EE, van Dijkman PR, de Roos A, et al. Diagnostic significance of gadolin-
cated myocardial infarction: A dobutamine stress echocardiographic study. Circulation. ium-DTPA (diethylenetriamine penta-acetic acid) enhanced magnetic resonance imaging
1998;98(11):1078–1084. in thrombolytic treatment for acute myocardial infarction: Its potential in assessing reper-
13. Previtali M, Fetiveau R, Lanzarini L, et al. Prognostic value of myocardial viability and fusion. Br Heart J. 1990;63(1):12–17.
ischemia detected by dobutamine stress echocardiography early after acute myocardial 42. van Dijkman PR, van der Wall EE, de Roos A, et al. Acute, subacute, and chronic myo-
infarction treated with thrombolysis. J Am Coll Cardiol. 1998;32(2):380–386. cardial infarction: Quantitative analysis of gadolinium-enhanced MR images. Radiology.
14. Kaul S. There may be more to myocardial viability than meets the eye. Circulation. 1991;180(1):147–151.
1995;92(10):2790–2793. 43. Fedele F, Montesano T, Ferro-Luzzi M, et al. Identification of viable myocardium in
15. Gibbons RJ, Miller TD, Christian TF. Infarct size measured by single photon emission patients with chronic coronary artery disease and left ventricular dysfunction: Role of
computed tomographic imaging with (99m)Tc-sestamibi: A measure of the efficacy of magnetic resonance imaging. Am Heart J. 1994;128(3):484–489.
therapy in acute myocardial infarction. Circulation. 2000;101(1):101–108. 44. Lima JA, Judd RM, Bazille A, et al. Regional heterogeneity of human myocardial
16. Miller TD, Christian TF, Hopfenspirger MR, et al. Infarct size after acute myocardial infarcts demonstrated by contrast-enhanced MRI. Potential mechanisms. Circulation.
infarction measured by quantitative tomographic 99mTc sestamibi imaging predicts sub- 1995;92(5):1117–1125.
sequent mortality. Circulation. 1995;92(3):334–341. 45. Judd RM, Lugo-Olivieri CH, Arai M, et al. Physiological basis of myocardial contrast
17. Kim HW, Farzaneh-Far A, Kim RJ. Cardiovascular magnetic resonance in patients with enhancement in fast magnetic resonance images of 2-day-old reperfused canine infarcts.
myocardial infarction: Current and emerging applications. J Am Coll Cardiol. [Review]. Circulation. 1995;92(7):1902–1910.
2009;55(1):1–16. 46. Ramani K, Judd RM, Holly TA, et al. Contrast magnetic resonance imaging in the assess-
18. Hillis LD, Smith PK, Anderson JL, et al. 2011 ACCF/AHA Guideline for Coronary ment of myocardial viability in patients with stable coronary artery disease and left ven-
Artery Bypass Graft Surgery: Executive summary: A report of the American College of tricular dysfunction. Circulation. 1998;98(24):2687–2794.
Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. 47. Rogers WJ, Jr., Kramer CM, Geskin G, et al. Early contrast-enhanced MRI predicts late func-
Circulation. [Consensus Development Conference]. 2011;124(23):2610–2642. tional recovery after reperfused myocardial infarction. Circulation. 1999;99(6):744–750.
19. Bonow RO. Myocardial viability and prognosis in patients with ischemic left ventricular 48. Kim RJ, Albert TS, Wible JH, et al. Performance of delayed-enhancement magnetic
dysfunction. J Am Coll Cardiol.. 2002;39(7):1159–1162. resonance imaging with gadoversetamide contrast for the detection and assessment of
20. Allman KC, Shaw LJ, Hachamovitch R, et al. Myocardial viability testing and impact of myocardial infarction: An international, multicenter, double-blinded, randomized trial.
revascularization on prognosis in patients with coronary artery disease and left ventricular Circulation. 2008;117(5):629–637.
dysfunction: A meta-analysis. J Am Coll Cardiol. 2002;39(7):1151–1158. 49. Choudhury L, Mahrholdt H, Wagner A, et al. Myocardial scarring in asymptom-
21. Hunt SA, Abraham WT, Chin MH, et al. 2009 focused update incorporated into the ACC/ atic or mildly symptomatic patients with hypertrophic cardiomyopathy. J Am Coll
AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults: A Cardiol. [Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S.].
report of the American College of Cardiology Foundation/American Heart Association 2002;40(12):2156–2164.
Task Force on Practice Guidelines: Developed in collaboration with the International 50. Ricciardi MJ, Wu E, Davidson CJ, et al. Visualization of discrete microinfarction after
Society for Heart and Lung Transplantation. Circulation. [Practice Guideline Review]. percutaneous coronary intervention associated with mild creatine kinase-MB elevation.
2009;119(14):e391–e479. Circulation. 2001;103(23):2780–2783.
22. Velazquez EJ, Lee KL, Deja MA, et al. Coronary-artery bypass surgery in patients with 51. Wagner A, Mahrholdt H, Holly TA, et al. Contrast-enhanced MRI and routine single
left ventricular dysfunction. N Engl J Med. 2011;364(17):1607–1616. photon emission computed tomography (SPECT) perfusion imaging for detection of sub-
23. Orn S, Manhenke C, Greve OJ, et al. Microvascular obstruction is a major determinant of endocardial myocardial infarcts: An imaging study. Lancet. 2003;361(9355):374–379.
infarct healing and subsequent left ventricular remodelling following primary percutane- 52. Wu E, Judd RM, Vargas JD, et al. Visualisation of presence, location, and transmural extent
ous coronary intervention. Eur Heart J. 2009;30(16):1978–1985. of healed Q-wave and non-Q–wave myocardial infarction. Lancet. 2001;357(9249):21–28.
24. Bonow RO, Maurer G, Lee KL, et al. Myocardial viability and survival in ischemic 53. Sievers B, Elliott MD, Hurwitz LM, et al. Rapid detection of myocardial infarction by
left ventricular dysfunction. N Engl J Med. [Research Support, N.I.H., Extramural]. subsecond, free-breathing delayed contrast-enhancement cardiovascular magnetic reso-
2011;364(17):1617–1625. nance. Circulation. 2007;115(2):236–244.
LWBK1209-ch17_p251-282.indd 280 17/05/13 5:10 PM

54. Sievers B, Rehwald WG, Albert TS, et al. Respiratory motion and cardiac arrhythmia 84. Grines CL, Browne KF, Marco J, et al. A comparison of immediate angioplasty with
effects on diagnostic accuracy of myocardial delayed-enhanced MR imaging in canines. thrombolytic therapy for acute myocardial infarction. The Primary Angioplasty in
Radiology. 2008;247(1):106–114. Myocardial Infarction Study Group. N Engl J Med. [Clinical Trial Comparative Study
55. Judd RM, Kim RJ. Imaging time after Gd-DTPA injection is critical in using delayed Multicenter Study Randomized Controlled Trial]. 1993;328(10):673–679.
enhancement to determine infarct size accurately with magnetic resonance imaging. 85. Zijlstra F, de Boer MJ, Hoorntje JC, et al. A comparison of immediate coronary angioplasty
Circulation. 2002;106(2):e6; author reply e6. with intravenous streptokinase in acute myocardial infarction. N Engl J Med. [Clinical Trial
56. Kellman P, Arai AE, McVeigh ER, et al. Phase-sensitive inversion recovery for detecting Comparative Study Randomized Controlled Trial]. 1993;328(10):680–684.
myocardial infarction using gadolinium-delayed hyperenhancement. Magn Reson Med. 86. Gersh BJ, Anderson JL. Thrombolysis and myocardial salvage. Results of clinical trials and the
Official Journal of the Society of Magnetic Resonance in Medicine/Society of Magnetic animal paradigm—paradoxic or predictable? Circulation. [Review]. 1993;88(1):296–306.
Resonance in Medicine. [Comparative Study]. 2002;47(2):372–383. 87. Gerber BL, Garot J, Bluemke DA, et al. Accuracy of contrast-enhanced magnetic reso-
57. Chung YC, Lee VS, Laub G, et al.. Inversion recovery cine trueFISP for optimizing TI in nance imaging in predicting improvement of regional myocardial function in patients after
myocardial infarct imaging. Proc Intl Soc Mag Reson Med. 2002;10:219. acute myocardial infarction. Circulation. 2002;106(9):1083–1089.
58. Scheffler K, Hennig J. T(1) quantification with inversion recovery TrueFISP. Magn Reson 88. Kitagawa K, Sakuma H, Hirano T, et al. Acute myocardial infarction: Myocardial viabil-
Med. Official Journal of the Society of Magnetic Resonance in Medicine/Society of ity assessment in patients early thereafter comparison of contrast-enhanced MR imaging
Magnetic Resonance in Medicine. 2001;45(4):720–723. with resting (201)Tl SPECT. Single photon emission computed tomography. Radiology.
59. Mahrholdt H, Wagner A, Holly TA, et al. Reproducibility of chronic infarct size measure- [Comparative Study]. 2003;226(1):138–144.
ment by contrast-enhanced magnetic resonance imaging. Circulation. 2002;106(18):2322– 89. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance imaging
2327. to identify reversible myocardial dysfunction. N Engl J Med. 2000;343(20):1445–1453.
60. Oshinski JN, Yang Z, Jones JR, et al. Imaging time after Gd-DTPA injection is critical in 90. Schvartzman PR, Srichai MB, Grimm RA, et al. Nonstress delayed-enhancement magnetic
using delayed enhancement to determine infarct size accurately with magnetic resonance resonance imaging of the myocardium predicts improvement of function after revascular-
imaging. Circulation. [Research Support, U.S. Gov’t, P.H.S.]. 2001;104(23):2838–2842. ization for chronic ischemic heart disease with left ventricular dysfunction. Am Heart J.
61. de Roos A, van Rossum AC, van der Wall E, et al. Reperfused and nonreperfused myo- 2003;146(3):535–541.
cardial infarction: Diagnostic potential of Gd-DTPA-enhanced MR imaging. Radiology. 91. Selvanayagam JB, Kardos A, Francis JM, et al. Value of delayed-enhancement cardiovas-
1989;172(3):717–720. cular magnetic resonance imaging in predicting myocardial viability after surgical revas-
62. Dendale P, Franken PR, Block P, et al. Contrast enhanced and functional magnetic res- cularization. Circulation. 2004;110(12):1535–1541.
onance imaging for the detection of viable myocardium after infarction. Am Heart J. 92. Bonow RO. Identification of viable myocardium. Circulation. 1996;94(11):2674–2680.
1998;135(5 pt 1):875–880. 93. Kim RJ, Shah DJ. Fundamental concepts in myocardial viability assessment revisited:
63. Pereira RS, Prato FS, Wisenberg G, et al. The determination of myocardial viability using When knowing how much is “alive” is not enough. Heart. 2004;90(2):137–140.
Gd-DTPA in a canine model of acute myocardial ischemia and reperfusion. Magn Reson 94. Ragosta M, Beller GA, Watson DD, et al. Quantitative planar rest-redistribution 201
Med. 1996;36(5):684–693. Tl imaging in detection of myocardial viability and prediction of improvement in left
64. Schaeffer S, Malloy CR, Katz J, et al. Gadolinium-DTPA–enhanced nuclear magnetic ventricular function after coronary bypass surgery in patients with severely depressed left
resonance imaging of reperfused myocardium identification of the myocardial bed at risk. ventricular function. Circulation. 1993;87(5):1630–1641.
J Am Coll Cardiol. 1988;12:1064–1072. 95. Force T, Kemper A, Perkins L, et al. Overestimation of infarct size by quantitative
65. Reimer KA, Jennings RB. The “wavefront phenomenon” of myocardial ischemic cell two-dimensional echocardiography: The role of tethering and of analytic procedures.
death. II. Transmural progression of necrosis within the framework of ischemic bed size Circulation. 1986;73(6):1360–1368.
(myocardium at risk) and collateral flow. Lab Invest. 1979;40(6):633–644. 96. Helak JW, Reichek N. Quantitation of human left ventricular mass and volume
66. Reimer KA, Jennings RB. Myocardial ischemia, hypoxia, and infarction. In: Fozzard HA, by two-dimensional echocardiography: In vitro anatomic validation. Circulation.
Haber E, Jennings RB, Katz AM, Morgan HE. eds. The Heart and Cardiovascular System. 1981;63(6):1398–1407.
2nd ed. New York, NY: Raven Press; 1992: 1875–1973. 97. Nicod P, Gilpin E, Dittrich H, et al. Short- and long-term clinical outcome after Q
67. Jennings RB, Schaper J, Hill ML, et al. Effect of reperfusion late in the phase of reversible wave and non-Q wave myocardial infarction in a large patient population. Circulation.
ischemic injury. Changes in cell volume, electrolytes, metabolites, and ultrastructure. Circ 1989;79(3):528–536.
Res. 1985;56(2):262–278. 98. Klein C, Nekolla SG, Bengel FM, et al. Assessment of myocardial viability with contrast-
68. Whalen DA, Hamilton DG, Ganote CE, et al. Effect of a transient period of ischemia on enhanced magnetic resonance imaging: Comparison with positron emission tomography.
myocardial cells. I. Effects on cell volume regulation. Am J Pathol. 1974;74(3):381–397. Circulation. 2002;105(2):162–167.
69. Reimer KA, Lowe JE, Rasmussen MM, et al. The wavefront phenomenon of ischemic cell 99. Ibrahim T, Bulow HP, Hackl T, et al. Diagnostic value of contrast-enhanced magnetic reso-
death. 1. Myocardial infarct size vs. duration of coronary occlusion in dogs. Circulation. nance imaging and single-photon emission computed tomography for detection of myocardial
1977;56(5):786–794. necrosis early after acute myocardial infarction. J Am Coll Cardiol. 2007;49(2):208–216.
70. Kowallik P, Schulz R, Guth BD, et al. Measurement of regional myocardial blood flow 100. Cwajg JM, Cwajg E, Nagueh SF, et al. End-diastolic wall thickness as a predictor of recov-
with multiple colored microspheres. Circulation. 1991;83(3):974–982. ery of function in myocardial hibernation: Relation to rest-redistribution T1-201 tomogra-
71. Kim RJ, Fieno DS, Parrish TB, et al. Relationhip of MRI delayed contrast enhancement to phy and dobutamine stress echocardiography. J Am Coll Cardiol. 2000;35(5):1152–1161.
irreversible injury, infarct age, and contractile function. Circulation. 1999;100(19):1992– 101. John AS, Dreyfus GD, Pennell DJ. Images in cardiovascular medicine. Reversible wall
2002. thinning in hibernation predicted by cardiovascular magnetic resonance. Circulation.
72. Saeed M, Wendland MF, Masui T, et al. Dual mechanisms for change in myocardial signal 2005;111(3):e24–e25.
intensity by means of a single MR contrast medium—dependence on concentration and 102. Shah DJ, Kim HW, Elliott M, et al. Contrast MRI predicts reverse remodeling and contrac-
pulse sequence. Radiology. 1993;186:175–182. tile improvement in akinetic thinned myocardium. Circulation. 2003;108(17 suppl): 697.
73. Koenig SH, Spiller M, Brown RD, 3rd, et al. Relaxation of water protons in the 103. Shah DJ, Kim HW, Olga James O, et al. Prevalence of regional myocardial thinning and
intra- and extracellular regions of blood containing Gd(DTPA). Magn Reson Med. 1986; relationship with myocardial scarring in patients with coronary artery disease. JAMA.
3(5):791–795. 2013;309(9):909–918.
74. Weinmann HJ, Brasch RC, Press WR, et al. Characteristics of gadolinium-DTPA complex: 104. Albert TS, Kim RJ, Judd RM. Assessment of no-reflow regions using cardiac MRI. Basic
A potential NMR contrast agent. AJR Am J Roentgenol. 1984;142(3):619–624. Res Cardiol. 2006;101(5):383–390.
75. Fieno DS, Kim RJ, Chen EL, et al. Contrast-enhanced magnetic resonance imaging of 105. Hombach V, Grebe O, Merkle N, et al. Sequelae of acute myocardial infarction regarding
myocardium at risk distinction between reversible and irreversible injury throughout cardiac structure and function and their prognostic significance as assessed by magnetic
infarct healing. J Am Coll Cardiol. 2000;36(6):1985–1991. resonance imaging. Eur Heart J. 2005;26(6):549–557.
76. Rehwald WG, Fieno DS, Chen EL, et al. Myocardial magnetic resonance imaging con- 106. Wu KC, Zerhouni EA, Judd RM, et al. Prognostic significance of microvascular obstruc-
trast agent concentrations after reversible and irreversible ischemic injury. Circulation. tion by magnetic resonance imaging in patients with acute myocardial infarction.
2002;105(2):224–229. Circulation. 1998;97(8):765–772.
77. Bolli R. Myocardial ‘stunning’ in man. Circulation. 1992;86(6):1671–1691. 107. Nijveldt R, Hofman MB, Hirsch A, et al. Assessment of microvascular obstruction and
78. Hillenbrand HB, Kim RJ, Parker MA, et al. Early assessment of myocardial salvage by prediction of short-term remodeling after acute myocardial infarction: Cardiac MR imag-
contrast-enhanced magnetic resonance imaging. Circulation. 2000;102(14):1678–1683. ing study. Radiology. 2009;250(2):363–370.
79. Choi KM, Kim RJ, Gubernikoff G, et al. Transmural extent of acute myocardial infarction 108. Rochitte CE, Lima JA, Bluemke DA, et al. Magnitude and time course of microvas-
predicts long-term improvement in contractile function. Circulation. 2001;104(10):1101– cular obstruction and tissue injury after acute myocardial infarction. Circulation.
1107. 1998;98(10):1006–1014.
80. Beek AM, Kuhl HP, Bondarenko O, et al. Delayed contrast-enhanced magnetic resonance 109. Mollet NR, Dymarkowski S, Volders W, et al. Visualization of ventricular thrombi with
imaging for the prediction of regional functional improvement after acute myocardial contrast-enhanced magnetic resonance imaging in patients with ischemic heart disease.
infarction. J Am Coll Cardiol. 2003;42(5):895–901. Circulation. 2002;106(23):2873–2876.
81. Ingkanisorn WP, Rhoads KL, Aletras AH, et al. Gadolinium delayed enhancement car- 110. Srichai MB, Junor C, Rodriguez LL, et al. Clinical, imaging, and pathological characteris-
diovascular magnetic resonance correlates with clinical measures of myocardial infarc- tics of left ventricular thrombus: A comparison of contrast-enhanced magnetic resonance
tion. J Am Coll Cardiol. [Comparative Study Research Support, U.S. Gov’t, P.H.S.]. imaging, transthoracic echocardiography, and transesophageal echocardiography with
2004;43(12):2253–2259. surgical or pathological validation. Am Heart J. 2006;152(1):75–84.
82. Nishimura T, Kobayashi H, Ohara Y, et al. Serial assessment of myocardial infarction by 111. Weinsaft JW, Kim HW, Shah DJ, et al. Detection of left ventricular thrombus by delayed-
using gated MR imaging and Gd-DTPA. AJR Am J Roentgenol. 1989;153(4):715–720. enhancement cardiovascular magnetic resonance prevalence and markers in patients with
83. Gruppo Italiano per lo Studio della Streptochinasi nell’Infarto Miocardico (GISSI). systolic dysfunction. J Am Coll Cardiol. [Comparative Study Multicenter Study Research
Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t Validation Studies].
Lancet. 1986;1(8478):397–402. 2008;52(2):148–157.
LWBK1209-ch17_p251-282.indd 281 17/05/13 5:10 PM

112. Weinsaft JW, Kim RJ, Ross M, et al. Contrast-enhanced anatomic imaging as compared 118. Nelson C, McCrohon J, Khafagi F, et al. Impact of scar thickness on the assessment of
to contrast-enhanced tissue characterization for detection of left ventricular thrombus. viability using dobutamine echocardiography and thallium single-photon emission computed
JACC Cardiovasc Imaging. [Comparative Study Research Support, Non-U.S. Gov’t]. tomography: A comparison with contrast-enhanced magnetic resonance imaging. J Am
2009;2(8):969–979. Coll Cardiol. 2004;43(7):1248–1256.
113. Mahrholdt H, Wagner A, Judd RM, et al. Delayed enhancement cardiovascular magnetic reso- 119. Samady H, Elefteriades JA, Abbott BG, et al. Failure to improve left ventricular func-
nance assessment of non-ischaemic cardiomyopathies. Eur Heart J. 2005;26(15):1461–1474. tion after coronary revascularization for ischemic cardiomyopathy is not associated with
114. Wellnhofer E, Olariu A, Klein C, et al. Magnetic resonance low-dose dobutamine test is worse outcome. Circulation. 1999;100(12):1298–1304.
superior to SCAR quantification for the prediction of functional recovery. Circulation. 120. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation
2004;109(18):2172–2174. and nomenclature for tomographic imaging of the heart: A statement for healthcare pro-
115. Perrone-Filardi P, Pace L, Prastaro M, et al. Assessment of myocardial viability in patients fessionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of
with chronic coronary artery disease. Rest-4-hour–24-hour 201Tl tomography versus the American Heart Association. Circulation. 2002;105(4):539–542.
dobutamine echocardiography. Circulation. 1996;94(11):2712–2719. 121. Kim RJ, Manning WJ. Viability assessment by delayed enhancement cardiovascular mag-
116. Lieberman AN, Weiss JL, Jugdutt BI, et al. Two-dimensional echocardiography and netic resonance: Will low-dose dobutamine dull the shine? Circulation. 2004;109(21):
infarct size: Relationship of regional wall motion and thickening to the extent of myocar- 2476–2479.
dial infarction in the dog. Circulation. 1981;63(4):739–746. 122. Fuster V, Kim RJ. Frontiers in cardiovascular magnetic resonance. Circulation. 2005;
117. Mahrholdt H, Wagner A, Parker M, et al. Relationship of contractile function to trans- 112(1):135–144.
mural extent of infarction in patients with chronic coronary artery disease. J Am Coll
Cardiol. 2003;42(3):505–512.
LWBK1209-ch17_p251-282.indd 282 17/05/13 5:10 PM

Chapter
Charles B. Higgins
Karen G. Ordovas 18
Delayed Enhancement in
Nonischemic Myocardial Disease
■■ Goals and Indications 1. Diagnosis of acute and chronic myocarditis

2. Recognition of infiltration and bulk fibrosis in cardiomy-
■■ Mechanism of Lge in Nonischemic
opathies
Myocardial Disease 3. Risk assessment and prognosis in dilated and hypertro-
■■ Viral Myocarditis phic cardiomyopathies
4. Diagnosis and risk assessment in cardiac sarcoid
■■ Dilated (Congestive) Cardiomyopathy
5. Diagnosis of arrhythmogenic right ventricular dysplasia
■■ Left Ventricular Partial Noncompaction (cardiomyopathy)
6. Recognition of myocardial involvement in systemic
■■ Peripartum-Dilated Cardiomyopathy
vasculitides
■■ Hypertrophic Cardiomyopathy 7. Recognition of myocardial involvement in chronic infec-
■■ Restrictive Cardiomyopathy tions
8. Recognition of myocardial fibrosis after repair of Tetralogy
■■ Specific Infiltrative Cardiomyopathies of Fallot
Cardiac Amyloidosis LGE occurs in many myocardial diseases other than isch-
Cardiac Sarcoidosis emic heart disease (1,2). While the distribution of LGE is
Systemic Sclerosis (Scleroderma) invariably subendocardial for acute and chronic myocar-
Anderson–Fabry Disease dial infarction, the pattern of enhancement has a variable
Danon Disease transmural distribution in nonischemic myocardial disease.
Endomyocardial Fibrosis LGE conforms to the distribution of one or more coronary
■■ Myocardial Vasculitides arteries in ischemic heart disease while in nonischemic
myocardial disease, it does not. With nonischemic myocar-
Churg–Strauss Syndrome dial disease the mural pattern is not uncommonly midwall
■■ Arrhythmogenic Right Ventricular Dysplasia or subepicardial and may be spotty in multiple regions of
(Cardiomyopathy) the ventricle. The pattern of distribution in some may be
almost diagnostic of a specific myocardial disease. In some
■■ Cardiac Transplant Rejection Arteriolopathy
nonischemic myocardial disease the mere presence as well as
■■ Chronic Infections with Myocardial the overall extent of LGE carries predictive implications for
Involvement morbidity and mortality.
Chagas Disease
Lyme Disease
Mechanism of Lge in Nonischemic
■■ Postoperative Congenital Heart Disease
Myocardial Disease
■■ Diffuse Myocardial Fibrosis
The mechanism of LGE in acute and chronic infarction is
the increase in the extracellular space caused by necrosis
Goals and Indications (loss of cell membrane integrity) with the former and the
larger extracellular space of scar compared to myocardium
The major goals and indications for late gadolinium enhance- with the latter (1). The mechanism in nonischemic myocar-
ment (LGE) magnetic resonance (MR) in nonischemic heart dial disease is also predominantly, if not exclusively, due to
disease are: increase in the distribution volume of gadolinium chelates
283
LWBK1209-ch18_p283-291.indd 283 5/18/13 12:53 PM

Figure 18.1. Acute myocarditis. Inversion recovery gradient-echo

(IRGRE) image in horizontal long-axis (HLA) plane acquired 15 minutes
after injection of gadolinium chelate (LGE image). There are multiple
sites of hyperenhancement in right ventricular (RV) and left ventricular
(LV) myocardium. Note hyperenhancement in midwall of septum and
subepicardium in lateral wall of the LV.
Figure 18.2. Myocarditis caused by parvovirus. LGE image in ver-

tical long-axis plane shows hyperenhancement of the lateral wall of the
due to an increase in the extracellular space of the myocar- LV in subepicardial layer. This site of LGE is considered to be charac-
dium (1). The increase in the extracellular space may be due teristic for parvovirus myocarditis (Compliment of Udo Sechtem, MD,
to necrosis of myocardial cells (myocarditis), replacement of Stuttgast, Germany). LV, left ventricle; RV, right ventricle.
myocardial cells by bulk fibrosis (hypertrophic cardiomy-
opathy [HCM]), or replacement of myocardium by various
infiltrates (amyloidosis, Anderson–Fabry disease, sarcoid-
osis), or a combination of two or three of the above (1). has no LGE (Fig. 18.3). On the other hand, LGE is observed
in about 30% of patients with alleged nonischemic-dilated
cardiomyopathy (9). The distribution is usually midwall
Viral Myocarditis rather than subendocardial (9).
The extent of fibrosis depicted on LGE images has been
Some but not all patients with acute or subacute myocarditis associated with increased intraventricular dyssynchrony in
show LGE (3–8) (Fig. 18.1). In acute myocarditis there is dilated cardiomyopathy (10). Patients with LGE have been
invariably myocardial hyperenhancement on T2-weighted shown to have more future adverse cardiac events (11).
imaging even when no LGE is evident (3). Moreover, LGE The presence and extent of LGE was more predictive of
may ensue later than T2 hyperintensity and persist after adverse cardiac events than left ventricular ejection frac-
resolution of T2 hyperenhancement. T2 hyperenhancement tion (11,12).
is attributed to myocardial edema while LGE indicates myo-
cardial necrosis and/or fibrosis. LGE is detected in about
70% of patients with chronic myocarditis (7,8). Left Ventricular Partial
The location of LGE in myocarditis is subepicardial Noncompaction
(38%) or midwall (68%) in distribution in either ventricle
or both (3–6). Some etiologic agents of viral myocarditis Partial noncompaction is now recognized as the cause of
may favor a specific location in the ventricle(s) such as the dilated cardiomyopathy especially in older children and young
posterolateral wall of the left ventricle with parvovirus (6) adults. It is classified as a primary genetic-dilated cardiomy-
(Fig. 18.2). opathy. The left ventricular volume is increased and ejection
fraction reduced. The diagnostic criterion on MR is estab-
lished on the cine MR image at end diastole in the short-axis
Dilated (Congestive) phase. A ratio of noncompacted to compacted wall thickness
Cardiomyopathy greater than 2.3 establishes the diagnosis. It is not yet clear
if lesser ratios identify milder forms. LGE of the trabecular
LGE is used to distinguish between ischemic-dilated and portion has been reported for more severe or late form of the
nonischemic-dilated cardiomyopathy. Subendocardial or disease (13). Because of stasis of gadolinium-enhanced blood
transmural LGE is indicative of ischemic etiology of a dilated pool within the trabecula, interpretation of this finding may
cardiomyopathy. Idiopathic-dilated cardiomyopathy usually be equivocal.
LWBK1209-ch18_p283-291.indd 284 16/05/13 9:35 PM

Chapter 18 ■ Delayed Enhancement in Nonischemic Myocardial Disease 285
Figure 18.3. Idiopathic-dilated

cardiomyopathy. LGE image in HLA
(left) and VLA (right) planes shows
severe LV enlargement and no delayed
hyperenhancement at 15 minutes after
injection of gadolinium chelate. LA,
left atrium; LV, left ventricle. (From
Kim RJ, Shah DJ, Mudd RM. How
we perform delayed enhancement
imaging. J Cardiovasc Magn Reson
2003;5(3):505–514, with permission.)
Peripartum-Dilated LGE usually corresponds to areas of most severe hypertro-

Cardiomyopathy phy with focal necrosis and/or fibrosis (16).
LGE has been repeated in 81% of patients with HCM
Peripartum-dilated cardiomyopathy is diagnosed when (17). The distribution is focal in the area of most severe hy-
heart failure and dilated left ventricle with reduced ejection pertrophy, usually the ventricular septum (Fig. 18.4), or at
fraction ensues in the last month of or first 5 months after the junction point between the septum and the anterior and
delivery in a previously healthy woman in the absence of an posterior walls of the ventricle (17) (Fig. 18.5). It is usually
identifiable etiology. Cine MR shows increases in LV end- midwall in distribution and never subendocardial.
diastolic and end-systolic volumes and decreased ejection LGE is important in risk assessment in patients with
fraction. In a minority of patients, LGEs have been identified HCM. Patients with HCM and any degree of LGE have
in a midwall distribution (14). a sevenfold higher risk of nonsustained ventricular tachy-
cardia (18). In addition to an increased incidence of ar-
rhythmias, LGE is also associated with poorer functional
Hypertrophic Cardiomyopathy class, decreased left ventricular function, conduction distur-
bances, abnormal Q-waves, and giant T-waves (19). Patients
Cardiac MR imaging is important in assessing the severity with more rapid progression of disease as reflected in left
and distribution of hypertrophy in HCM. It is the most effec- ventricular dilatation have more extensive LGE (15). The
tive modality in the latter, especially in disclosing unusual same study showed greater extent of LGE in patients with
hypertrophy patterns such as midventricular and apical two or more risk factors for sudden death. Thus, CVMR
forms of HCM. Moreover, the quantification of LV mass with LGE imaging is probably indicated in all patients with
and identification of LGE by CVMR constitute two of the HCM not necessarily for diagnosis but for risk assessment
most important prognostic indicators in this disease (15). and prognosis.
Figure 18.4. Hypertrophic car-

diomyopathy. Cine MR (left) and
LGE (right) images show asymmetric
septal hypertrophy (asterisk). The
hyperenhancement (arrow) is at the
site of maximum septal hypertrophy.
(From Kim RJ, Shah DJ, Mudd RM.
How we perform delayed enhancement
2003;5(3):505–514, with permission.)
LWBK1209-ch18_p283-291.indd 285 16/05/13 9:35 PM

Figure 18.5. Hypertrophic cardiomyopathy. LGE image in the Figure 18.6. Cardiac amyloidosis. LGE image in short-axis plane
short-axis plane shows hyperenhancement at sites of connections shows circumferential subendocardial hyperenhancement at 5 to
(hinge points, arrows) of ventricular septum to anterior and posterior 6 minutes after injection of gadolinium chelate. (From Kim RJ, Shah
walls of the LV. LV, left ventricle; RV, right ventricle. DJ, Mudd RM. How we perform delayed enhancement imaging.
J Cardiovasc Magn Reson 2003;5(3):505–514, with permission.)
Restrictive Cardiomyopathy have been identified—patchy midwall, subendocardial, or

subepicardial. DCE of the tips of the papillary muscles, right
Restrictive cardiomyopathy (RCM) is characterized typi- atrial wall, and atrioventricular valve leaflets may be recog-
cally on CVMR by dilated atria in the presence of normal- nized, usually in association with the global subendocardial
or small-sized ventricles. Functionally, there is diastolic dys- pattern.
function with normal or only mildly reduced left ventricular An ensemble of multiple inversion times may be em-
systolic function. Some cases of RCM are due to myocar- ployed to determine the nulling point of the amyloid-
dial infiltrative diseases or storage diseases. The infiltrative infiltrated myocardium compared to cavitary blood (21).
types sometimes can be recognized by the presence of even Normally with increasing inversion times, the blood pool
a pathognomonic pattern of LGE such as that occurs in nulls before the myocardium. In amyloidosis, the myocar-
cardiac amyloidosis. The LGE in these entities is described dium nulls slightly before the blood pool (Fig. 18.7). When
below under the several specific entities. Other causes of the diagnosis of cardiac amyloidosis is suspected, a train of
RCM involve fibrosis of the myocardium. These entities may five to eight inversion times at one short-axis level is done
or may not demonstrate LGE. and then higher-resolution inversion recovery gradient-
echo images of the entire heart are done in the short-axis
and horizontal long-axis planes at 6 to 8 minutes and re-
Specific Infiltrative peated at 10 to 15 minutes after gadolinium injection.
Cardiomyopathies Contrast between sites of amyloid deposition and the
remainder of the myocardium starts to diminish at about
Cardiac Amyloidosis
8 minutes.
The mechanism of LGE in amyloidosis is not entirely
Cardiac amyloidosis is characterized on echocardiogra- clear and probably multifactorial. Some have alleged that it
phy and cine MR by thickened myocardial walls (usually is mostly due to fibrosis accompanying amyloid infiltration
>15 mm) and reduced left ventricular ejection fraction. The while others hold expansion of the extracellular space by
differential diagnoses are usually compensatory left ventric- the amyloid deposition as the dominant factor (22,23). It is
ular hypertrophy and HCM. Both of the latter have high likely that both factors are contributory. Moreover, earlier
normal or supranormal left ventricular ejection fraction. The nulling of the amyloid-containing myocardium compared to
diagnosis of cardiac amyloidosis can be made with a sensi- the blood pool raises the question of active uptake of gado-
tivity of 80% and specificity of 94% with DCE-MR imaging linium by the amyloid.
using myocardial biopsy as the gold standard (20). Global The pattern of LGE in a global subendocardial pattern
subendocardial pattern of DCE is nearly pathognomonic for has also been reported in systemic sclerosis (scleroderma)
amyloidosis (20–22) (Fig. 18.6). Other less specific patterns and chronic cardiac transplant rejection arteriolopathy (24).
LWBK1209-ch18_p283-291.indd 286 16/05/13 9:35 PM

Figure 18.7. Series of increasing

inversion times for short-axis IRGRE
images (top panel) shows the myo-
cardial signal starting to null (arrow)
before cavitary blood. Higher-resolution
LGE at optimal inversion time for null-
ing normal myocardial signal (lower
right) demonstrates circumferential sub-
endocardial hyperenhancement. Note
signal of cavitary blood is nulled with
normal myocardium. The near simulta-
neous nulling of blood pool and normal
myocardium is characteristic for myo-
cardial amyloidosis. LV, left ventricle.
(From Kim RJ, Shah DJ, Mudd RM.
How we perform delayed enhancement
2003;5(3):505–514, with permission.)
Any vasculitis that diffusely involves the myocardium might

cause subendocardial ischemic necrosis and or fibrosis and
induce this pattern.
Primary amyloidosis (immunoglobulin light chain dis-
ease) and heredity transthyretin types cause cardiomyopa-
thies. Restrictive physiology with diastolic dysfunction is a
consequent of associated endomyocardial fibrosis.
Cardiac Sarcoidosis
The diagnosis of cardiac sarcoidosis is not made in the
majority of patients with systemic sarcoidosis with cardiac
involvement; approximately 7% of patients have recognized
cardiac symptoms compared to 20% to 30% of patients
with cardiac involvement on postmortem examination (25–
28). The clinical features of heart block, ventricular arrhyth-
mias, or congestive heart failure in a patient with known
pulmonary sarcoidosis suggest cardiac involvement; how-
ever, even symptomatic cardiac sarcoidosis in the absence of
known pulmonary sarcoidosis renders this diagnosis nonob-
vious. The pattern of LGE may be relatively nonspecific but
some patterns such as DCE of the right side of the ventricu-
Figure 18.8. Cardiac sarcoid. LGE image in short-axis plane
lar septum or right ventricular free wall suggest myocardial shows spotty hyperenhancement at several sites in the LV and the right
sarcoidosis (Fig. 18.8). side (arrowheads) of the ventricular septum. (Compliments of Peter
In suspected cardiac sarcoidosis, demonstration of LGE Buser, M.D., Basle, Switzerland; From Kim RJ, Shah DJ, Mudd RM.
is more sensitive for the diagnosis than all other clinical and How we perform delayed enhancement imaging. J Cardiovasc Magn
electrocardiographic features combined (27). The pattern of Reson 2003;5(3):505–514, with permission).
LWBK1209-ch18_p283-291.indd 287 16/05/13 9:35 PM

Figure 18.9. Cardiac sarcoid. LGE

image in short-axis (left) and HLA
(right) planes show hyperenhancement
of the free wall of the RV. LV, left ven-
tricle; RV, right ventricle.
LGE is variable; diffuse or focal patterns in the midwall or Danon Disease

subepicardium occur and are nonspecific (27,28). However,
This is a rare X-linked genetic disorder due to deficiency
involvement of the right side of the ventricular septum is
of a lysosome. It usually affects men early in life causing
considered more characteristic (27). Transmural LGE of the
heart failure, skeletal myopathy, and mental retardation
right ventricular free wall suggests the diagnosis of either
(32,34). In the heart, there is symmetric wall thickening
cardiac sarcoidosis or arrhythmogenic right ventricular car-
of a severe degree mimicking concentric form of HCM but
diomyopathy (dysplasia) (Fig. 18.9).
with reduced left ventricular function (34). LGE has been
The presence and extent of LGE has prognostic sig-
reported to be subendocardial rather than midwall as in
nificance in cardiac sarcoidosis (28,29). The presence
HCM (35).
of LGE has been related to ventricular dimensions and
function, ventricular arrhythmias, and severity of mitral
regurgitation (28,29). A significant correlation has been
Endomyocardial Fibrosis
shown between the global extent of myocardial LGE and
increased left ventricular volumes and decreased contrac- The chronic fibrotic stage of endomyocardial fibrosis has
tility (29). bulk fibrosis (scar) in the subendocardium of either the right
or left ventricle or both (36,37). This may be associated with
or preceded by eosinophilic infiltration of the subendocar-
Systemic Sclerosis (Scleroderma) dium and peripheral eosinophilia (Loeffler’s eosinophilic
endomyocardial fibroplasia). In infants, a similar process
Myocardial fibrosis is not uncommon at postmortem exami-
occurs in primary or secondary fibroelastosis. Congenital
nation (30); however, cardiac involvement is seldom docu-
aortic stenosis in utero can cause secondary endocardial
mented during life. Progressive right and/or left ventricular
fibroelastosis (38).
failure or ventricular arrhythmias suggests the diagnosis.
LGE-MR demonstrates extensive subendocardial en-
Cine MR can be used to document reduced ventricular func-
hancement in the left ventricle and/or right ventricle (36–38)
tion. The LGE-MR imaging displays the sites of bulk fibro-
(Fig. 18.10). Global circumferential LGE of the left ventricle
sis; the pattern is linear midwall pattern in predominantly
has been reported in secondary endocardial fibroelastosis in
basal and midventricular levels (31).
an infant with severe aortic stenosis in utero (38).
Anderson–Fabry Disease
Myocardial Vasculitides
This is an X-linked recessive genetic disease due to lack of
lysosomal enzyme, galactosidase-A, which metabolizes gly- Myocardial vasculitis of the distal branches (end arteries
cosphingolipids (32). This causes intracellular lysosomal and arterioles) causes ischemic injury and fibrosis of the
infiltration of ceramide trihexoside in mostly skin, kid- subendocardium typically, but apparently may also involve
neys, and heart. It occurs mostly in men. Cardiac involve- the midwall and subepicardium. Myocardial vasculitis may
ment can mimic morphologic and clinical features of HCM. occur in Churg–Strauss syndrome and Wegener’s syndrome.
Concentric wall thickening is usual.
Cine MR demonstrates increase in left ventricular wall
Churg–Strauss Syndrome
thickness, especially or even exclusively in the inferolateral
region. LGE characteristically occurs focally in the midwall Several case reports have described LGE of the left ventricu-
of the inferolateral region (33). lar myocardium as an indication of cardiac involvement in
LWBK1209-ch18_p283-291.indd 288 16/05/13 9:35 PM

Figure 18.10. Loeffler’s eosino-

philic subendocardial fibrosis. LGE
(left) and cine (right) images in axial
planes show extensive subendocardial
hyperenhancement throughout the
RVE, and severe enlargement of the RV
and right atrium. RA, right atrium; RV,
right ventricle.
the syndrome (39–41). The LGE is most typically patchy the right ventricular free wall. Tandri et al. (42) found that
distribution in the subendocardium but midwall or subepi- LGE is predictive of inducible ventricular tachycardia by
cardium distribution may also be encountered (39,40). The programmed electric stimulation of the ventricle.
most frequent site is the ventricular septum. Diffuse suben-
docardial pattern including the papillary muscles has been
reported (40). Cardiac Transplant Rejection
Arteriolopathy
Arrhythmogenic Right Severe chronic transplant rejection may cause a diffuse coro-
Ventricular Dysplasia nary arteriolopathy and arteriolopathy resulting in global
(Cardiomyopathy) subendocardial ischemic injury and fibrosis. Cine MR dem-
onstrates global hypokinesis and reduced ejection fraction of
LGE has been recognized at the site of fibrosis or fibropathy the left ventricle. There may be global subendocardial LGE
replacement of the right ventricle in patients with arrhyth- similar to that found in cardiac amyloidosis.
mogenic right ventricular dysplasia (ARVD). It has also
been reported in the left ventricle in these patients. It usually
involves part or the entire free wall of the right ventricle (Fig. Chronic Infections with
18.11). It has also been shown in the right side of the ven-
Myocardial Involvement
tricular septum. The distribution of LGE in ARVD is similar
to the distribution in some patients with cardiac sarcoid.
Chagas Disease
The incidence of LGE in ARVD is unknown, but in one
report (41), it was identified in seven of eight patients. At Cardiac involvement is frequent in Chagas disease. Left ven-
the site of LGE, there is usually akinesis or dyskinesis of tricular segments with fibrosis show LGE with a variable mural
Figure 18.11. Arrhythmogenic right

ventricular dysplasia (cardiomyopathy).
Cine (left) and LGE (right) shows wall
thinning and dyskinesis (arrow) of
RV free wall and hyperenhancement
(arrowhead). LV, left ventricle; RV,
right ventricle.
LWBK1209-ch18_p283-291.indd 289 16/05/13 9:35 PM

Figure 18.12. Postoperative

Tetralogy of Fallot. LGE (left) and
spin-echo (right) images show hyperen-
hancement at the site of infundibular
resection in Tetralogy of Fallot. Also
note focal enhancement of the ventricu-
lar septum at a hinge point (arrow).
MPA, main pulmonary artery; RV,
right ventricle.
distribution: Subepicardial, midwall, and transmural (43,44). has also been shown between LGE and adverse clinical mark-
The regional distribution is typically inferolateral basal and/or ers in adults after repair of Tetralogy of Fallot. Likewise,
apical. Apical left ventricular aneurysm with transmural LGE others have reported correlation between LGE and symp-
is very characteristic of chronic Chagas disease. toms of heart failure in patients after repair of Tetralogy of
LGE may have prognostic implications. It is seen more Fallot (51).
often in patients with advanced disease and worse clinical
status. Patients with ventricular tachycardia have LGE of the
left ventricular myocardium. Diffuse Myocardial Fibrosis
The application of LGE-MR has been directed at demon-
Lyme Disease
strating bulk (macroscopic) myocardial fibrosis. A technique
Cardiac involvement in Lyme disease is infrequent but may has also been devised for demonstrating diffuse (micro-
be heralded by the onset of heart block (45). The LGE in this scopic) fibrosis. This technique involves the estimation of
disease usually is midwall in the basal anteroseptal region, the distribution volume of gadolinium chelate in the myo-
which is the site of the atrioventricular node and proximal cardium based on the ratio of T1 relaxivity of myocardium
atrioventricular conducting fibers (46). Decrease or even to blood over time (52,53). Since the extracellular volume
resolution of LGE has been shown after treatment and reso- of fibrosis is greater than normal myocardium, the distribu-
lution of atrioventricular block (46,47). tion volume of diffusely fibrotic myocardium is greater than
normal myocardium. Diffuse myocardial fibrosis has been
identified with this technique in patients with heart failure
Postoperative Congenital and dilated cardiomyopathy, and in adults with congenital
Heart Disease heart disease and ventricular dysfunction.
Fibrosis at the site of surgical resection in congenital heart

disease not uncommonly demonstrates LGE. This is typical References
in the right ventricular outlet region after infundibulectomy
1. Ordovas KG, Higgins CB. Delayed enhancement on MR images of the myocardium: Past,
or outflow tract patch placement in Tetralogy of Fallot (Fig. present, future. Radiology. 2011;261:358–374.
18.12). However, LGE has also been found at sites remote 2. O’Donnell DH, Abbara S, Chaithirapan V, et al. Cardiac MR imaging of nonischemic
from the surgical site. In postoperative patients with exces- cardiomyopathies. Radiology. 2012;262:403–422.
3. Abdel-Aty H, Boye P, Zagrosek A, et al. Diagnostic performances of CMR in patients
sive septal motion in diastole and/or systole due pulmonary with suspected acute myocarditis; comparison of different methods. J Am Coll Cardiol.
regurgitation and/or stenosis, LGE may be sometimes seen 2005;45:1815–1822.
at the ventricular septal hinge points (Fig. 18.12). 4. Friedrich M, Strohm O, Schulz-Menger J, et al. MRI for the diagnosis and follow-up of
myocarditis. Circulation. 1999;99:457–460.
After repair of Tetralogy of Fallot, Oosterhof et al. (48) 5. Mahrholdt H, Goedecke C, Wagner A, et al. CMR assessment of human myocarditis; a
found LGE in the right ventricular outlet region in 71% comparison to histology and molecular pathology. Circulation. 2004;109:1250–1258.
6. Mahrholdt H., Wagner A., Deluigi CC, et al. Presentation, patterns of myocardial damage,
of 24 patients. Harris et al. (49) found LGE in 91% of and clinical course of viral myocarditis. Circulation. 2006;114:1581–1590.
34 patients after correction of conotruncal abnormalities. 7. Gutberlet M, Spors B, Thoma T, et al. Suspected chronic myocarditis at CMR; diagnostic
Sites of LGE remote from surgical trauma especially at the accuracy and association with immunohistological detected inflammation and viral persis-
tence. Radiology. 2008;246:401–409.
inferior right ventricular insertion point have been recog- 8. De Cobelli F, Pieroni M, Esporsito A, et al. Delayed gadolinium enhanced CMR in patients
nized (50). with chronic myocarditis presenting with heart failure or recurrent arrhythmias. J Am Coll
The presence of LGE in the outflow tract after repair of Cardiol. 2006;47:1649–1654.
9. McCrohon JA, Moon JC, Prasad SK, et al. Differentiation of heart failure related to
Tetralogy of Fallot has been associated with increased right dilated cardiomyopathy and coronary artery diseases using gadolinium enhanced CMR.
ventricular volumes and worse function (48). An association Circulation. 2003;108:54–59.
LWBK1209-ch18_p283-291.indd 290 16/05/13 9:35 PM

10. Tigen K, Karaahmet T, Kirma C, et al. Diffuse LGE by CMR predicts significant intraven- 32. Seward JB, Casaclang-Verzosa G. Infiltrative cardiovascular diseases: Cardiomyopathies
tricular systolic dyssynchrony in patients with nonischemic dilated cardiomyopathy. J Am that look alike. J Am Coll Cardiol. 2010;55:1769–1779.
Soc Echocardiogr. 2010;23:416–422. 33. DeCobelli F, Esposito A, Belloni E, et al. Delayed enhanced cardiac MRI for differentia-
11. Assomull RG, Prasad SK, Lyne J, et al. Cardiovascular MR, fibrosis and prognosis in tion of Fabry’s disease and symmetric hypertrophic cardiomyopathy. Am J Roentgenol.
dilated cardiomyopathy. J Am Coll Cardiol. 2006;48:1977–1985. 2009;192:W97–W102.
12. Wu KC, Weiss RG, Thiemann DR, et al. Late gadolinium enhancement by CMR herald 34. Arad M, Baron BJ, Gorham JM, et al. Glycogen storage diseases presenting as HCM. N
adverse prognosis in nonischemic cardiomyopathy. J Am Coll Cardiol. 2008;51:2414– Eng J Med. 2005;352:262–272.
2421. 35. Piotrowska-Kownacka D, Kownacki L, Kuch M, et al. CMR findings in a case of Danon
13. Dodd JD, Holmvang G, Hoffman U, et al. Quantification of LV noncompaction and tra- disease. J Cardiovasc Magn Reson. 2009;11:12.
becular delayed enhancement with cardiac MRI: Correlation with clinical severity. AJR. 36. Paydar A, Ordovas KG, Reddy GP. MRI for endomyocardial fibrosis. Pediatr Cardiol.
2007;189:974–980. 2008;29:1004–1005.
14. Marmursztejn J, Vignaux O, Goffinet F, et al. Delayed enhanced CMR imaging features in 37. Syed IS, Martinez MW, Feng DL, et al. Cardiac MRI of eosinophilic endomyocardial
peripartum cardiomyopathy. Int J Cardiol. 2009;137:63–64. disease. Int J Cardiol. 2008;126:e50–e52.
15. Moon JC, McKenna WJ, McCrohon JA, et al. Toward clinical risk assessment in HCM 38. Tworetzky W, del Nido PJ. Powell AJ, et al. Usefulness of MRI of left ventricular endocar-
with gadolinium CMR. J Am Coll Cardiol. 2003;41:1561–1567. dial fibroelastosis in infants after fetal intervention for aortic valve stenosis. Am J Cardiol.
16. Moon JC, Reed E, Sheppard MN, et al. The histologic basis of late gadolinium enhance- 2005;96:1568–1570.
ment CMR in HCM. J Am Coll Cardiol. 2004;43:2260–2264. 39. Baccouche H, Yilmaz A, Alscher D, et al. MR assessment and therapy monitoring of car-
17. Choudhury L, Mahrholdt H, Wagner A, et al. Myocardial scarring in asymptomatic or diac involvement in Chung-Strauss syndrome. Circulation. 2008;117:1745–1749.
mildly symptomatic patients with HCM. J Am Coll Cardiol. 2002;40:2156–2164. 40. Petersen SE, Kardos A, Neubauer S. Subendocardial and papillary muscle involvement
18. Adabag AS, Maron BJ, Appelbaum E, et al. Occurrence and frequency of arrhythmias in in a patient with Churg-Strauss syndrome, detected by contrast enhanced CMR. Heart.
HCM in relation to delayed enhancement on CMR. J Am Coll Cardiol. 2008;51:1369–1374. 2005;91:e9.
19. Satoh H, Matoh F, Shiraki K, et al. Delayed enhancement on CMR and clinical morpho- 41. Pfluger HB, Phrommintikul A, Mariani JA, et al. Utility of myocardial fibrosis and fatty
logic, and electrocardiographic features in HCM. J Card Fail. 2009;15:419–427. infiltration detected by CMR in the diagnosis of ARVD—a single center experience. Heart
20. Vogelsberg H, Mahrholdt H, Deluigi CC, et al. Cardiovascular MR in clinically sus- Lung Circ. 2008;17:478–483.
pected amyloidosis: Noninvasive imaging compared to endomyocardial biopsy. J Am Coll 42. Tandri H, Saranathan M, Rodriguez ER, et al. Noninvasive detection of myocardial fibro-
Cardiol. 2008;51:1022–1030. sis in ARVC using delayed enhanced MRI. J Am Coll Cardiol. 2005;45:98–103.
21. Maciera AM, Joshi J, Prasad SK, et al. Cardiovascular MR in cardiac amyloidosis. 43. Rochitte CE, Oliveira PF, Andrade JM, et al. Myocardial delayed enhancement by
Circulation. 2005;111:186–193. MRI in patients with Chagas disease; a marker of disease severity. J Am Coll Cardiol.
22. Perugini E, Rapezzi C, Piva T, et al. Non-invasive evaluation of the myocardial substrate of 2005;46:1553–1558.
cardiac amyloidosis by gadolinium CMR. Heart. 2006;92:343–349. 44. Rochitte CE, Nacif MS, de Oliveira Júnior AC, et al. Cardiac magnetic resonance in
23. Hosch W, Kristen AV, Libicher M, et al. Late enhancement in cardiac amyloidosis: Chagas Disease. Artif Organs. 2007;31:259–267.
Correlation of MRI enhancement pattern with histopathological findings. Amyloid. 45. Lelovas P, Dontas I, Bassiakou E, et al. Cardiac implications of Lyme disease, diagnosis and
2008;15:196–204. therapeutic approach. Int J Cardiol. 2008;129:15–21.
24. Mahrholdt H, Wagner A, Judd RM, et al. Delayed enhanced CMR assessment of non- 46. Naik M, Kim D, O’Brien F, et al. Images in cardiovascular medicine: Lyme Carditis.
ischemic cardiomyopathies. Eur Heart J. 2005;26:1461–1474. Circulation. 2008;118:1881–1884.
25. Smedema JP, Snoep G, van Kroonenburgh MP, et al. Evaluation of the accuracy of 47. Karadag B, Spieker LE, Schwitter J, et al. Lyme Carditis: restituto ad integrum documented
gadolinium enhanced CMR in the diagnosis of cardiac sarcoidosis. J Am Coll Cardiol. by CMR. Cardiol Rev. 2004;12:185–187.
2005;45:1683–1690. 48. Oosterhof T, Mulder BJ, Vliegen HW, et al. Corrected tetralogy of Fallot: Delayed enhance-
26. Vignaux O. Cardiac sarcoidosis: Spectrum of MRI features. Am J Roentgenol. 2005; ment in the right ventricular outflow tract. Radiology. 2005;237:868–871.
184:249–254. 49. Harris MA, Johnson TR, Weinberg PM, et al. Delayed enhancement CMR identifies
27. Patel MR, Cawley PJ, Heitner JF, et al. Detection of myocardial sarcoid damage in patients fibrous tissue in children after surgery for congenital heart disease. J Thorac Cardiovasc
with sarcoidosis. Circulation. 2009;120:1969–1977. Surg. 2007;133:676–681.
28. Smedema JP, Snoep G, van Kroonenburgh MP, et al. The additional value of gadolinium 50. Babu-Narayan SV, Kilner PJ, Li W, et al. Ventricular fibrosis suggested by CMR in adults
enhanced MRI to standard assessment for cardiac involvement in patients with pulmonary with repaired tetralogy of Fallot and its relationship to adverse marker of clinical outcome.
sarcoidosis. Chest. 2005;128:1629–1637. Circulation. 2006;113:405–413.
29. Ichinose A, Otani H, Oikawa M, et al. MRI of cardiac sarcoidosis: Basal and subepicar- 51. Wald RM, Haber I, Wald R, et al. Effects of regional dysfunction and LGE on global right
dial localization of myocardial lesions and their effect on LV function. Am J Roentgenol. ventricular function and exercise capacity in patients with repaired tetralogy of Fallot.
2008;191:862–869. Circulation. 2009;119:1370–1377.
30. Kahan A, Allanore Y. Primary myocardial involvement in systemic sclerosis. Rheumatology. 52. Iles L, Pfluger H, Phrommintikul A, et al. Evaluation of diffuse myocardial fibrosis in heart
2006;45(Suppl 4): iv14–iv17. failure with CMR contrast enhanced T1 mapping. J Am Coll Cardiol. 2008;52:1574–1580.
31. Tzelepis GE, Kelekis NL, Plastiras SC, et al. Pattern and distribution of myocardial 53. Jerosch-Herold M, Sheridan DC, Ksuhner JD, et al. CMR imaging of myocardial contrast
fibrosis in systemic sclerosis: a delayed enhanced MR imaging study. Arthritis Rheum. uptake and blood flow in patients affected with idiopathic or familial dilated cardiomyopa-
2007;56:3827–3836. thy. Am J Physiol. 2008;295:H1234–H1242.
LWBK1209-ch18_p283-291.indd 291 16/05/13 9:35 PM

Chapter
19 Stefano Muzzarelli
Magnetic Resonance Imaging Risk

Assessment in Ischemic Heart Disease
■■ Introduction detection of myocardial ischemia and scar tissue, as well as

a precise quantification of the cardiac function are impor-
■■ Left Ventricular Systolic Function tant for the risk stratification and the treatment guidance of
■■ Right Ventricular Systolic Function patients with CAD.
■■ Myocardial Ischemia Cardiovascular magnetic resonance imaging (CMR) has
emerged as a pivotal technique in the arena of cardiovascu-
■■ Myocardial Tissue Characterization lar imaging as an alternative, complementary, and frequently
Myocardial Necrosis and Scar superior imaging modality, allowing for a comprehensive
Microvascular Obstruction and Myocardial evaluation of all the above-mentioned aspects, without ex-
Hemorrhage posure to iodinated contrast agents or ionizing radiations.
Peri-infarct Zone This chapter will provide an overview about the present role
Myocardial Edema of CMR for the risk stratification of patients with CAD.
■■ Conclusions
Left Ventricular Systolic

Introduction Function
Coronary artery disease (CAD) is a highly prevalent condi- The left ventricular ejection fraction is a quantitative mea-
tion in the industrialized countries. About 7% of adults in surement of the global left ventricular systolic function,
the United States are estimated as being affected by CAD and it is a well-known parameter for risk stratification of
and a substantial increase is foreseen in next decades (1). patients with CAD, since cardiovascular mortality, but
CAD is associated with a high morbidity and mortality, also more specific outcomes such as malignant ventricular
resulting in a major burden for the individual patient, for arrhythmias and sudden cardiac death are tightly related
the health care system, and for the society (1). to it (4–6). Several mechanisms may cause left ventricular
Primary prevention of CAD with management of the car- systolic dysfunction in patients with CAD, including: (1)
diovascular risk factors is the main component of disease Myocardial infarction (scar tissue), (2) myocardial hypo-
control. However, once CAD is present, risk stratification perfusion (myocardial stunning and/or hibernating myocar-
is important for a tailored and efficient treatment (2,3). dium), and (3) negative postinfarction remodeling process.
Determining the prognosis in CAD is complex. In addition Myocardial stunning is a reversible myocardial contractile
to the specific cardiac situation, age, comorbidities, and vari- abnormality caused by single or repetitive brief episodes of
ation in individual disease progression are important factors myocardial ischemia. Hibernating myocardium is caused by
influencing the outcome of patients with CAD. Several as- a persistent hypoperfusion, which is not severe enough to
pects of the cardiac pathophysiology, which are important cause myocardial cell death, but is sufficient to cause con-
for risk stratification, can be assessed with cardiac imaging tractile dysfunction (7). Both forms are reversible after resto-
modalities. The severity of myocardial ischemia is one of the ration of myocardial perfusion, but weeks may ensue before
most important risk predictors and therefore represents a functional recovery occurs. Postinfarction remodeling is a
therapeutic target in CAD. In patients who already suffered process characterized by a progressive maladaptation of the
a myocardial infarction (irreversible end-organ damage), the myocardial tissue after an acute ischemic myocardial injury,
extent and the composition of the myocardial scar, as well as causing fibrosis, apoptosis, and myocyte hypertrophy. This
its functional impact on the cardiac performance, are impor- remodeling process is mainly driven by overstimulation of
tant additional prognostic factors. Therefore, an accurate the neuroendocrine system and contributes to progressive
292
LWBK1209-ch19_p292-304.indd 292 16/05/13 8:15 PM

Chapter 19 ■ Magnetic Resonance Imaging Risk Assessment in Ischemic Heart Disease 293
ventricular wall thinning, dilation, and functional com-

promise even in noninfarcted myocardium (8). In a clinical
perspective, the left ventricular ejection fraction is a “surro-
gate” of these different mechanisms of disease and therefore
represents a “simple” and powerful parameter for risk strat-
ification and patient selection for specific medical therapies
or device implantation, such as internal cardiac defibrillator
(ICD) or cardiac resynchronization therapy (CRT) (4,5). So,
for example, a primary prophylactic ICD implantation after
myocardial infarction is recommended in case of persistently
reduced left ventricular ejection fraction ≤35% in patients
in New York Heart Association (NYHA) class II–III accord-
ing to the most recent guidelines (4,5). Therefore, a precise
quantification of the left ventricular ejection fraction is of
utmost importance, since it is a major determinant of prog-
nosis and clinical decision making.
A detailed description of the technical aspects about
the quantification of the ventricular systolic function by
CMR is extensively reviewed in Chapter 6. The ability to
obtain images in all plane directions, the good temporal
and spatial resolutions, the reliable endocardial border
detection, and the independency of geometrical assump-
tions for the quantification of the ventricular volumes are
main advantages of CMR, allowing for a more robust and Figure 19.1. Delayed gadolinium enhancement short-axis image
less observer-variable quantification of ejection fraction as showing an inferior myocardial infarction (arrowhead) with extension
compared to other imaging modalities. Therefore, CMR is to the basal aspect of the right ventricle (arrow). LV, left ventricle; RV,
right ventricle.
now recognized as the gold standard method for the quan-
tification of the ventricular systolic function, and a precise
quantification of the left ventricular volume and ejection Myocardial Ischemia
fraction should be performed in every patient evaluated for
CAD (9). In the presence of a hemodynamically significant coro-
nary stenosis, the ability of increasing the myocardial
blood flow on stress conditions is reduced, resulting in an
Right Ventricular Systolic imbalance between the demand and the delivery of oxygen
Function and nutrients to the myocardial tissue, causing metabolic
abnormalities of the cardiomyocytes with a consequent
Even though a myocardial infarction mainly involves the impairment of the contractile function. The assessment
left ventricle, the right ventricle too may be affected by an of myocardial ischemia by CMR is performed with two
ischemic injury, especially in case of occlusion of the right main techniques: (1) Myocardial perfusion imaging and (2)
coronary artery, which usually supplies the majority of the stress-induced myocardial wall motion imaging, which are
right ventricular myocardium. In the acute setting, an exten- extensively reviewed in Chapters 15 and 16, respectively.
sive ischemic injury of the right ventricle is associated with a Myocardial perfusion imaging evaluates the “first step” of
poor prognosis and needs specific treatment (10). However, the myocardial ischemic cascade, whereas stress-induced
thanks to the lower oxygen consumption of the right ven- wall motion imaging evaluates the functional consequences
tricular myocardium and the direct oxygen diffusion from of an insufficient perfusion on myocardial contractile per-
the ventricular cavity, the likelihood of functional recovery formance. Considering that a physical stress test is not eas-
after an acute ischemic injury is higher as compared to the ily performable in a CMR scanner environment, the stress
left ventricle (11). Considering the complex geometry of the test is typically performed with positive chronotropic drugs
right ventricle, CMR offers clear advantages for the quantifi- (dobutamine) for myocardial wall motion imaging, and
cation of the right ventricular volumes and ejection fraction with vasodilator drugs (dipyridamole, adenosine, or regad-
as compared with other imaging techniques. This is particu- enoson) in case of myocardial perfusion testing. Both tech-
larly true for two-dimensional echocardiography. niques are supported by a large body of evidence, showing
Multiple CMR studies provided the evidence that persis- a high diagnostic accuracy for the detection of hemody-
tent right ventricular dysfunction and/or necrosis of the right namically significant CAD in a pooled meta-analysis of
ventricular myocardium after acute myocardial infarction multiple studies evaluating the diagnostic performance of
are independent predictors of mortality and cardiovascular stress CMR (sensitivity and specificity of 83% to 86% for
events, underscoring the importance of the assessment of the wall motion imaging and 91% to 81% for myocardial per-
right ventricle for the global risk stratification of patients fusion imaging, respectively) (14). Here too, myocardial
suffering an acute myocardial infarction (12,13). Figure perfusion imaging was shown to provide a better diagnos-
19.1 shows a typical example of inferior myocardial infarc- tic accuracy as compared to myocardial perfusion SPECT
tion, involving the basal aspect of the right ventricle. in multiple large comparative multicenter studies, underscoring
LWBK1209-ch19_p292-304.indd 293 16/05/13 8:15 PM

1.0 1.00
RevPD and LGE
both absent
P = 0.01
0.8
0.75
Event-free survival (%)
RevPD present
Event-free survival
Normal MR perfusion LGE absent
0.6 Abnormal MR perfusion P = 0.04
0.50
0.4 RevPD present

LGE present
0.25
0.2
D
P < 0.001
0.00
0.0 0 1 2 3 4 5
0 1 2 3 4 5
Follow-up time (years)
Time (years)
#at risk Figure 19.3. Kaplan–Meier curve showing the impact of myocar-
Normal MR perfusion 302 272 205 84 16 0
Abnormal MR perfusion 159 123 73 33 7 0
dial ischemia and scar on the prognosis (major adverse cardiovascular
events) of patients without prior history of CAD referred for CMR
Figure 19.2. Kaplan–Meier curve showing the excellent prognosis evaluation. Note the complementary prognostic value of myocardial
of patients without evidence of myocardial ischemia. Note that the
scar and ischemia testing. (Reprinted with permission from Steel K,
3-year event-free survival was of 99.2% in ischemia-negative patients
Broderick R, Gandla V, et al. Complementary prognostic values of
referred for CMR testing. (Reprinted with permission from Jahnke C,
stress myocardial perfusion and late gadolinium enhancement imaging
Nagel E, Gebker R, et al. Prognostic value of cardiac magnetic reso-
by cardiac magnetic resonance in patients with known or suspected
nance stress tests: Adenosine stress perfusion and dobutamine stress
coronary artery disease. Circulation. 2009;120:1390–1400).
wall motion imaging. Circulation. 2007;115:1769–1776).
the robustness and the validity of this technique (15,16). involving the entire anterior and septal left ventricular
Furthermore, the diagnostic performance and the prog- myocardium, related to a severe stenosis in the proximal
nostic value of the different CMR stress test modalities left anterior descending coronary artery. Figure 19.5 dem-
were proven in several clinical scenarios and in several onstrates a small ischemia located in the inferior aspect
patient populations, including patients suspected of CAD of the ventricular apex, related to a stenosis of the distal
(as a test to role out CAD), patients with known CAD, left anterior descending coronary artery with a transapi-
patients with prior myocardial revascularization (including cal course. These two examples illustrate that CMR allows
coronary artery bypass grafting (17,18) and percutaneous for a quantitative assessment of the ischemic myocardium.
coronary angioplasty (16,17)), and patients presenting at This is a central aspect, because the severity and the extent
the emergency department with acute chest pain and nega- of myocardial ischemia are powerful outcome predictor in
tive cardiac biomarkers (19–21). Consistently, patients patients with CAD, and there is a rising evidence that the
with negative myocardial stress test (without evidence for prognostic benefit of revascularization procedures among
myocardial ischemia) had an excellent cardiac prognosis patients with stable CAD may depend on the extent of
with a low cardiovascular event rate during the follow-up ischemic myocardium (33–35). In fact, in the current era
(20–29) (Fig. 19.2). Conversely, patients with evidence of where potent antianginal, lipid-lowering drugs, and anti-
myocardial ischemia have a worse cardiac prognosis. This platelet therapies are available for secondary prevention of
also seems true for patients with prior myocardial infarc- CAD, the results of multiple studies suggested that a treat-
tion, where the presence and the extent of myocardial isch- ment with myocardial revascularization is not superior to
emia provide additional prognostic information (30–32) an optimal pharmacologic therapy in patients with mild-to-
(Fig. 19.3). Notably, patients with prior myocardial infarc- moderate stable CAD (36,37). However, large retrospective
tion are exposed at a higher risk of malignant ventricu- data and post hoc analysis of randomized trials suggested
lar arrhythmias, because the myocardial scar represents that myocardial revascularization had greater survival ben-
the morphologic substrate for ventricular tachycardia. On efit as compared with medical therapy alone in patients
top of that, ischemia of the peri-infarct zone may further with a severe inducible ischemia, suggesting that the ex-
destabilize the electrical membrane stability of the cardio- tent of myocardial ischemia as assessed with a myocardial
myocytes and therefore acts as an important trigger of such stress test may be the key component to identify low-risk
malignant arrhythmias. patients, who can safely be treated medically, from high-
Importantly, myocardial ischemia testing with an imag- risk patients, who may benefit from coronary angiography
ing modality enables identifying the localization and the and revascularization procedure (33–35). The concept of
extent of the ischemic myocardium. Indeed, all segments ischemia-guided revascularization was also tested in a large
of the left ventricle are represented in both CMR perfu- multicenter trial, where patients undergoing percutaneous
sion and stress-induced myocardial wall motion imaging. coronary intervention for multivessel CAD were randomly
Figure 19.4 shows an example of an extensive ischemia allocated to angiography-guided revascularization (where
LWBK1209-ch19_p292-304.indd 294 16/05/13 8:15 PM

A B A B
C D C D
E E
Figure 19.4. Adenosine stress first-pass perfusion imaging acquired Figure 19.5. Adenosine stress first-pass perfusion imaging acquired
in four short-axis views (A, basal slice; B,C, midventricular slices; and in four short-axis views (A, basal slice; B,C, midventricular slices; and
D, apical slice), showing an extensive perfusion defect in the anterior D, apical slice), showing a small perfusion defect in the inferior aspect
and septal left ventricular segments (arrows). The coronary angi- of the left ventricular apex (one ventricular segment, arrow). The
ography (E) showed serial significant stenoses (black arrows) in the coronary angiography (E) showed a significant stenosis (black arrow)
proximal left anterior descending artery (LAD) in a patient with prior in the very distal segment of the left anterior descending artery (LAD).
coronary bypass graft surgery (the left internal mammary artery to the Note that the localization of the ischemia is explained by the transapi-
distal LAD was occluded [image not shown]). LV, left ventricle; RV, cal course of the LAD that perfuses the inferoapical segment. LV, left
right ventricle. ventricle; RV, right ventricle.
coronary arteries with a stenosis >50% were treated), or recent guidelines recommend to perform a functional stress
ischemia-guided revascularization, where the hemody- test before a coronary invasive procedure in patients with
namic significance of the coronary stenosis was evaluated stabile CAD, preferably using noninvasive testing before
invasively by measuring the fractional flow reserve. In line invasive angiography (40).
with the above-mentioned results, this study demonstrated Notably, the presence and the extent of myocardial isch-
that “functional,” ischemia-guided therapy was associ- emia adds incremental prognostic value as compared to the
ated with a better cardiac prognosis and cost-effectiveness clinical profile, the left ventricular systolic function at rest,
as compared with a “morphologic,” coronary angiogra- and the presence of myocardial scar (30,32). Therefore,
phy-guided therapy (38, 39). All together, the functional CMR stress testing represents a valuable diagnostic tool
evaluation of the coronary arteries by the different avail- for the initial evaluation and the surveillance of patients
able stress tests is gaining a wide acceptance, and the most with known or suspected CAD, as it allows for a reliable
LWBK1209-ch19_p292-304.indd 295 16/05/13 8:15 PM

Table 19.1 Prognostic Value of CMR Ischemia Testing
Number of Patients
(Reference) Stress Test Modality Setting Main Outcome
405 (24) Stress perfusion CMR Patients with suspected or Annual MACE rate in ischemia-negative patients
known stable CAD was of 0.3% in women and 1.1% in men
1,532 (25) Dobutamine stress wall Patients with suspected or The cardiac event rate was of 0.96% during a
motion CMR known stable CAD median follow-up of 2.1 years in ischemia-
negative patients
1,463 (27) Dobutamine stress wall Patients with suspected or The cardiac event rate was of 4% during a median
motion CMR known stable CAD follow-up of 6 years in ischemia-negative patients
1,493 (28) Dobutamine stress wall Patients with suspected or The cardiac event rate was of <0.5% in patients
motion and perfusion known stable CAD with low to intermediate pre-test probability
CMR of CAD over a median follow-up of 2 years in
ischemia-negative patients
601 (22) Dipyridamole stress wall Patients with suspected or The MACE rate was of 4% over a median
motion and perfusion known stable CAD follow-up of 553 days in ischemia-negative
CMR patients
103 (21) Adenosine stress perfusion Patients with low-risk acute The cardiac event rate was of 0% over a median
chest pain follow-up of 277 days in ischemia-negative
patients
218 (29) Adenosine stress perfusion Patients with suspected stable The annual MACE rate was of 0.9% in ischemia-
CAD negative patients
420 (23) Dipyridamole stress wall Patients with suspected or The MACE rate was of 4% over a median
motion and perfusion known stable CAD follow-up of 420 days in ischemia-negative
CMR patients
513 (26) Adenosine stress perfusion Patients with suspected or The 3-year cardiac event-free survival was of 99.2%
and dobutamine stress known stable CAD in ischemia-negative patients
wall motion CMR
135 (20) Adenosine stress perfusion Patients with-low risk acute The MACE rate was of 0% over a median
chest pain follow-up of 467 days in ischemia-negative
patients
Total number of patients = 6,883.

MACE, major adverse cardiovascular events; CAD, coronary artery disease.
etection/role out of significant CAD, but also for monitor-

d (scar) or fibrosis, myocardial cells are replaced by extracel-
ing of disease progression and risk stratification. Tables 19.1 lular collagen matrix. Therefore, the distribution volume of
and 19.2 provide an overview of the literature about the gadolinium is increased in both acute and chronic myocar-
prognostic value of CMR ischemia testing. dial infarctions (44,45); (4) the kinetics of contrast enhance-
ment is characterized by a delayed wash-out of gadolinium
Myocardial Tissue in infarcted as compared to normal myocardium (46). These
principles explain why areas of fibrosis or necrosis appear
Characterization
as a bright signal as compared to normal myocardium on
delayed T1-weighted CMR sequences after administration of
Myocardial Necrosis and Scar
gadolinium (47). Inversion recovery sequences with inver-
CMR has evolved as an excellent diagnostic tool for the sion time set to null the signal of normal myocardium are
characterization of the myocardial tissue. Acute myocardial used to maximize the contrast between normal myocardium
necrosis, chronic scar tissue, and myocardial fibrosis can and areas of enhancement.
be visualized after application of gadolinium chelates with This imaging technique was extensively validated in ani-
excellent contrast characteristics and a millimetric spatial mal model and in humans, showing that the detection and
resolution (41,42). Technical details about contrast-delayed the quantification of myocardial necrosis and scar tissue
enhancement imaging are comprehensively reviewed in can be performed very precisely (48,49). Notably, thanks
Chapter 17. In short, the gadolinium hyperenhancement of to the high spatial resolution of delayed gadolinium en-
myocardial necrosis and scar tissue is based on the following hancement imaging, not only the global extent (mass) of
principles: (1) Gadolinium chelates are extracellular contrast the myocardial necrosis/scar tissue, but also its transmural-
agents, which are not able to cross an intact cellular mem- ity can be established (Fig. 19.6). Considering that myocar-
brane; (2) both, the T1 and T2 relaxation times are shortened dial necrosis represents an irreversible ischemic injury, this
by gadolinium (43); (3) in acute myocardial necrosis, the imaging technique has acquired a wide acceptance in clini-
myocardial cellular membranes are disrupted and permeable cal practice, because it allows establishing the myocardial
to gadolinium. In areas of chronic myocardial infarction viability in acute and chronic myocardial infarcts. Multiple
LWBK1209-ch19_p292-304.indd 296 16/05/13 8:15 PM

Table 19.2 Prognostic Value of Multiparametric CMR Imaging
Number of Patients
(Reference) Imaging Modality Setting Main Outcome
908 (30) Adenosine stress perfusion Patients with suspected or CMR analysis of myocardial perfusion and
and delayed gadolinium known stable CAD scar add incremental prognostic value at
enhancement imaging predicting adverse cardiac events. The
cardiac mortality rate was of 0.4% in
patients without ischemia and scar (median
follow-up = 2.6 years)
1,232 (31) Adenosine stress perfusion Patients with suspected or Myocardial ischemia and scar are predictors of
and delayed gadolinium known stable CAD MACE (median follow-up = 35 months)
enhancement imaging
254 (32) Adenosine or dipyridamole Patients with suspected or Patients without ischemia and scar had a
stress perfusion and delayed known stable CAD 98.1% negative annual event rate for death
gadolinium enhancement or myocardial infarction (median follow-up
imaging = 17 months)

studies demonstrated that the left ventricular functional r ecovery after revascularization procedures is important
recovery after reperfused acute myocardial infarction was in both clinical scenarios—acute reperfused myocardial
inversely related to the transmurality of the myocardial in- infarction and chronic ischemic heart disease. In the for-
farction (50,51). This is also true in patients with chronic mer, patients with extensive irreversible myocardial injury
heart failure and left ventricular systolic dysfunction due are more likely to develop a negative remodeling of left
to CAD. In dysfunctional myocardial segments without ventricle leading to a persistent severe cardiac dysfunc-
evidence of scar, the rate of functional recovery was 75% tion, which is associated with a worse cardiac prognosis.
to 80% after coronary revascularization. In contrast, the Therefore such patients require particular attention during
chance of recovery decreased to 60%, 40%, less than 20%, the follow-up after the acute event, in order to optimize
and finally 1% to 3% as transmural extent of scar tissue as the pharmacologic and the cardiac device treatment. In
assessed with delayed gadolinium enhancement increased patients with chronic ischemic heart disease and left ven-
in quartiles from 1% to 25%, 26% to 50%, 51% to 75%, tricular systolic dysfunction, myocardial viability testing is
and 76% to 100%, respectively (52,53). In addition, the recommended to plan the revascularization procedure and
absolute thickness of the viable myocardium is important identify coronary territories with viable myocardium that
for predicting functional recovery after revascularization, represents the functional target of revascularization (40).
since a minimal amount of 4 to 5 mm of functional myo- Moreover, multiple studies demonstrated that the presence
cardium is required to produce enough power to overcome and the extent of myocardial scar tissue as assessed with
intramyocardial wall tension, which in turn is related CMR-delayed gadolinium enhancement is a strong predic-
to the ventricular radial dimension and to the myocar- tor of mortality and cardiovascular events among patients
dial wall thickness (54,55). The prediction of functional with CAD (30–32,56–61) (Table 19.3).
Figure 19.6. A: Delayed gadolin-

ium enhancement image (four-chamber
view) showing a sub-endocardial scar
in the anterolateral wall (arrowhead).
Note the presence of a wide subepi-
cardial rim of viable myocardium.
B: Delayed gadolinium enhancement
image (three-chamber view) showing
a transmural scar of the anteroseptal
wall, without residual myocardial via-
bility (arrowhead). Note the presence
of an extensive intraventricular throm-
bus, which is adherent to the infarcted
segments (asterisk). LV, left ventricle;
RV, right ventricle; LA, left atrium; RA,
A B
right atrium; Ao, Aorta.
LWBK1209-ch19_p292-304.indd 297 16/05/13 8:15 PM

Table 19.3 Diagnostic Value of Myocardial Scar Imaging
Number of Patients
100 (56) Infarct size measurement Known CAD The extent of myocardial scar is an independent
by delayed gadolinium predictor of death in patients with CAD
enhancement imaging (median follow-up = 4.8 years)
857 (57) Presence and extent of Consecutive patients referred The presence and extent of scar predicts
myocardial scar by delayed for delayed gadolinium all-cause mortality and need of cardiac
gadolinium enhancement enhancement imaging with transplantation (median follow-up = 4.4
imaging known or unknown CAD years)
177 (58) Infarct size measurement Known CAD postmyocardial Myocardial infarct size is a predictor of
by delayed gadolinium infarction mortality (median follow-up = 20 months)
enhancement imaging
349 (59) Infarct size measurement Patients with CAD and severe The extent of myocardial scar is associated
by delayed gadolinium reduced left ventricular with increased mortality and need for cardiac
enhancement imaging ejection fraction transplantation in patients with severe
ischemic cardiomyopathy (median follow-up
= 2.6 years)
195 (60) Presence of myocardial scar Patients with clinical The presence of prior unrecognized myocardial
by delayed gadolinium suspicion of CAD without scar is a predictor of MACE (median
enhancement imaging prior history of myocardial follow-up = 16 months)
infarction
86 (61) Infarct size measurement Patients with CAD and severe The extent of the myocardial scar is an
by delayed gadolinium reduced left ventricular independent predictor of cardiovascular
enhancement imaging ejection fraction events (median follow-up = 20 months)

Microvascular Obstruction and association between microvascular obstruction and absence

Myocardial Hemorrhage of myocardial viability. Multiple CMR imaging techniques
are available to evaluate the presence and the extent of
The microvascular obstruction is defined as a failure of microvascular obstruction including, first-pass perfusion
tissue perfusion in the core-zone of the acute myocardial imaging, early and delayed gadolinium enhancement imag-
infarct. This phenomenon of “no-reflow” is a consequence ing. First-pass perfusion imaging typically shows a hypoper-
of capillaries compression due to extensive myocardial fusion in the core-zone of the acute myocardial infarct in
edema, capillary occlusion with microthrombotic mate- case of microvascular obstruction (63). Early and delayed
rial, or capillary necrosis (62). Therefore, microvascular gadolinium enhancement both show a dark, nonenhanced,
obstruction may be observed even in spite of a successful core-zone surrounded by hyperenhanced myocardium (Fig.
recanalization of the culprit coronary vessel. This patho- 19.7). Notably, microvascular obstruction, as assessed with
physiologic background is important for understanding the delayed gadolinium enhancement, has been shown to be a
A B
Figure 19.7. A: Delayed gadolinium enhancement image (two-chamber view) showing an extensive trans-
mural acute anterior myocardial infarction (arrows) with evidence of microvascular obstruction (asterisk),
resulting as a dark, nonenhanced core-zone surrounded by hyperenhanced myocardium. B: T2-weighted image
in the same slice location depicting a hypoenhanced zone within the core of the myocardial infarction (aster-
isk), consistent with an intramyocardial hemorrhage. LV, left ventricle; LA, left atrium.
LWBK1209-ch19_p292-304.indd 298 16/05/13 8:15 PM

1.00 1.00
0.8 0.8

0.6 0.6
P < 0.01 P = 0.01

0.4 0.4
0.2 0.2
No late MO No late MO
0.0 Late MO 0.0 Late MO
0 10 20 30 40 0 10 20 30 40
Time to event (month) Time to event (month)
No. at rick No. at rick
No late MO 126 104 73 31 No late MO 126 107 74 31
A Late MO 296 215 119 50 Late MO 296 241 131 53 B
Figure 19.8. Kaplan–Meier curves for the cumulative incidence of major adverse cardiac events (A) and
death (B) in patients with or without microvascular obstruction. (Reprinted with permission from de Waha S,
Desch S, Eitel I, et al. Impact of early vs. late microvascular obstruction assessed by magnetic resonance imag-
ing on long-term outcome after ST-elevation myocardial infarction: A comparison with traditional prognostic
markers. Eur Heart J. 2010;31:2660–2668).
strong predictor of adverse left ventricular remodeling and re-entrant tachycardia after myocardial infarction, and is
cardiac prognosis after acute myocardial infarction, that characterized by a heterogeneous and jeopardized tissue
performs better as compared with early enhancement and containing both myocardial fibrosis and viable myocardial
first-pass perfusion imaging (63–68) (Fig. 19.8). cells (75,76). Sudden cardiac death related to ventricular
Restoration of the coronary and myocardial blood flow tachycardia or ventricular fibrillation is a frequent cause of
within the infarct-related coronary artery is the main ther- mortality among patients with ischemic heart disease. As
apeutic target in acute myocardial infarction. However, if discussed above, left ventricular ejection fraction is one of
the restoration of the myocardial blood flow does not occur the most powerful risk factors for malignant arrhythmias,
early (<12 hours), the reperfusion may cause an additional and an ICD is recommended in primary prophylaxis in case
damage of the vascular endothelium and the myocardium. of persistent, severely depressed left ventricular ejection frac-
This phenomenon is known as reperfusion injury (62). The tion after myocardial infarction (4,5). However, current risk
reperfusion injury is often associated with intramyocardial assessment after myocardial infarction remains suboptimal,
hemorrhage, which is a consequence of the functional and and the need for more accurate prognostication of malig-
structural loss of integrity of the coronary microcircula- nant arrhythmias and sudden cardiac death is evident. Bello
tion, resulting in extravasation of erythrocytes (69,70). et al. (77) first tested the concept of myocardial scar quan-
Intramyocardial hemorrhage can be visualized by T2- tification by delayed gadolinium enhancement to identify
weighted imaging that typically shows a hypointense zone patients at highest risk of ventricular arrhythmias after myo-
within the infarcted myocardium due to the paramagnetic cardial infarction, showing that the amount of myocardial
properties of the break-down products of the extravasated scar predicted the inducibility of ventricular tachycardia on
hemoglobin (Fig. 19.7). Multiple studies demonstrated that electrophysiologic study. Considering that myocardial scar,
acute myocardial infarction with evidence of intramyocar- and more precisely the peri-infarct zone, represents the mor-
dial hemorrhage as assessed with T2-weighted CMR imag- phologic substrate of re-entrant ventricular tachycardia, this
ing is associated with adverse left ventricular recovery and concept was further developed by testing the impact of the
clinical outcomes (71–74). An overview about the prognos- peri-infarct zone extent on clinical and electrophysiologic
tic value of microvascular obstruction and intramyocardial outcomes after myocardial infarction. Yan et al. (78) dem-
hemorrhage in patients with acute myocardial infarction is onstrated that the extent of the peri-infarct zone provided
provided in Table 19.4. incremental prognostic value beyond the left ventricular ejec-
tion fraction for predicting mortality in a cohort of patients
with prior myocardial infarction. The pathophysiologic
Peri-Infarct Zone
background of this finding was then provided by Schmidt
Areas of chronic myocardial infarction are composed by et al. (79), showing that the extent of the peri-infarct zone
a core-zone of scar tissue and a peri-infarct zone. The lat- was associated with the inducibility of ventricular tachycar-
ter represents the morphologic substrate of ventricular dia in patients with ischemic cardiomyopathy and severe
LWBK1209-ch19_p292-304.indd 299 16/05/13 8:15 PM

Table 19.4 Prognostic Value of Microvascular Obstruction and Intramyocardial Hemorrhage
Number of Patients
438 (65) Delayed gadolinium Acute reperfused ST-elevation Microvascular obstruction was an independent
enhancement imaging myocardial infarction predictor of MACE (median follow-up
= 19 months)
184 (64) Delayed gadolinium Acute reperfused myocardial Microvascular obstruction was an independent
enhancement imaging infarction predictor of MACE (median follow-up
= 1 year)
110 (66) Delayed gadolinium Acute reperfused myocardial Microvascular obstruction was an independent
enhancement imaging infarction predictor of MACE (median follow-up
= 1 year)
44 (68) Early gadolinium enhancement Acute reperfused myocardial Microvascular obstruction predicted
imaging infarction cardiovascular complication (median
follow-up = 16 months)
346 (72) T2-weighted CMR Acute reperfused myocardial The presence of a hypointense core zone within
infarction the myocardial infarct was associated with
increased cardiac events (median follow-up
= 6 months)

MACE, major adverse cardiovascular events.
reduced left ventricular systolic function. Finally, Roes et al. To date, two different methods for the quantification
(80) identified that the extent of the peri-infarct zone was of the peri-infarct zone are reported in the literature, both
the strongest predictor of spontaneous ventricular arrhyth- of which are based on the quantification of the signal in-
mia and adequate delivery of ICD therapy among patients tensity on delayed gadolinium images (Fig. 19.10) (78,79).
getting an ICD implanted after myocardial infarction (Fig. Although both methods are based on signal intensity thresholds
19.9). These findings suggest that the quantification of the
peri-infarct zone provides further risk stratification even
in patients with highest risk of sudden cardiac death, who 1.0
p - value 0.003
qualify for primary prophylactic ICD therapy according to
the most recent guidelines. Therefore, the quantification of
myocardial scar tissue heterogeneity may be useful for fur- 0.8
ther risk stratification and more targeted ICD implantation. Infarct gray zone > 16.7 g
Cumulative event rate
However, these concepts need to be tested in specifically

designed trials.
Technically, the quantitative evaluation of the peri-infarct 0.6
zone by CMR is based on delayed gadolinium enhancement
images. The core-zone of a myocardial infarct is character-
ized by the highest distribution volume of extracellular gad- 0.4
olinium chelates because myocardial cells are replaced by
dense bundles of extracellular collagen. Therefore, the core-
zone shows the highest signal intensity on inversion-recov- 0.2
ery T1-weighted imaging sequences after administration of Infarct gray zone 16.7 g
gadolinium. Conversely, the peri-infarct zone has an inter-
mediate signal intensity due to an intermediate distribution
0.0
volume and wash-out kinetic of gadolinium as compared to
the core-zone and the remote (noninfarcted) myocardium 0 10 20 30 40 50
(81,82). It may be argued that the intermediate signal inten- Follow-up (months)
sity of the peri-infarct zone is solely due to a partial volume
effect creating voxels with intermediate signal intensity on Figure 19.9. Kaplan–Meier curve analysis showing the difference
in appropriate ICD therapy when patients are stratified according the
the boundaries of the infarct. However, the existence of a
median value of the peri-infarct zone (16.7 g). (Reprinted with permis-
peri-infarct zone is supported by histopathologic data, and sion from Roes SD, Borleff CJ, van der Geest RJ, et al. Infarct tissue
prior investigations demonstrated that a peri-infarct zone heterogeneity assessed with contrast-enhanced MRI predicts spontane-
with intermediate distribution volume of gadolinium was ous ventricular arrhythmia in patients with ischemic cardiomyopathy
reproducible even with high-resolution imaging techniques and implantable cardioverter-defibrillator. Circ Cardiovasc Imaging.
(81,83,84). 2009;2:183–190.
LWBK1209-ch19_p292-304.indd 300 16/05/13 8:15 PM

A B C
Figure 19.10. Delayed gadolinium enhancement images of a patient with two prior myocardial infarctions
in the anteroseptal and the inferior wall (A). (B,C) show the quantification of the peri-infarct zone. First, the
signal intensity profile in a remote zone of healthy myocardium is defined. The core-zone volume (yellow zone,
C) and total infarct volume (yellow zone, B) are usually determined by applying a threshold of >3 standard
deviations and >2 standard deviations in signal intensity above the remote (noninfracted) myocardial region,
respectively. The peri-infarct zone volume is then calculated by subtracting core-zone volume from the total
infarct volume. A second method for measuring the peri-infarct zone (images not shown) starts with placing a
region of interest within the zone having the highest signal intensity (core-zone), and determining the maximal
signal intensity within this region. The infarct core-zone is then defined as myocardium with a signal intensity
>50% of the maximal measured signal intensity. A second region of interest is then placed into a remote zone
to define the signal intensity profile of noninfarcted myocardium. The peri-infarct zone is then defined as a
zone with a signal intensity which is higher than the peak signal intensity of the remote zone and <50% to the
maximal signal intensity of the scar region of interest. LV, left ventricle; RV, right ventricle.
for defining the extent of the peri-infarct zone, it is unknown Myocardial Edema
whether results of these different methods are equivalent
Edema of the myocardial tissue is a relevant component of
and interchangeable. In addition, it is unclear whether the
an acute ischemic injury. In case of prolonged and severe
dosage of gadolinium, the timing of the image acquisition,
ischemia causing an irreversible myocardial necrosis, the area
or the utilization of different read-out sequences for delayed
of myocardial edema may extend the area of myocardial
gadolinium imaging may alter the aspect and size of the peri-
necrosis. Following the principle of the “wavefront phenom-
infarct zone. Therefore, efforts should be made to standard-
enon” of myocardial ischemia, areas of myocardial necrosis
ize the diagnostic procedures defining the peri-infarct zone
are subendocardial and show different degrees of transmu-
and to provide robust data about its prognostic value on
rality depending on the severity and duration of ischemia,
larger collectives of patients. This may be of particular inter-
whereas edema typically shows a transmural extension in the
est because the extension of the peri-infarct zone has a great
infarct-related coronary vascular bed.
potential for predicting mortality and major arrhythmic
Therefore, in acute myocardial infarction, the myocar-
events after myocardial infarction (Table 19.5).
dial edema may extend over the area of irreversible injury
Table 19.5 Prognostic Value of the Peri-infarct Zone
Number of Patients
91 (80) Quantification of the peri- Patient with prior myocardial The extent of the peri-infarct zone was
infarct zone by delayed infarction scheduled for ICD the strongest predictor of adequate
gadolinium enhancement implantation delivery of ICD therapy (median
imaging follow-up = 8.5 months)
144 (78) Quantification of the peri- Patient with prior myocardial The extent of the peri-infarct zone
infarct zone by delayed infarction was an independent predictor of
gadolinium enhancement cardiovascular mortality (median
imaging follow-up = 2.4 years)
Total number of patients = 235.
LWBK1209-ch19_p292-304.indd 301 16/05/13 8:15 PM

A B
Figure 19.11. T2-weighted image (A) and delayed gadolinium enhancement (B) short-axis images,
showing an example of acute, revascularized anteroseptal myocardial infarction. In the T2-weighted
image (A), the myocardial edema (hyperintense area at risk within the two white lines) is bigger (in terms
of both lateral and transmural extensions) as compared with the irreversible ischemic injury (necrosis
on delayed gadolinium enhancement image B, arrow). The salvaged myocardium can be calculated by
subtracting the myocardial volume with hyperintense signal on T2-weighted images and the volume of
myocardial necrosis as assessed with delayed gadolinium enhancement images. LV, left ventricle; RV,
right ventricle.
in case of prompt revascularization. Hence, T2-weighted cardiac imaging. This concept was tested by Cury et al.
imaging may identify the area of myocardial salvage in re- (19), showing that CMR is feasible and safe in patients
perfused acute myocardial infarction, assessed as the dif- with stable acute coronary syndrome, and that integration
ference between T2-hyperintense (edema) and the delayed on T2-weighted imaging increases the diagnostic accuracy
enhancement positive zone (necrosis, irreversible injury) as compared with a “routine” CMR protocol and a clinical
(Fig. 19.11). This concept was first investigated by Aletras evaluation. Therefore, T2-weighted CMR imaging may be-
et al. (85) in an animal model of reperfused myocardial come relevant for the clinical management and risk stratifi-
infarcts, showing that the area at risk was comparable cation of patients presenting at the emergency department
to the size of the hyperintense zone on T2-weighted im- with chest pain of unclear origin.
ages, whereas the size of irreversible necrotic damage as Moreover, this imaging technique allows assessing the ef-
assessed with delayed gadolinium imaging was smaller in ficacy of reperfusion therapies by measuring the amount of
size. Here too, functional recovery was observed in edema- salvaged myocardium, suggesting great potentials to evalu-
tous myocardium without evidence of necrosis. These find- ate the impact of different revascularization techniques, such
ings were confirmed by multiple clinical studies, showing as thrombus aspiration devices, embolic protection devices,
that areas of reversible injury with T2-hyperintensity ex- and pharmacologic therapy aimed to reduce the ischemic
tended areas of irreversible ischemic damage as assessed and reperfusion myocardial damage in clinical trials.
with delayed gadolinium imaging in case of successful and Therefore, the quantification of the salvaged myocardium
prompt revascularization of acute myocardial infarction, by CMR has potential implications on patient prognosis and
meaning that CMR allows for quantitative assessment of for the design of future trials intended to test the efficacy of
salvaged myocardium, which is inversely proportional to new therapies in acute myocardial infarction.
the time delay between onset of ischemic symptoms and
reperfusion (86–91). Very recent studies demonstrated that
the extent of salvaged myocardium as assessed with CMR Conclusions
predicts clinical outcomes in patients with ST-segment
elevation myocardial infarction (STEMI) (87), as well as CMR has become an important tool for diagnosis, risk strat-
in non-ST–segment elevation acute coronary syndrome ification, and therapeutic guidance of patients with CAD.
(NSTEMI) (92). Furthermore, CMR with T2-weighted im- The evaluation of ventricular function, myocardial isch-
aging may improve early detection of acute coronary syn- emia, and tissue characterization by CMR are well-estab-
drome in patients presenting with chest pain. As discussed lished techniques that are supported by a robust and ex-
above, myocardial edema is an early component follow- tensive literature. These different pathophysiologic aspects
ing an acute myocardial ischemic injury, meaning that also provide complimentary rather than overlapping informa-
patients with only a short episode of myocardial ischemia tion, underscoring the importance of a comprehensive
not causing necrosis or rise in cardiac enzymes, or patients assessment, including the ventricular systolic function,
who present after spontaneous recanalization of the cul- ischemia, and myocardial viability in all patients evaluated
prit vessel, may have pathologic findings on T2-weighted for CAD.
LWBK1209-ch19_p292-304.indd 302 16/05/13 8:15 PM

References 25. Gebker R, Jahnke C, Manka R, et al. The role of dobutamine stress cardiovascular mag-
netic resonance in the clinical management of patients with suspected and known coronary
artery disease. J Cardiovasc Magn Reson. 2011;13:46.
1. Roger VL, Go AS, Lloyd-Jones DM, et al. Executive summary: Heart disease and stroke 26. Jahnke C, Nagel E, Gebker R, et al. Prognostic value of cardiac magnetic resonance stress
statistics—2012 update: A report from the American Heart Association. Circulation. tests: Adenosine stress perfusion and dobutamine stress wall motion imaging. Circulation.
2012;125:188–197. 2007;115:1769–1776.
2. Fox K, Garcia MA, Ardissino D, et al. Guidelines on the management of stable angina 27. Kelle S, Chiribiri A, Vierecke J, et al. Long-term prognostic value of dobutamine stress
pectoris: Executive summary: The task force on the management of stable angina pectoris CMR. JACC Cardiovasc Imaging. 2011;4:161–172.
of the European Society of Cardiology. Eur Heart J. 2006;27:1341–1381. 28. Korosoglou G, Elhmidi Y, Steen H, et al. Prognostic value of high-dose dobutamine stress
3. Graham I, Atar D, Borch-Johnsen K, et al. European guidelines on cardiovascular disease magnetic resonance imaging in 1,493 consecutive patients: Assessment of myocardial wall
prevention in clinical practice: Executive summary: Fourth joint task force of the European motion and perfusion. J Am Coll Cardiol. 2010;56:1225–1234.
Society of Cardiology and other societies on cardiovascular disease prevention in clinical 29. Pilz G, Jeske A, Klos M, et al. Prognostic value of normal adenosine-stress cardiac magnetic
practice (Constituted by representatives of nine societies and by invited experts). Eur Heart resonance imaging. Am J Cardiol. 2008;101:1408–1412.
J. 2007;28:2375–2414. 30. Bingham SE, Hachamovitch R. Incremental prognostic significance of combined cardiac
4. Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC Guidelines for the diagnosis and magnetic resonance imaging, adenosine stress perfusion, delayed enhancement, and left
treatment of acute and chronic heart failure 2008: The task force for the diagnosis and ventricular function over preimaging information for the prediction of adverse events.
treatment of acute and chronic heart failure 2008 of the European Society of Cardiology. Circulation. 2011;123:1509–1518.
Developed in collaboration with the Heart Failure Association of the ESC (HFA) and 31. Krittayaphong R, Chaithiraphan V, Maneesai A, et al. Prognostic value of combined mag-
endorsed by the European Society of Intensive Care Medicine (ESICM). Eur Heart J. netic resonance myocardial perfusion imaging and late gadolinium enhancement. Int J
2008;29:2388–2442. Cardiovasc Imaging. 2011;27:705–714.
5. Hunt SA, Abraham WT, Chin MH, et al. 2009 Focused update incorporated into the ACC/ 32. Steel K, Broderick R, Gandla V, et al. Complementary prognostic values of stress myocardial
AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults A perfusion and late gadolinium enhancement imaging by cardiac magnetic resonance in patients
Report of the American College of Cardiology Foundation/American Heart Association with known or suspected coronary artery disease. Circulation. 2009;120:1390–1400.
Task Force on Practice Guidelines Developed in Collaboration With the International 33. Hachamovitch R, Hayes SW, Friedman JD, et al. Comparison of the short-term survival
Society for Heart and Lung Transplantation. J Am Coll Cardiol. 2009;53:e1–e90. benefit associated with revascularization compared with medical therapy in patients with
6. Zipes DP, Camm AJ, Borggrefe M, et al. ACC/AHA/ESC 2006 guidelines for management no prior coronary artery disease undergoing stress myocardial perfusion single photon
of patients with ventricular arrhythmias and the prevention of sudden cardiac death— emission computed tomography. Circulation. 2003;107:2900–2907.
executive summary: A report of the American College of Cardiology/American Heart 34. Hachamovitch R, Rozanski A, Shaw LJ, et al. Impact of ischaemia and scar on the
Association Task Force and the European Society of Cardiology Committee for Practice therapeutic benefit derived from myocardial revascularization vs. medical therapy
Guidelines (Writing Committee to Develop Guidelines for Management of Patients with among patients undergoing stress-rest myocardial perfusion scintigraphy. Eur Heart J.
Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death). Developed in col- 2011;32:1012–1024.
laboration with the European Heart Rhythm Association and the Heart Rhythm Society. 35. Shaw LJ, Berman DS, Maron DJ, et al. Optimal medical therapy with or without per-
Eur Heart J. 2006;27:2099–2140. cutaneous coronary intervention to reduce ischemic burden: Results from the Clinical
7. Camici PG, Prasad SK, Rimoldi OE. Stunning, hibernation, and assessment of myocardial Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial
viability. Circulation. 2008;117:103–114. nuclear substudy. Circulation. 2008;117:1283–1291.
8. McMurray JJ. Clinical practice. Systolic heart failure. N Engl J Med. 2010;362:228–238. 36. Boden WE, O’Rourke RA, Teo KK, et al. Optimal medical therapy with or without PCI for
9. Maceira AM, Prasad SK, Khan M, et al. Reference right ventricular systolic and diastolic stable coronary disease. N Engl J Med. 2007;356:1503–1516.
function normalized to age, gender and body surface area from steady-state free precession 37. Frye RL, August P, Brooks MM, et al. A randomized trial of therapies for type 2 diabetes
cardiovascular magnetic resonance. Eur Heart J. 2006;27:2879–2888. and coronary artery disease. N Engl J Med. 2009;360:2503–2515.
10. Pfisterer M. Right ventricular involvement in myocardial infarction and cardiogenic shock. 38. Fearon WF, Bornschein B, Tonino PA, et al. Economic evaluation of fractional flow
Lancet. 2003;362:392–394. reserve-guided percutaneous coronary intervention in patients with multivessel disease.
11. Konishi T, Ichikawa T, Yamamuro M, et al. Incidence and clinical course of right ventricular Circulation. 2010;122:2545–2550.
infarction: Assessment with radionuclide ventriculography. Angiology. 1987;38:741–749. 39. Tonino PA, De Bruyne B, Pijls NH, et al. Fractional flow reserve versus angiography for
12. Miszalski-Jamka T, Klimeczek P, Tomala M, et al. Extent of RV dysfunction and myocar- guiding percutaneous coronary intervention. N Engl J Med. 2009;360:213–224.
dial infarction assessed by CMR are independent outcome predictors early after STEMI 40. Wijns W, Kolh P, Danchin N, et al. Guidelines on myocardial revascularization. Eur Heart
treated with primary angioplasty. JACC Cardiovasc Imaging. 2010;3:1237–1246. J. 2010;31:2501–2555.
13. Larose E, Ganz P, Reynolds HG, et al. Right ventricular dysfunction assessed by cardiovas- 41. Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed contrast enhance-
cular magnetic resonance imaging predicts poor prognosis late after myocardial infarction. ment to irreversible injury, infarct age, and contractile function. Circulation. 1999;100:
J Am Coll Cardiol. 2007;49:855–862. 1992–2002.
14. Nandalur KR, Dwamena BA, Choudhri AF, et al. Diagnostic performance of stress cardiac 42. Wagner A, Mahrholdt H, Holly TA, et al. Contrast-enhanced MRI and routine single pho-
magnetic resonance imaging in the detection of coronary artery disease: A meta-analysis. J ton emission computed tomography (SPECT) perfusion imaging for detection of subendo-
Am Coll Cardiol. 2007;50:1343–1353. cardial myocardial infarcts: An imaging study. Lancet. 2003;361:374–379.
15. Greenwood JP, Maredia N, Younger JF, et al. Cardiovascular magnetic resonance and 43. Wesbey GE, Higgins CB, McNamara MT, et al. Effect of gadolinium-DTPA on the
single-photon emission computed tomography for diagnosis of coronary heart disease magnetic relaxation times of normal and infarcted myocardium. Radiology. 1984;153:
(CE-MARC): A prospective trial. Lancet. 2012;379:453–460 165–169.
16. Schwitter J, Wacker CM, van Rossum AC, et al. MR-IMPACT: Comparison of perfusion- 44. Rehwald WG, Fieno DS, Chen EL, et al. Myocardial magnetic resonance imaging con-
cardiac magnetic resonance with single-photon emission computed tomography for the trast agent concentrations after reversible and irreversible ischemic injury. Circulation.
detection of coronary artery disease in a multicentre, multivendor, randomized trial. Eur 2002;105:224–229.
Heart J. 2008;29:480–489. 45. Wendland MF, Saeed M, Lauerma K, et al. Alterations in T1 of normal and reperfused
17. Bernhardt P, Spiess J, Levenson B, et al. Combined assessment of myocardial perfusion infarcted myocardium after Gd-BOPTA versus GD-DTPA on inversion recovery EPI. Magn
and late gadolinium enhancement in patients after percutaneous coronary intervention or Reson Med. 1997;37:448–456.
bypass grafts: A multicenter study of an integrated cardiovascular magnetic resonance pro- 46. Kim RJ, Chen EL, Lima JA, et al. Myocardial Gd-DTPA kinetics determine MRI contrast
tocol. JACC Cardiovasc Imaging. 2009;2:1292–1300. enhancement and reflect the extent and severity of myocardial injury after acute reperfused
18. Klein C, Nagel E, Gebker R, et al. Magnetic resonance adenosine perfusion imag- infarction. Circulation. 1996;94:3318–3326.
ing in patients after coronary artery bypass graft surgery. JACC Cardiovasc Imaging. 47. Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imaging technique for the visual-
2009;2:437–445. ization of myocardial infarction. Radiology. 2001;218:215–223.
19. Cury RC, Shash K, Nagurney JT, et al. Cardiac magnetic resonance with T2-weighted 48. Amado LC, Gerber BL, Gupta SN, et al. Accurate and objective infarct sizing by contrast-
imaging improves detection of patients with acute coronary syndrome in the emergency enhanced magnetic resonance imaging in a canine myocardial infarction model. J Am Coll
department. Circulation. 2008;118:837–844. Cardiol. 2004;44:2383–2389.
20. Ingkanisorn WP, Kwong RY, Bohme NS, et al. Prognosis of negative adenosine stress mag- 49. Kim RJ, Albert TS, Wible JH, et al. Performance of delayed-enhancement magnetic
netic resonance in patients presenting to an emergency department with chest pain. J Am resonance imaging with gadoversetamide contrast for the detection and assessment of
Coll Cardiol. 2006;47:1427–1432. myocardial infarction: An international, multicenter, double-blinded, randomized trial.
21. Lerakis S, McLean DS, Anadiotis AV, et al. Prognostic value of adenosine stress cardiovas- Circulation. 2008;117:629–637.
cular magnetic resonance in patients with low-risk chest pain. J Cardiovasc Magn Reson. 50. Beek AM, Kuhl HP, Bondarenko O, et al. Delayed contrast-enhanced magnetic resonance
2009;11:37. imaging for the prediction of regional functional improvement after acute myocardial
22. Bodi V, Sanchis J, Lopez-Lereu MP, et al. Prognostic value of dipyridamole stress cardio- infarction. J Am Coll Cardiol. 2003;42:895–901.
vascular magnetic resonance imaging in patients with known or suspected coronary artery 51. Choi KM, Kim RJ, Gubernikoff G, et al. Transmural extent of acute myocardial infarc-
disease. J Am Coll Cardiol. 2007;50:1174–1179. tion predicts long-term improvement in contractile function. Circulation. 2001;104:
23. Bodi V, Sanchis J, Lopez-Lereu MP, et al. Prognostic and therapeutic implications of dipyri- 1101–1107.
damole stress cardiovascular magnetic resonance on the basis of the ischaemic cascade. 52. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonance imaging
Heart. 2009;95:49–55. to identify reversible myocardial dysfunction. N Engl J Med. 2000;343:1445–1453.
24. Coelho-Filho OR, Seabra LF, Mongeon FP, et al. Stress myocardial perfusion imaging by 53. Selvanayagam JB, Kardos A, Francis JM, et al. Value of delayed-enhancement cardiovascu-
CMR provides strong prognostic value to cardiac events regardless of patient’s sex. JACC lar magnetic resonance imaging in predicting myocardial viability after surgical revascular-
Cardiovasc Imaging. 2011;4:850–861. ization. Circulation. 2004;110:1535–1541.
LWBK1209-ch19_p292-304.indd 303 16/05/13 8:15 PM

54. Ichikawa Y, Sakuma H, Suzawa N, et al. Late gadolinium-enhanced magnetic resonance 74. Mather AN, Fairbairn TA, Ball SG, et al. Reperfusion haemorrhage as determined by car-
imaging in acute and chronic myocardial infarction. Improved prediction of regional myo- diovascular MRI is a predictor of adverse left ventricular remodelling and markers of late
cardial contraction in the chronic state by measuring thickness of nonenhanced myocar- arrhythmic risk. Heart. 2011;97:453–459.
dium. J Am Coll Cardiol. 2005;45:901–909. 75. Ursell PC, Gardner PI, Albala A, et al. Structural and electrophysiological changes in the
55. Knuesel PR, Nanz D, Wyss C, et al. Characterization of dysfunctional myocardium by epicardial border zone of canine myocardial infarcts during infarct healing. Circ Res.
positron emission tomography and magnetic resonance: Relation to functional outcome 1985;56:436–451.
after revascularization. Circulation. 2003;108:1095–1100. 76. Yao JA, Hussain W, Patel P, et al. Remodeling of gap junctional channel function in epicar-
56. Bello D, Einhorn A, Kaushal R, et al. Cardiac magnetic resonance imaging: Infarct size dial border zone of healing canine infarcts. Circ Res. 2003;92:437–443.
is an independent predictor of mortality in patients with coronary artery disease. Magn 77. Bello D, Fieno DS, Kim RJ, et al. Infarct morphology identifies patients with substrate for
Reson Imaging. 2011;29:50–56 sustained ventricular tachycardia. J Am Coll Cardiol. 2005;45:1104–1108.
57. Cheong BY, Muthupillai R, Wilson JM, et al. Prognostic significance of delayed-enhance- 78. Yan AT, Shayne AJ, Brown KA, et al. Characterization of the peri-infarct zone by contrast-
ment magnetic resonance imaging: Survival of 857 patients with and without left ventricu- enhanced cardiac magnetic resonance imaging is a powerful predictor of post-myocardial
lar dysfunction. Circulation. 2009;120:2069–2076. infarction mortality. Circulation. 2006;114:32–39.
58. Kelle S, Roes SD, Klein C, et al. Prognostic value of myocardial infarct size and contractile 79. Schmidt A, Azevedo CF, Cheng A, et al. Infarct tissue heterogeneity by magnetic resonance
reserve using magnetic resonance imaging. J Am Coll Cardiol. 2009;54:1770–1777. imaging identifies enhanced cardiac arrhythmia susceptibility in patients with left ventricu-
59. Kwon DH, Halley CM, Carrigan TP, et al. Extent of left ventricular scar predicts out- lar dysfunction. Circulation. 2007;115:2006–2014.
comes in ischemic cardiomyopathy patients with significantly reduced systolic function: A 80. Roes SD, Borleffs CJ, van der Geest RJ, et al. Infarct tissue heterogeneity assessed with
delayed hyperenhancement cardiac magnetic resonance study. JACC Cardiovasc Imaging. contrast-enhanced MRI predicts spontaneous ventricular arrhythmia in patients with isch-
2009;2:34–44. emic cardiomyopathy and implantable cardioverter-defibrillator. Circ Cardiovasc Imaging.
60. Kwong RY, Chan AK, Brown KA, et al. Impact of unrecognized myocardial scar detected 2009;2:183–190.
by cardiac magnetic resonance imaging on event-free survival in patients presenting with 81. Arheden H, Saeed M, Higgins CB, et al. Measurement of the distribution volume of gado-
signs or symptoms of coronary artery disease. Circulation. 2006;113:2733–2743. pentetate dimeglumine at echo-planar MR imaging to quantify myocardial infarction:
61. Yokota H, Heidary S, Katikireddy CK, et al. Quantitative characterization of myocardial Comparison with 99mTc-DTPA autoradiography in rats. Radiology. 1999;211:698–708.
infarction by cardiovascular magnetic resonance predicts future cardiovascular events in 82. Arheden H, Saeed M, Higgins CB, et al. Reperfused rat myocardium subjected to vari-
patients with ischemic cardiomyopathy. J Cardiovasc Magn Reson. 2008;10:17. ous durations of ischemia: Estimation of the distribution volume of contrast material with
62. Bekkers SC, Yazdani SK, Virmani R, et al. Microvascular obstruction: Underlying patho- echo-planar MR imaging. Radiology. 2000;215:520–528.
physiology and clinical diagnosis. J Am Coll Cardiol. 2010;55:1649–1660. 83. Freeman I, Grunwald AM, Robin B, et al. Effect of early reperfusion on use of triphen-
63. Nijveldt R, Hofman MB, Hirsch A, et al. Assessment of microvascular obstruction and yltetrazolium chloride to differentiate viable from non-viable myocardium in area of risk.
prediction of short-term remodeling after acute myocardial infarction: Cardiac MR imag- Cardiovasc Res. 1990;24:109–114.
ing study. Radiology. 2009;250:363–370. 84. Schelbert EB, Hsu LY, Anderson SA, et al. Late gadolinium-enhancement cardiac magnetic
64. Cochet AA, Lorgis L, Lalande A, et al. Major prognostic impact of persistent microvascu- resonance identifies postinfarction myocardial fibrosis and the border zone at the near cel-
lar obstruction as assessed by contrast-enhanced cardiac magnetic resonance in reperfused lular level in ex vivo rat heart. Circ Cardiovasc Imaging. 2010;3:743–752.
acute myocardial infarction. Eur Radiol. 2009;19:2117–2126. 85. Aletras AH, Tilak GS, Natanzon A, et al. Retrospective determination of the area at risk
65. de Waha S, Desch S, Eitel I, et al. Impact of early vs. late microvascular obstruction for reperfused acute myocardial infarction with T2-weighted cardiac magnetic resonance
assessed by magnetic resonance imaging on long-term outcome after ST-elevation myocar- imaging: Histopathological and displacement encoding with stimulated echoes (DENSE)
dial infarction: A comparison with traditional prognostic markers. Eur Heart J. 2010;31: functional validations. Circulation. 2006;113:1865–1870.
2660–2668. 86. Berry C, Kellman P, Mancini C, et al. Magnetic resonance imaging delineates the isch-
66. Hombach V, Grebe O, Merkle N, et al. Sequelae of acute myocardial infarction regarding emic area at risk and myocardial salvage in patients with acute myocardial infarction. Circ
cardiac structure and function and their prognostic significance as assessed by magnetic Cardiovasc Imaging. 2010;3:527–535.
resonance imaging. Eur Heart J. 2005;26:549–557. 87. Eitel I, Desch S, Fuernau G, et al. Prognostic significance and determinants of myocar-
67. Nijveldt R, Beek AM, Hirsch A, et al. Functional recovery after acute myocardial infarc- dial salvage assessed by cardiovascular magnetic resonance in acute reperfused myocardial
tion: Comparison between angiography, electrocardiography, and cardiovascular magnetic infarction. J Am Coll Cardiol. 2010;55:2470–2479.
resonance measures of microvascular injury. J Am Coll Cardiol. 2008;52:181–189. 88. Francone M, Bucciarelli-Ducci C, Carbone I, et al. Impact of primary coronary angio-
68. Wu KC, Zerhouni EA, Judd RM, et al. Prognostic significance of microvascular obstruction plasty delay on myocardial salvage, infarct size, and microvascular damage in patients
by magnetic resonance imaging in patients with acute myocardial infarction. Circulation. with ST-segment elevation myocardial infarction: Insight from cardiovascular magnetic
1998;97:765–772. resonance. J Am Coll Cardiol. 2009;54:2145–2153.
69. Fishbein MC, Y-Rit J, Lando U, et al. The relationship of vascular injury and myocardial 89. Friedrich MG, Abdel-Aty H, Taylor A, et al. The salvaged area at risk in reperfused acute
hemorrhage to necrosis after reperfusion. Circulation. 1980;62:1274–1279. myocardial infarction as visualized by cardiovascular magnetic resonance. J Am Coll
70. Higginson LA, White F, Heggtveit HA, et al. Determinants of myocardial hemorrhage after Cardiol. 2008;51:1581–1587.
coronary reperfusion in the anesthetized dog. Circulation. 1982;65:62–69. 90. Masci PG, Ganame J, Strata E, et al. Myocardial salvage by CMR correlates with LV
71. Beek AM, Nijveldt R, van Rossum AC. Intramyocardial hemorrhage and microvascular remodeling and early ST-segment resolution in acute myocardial infarction. JACC
obstruction after primary percutaneous coronary intervention. Int J Cardiovasc Imaging. Cardiovasc Imaging. 2010;3:45–51.
2010;26:49–55. 91. Ubachs JF, Engblom H, Erlinge D, et al. Cardiovascular magnetic resonance of the myo-
72. Eitel I, Kubusch K, Strohm O, et al. Prognostic value and determinants of a hypointense cardium at risk in acute reperfused myocardial infarction: Comparison of T2-weighted
infarct core in T2-weighted cardiac magnetic resonance in acute reperfused ST-elevation- imaging versus the circumferential endocardial extent of late gadolinium enhancement with
myocardial infarction. Circ Cardiovasc Imaging. 2011;4:354–362. transmural projection. J Cardiovasc Magn Reson. 2010;12:18.
73. Ganame J, Messalli G, Dymarkowski S, et al. Impact of myocardial haemorrhage on left 92. Raman SV, Simonetti OP, Winner MW 3rd, et al. Cardiac magnetic resonance with edema
ventricular function and remodelling in patients with reperfused acute myocardial infarc- imaging identifies myocardium at risk and predicts worse outcome in patients with non-ST-
tion. Eur Heart J. 2009;30:1440–1449. segment elevation acute coronary syndrome. J Am Coll Cardiol. 2010;55:2480–2488.
LWBK1209-ch19_p292-304.indd 304 16/05/13 8:15 PM

Chapter
Oliver M. Weber 20
Coronary Magnetic Resonance
Angiography: Technical Approaches
■■ Technical Considerations Multidetector computed tomography (MDCT) has

Suppression of Motion Artifacts recently gained great popularity and experienced great
Contrast Enhancement success in clinical use for coronary angiography (3).
Three-dimensional (3D) data sets can be obtained during
Spatial Resolution
intravenous injection of contrast media in a single breath-
Dedicated Coils hold. The diagnostic accuracy for the detection of significant
Scout Scanning and Imaging Volume Definition stenoses is good, showing a high negative predictive value
Two-dimensional versus Three-dimensional Data and a somewhat lower positive predictive value (3). Recent
Acquisition technical developments such as dual-source CT and higher
Parallel Imaging Techniques rotation speed of the gantry, as well as iterative reconstruc-
Magnetic field strength: 1.5 T vs. 3 T tion algorithms, eliminated a substantial part of the prob-
Visualization lems previously encountered in patients with higher heart
rates or severe calcifications and may be expected to further
■■ Coronary Mra Acquisition Sequences
improve the diagnostic accuracy (4). MDCT has therefore
Bright Blood Coronary MRA been found to be an appropriate method in the diagnosis of
Black Blood Coronary MRA coronary artery disease in a number of clinical indications
■■ Coronary Mra: Future Perspective (5). However, despite recent efforts to reduce doses, MDCT
Stents exposes the patients to significant levels of ionizing radia-
Interventional Coronary MRA tion, which bears potential long-term risk (6). In some pa-
tients, beam-hardening effects, caused by metallic implants
High Field Strength Coronary MRA
or severe calcifications, continue to cause severe image arti-
■■ Summary facts (7).
Coronary magnetic resonance angiography (MRA) com-
bines several advantages and great potential. It is noninva-
Coronary artery disease is one of the major causes of sive and can survey the heart in arbitrary image planes, and
morbidity and mortality in the Western world, where it is it does not involve the use of possibly harmful ionizing ra-
responsible for about one-sixth of all deaths (1). The current diation or iodinated contrast media. In addition to provid-
gold standard for the diagnosis of coronary disease is selec- ing a high degree of spatial resolution, magnetic resonance
tive x-ray coronary angiography. Even though the number (MR) is not associated with any known short- or long-term
of procedures performed has slightly declined over recent side effects. The utility of MRA for visualizing the coronary
years, more than 1,000,000 of these diagnostic procedures anatomy has been investigated since the late 1980s (8,9).
are performed each year in the United States (1) and Europe. Although no coronary stenoses were identified in these early
X-ray coronary angiography is used to define coronary anat- studies, demonstrations of the potential of magnetic reso-
omy and guide patient therapy. However, x-ray coronary nance imaging (MRI) to assess the anatomy of the coronary
angiography is expensive and invasive, exposes both patient vessels triggered intense and ongoing interest in the field.
and operator to potentially harmful ionizing radiation, and Successful coronary MRA data acquisition is technically
carries a small risk for serious complications. Furthermore, demanding because of the small caliber and tortuosity of the
a significant minority of patients undergoing x-ray angiog- coronary arteries and the presence of signal from surround-
raphy are found not to have any significant disease (2), but ing epicardial fat and myocardium. In addition, cardiac and
remain exposed to the costs and risks of this invasive proce- respiratory motion affects the position of the coronary ar-
dure. Thus, a more cost-effective, noninvasive approach for teries by a multiple of their diameter. Efficient strategies to
defining luminographic disease is urgently needed. suppress motion must therefore be applied. Furthermore,
305
LWBK1209-ch20_p305-323.indd 305 17/05/13 5:11 PM

enhancement of the contrast between the coronary vessel motion is the bulk motion of the heart induced by respi-
lumen and the surrounding tissue (myocardium, epicardial ration. Both extrinsic and intrinsic myocardial motion may
fat) is mandatory for a successful visualization of the coronary greatly exceed the coronary artery diameter and so cause
anatomy. Despite considerable technical progress and promising blurring and ghosting of the coronary vessels on images.
early clinical results, coronary MRA has not yet found wide- Therefore, strategies that minimize the adverse effects of
spread clinical acceptance, and is currently considered appropri- both motion components are needed. Suppression of intrin-
ate only in the evaluation of suspected coronary anomalies, but sic motion is routinely achieved by synchronization with the
not in the evaluation of chest pain syndrome (5). electrocardiogram. Methods for the suppression of extrin-
The general approaches described in this chapter are sic motion include patient breath-holds, signal averaging to
available on current state-of-the-art cardiac MR units from reduce the effects of respiration, and detection of respira-
all vendor platforms, but some nuances may be vendor- tory status by external sensors such as respiratory belts or
specific (e.g., navigator implementation). Established and by means of MR methods.
advanced coronary MRA methods are reviewed. Specific
strategies for motion suppression and contrast enhancement
Suppression of Intrinsic Myocardial Motion
are discussed, and representative image material is displayed.
For submillimeter coronary MRA, the image data cannot be
collected during a single R–R interval. Therefore, the coronary
Technical Considerations
MRA data acquisition must be synchronized with the cardiac
cycle, and k-space segmentation must be applied (Fig. 20.1).
The main part of research studies and method development in
For segmented k-space techniques, an accurate electrocar-
coronary MRA has been conducted on 1.5-Tesla (T) whole-
diographic (ECG) synchronization is mandatory. Less robust
body systems. More recently, many of these methods were
peripheral pulse detection methods yield inferior results. Even
implemented and investigated on 3-T systems as well with
though ECG triggering is superior to peripheral pulse detec-
the goal to exploit the higher signal-to-noise ratio (SNR)
tion, reliable R-wave detection in the presence of a strong
available at higher field strength (10,11). However, high-
static magnetic field (causing the “hydrodynamic effect”) and
quality coronary MRA data have also been obtained at
switching magnetic field gradients is technically challenging.
lower field strength on 0.5-T systems (12–14).
A four-lead vector ECG approach has been shown to provide
more robust ECG triggering than three-lead recording (15),
Suppression of Motion Artifacts
especially at higher magnetic field strengths, and signal filter-
The heart is subject to intrinsic and extrinsic motion. ing can further improve the triggering quality.
Intrinsic myocardial motion is the rhythmic contraction and During the cardiac cycle, the coronary arteries can move
relaxation during the R–R interval. Extrinsic myocardial more than 1 cm, a multiple of their own diameter (Fig. 20.2).
Figure 20.1. Principle of electrocardio-

gram-triggering, k-space segmentation, and
diastolic data acquisition. After data collec-
tion in k-space is completed, images are gen-
erated by means of fast Fourier transform
(FFT).
LWBK1209-ch20_p305-323.indd 306 17/05/13 5:11 PM

Chapter 20 ■ Coronary Magnetic Resonance Angiography: Technical Approaches 307
Figure 20.2. Three frames (out of 40 frames covering the R–R interval) of an axial steady-state free
precession cine MRI. Motion of the right (RCA) and the left anterior descending (LAD) coronary arteries
can be seen. The RCA undergoes more pronounced motion than the LAD. In mid-diastole, both vessels
are quiescent for a short interval (in this subject, approximately 160 ms). For optimal coronary MRA, an
acquisition window during minimal motion should be chosen (i.e., trigger delay of 620 ms in this sub-
ject). The dashed circles represent the position at systole.
The displacement of the right coronary artery (RCA) is larger sharpness of the resulting images. Applying intra-RR mo-
than that of the left anterior descending (LAD) or the left tion correction based on partial data sets and motion correc-
circumflex (LCX) artery. The coronary arteries are relatively tion during postprocessing, the acquisition window could be
quiescent for short periods after completion of ventricular extended even further, cutting the total acquisition time in
systole and, for a longer period, at mid-diastole (16,17). To half (21).
minimize motion artifacts in the images, the coronary MRA
data should be acquired during a period of minimal myocar-
Suppression of Extrinsic Myocardial Motion
dial motion. Mid-diastole offers a longer acquisition oppor-
tunity and represents a period of rapid coronary blood flow breath-hold techniques. Among the major difficulties
resulting in inflow of unsaturated spins, a major mechanism encountered in coronary MRA is the bulk cardiac motion as-
of contrast in many bright blood (gradient-echo) sequences. sociated with respiration (Fig. 20.3). Early compensation of
Both duration and time point after R-wave of the quiescent respiratory motion consisted of breath-holding (22). Two-
phases can be roughly estimated based on simple models dimensional (2D) breath-hold coronary MRA relied on the
and by taking into account the patient’s heart rate. However, acquisition of contiguous parallel images, with the goal of
both values are strongly subject-specific (17), and for opti- surveying the proximal segments of the coronary arteries
mal results, individual trigger delays and optimal acquisition during serial breath-holds (23). Subsequently, 3D breath-
windows may thus have to be defined by visual inspection of hold techniques for coronary MRA were implemented (24–
cine images or more advanced automated approaches (18– 26). In combination with parallel imaging, even whole-heart
20). Selection of the acquisition window duration is a trade- coverage can be achieved, however, with the compromise of
off between the competing interests to freeze coronary mo- reduced spatial resolution, an increased acquisition window,
tion (requiring a short acquisition window) and to achieve and a breath-hold duration that is beyond the capacity of
an acceptable overall acquisition duration (requiring a long most patients (27).
acquisition window). The acquisition window in each R–R Breath-hold approaches offer the advantage of rapid im-
interval should therefore be chosen as long as the quiescent aging and are technically easy to implement in compliant pa-
phase. Depending on the sampling strategy in k-space, it can tients. However, they also have several practical limitations.
also be chosen slightly longer without negative effect on the Major patient and operator involvement is required for serial
Figure 20.3. End-expiratory (left)

and end-inspiratory (right) coronal
single-shot SSFP images from a series of
images acquired during free breathing.
The images demonstrate the extent of cra-
niocaudal diaphragm motion (dark gray
lines) and the associated motion of the left
main coronary origin (light gray lines).
The solid lines reflect positions in expira-
tion; the dashed lines, in inspiration.
LWBK1209-ch20_p305-323.indd 307 17/05/13 5:11 PM

Figure 20.4. Free-breathing, sin-

gle-shot scout images in three orthog-
onal planes show navigator placement
on the right hemidiaphragm. For opti-
mal detection of the lung–liver inter-
face, the navigator should be placed
through the top of the right hemidia-
phragm dome, its length covering the
full range of diaphragm motion.
breath-holds. Patients with cardiac or pulmonary disease with breath-holding (34,38). Enhancements included respi-
frequently have difficulty sustaining adequate breath-holds, ratory feedback monitoring (31). However, respiratory belt
particularly when the duration exceeds 5 to 10 seconds. For gating is often not reliable due to poor correlation between
sufficient anatomic coverage, a considerable number of se- abdominal motion and bulk motion of the heart. A time
rial breath-holds is often needed. Alternative breath-hold- delay between chest wall expansion and diaphragmatic mo-
ing techniques, including serial brief breath-holds (28) and tion may occur, introducing additional motion artifacts to
coached breath-holding with visual or audible feedback (29– the images. Subsequently, more accurate and flexible MR
31), have been used to minimize respiratory motion artifacts navigators have replaced belt gating.
and patient inconvenience, but these are practical only for
highly motivated subjects. Even with cooperative patients, free-breathing navigator approaches. The use of
breath-holding may be problematic. With a sustained breath- free breathing with respiratory navigators, first proposed
hold, cranial diaphragmatic drift, which is often substantial by Ehman and Felmlee (38), serves to overcome the time
(∼1 cm), may occur (22,32–34). During serial breath-holds, constraints and cooperation imposed by breath-hold ap-
the diaphragmatic and cardiac positions frequently vary by proaches. Free-breathing navigator methods are particularly
up to 1 cm, so that image registration errors result unless well suited for prolonged 3D coronary MRA approaches,
they are corrected for (35). Misregistration causes apparent which combine the benefits of thin, adjacent slices with the
gaps between the segments of the visualized coronary arter- submillimeter spatial resolution afforded by an improved
ies, which can be misinterpreted as signal voids from coro- SNR. The use of navigator gating has received considerable
nary stenoses. Finally, breath-holds of reasonable duration attention, and implementations vary from relatively simple
(15 seconds) severely limit the options for improving spatial retrospective gating based on a single right-diaphragm navi-
resolution, SNR, and volume coverage. gator (39) to complex prospective affine motion correction
based on multiple navigator signals (40). In principle, the
signal averaging and respiratory belts. Initial free- MR navigator monitors the motion of an interface such as
breathing coronary MRA approaches used signal averaging the lung–diaphragm interface (Fig. 20.4) or the lung–myo-
to minimize motion artifacts (36,37). This averaging ap- cardium interface. Data are accepted only when the selected
proach is reasonable for relatively low spatial resolutions interface falls within a user-defined window (usually 3 to
(>2-mm pixel size), but inadequate for reliable detection of 7 mm) positioned around the end-expiratory level of the
stenoses. As an alternative early free-breathing approach, interface (Fig. 20.5). Although the need for patient coop-
thoracic respiratory belts were used to monitor chest wall eration and operator involvement is reduced with navigator/
expansion and thereby gate image acquisition to the end- free-breathing methods, diaphragmatic drift and patient mo-
expiratory position. Results were promising in comparison tion remain relevant issues (34,41). They may be taken into
Figure 20.5. Free-breathing naviga-

tor display, as visualized in real time on
the scanner console. The X-axis repre-
sents time; the Y-axis shows the MR sig-
nal of the navigator pencil beam. During
a preparation phase, the end-expiratory
position of the lung–liver interface is
determined. A 5-mm gating window is
then automatically calculated, and sub-
sequent acquisitions are only accepted if
the detected interface position is within
this window. Typical navigator efficiency
is approximately 50% for a 5-mm gating
window.
LWBK1209-ch20_p305-323.indd 308 17/05/13 5:11 PM

a ccount by manual or automated adjustment of the position tion and/or of the demodulator phase and frequency. With
of the acceptance window. real-time tracking, a 5-mm diaphragmatic gating window is
often used with a navigator efficiency close to 50% (49).
Navigator Localization and Geometry. Navigators can Coronary MRA with real-time navigator tracking has been
be positioned at any interface that accurately reflects respira- shown to minimize registration errors (in comparison with
tory motion, including the dome of the right hemidiaphragm breath-holding), and image quality is maintained or im-
(30,37,42), the left hemidiaphragm, and the anterior chest proved in both 2D and 3D approaches (34,43).
wall, or directly through the anterior free wall of the left A more elaborate tracking algorithm proposed by Manke
ventricle (34,43). Navigators have been implemented as two et al. (40) accounts not only for craniocaudal translation but
intersecting planes (44,45) and as 2D selective pencil beam also for translation in the other two directions, as well as
excitations (46). Although the intersecting planes are easier rotation, scaling, and shear transformation (summarized as
to implement, they may compromise magnetization in the affine transformation). Transformation parameters are first
volume of interest due to saturation effects. In contrast, 2D determined in a low-resolution dynamic scan during free
selective pencil beam excitations can be implemented with breathing, with simultaneous recording of one to three navi-
the use of shallow radiofrequency (RF) excitation angles, gator signals. During the high-resolution coronary scan, the
so that they only minimally affect the magnetization in the imaged volume is then prospectively adapted to compensate
region of interest. Studies suggest that cardiac motion re- motion. The affine motion model offers the potential to in-
lated to respiration is predominantly in the superior–inferior crease the navigator acceptance window and thus to increase
axis (47,48) and that most single-navigator locations yield the scan efficiency.
similar image quality (34,43). The right hemidiaphragm is
therefore the preferred location for the navigator because of Navigator Tracking Alone. Limitations of breath-hold
the relative ease of identifying the interface from a series of techniques include diaphragmatic drift and inconsistent end-
coronal, sagittal, and transverse scout images (Fig. 20.4). expiratory positions between serial breath-holds. To over-
come these limitations, breath-hold coronary MRA has been
Navigator Gating. The gating process can be retrospec- combined with navigator tracking (50,51). With these tech-
tive or prospective. In retrospective gating, actual MRA niques, coronary MRA data can be acquired in serial breath-
data are recorded along with the respiratory position they holds, and adverse effects of diaphragmatic drift during or
are acquired in. After completion of data acquisition, only between serial breath-holds can be minimized, while main-
data segments acquired in a narrow acceptance window are taining the high data acquisition efficiency of a breath-hold.
used for the reconstruction of the MRA images. To ensure Breath-holding with tracking during acquisition of central
availability of all data segments at an acceptable respira- k-lines and subsequent free breathing with gating and track-
tory position, k-space must be oversampled several times. In ing during acquisition of outer k-space data were also suc-
prospective gating, data segments are stored only if they are cessfully combined to accelerate data acquisition or for use
acquired in an acceptable respiratory position. Otherwise, in combination with contrast media (52).
the pulse sequence is played out (to maintain steady state),
but data are not recorded. Exactly one copy of each data Advanced Navigator Techniques. Sophisticated naviga-
segment is thus acquired at an acceptable position, making tor algorithms have been implemented to collect important
this approach more efficient than retrospective gating. With k-space profiles more efficiently based on the navigator-detected
navigator gating (without tracking), a 3-mm end-expiratory interface position. Implementations of such k-space–reordered
diaphragmatic window is typically used, and data are col- techniques include motion adaptive gating (53), the diminish-
lected on average from one-third of R–R intervals (33% ing variant algorithm (54), phase ordering with automated
navigator efficiency) (34). window selection (55), and the zonal motion adaptive reor-
dering technique (56).
Navigator Gating and Tracking. From several MR stud-
Image-based Motion Compensation. As an alternative to
ies, it was noted that the dominant impact of respiration
assess diaphragm motion, a number of techniques have been
on cardiac position is in the superior–inferior direction
presented that determine the position of the heart in each
(47,48). The correction factor between diaphragm displace-
cardiac beat based on MR signal originating from the heart
ment and craniocaudal displacement of the RCA and left
itself. These techniques include single-shot low-resolution
coronary artery (LCA) was early on reported to be approxi-
images acquired in real time (57), tracking of epicardial fat
mately 0.6 and 0.7, respectively (47). Newer results suggest
(58), or determination of heart position from changes in the
that the actual factors might be smaller and show consider-
gradient-echo (59,60). All these techniques have in common
able intersubject variability and that motion along the other
that they do not rely on a model for the correlation between
two axes might not be negligible (48). Nevertheless, a great
respiratory motion and heart motion. Instead, they assess
number of successful studies were performed with a dis-
directly the position of the heart, which allows for more ac-
placement correction factor of 0.6 in craniocaudal direction
curate gating and tracking. These methods are thus more
only. Knowledge of this relationship offers the opportunity
efficient in data collection and result in shorter acquisition
for prospective navigator gating with real-time tracking of
duration.
the imaged volume position (42). This facilitates the use of
wider gating windows and shortens scan time through in- Navigators, Prepulses, and Imaging Sequences. Most
creased navigator efficiency. The correction of the imaged of the navigator concepts described can be freely combined
volume position is obtained by a prospective run-time ad- with prepulses and 2D or 3D imaging sequences (Fig. 20.6).
aptation of the frequency of the slice-selective RF excita- It has been found that navigator accuracy is improved by
LWBK1209-ch20_p305-323.indd 309 17/05/13 5:11 PM

Trigger delay tion, such as nonselective inversion or dual-inversion pulses,

compromise navigator detection of the interface position.
Countermeasures such as local reinversion of the magneti-
zation after the inversion or dual-inversion prepulses (Nav-
Restore) were developed and successfully applied (62).
Contrast Enhancement
The coronary arteries are surrounded by both epicardial fat
and myocardium. For successful visualization of the coro-
nary arteries, a high level of contrast between the coronary
lumen and the surrounding tissue is desirable. The contrast
T2 prep FGE between the coronary blood pool and the surrounding tis-
sue can be manipulated by using the inflow effect (unsatu-
rated protons entering the imaging field between successive
Inversion
EPI RF pulses), by the application of MR prepulses (endogenous
contrast enhancement), or by the administration of contrast
FGE–EPI agents (exogenous contrast enhancement) with or without
preparatory pulses. Preparatory pulses such as fat satura-
Dual IR tion (63), magnetization transfer contrast (36), T2 prepara-
SSFP tion (64,65), local saturation bands, and inversion (66,67)
and dual-inversion (68) prepulses were all shown to enhance
contrast in coronary MRA.
Navigator
Saturation
FSE
Endogenous Contrast Enhancement

Labeling FSE–EPI
fat saturation in bright blood coronary mag-
netic resonance angiography. In most subjects, the
GRASE coronary arteries are surrounded by epicardial fat. Fat has a
MTC relatively short T1 (250 milliseconds at 1.5 T and 400 mil-
Spiral liseconds at 3 T) and a resultant MR signal intensity similar
to that of flowing blood. Fat saturation prepulses are used
Fat Sat to suppress signal from surrounding fat selectively to allow
Radial FGE visualization of the underlying coronary arteries. This is
often accomplished with a frequency-selective prepulse that
Sat Band
minimizes the fat signal and thereby allows visualization of
Radial SSFP
the coronary vessels. Alternatively, a frequency-selective RF
Figure 20.6. Modular blocks used for navigator-controlled coro- pulse, aiming at exciting exclusively the water resonance,
nary MRA. Electrocardiogram triggering is used for the suppression may be used. However, for conventional segmented gradi-
of intrinsic motion. Contrast is generated by prepulses (red boxes), ent-echo sequences, the duration of these pulses is too long
such as T2 preparation, inversion, dual-inversion, saturation, spin- for practical use. On the other hand, in the case of spiral
labeling, magnetic transfer contrast, fat saturation, or local saturation imaging, typically, only a single RF pulse is applied in each
bands. Some of these contrast mechanisms may be combined with the heartbeat. In this case a spectral–spatial excitation RF pulse
administration of contrast media. Candidate imaging sequences (green targeted at water can be used efficiently (69).
boxes) include fast gradient-echo, echo planar imaging, fast gradient–
echo planar imaging, steady-state free precession, fast spin-echo, fast
signal from the myocardium in bright blood
spin–echo planar imaging, gradient- and spin-echo, spiral imaging,
coronary magnetic resonance angiography. The
radial fast gradient-echo, and radial steady-state free precession. In
addition, all sequences can theoretically be combined with parallel coronary arteries run in close proximity to the epimyocar-
imaging techniques. dium. The relatively similar T1 relaxation values of myo-
cardium and coronary blood (1,000 and 1,200 to 1,400
milliseconds, respectively, at 1.5 T; 1,450 and 1,550 to
1,900 milliseconds at 3 T, respectively (70)) complicate
preceding it with a fat saturation pulse because of reduced the differentiation of the coronary arteries for 3D coro-
effect of the excitation sidebands. To provide optimal fat nary MRA because blood exchange (inflow effect) is re-
suppression during acquisition of the coronary MR data, a duced between successive RF excitations as compared to
second fat suppression pulse immediately preceding data ac- 2D acquisitions. Different methods can be used to enhance
quisition is beneficial. Since the navigator data are intended the contrast between the coronary arteries and myocar-
to reflect the position of the heart during the subsequent ac- dium. The most frequently used ones are prepulses such
quisition period, a short delay between the navigator and as T2 preparation (64,65) and magnetization transfer con-
the data acquisition block and rapid navigator analysis is trast (36). Since the T2 relaxation times of coronary ar-
crucial (61). Some prepulses preceding the navigator excita- terial blood (250 to 300 milliseconds) and myocardium
LWBK1209-ch20_p305-323.indd 310 17/05/13 5:11 PM

Figure 20.7. Principle of T2 prepara-

tion for T2 contrast enhancement between
arterial blood (T2 = 250 ms) and the
myocardium (T2 = 50 ms). T2 affects only
magnetization in the XY-plane, but not
magnetization along the Z-axis (the direc-
tion of the main magnetic field). An initial
90-degree pulse thus rotates the equilib-
rium Mz magnetization of blood (red) and Blood (T2 = 250 ms)
myocardium (green) into the XY-plane,
where it is subject to T2 decay. Due to its
180°
longer T2 time, the magnetization of arte-
rial blood undergoes slower decay. After a
number of refocusing 180-degree pulses, 90° −90°
Myocardium
the magnetization is rotated back along
(T2 = 50 ms)
the Z-axis and is now available for the
subsequent imaging part. Since less mag-
netization is available for the myocardium,
0 20 40 60 80 100
it will show up darker in the MR images
TE
than the blood. TE [ms]
(40 to 50 milliseconds)) are substantially different, the ap- intravascular T1 relaxation rate and therefore may allow for
plication of a T2 preparation prepulse serves to suppress true lumen imaging. With the use of MR contrast agents,
myocardial signal, with relative preservation of the sig- the T1 relaxation of blood can be markedly shortened to
nal from coronary arterial blood (Fig. 20.7). As an added increase the contrast-to-noise ratio (CNR) for coronary
benefit, the signal of deoxygenated blood in the cardiac MRA (66,67,73–75). For this purpose, a number of extra-
veins, which has a shorter T2 than arterial blood, is also cellular and intravascular contrast agents are available.
suppressed when T2 preparation is used (65). Similarly, Extracellular agents quickly extravasate into the extravas-
magnetization transfer may be exploited to partially sup- cular space; therefore, their use requires either rapid first-
press the signal of myocardium. The signal of blood on pass imaging in combination with breath-hold techniques
the other hand shows virtually no magnetization transfer (24,76) or slow infusion of contrast agent over a prolonged
effects. Thus, with the use of fat saturation and T2 prepa- period of time (77,78). However, first-pass coronary MRA
ration or magnetization transfer contrast prepulses, the with extravascular contrast agents is limited by the need for
coronary lumen appears bright, and the signal intensity repeated injections of contrast when more than one slab is
of the surrounding tissue (including fat, myocardium, and imaged. With each subsequent injection, the CNR becomes
veins) is reduced (Fig. 20.8). lower as the signal from the extracellular space continu-
ously increases (because of a progressively decreasing T1
signal from the myocardium and epicardial fat in due to accumulation of contrast agent) after initial contrast
black blood coronary magnetic resonance angiog- administration. The use of intravascular agents has the
raphy. In black blood coronary MRA, a signal-enhanced inherent advantage of allowing image acquisition for lon-
myocardium and a signal-attenuated coronary lumen are de- ger periods of time without changes in contrast behavior or
sirable. For this purpose, a dual-inversion prepulse consist-
ing of a nonselective inversion followed by a slice-selective
inversion is used to re-establish the initial magnetization of
the myocardium at the slice of interest (71). For black blood
coronary MRA, a high signal from the myocardium and epi-
cardial fat is necessary to maximize the contrast between the
signal-attenuated coronary blood pool and the surrounding
tissue. Therefore, no fat saturation is used in this approach
(68).
Exogenous Contrast Enhancement

Bright blood, time-of-flight coronary MRA methods A B
depend heavily on the inflow of unsaturated protons/blood
Figure 20.8. Curved reformats of double-oblique, T2-prepared
into the imaging plane. If, however, flow is slow, saturation coronary MRA’s of the right (A) and left (B) system. A prospective
effects will cause a loss of signal. Furthermore, vessel wall, navigator with gating and tracking was used to image 20 overlapping
plaque, and thrombus can have signal intensities similar to slices during free breathing. In-plane resolution is 0.7 × 1 mm2, slice
that of coronary blood (72). In contrast-enhanced MRA, thickness 3 mm. Note the reduced signal of the great cardiac vein due
enhancement of the blood signal is based primarily on the to the short T2 time of venous blood.
LWBK1209-ch20_p305-323.indd 311 17/05/13 5:11 PM

the need for repeated injections (67). In conjunction with Dedicated Coils
navigator technology, 3D, free-breathing, high-resolution
Spatial resolution requirements for coronary MRA mandate
coronary MRA data acquisition has been demonstrated
maximizing the SNR with the use of appropriate cardiac
(67,79), resulting in improved CNR compared with
receiver coils. Since the SNR decreases with the distance
noncontrast-enhanced approaches.
from the receiver coil, as well as with increasing volume that
the coil is sensitive to, cardiac-specific coils have been opti-
Spatial Resolution mized for the size of the heart and the distance of the heart
from the chest wall. Current commercial products feature
The spatial resolution requirements for coronary MRA
up to 32 elements, typically 16 anterior and 16 posterior
depend on whether the goal is simply to identify the ostial
elements (82). The use of these cardiac-specific phased-array
take-off and proximal course of the coronary artery (as in
coils allows flexibility in volume selection and selection of
suspected cases of congenital anomalous coronary arteries)
parallel imaging techniques. The SNR these specific coils
or to identify focal stenoses. Figure 20.9 displays an x-ray
offer is thus considerably higher than less suitable coils are
coronary angiogram at 0.3-mm spatial resolution, along
able to provide. The use of cardiac-specific coils should be
with simulated resolutions of 0.5, 1, and 2 mm. At resolu-
standard for all coronary MRA examinations.
tions of 0.5 and 1 mm, the focal coronary stenoses are read-
ily detectable, whereas at in-plane resolutions of more than
1 mm, focal stenoses are not discernible. Thus, a spatial Scout Scanning and Imaging
resolution of 1 mm or preferably better is necessary for the Volume Definition
identification of focal stenoses.
Although submillimeter in-plane spatial resolution is The shape, geometry, and orientation of the heart and the
routinely achieved in targeted coronary MRA, typically coronary arteries are highly variable among individuals.
relatively thick (1.5- to 3-mm) slices are acquired, result- As a consequence, the imaging volumes have to be planned
ing in anisotropic voxels. This anisotropic voxel size may carefully. For the first scout, we use an untriggered, free-
lead to vessel blurring. The use of isotropic voxel sizes has breathing, 2D, thoracic acquisition with 17 transverse, 17
been shown to be advantageous for coronary MRA, with coronal, and 25 sagittal single-shot acquisitions. At 1.5 T,
improved vessel sharpness (80), but at the cost of a reduced we use a steady-state free precession (SSFP) sequence; at 3
SNR. Whole-heart coronary MRA inherently suffers from T we prefer a spoiled gradient-echo sequence. Total acquisi-
prolonged acquisition time. To shorten acquisition time, tion time for the entire data set is less than 20 seconds. In
spatial resolution is, in practice, frequently accepted to be this data set, coil positioning can be checked and the loca-
somewhat above the ideal 1-mm in-plane resolution, with tion of the heart and the dome of the right hemidiaphragm
promising results (81). can be identified (Fig. 20.4).
In practice, whole-heart coronary MRA (83) has replaced
the previously used targeted approach (12,25,84), in which
the RCA and LCA systems were imaged in separate volumes
that had to be oriented parallel to the main axis of the ves-
sels (Fig. 20.10). The whole-heart volume can be planned
directly on the survey images. Care must be taken to choose
the volume size and its position to cover the entire coronary
tree. Typically, the volume is prescribed in axial (transverse)
orientation, which yields source images most akin to the
ones from CT. Depending on the patient, a volume size of 10
A B to 15 cm in craniocaudal orientation is needed. The whole-
heart approach offers complete coverage of the coronary
tree in a single measurement and is easy to plan. Depending
on heart rhythm and navigator efficiency, a whole-heart cor-
onary scan takes in the order of 10 minutes (81).
Two-dimensional versus Three-

dimensional Data Acquisition
C D In the early years of coronary MRA, only the acquisition

of a single 2D image during fewer than 20 heartbeats and
Figure 20.9. A: Conventional x-ray coronary angiogram acquired therefore within a single breath-hold was possible. This tech-
with a spatial resolution of 0.3 mm. Focal lesions of the left main, the nique, first described in humans by Edelman et al. (22), took
left anterior descending, and the proximal left circumflex coronary full advantage of the inflow-related contrast between coro-
arteries are visible. Simulated spatial resolutions of (B) 0.5 mm; (C)
nary blood and surrounding stationary tissues. However,
1 mm; and (D) 2 mm demonstrate the need for a spatial resolution
of 1 mm or better for the ability to detect stenoses. (With permission
2D approaches were limited by the tortuous nature of the
from: Daniel K. Sodickson, Bernard and Irene Schwartz Center for coronary vessels; focal signal loss related to the deviation of
Biomedical Imaging, Department of Radiology, NYU Langone Medical the coronary artery out of the image plane was prone to be
Center, New York, NY.) misinterpreted as a focal stenosis. SNR constraints, limited
LWBK1209-ch20_p305-323.indd 312 17/05/13 5:11 PM

A B C
D E
Figure 20.10. Low-resolution scout images (A–C) showing the origin and the proximal segments of
the RCA (arrows). Imaging sequence was a free-breathing, navigator-gated, T2-prepared SSFP sequence
with a spatial resolution of 1.3 × 1.4 × 5 mm3. Imaging time was <3 minutes. Selecting three points along
the vessels allows prescription of the double oblique volume (box in panel D) containing the proximal
segments of the RCA in a single plane (E).
spatial resolution, registration errors related to inconsis- Parallel Imaging Techniques

tency between serial breath-holds, and diaphragmatic drift
Parallel imaging techniques such as SMASH (85), SENSE
during sustained respiration were additional impediments to
(86), UNFOLD (87), and GRAPPA (88) offer considerable
2D breath-hold coronary MRA.
acceleration of image acquisition with minimal hardware
With the development of advanced gradient systems, the
modifications. These methods take advantage of the locally
availability of coils with a high number of receiver elements,
differing sensitivities of the receiver coil arrays, enabling
the implementation of parallel imaging techniques, as well as
artifact-free image reconstruction after the acquisition of
other improvements in scanner software, both breath-held
a fraction of k-space. Routinely, image acquisition is accel-
(27) and free-breathing 3D coronary MRA approaches, have
erated by a factor of typically 2 to 4 (82), at the cost of
become widely available (45,65). Three-dimensional coronary
slightly reduced SNR (Fig. 20.11). Higher acceleration fac-
MRA overcomes several of the limitations of 2D techniques
tors are possible (89) with modern coils containing a high
(12,36) and offers significant advantages, including enhanced
number (32 and more) of relatively small elements, but is
SNR, improved spatial resolution, and improved slice registra-
accompanied by a considerable reduction in SNR. In prin-
tion. Because of the reduced inflow effect in 3D approaches,
ciple, parallel imaging techniques can be combined with
alternative contrast mechanisms as discussed are mandatory.
Figure 20.11. Coronary MRA’s of the RCA obtained at 3 T using a free-breathing, T2-prepared gra-
dient-echo sequence. Using the parallel imaging technique SENSE, imaging time could be halved without
dramatic loss in image quality. Image quality at an acceleration factor of 3 (one-third of the original acqui-
sition time) is slightly reduced, but still acceptable. (From Huber ME, Kozerke S, Pruessmann KP,
et al. Sensitivity-encoded coronary MRA at 3T. Magn Reson Med. 2004;52:221–227, with permission.)
LWBK1209-ch20_p305-323.indd 313 17/05/13 5:11 PM

Figure 20.12. Visualizations of whole-heart coronary data sets. Volume rendering (left) provides a “realis-
tic” visualization of the LAD with a side branch, the LCX, and part of the RCA. In a different subject, maxi-
mum intensity projection of a thin deformed sphere (right) shows all major vessels and several side branches in
a 3D view, conserving the relative signal intensities of the original MR data.
any of the described imaging sequences, and promising few short segments. This is true for stacks of both 2D and
results were obtained in combination with a free-breathing, 3D scans, but the problem is aggravated in 3D data sets
T2-prepared coronary MRA technique at 1.5 T (90) and 3 because of thinner slices. To facilitate visual inspection,
T (11). a number of strategies have been investigated. A simple
approach consists in creating overlapping maximum inten-
sity projections (MIPs) throughout the data set (or, for
Magnetic Field Strength: 1.5 T versus 3 T
black blood approaches, minimum intensity projections).
With an increasing number of 3-T MR imagers installed at This shows continuation of the vessels and facilitates iden-
clinical sites, coronary MRA has attracted considerable inter- tification of stenoses. Curved reformats represent a some-
est at this field strength with the goal to profit from the double what more complex approach. Here, an MIP is performed
SNR available as compared to 1.5 T. However, for a number along a user-defined path. The path must be carefully
of reasons, coronary MRA at 3 T has proved to be more chal- defined to avoid artifacts that could be misinterpreted as
lenging, and it was soon found that the methods optimized signal voids from coronary stenoses. Alternatively, a thin
for 1.5 T could not be used without further optimization MIP can be performed of an arbitrarily defined curved sur-
(10,11,91). Increased magnetic field inhomogeneities, longer face (92) (Fig. 20.8). With the availability of whole-heart
T1 relaxation times, and limitations on deposited power make approaches (83), more sophisticated visualization strate-
3-T coronary MRA more demanding. Particularly the SSFP gies are required. Volume rendering is used to create 3D
sequence, which had proven to be extremely useful at 1.5 T, views of the heart and the coronary arteries (Fig. 20.12),
requires a number of steps for optimal image quality at 3 T, similar to the views known from MDCT. Alternatively,
including volume-selective shimming and adaptation of flip MIP onto a deformed sphere can be performed (Fig. 20.12).
angles. Even so, the images provide typically less contrast With all these approaches, it must be kept in mind that a
between the vessel and the background. Whereas coronary number of postprocessing steps are required, potentially
MRA at 1.5 T nowadays may be considered routine and is introducing artifacts, and that ultimately the source data
successfully performed at many hospitals, at 3 T it remains are closest to reality and should be inspected when in doubt.
challenging and provides unsatisfactory results in a higher
percentage of cases than at 1.5 T.
Coronary MRA Acquisition
Visualization
Sequences
Even when planning a volume along the vessels, the tortu- Although much progress has been made during the past
osity of the coronary arteries causes the vessels typically to two decades, a consensus regarding the “ideal” coronary
show up in several slices, each slice showing only one or a MRA sequence has not been established. Coronary MRA
LWBK1209-ch20_p305-323.indd 314 17/05/13 5:11 PM

sequences can be conceptualized as consisting of compo- data-acquisition period (far longer than the time that most
nents (Fig. 20.6) that in most cases include the following: patients can sustain a breath-hold) and reduced contrast
because of attenuation of the blood inflow effect hampered
•• Cardiac (ECG) gating to suppress cardiac motion.
the development of 3D coronary MRA. These hurdles were
•• Suppression of respiratory motion (breath-holding, respi-
removed with the implementation of free-breathing naviga-
ratory belts, navigators).
tor methods and the application of magnetization transfer
•• Prepulses and/or contrast agents to enhance the CNR of
contrast (36) and T2 preparation prepulses (65). Three-
the coronary arterial blood from surrounding tissue (fat
dimensional coronary MRA is now widely accepted as the
saturation, T2 preparation, magnetization transfer con-
current standard.
trast, spin-labeling, dual inversion, exogenous contrast
For targeted imaging of the LCA and RCA systems
agents).
separately, using free-breathing navigator-gated and nav-
•• Image acquisition that optimizes coronary artery SNR.
igator-corrected 3D coronary MRA, two 3D volumes are
A wide variety of both bright blood and black blood prescribed in parallel to the LCA and RCA systems (84).
techniques have been investigated in 2D or 3D implementa- A 30-mm slab with 20 overlapping slices is acquired in a
tions. A few of the techniques will be discussed in the fol- segmented k-space gradient-echo acquisition with submilli-
lowing paragraphs. meter in-plane resolution and an acquisition window of less
than 80 milliseconds per heartbeat (65,84) (Fig. 20.8). Each
3D segmented gradient-echo acquisition typically has a dura-
Bright Blood Coronary MRA tion of 8 to 10 minutes, and a navigator efficiency of approx-
The majority of reported coronary MRA methodologies imately 50% is obtained when a navigator gating window of
have used bright blood approaches with 2D or 3D Cartesian 5 mm is used. This sequence has successfully been used with
segmented k-space gradient-echo sequences. Particularly at promising results in a prospective multicenter trial to evalu-
1.5 T, SSFP sequences (93) have gained considerable interest. ate the clinical accuracy of coronary MRA for the detection
Methods of rapid acquisition, such as echo-planar imaging or exclusion of significant proximal coronary artery stenosis
(94), spiral imaging (95), and 3D radial imaging (21), have (97) (Fig. 20.13).
received and will continue to receive attention.
Three-dimensional Steady-state Free-precession
Coronary MRA
Two-dimensional Segmented K-space Gradient-echo
Coronary MRA Advances in MR hardware and software made SSFP (93)
(also known as TrueFISP, FIESTA, and balanced FFE) widely
The first robust approach to coronary MRA was the 2D available. Providing high SNR and good contrast between
segmented k-space gradient-echo acquisition scheme, first blood and myocardium, and not depending on inflow of
described in an isolated heart and an in vivo animal model unsaturated spins for contrast, SSFP was initially primarily
by Burstein (96) and subsequently in humans by Edelman used for cardiac functional studies. More recently, SSFP has
et al. (22). With this approach, multiple phase-encoding also been demonstrated to be useful for coronary MRA, in
steps were acquired during each cardiac cycle. As initially
implemented, eight phase-encoding steps were acquired during
each of 16 successive heartbeats with the use of an incremental
flip angle series and a fat saturation prepulse. Mid-diastolic
data were acquired during a 12- to 16-second (16 heart-
beats) breath-hold. Because of variability in the position of
the diaphragm in the serial breath-holds, repetitive images
acquired at the same position risked displaying different
regions of the coronary artery. Breath-hold variability and
coronary vessel tortuosity made 30 to 40 breath-holds nec-
essary for a complete examination.
Subsequent improvements in gradient strength and slew
rate made it possible to implement this 2D approach with
shorter repetition times. As a result, other options were fa-
cilitated, such as shorter breath-holds, shorter acquisition
windows, and enhanced spatial resolution. Two-dimensional
breath-hold segmented k-space coronary MRA is relatively
easy to implement, and good results can be obtained in well-
motivated subjects.
Figure 20.13. Free-breathing, T2-prepared MRA of the RCA

Three-dimensional Segmented K-space Gradient-echo shows three focal stenoses (left), in agreement with the selective x-ray
Coronary MRA angiogram (right). (From: Stuber M, Botnar RM, Danias PG, et al.
Double-oblique free-breathing high-resolution three-dimensional
The superior SNR and spatial resolution of 3D coronary coronary magnetic resonance angiography. J Am Coll Cardiol.
MRA make it particularly attractive, but the prolonged 1999;34:524–531, with permission.)
LWBK1209-ch20_p305-323.indd 315 17/05/13 5:11 PM

imaging readout train (e.g., five to nine echoes) to take

advantage of echo-planar imaging speed and keep the echo
and acquisition time short so that artifacts related to motion
and blood flow are minimized (103).
For patients with an irregular breathing pattern, or if
navigator gating is not robust, Wielopolski et al. (25) im-
plemented breath-hold 3D segmented echo-planar imaging.
The spatial resolution of this approach is inferior to that de-
scribed for free-breathing approaches, and data are acquired
during a longer period (100 milliseconds) in each R–R inter-
val, but the technique allows images of the major coronary
vessels to be acquired in fewer than 13 breath-holds. The
addition of navigator tracking to prolonged breath-holding
approaches may prove beneficial (51).
Two- and Three-dimensional Spiral Coronary MRA

The use of spiral coronary MRA was first reported by Meyer
et al. (95). Advantages of spiral acquisitions include efficient
filling of k-space, good SNR, and favorable flow properties.
As in echo-planar imaging, a single-shot k-space trajectory can
be used, but interleaved spiral imaging appears favorable and
Figure 20.14. Curved-surface maximum intensity projection of a is well suited for coronary artery imaging. Such an approach
free-breathing Cartesian SSFP data set with an in-plane resolution of 1 may be implemented in a breath-hold (30) or with free-breath-
× 1 mm2. Two focal stenoses in the proximal RCA are clearly depicted. ing navigator gating (104). The 3D navigator-gated and nav-
igator-corrected spiral technique showed significant increase
in SNR over conventional Cartesian approaches (105) (Fig.
both a breath-hold approach (98) and a free-breathing, T2- 20.17). However, magnetic field inhomogeneities result in
prepared approach (Fig. 20.14) (99). The SSFP technique image blurring, and distal segments showed lower image qual-
was found to provide good SNR and CNR, excellent vessel ity than with other methods (Fig. 20.15) (99). Nevertheless, the
delineation (Fig. 20.15), and short acquisition times (∼3 to increase in SNR makes spiral imaging a promising approach,
5 minutes for the targeted free-breathing approach). particularly for MRA of the proximal segments.
These properties made possible the implementation of the
whole-heart approach (100). A free-breathing, T2-prepared Three-dimensional Radial Coronary MRA
SSFP sequence was used to acquire a high-resolution, near- Radial imaging shows inherently a low sensitivity to motion.
isotropic data set (1 × 1 × 1.5 mm3) covering the entire heart. Respiratory motion can thus be compensated for by a sim-
A SENSE-acceleration factor of 2 was used and a short du- ple rigid-body motion correction of the individual profiles,
ration of 140 milliseconds per R–R interval was accepted allowing for the selection of a large gating window or even
to achieve an acquisition duration of approximately 14 acquisition without gating. In one implementation, the first
minutes. The resulting data sets allowed reformatting along echo in each gradient-echo train was oriented in the cranio-
the major vessels and 3D reconstructions (Fig. 20.12) show- caudal direction. This permitted detection of position of the
ing major vessels and side branches. The whole-heart MRA heart and subsequent motion correction based on informa-
technique is similar to MDCT insofar as both are able to tion extracted from this profile. The subsequent echoes of
provide 3D data sets of the coronary system. However, at each echo train were used to fill the entire k-space in an
present, MRA is inferior to MDCT with respect to spatial interleaved 3D radial fashion (59). Resulting images showed
resolution and CNR, resulting in slightly poorer clinical per- improved sharpness and overall image quality, while provid-
formance (101). Additionally, MRA requires considerably ing shorter scan times due to higher efficiency.
longer acquisition times than MDCT.
Contrast-enhanced Coronary MRA
Three-dimensional Segmented K-space Echo-planar
With contrast-enhanced MRA, enhancement of the blood
Coronary MRA
signal is based primarily on the intravascular T1 relaxation
One of the advantages of echo-planar imaging (94) is the rate and therefore may allow true lumen imaging also in the
time-efficient coverage of k-space. Therefore, early 3D presence of turbulent flow. Exogenous MR contrast agents
coronary MRA was performed with echo-planar imaging can be subcategorized into extracellular and intravascular
approaches (12). However, flow-related phase errors some- (blood pool) agents. Extracellular paramagnetic contrast
times severely degraded image quality. With the introduction agents (gadolinium chelates) have been used for first-pass
of hybrid segmented echo-planar imaging techniques (102), coronary MRA (24,25,106,107), as well as for imaging dur-
these phase errors could be minimized. For fast breath-hold ing slow infusion of contrast agents (77,78). The effective
or free-breathing 3D coronary MRA (Fig. 20.16), two to T1 relaxation rate depends on the relaxivity of gadolinium
four excitation pulses are followed by a short echo-planar and its local concentration with prominent, but transient,
LWBK1209-ch20_p305-323.indd 316 17/05/13 5:11 PM

Figure 20.15. The RCA was imaged with six acquisition techniques in combination with fat saturation,
T2 preparation, and prospective navigator gating. The sequences used were Cartesian-spoiled gradient-echo
(C-SGE); Cartesian steady-state free precession (C-SSFP); spiral gradient-echo with a single interleaf (S-GE);
spiral gradient-echo with two interleaves (S-SGE); radial spoiled gradient echo (R-SGE); and radial SSFP
(R-SSFP). Note the excellent delineating of the RCA in the SSFP approaches, and the excellent SNR in the spi-
ral images. (From Weber OM, Pujadas S, Martin AJ, et al. Free-breathing, three-dimensional coronary artery
magnetic resonance angiography: Comparison of sequences. J Magn Reson Imaging. 2004;20:395–402, with
permission.)
shortening of the blood T1 relaxation (T1 of 1,200 millisec- quality over noncontrast-enhanced implementations at 3
onds at 1.5 T) to less than 100 milliseconds during passage T (77). An inversion-recovery segmented gradient-echo se-
of the bolus. Rapid vascular equilibration and extravasation quence was played out during infusion of 0.2 mmol/kg body
into the extravascular space subsequently take place, which weight gadolinium-based contrast agent at a rate of 0.3 mL/
decrease the T1 of myocardium. To identify the timing of min. It was found that under these circumstances, an inver-
the peak gadolinium concentration (minimal T1), a test dose sion time of 200 milliseconds yielded effective myocardium
may be administered during dynamic imaging at the aortic signal suppression and maintained high blood signal intensity.
root (106). Full data sets were acquired in less than 5 minutes, and dur-
Whole-heart MR angiography during slow infusion of ing roughly half of this time the infusion was running. Thanks
contrast agent was demonstrated to provide improved image to a linear encoding scheme, the infusion could be timed to
LWBK1209-ch20_p305-323.indd 317 17/05/13 5:11 PM

the range of 60% to 160%, and SNR improvements in

the range of 20% to 60%, over conventional (T2-prepared)
methods were reported (Fig. 20.18). Taylor et al. (109) per-
formed 2D segmented k-space gradient-echo imaging with
a contrast agent consisting of ultrasmall superparamagnetic
iron oxide particles. The agent effectively decreased the T1
relaxation rate of blood to less than 100 milliseconds for
more than 2 hours and provided superior SNR and CNR in
comparison with noncontrast methods. Intravascular agents
remain under investigation at this time.
Coronary MRA Spin-labeling Methods

Conventional coronary MRA methods display the coronary
blood pool along with the surrounding structures, includ-
ing the coronary vessel wall, myocardium, ventricular and
atrial blood pool, and great vessels. This representation of
the coronary anatomy is not directly analogous to the infor-
mation provided by selective x-ray coronary angiography, in
which only the coronary lumen displayed by the radiopaque
contrast agent is seen. Analogous luminographic data was
obtained with MRI by means of selective tagging of blood
in the aortic root using a 2D selective “pencil” inversion
Figure 20.16. Image of the left coronary system, acquired with a pulse (110). After a wash-in time of 300 to 600 millisec-
segmented hybrid gradient-echo planar imaging technique (TF-EPI) in onds, the labeled blood entered the proximal coronary ves-
combination with T2 preparation and a prospective navigator. sels and was imaged with a 2- to 3-cm-thick slab in either
the transverse or oblique projection. The entire acquisition
occur during acquisition of the central part of k-space, pro- occurred during a 24-second breath-hold. This principle was
viding optimal signal enhancement in the coronaries. re-examined in a free-breathing, navigator-gated and nav-
Alternatively, intravascular MR contrast agents stay igator-corrected 3D interleaved segmented spiral approach
within the vascular system and do not leak into the inter- (111). The technique enables a 3D luminographic display of
stitial space. They thus afford longer scan times for free the coronary tree under various viewing angles (Fig. 20.19).
breathing and navigator technology without enhancement For visualization of the data, no user-assisted postprocess-
of the background tissue. By applying a nonselective 180-de- ing is required. Although dependent on blood-flow veloc-
gree inversion pulse, image acquisition can be timed to ity, such an approach may be an alternative for assessing
occur when the longitudinal magnetization of myocardium the proximal coronary arteries and antegrade blood flow
crosses the null point (67,79,108). CNR improvements in through intracoronary stents.
Figure 20.17. Curved reformats

of Cartesian MRA (A,D), spiral MRA
with a single interleaf (B,E), and spiral
MRA with two interleaves (C,F). All
3D data sets were acquired using a
T2 preparation pulse and a prospec-
tive navigator for respiratory control.
Imaging times were identical for the
A B C
Cartesian and the single spiral inter-
leaf approaches; the double-interleaf
approach was about 50% faster, at
the cost of 30% reduction in SNR as
compared to the single-interleaf image.
Both spiral images compare favorably
to the Cartesian image with respect
to SNR. (From Bornert P, Stuber M,
Botnar RM, et al. Direct comparison of
3D spiral vs. Cartesian gradient-echo
coronary magnetic resonance angiogra-
phy. Magn Reson Med. 2001;46:789–
D E F 794, with permission.)
LWBK1209-ch20_p305-323.indd 318 17/05/13 5:11 PM

Figure 20.18. Multiplanar reformatted images of T2-prepared (left) and intravascular contrast agent-
enhanced (right) acquisitions obtained in the same volunteer. The combination of the contrast agent and 3D
inversion-recovery MRA improves visibility of the coronary arteries and the great cardiac vein (GCV), and
depicts additional small-diameter branches of the LCA (short arrows). Ao, aorta; RVO, right ventricular
outflow tract. (From Huber ME, Paetsch I, Schnackenburg B, et al. Performance of a new gadolinium-based
intravascular contrast agent in free-breathing inversion-recovery 3D coronary MRA. Magn Reson Med.
Black Blood Coronary MRA assessing anomalous vessels or diseases. In one of the earliest
studies, an ECG-gated spin-echo technique was able to visual-
Two-dimensional Spin-echo Coronary MRA
ize portions of the native coronary arteries in only 7 (30%)
Early attempts to image the coronary arteries included the of 23 subjects (8). Subsequently, a similar methodology was
use of conventional ECG-gated spin-echo coronary MRA. used in six patients who had undergone x-ray coronary angi-
On these images, the signal of the blood pool appeared dark, ography (9). Even though the data were acquired during ven-
whereas that of the surrounding tissue, including myocardium tricular systole, respiratory motion was not suppressed, and
and epicardial fat, was bright. Although occasionally successful the data were acquired over several minutes, the origin of the
in identifying coronary ostia, this approach was not reliable for left main coronary artery was seen in all six subjects and the
ostium of the RCA in four of the six subjects. No stenoses
were visualized in either report. Subsequently, bright blood
gradient-echo techniques came into greater use and made it
possible to visualize proximal to middle portions of the coro-
nary arteries in healthy and diseased states.
Two- and Three-dimensional Fast Spin-echo Black Blood

Coronary MRA
Although used successfully for the detection of coronary
stenoses (81,97), bright blood sequences are prone to over-
estimate lesions because of artificial darkening caused by
focal turbulent flow (112). The vessel luminal diameter may
therefore be underestimated, and results appear biased in
comparison with those of conventional x-ray angiography.
Figure 20.19. Two orthogonal views of 3D visualizations of the On the other hand, the signal intensity of thrombus, vessel
left coronary artery system after spin labeling in the aortic root. The
complete absence of surrounding tissue signal enables 3D visualiza-
wall, and various components of plaque may appear high on
tion without need for user-directed postprocessing. (From Stuber M, bright blood coronary MRA (72), obscuring focal stenoses.
Bornert P, Spuentrup E, et al. Selective three-dimensional visualization A black blood spin-echo-based coronary MRA technique
of the coronary arterial lumen using arterial spin tagging. Magn Reson that exclusively displays the coronary blood pool may offer
Med. 2002;47:322–329, with permission.) advantages in coronary MRA. Methods based on spin-echo
LWBK1209-ch20_p305-323.indd 319 17/05/13 5:11 PM

vessel wall noninvasively and to detect subclinical athero-

sclerotic plaque so that new insights may be acquired on the
development and progress of subclinical atherosclerosis.
Stents
The number of patients with cardiac disease who undergo
percutaneous revascularization with intracoronary stent
deployment is increasing. Stents are metallic implants that
locally distort the magnetic field, and they may appear as sig-
nal voids on MR images. As a consequence, the direct assess-
ment of stent patency with MR is not practical. In compari-
son with gradient-echo approaches, techniques based on fast
spin-echo imaging may minimize such artifacts; however, no
large patient studies have been reported. The stent geometry,
material, and orientation with respect to the main magnetic
field are variables that cannot be controlled and that sub-
stantially influence the appearance of artifacts (113–115).
MR-compatible stents were developed (116,117), demon-
strating complete absence of artifacts and full visualization
within the stent, potentially solving the problem in the future.
Interventional Coronary MRA

With the availability of short-bore MR imagers with real-
time imaging capability and interactive user interfaces,
Figure 20.20. Multiplanar reformatted 3D black blood coronary interventional MRI has gained considerable interest during
MR angiogram acquired with a fast spin-echo sequence during free recent years (118). Coronary artery catheterization (119),
respiration. The view shows the RCA in-plane, a cross-section of the selective contrast injection (120) (Fig. 20.21), and coronary
LAD, and a proximal section of the LCX. (From Stuber M, Botnar
RM, Spuentrup E, et al. Three-dimensional high-resolution fast spin-
echo coronary magnetic resonance angiography. Magn Reson Med.
imaging also have the potential for enhancing CNR in com-

parison with gradient-echo approaches.
Submillimeter black blood coronary MRA images have
been acquired successfully with the use of ECG-triggered
navigator-gated free-breathing dual-inversion 2D (68) and
3D fast spin-echo MRA (62) (Fig. 20.20). Black blood meth-
ods appear to be particularly advantageous for patients
with metallic implants, such as vascular clips, markers, and
sternal wires. These metallic objects are a source of local
magnetic field inhomogeneities. The size of the artifacts is
increased with gradient-echo bright blood coronary MRA
but minimized with black blood approaches.
Coronary MRA: Future

Perspective
Current research in coronary MRA is focused on the clinical
assessment of many of the advanced methods described in this
chapter, as well as many other methods. The goal is to provide
a novel noninvasive test that makes it possible to screen for Figure 20.21. Left coronary artery system imaged during selective
injection of diluted contrast media in the left main coronary artery
major diseases of the proximal and middle coronary arteries. after catheterization under MRI guidance. A thick-slice (3 cm), mag-
Despite many advances during the past decade, the SNR and netization-prepared SSFP sequence was used. The LAD, the LCX, and
speed of data acquisition still must be improved for routine several marginal arteries are well depicted. (From Green JD, Omary
coronary MRA. Research and development of novel methods RA, Schirf BE, et al. Catheter-directed contrast-enhanced coronary
is therefore ongoing. Although beyond the scope of this chap- MR angiography in swine using magnetization-prepared True-FISP.
ter, cardiac MR also has the potential to image the coronary Magn Reson Med. 2003;50:1317–1321, with permission.)
LWBK1209-ch20_p305-323.indd 320 17/05/13 5:11 PM

angioplasty and stent placement (121) under real-time MR 11. Huber ME, Kozerke S, Pruessmann KP, et al. Sensitivity-encoded coronary MRA at 3T.
Magn Reson Med. 2004;52:221–227.
guidance were all successfully demonstrated in experimen- 12. Bornert P, Jensen D. Coronary artery imaging at 0.5 T using segmented 3D echo planar
tal models. However, the benefit of interventional coronary imaging. Magn Reson Med. 1995;34:779–785.
13. Yang GZ, Gatehouse PD, Keegan J, et al. Three-dimensional coronary MR angiography
MRA will have to be defined in comparison with that of using zonal echo planar imaging. Magn Reson Med. 1998;39:833–842.
conventional x-ray coronary angiography. 14. Jhooti P, Keegan J, Gatehouse PD, et al. 3D coronary artery imaging with phase reorder-
ing for improved scan efficiency. Magn Reson Med. 1999;41:555–562.
15. Fischer SE, Wickline SA, Lorenz CH. Novel real-time R-wave detection algorithm based
High Field Strength Coronary MRA on the vectorcardiogram for accurate gated magnetic resonance acquisitions. Magn
Reson Med. 1999;42:361–370.
In recent years, 3-T high-field MR imagers have become 16. Hofman MB, Wickline SA, Lorenz CH. Quantification of in-plane motion of the coro-
nary arteries during the cardiac cycle: Implications for acquisition window duration for
more widely available for routine cardiac applications, even MR flow quantification. J Magn Reson Imaging. 1998;8:568–576.
though coronary artery imaging still poses some problems. 17. Johnson KR, Patel SJ, Whigham A, et al. Three-dimensional, time-resolved motion of the
coronary arteries. J Cardiovasc Magn Reson. 2004;6:663–673.
In parallel, even higher field strengths are explored, and first 18. Wang Y, Watts R, Mitchell I, et al. Coronary MR angiography: Selection of acquisition
results of coronary MRA acquired at 7 T (122) look promis- window of minimal cardiac motion with electrocardiography-triggered navigator cardiac
ing and confirm the potential to profit from the higher intrin- motion prescanning—initial results. Radiology. 2001;218:580–585.
19. Ustun A, Desai M, Abd-Elmoniem KZ, et al. Automated identification of minimal
sic SNR available at higher field strength (123). However, myocardial motion for improved image quality on MR angiography at 3 T. AJR Am J
currently the number of installed 7-T MR imagers is small, Roentgenol. 2007;188:W283–W290.
and their clinical benefits remain to be demonstrated for car- 20. Jahnke C, Paetsch I, Nehrke K, et al. A new approach for rapid assessment of the cardiac
rest period for coronary MRA. J Cardiovasc Magn Reson. 2005;7:395–399.
diac MRI in general and coronary MRA in particular. 21. Stehning C, Bornert P, Nehrke K, et al. Free breathing 3D balanced FFE coronary mag-
netic resonance angiography with prolonged cardiac acquisition windows and intra-RR
motion correction. Magn Reson Med. 2005;53:719–723.
22. Edelman RR, Manning WJ, Burstein D, et al. Coronary arteries: Breath-hold MR angiog-
Summary raphy. Radiology. 1991;181:641–643.
23. Manning WJ, Edelman RR. Magnetic resonance coronary angiography. Magn Reson Q.
1993;9:131–151.
State-of-the-art coronary MRA methodologies allow a suc- 24. Goldfarb JW, Edelman RR. Coronary arteries: Breath-hold, gadolinium-enhanced, three-
cessful suppression of intrinsic and extrinsic myocardial dimensional MR angiography. Radiology. 1998;206:830–834.
motion. At the same time, the contrast between the coro- 25. Wielopolski PA, van Geuns RJ, de Feyter PJ, et al. Breath-hold coronary MR angiography
with volume-targeted imaging. Radiology. 1998;209:209–219.
nary blood pool and the surrounding tissue can be enhanced 26. Stuber M, Botnar RM, Danias PG, et al. Breathhold three-dimensional coronary magnetic
with the use of either exogenous or endogenous mechanisms resonance angiography using real-time navigator technology. J Cardiovasc Magn Reson.
1999;1:233–238.
of contrast enhancement. These techniques enable the visu- 27. Okada T, Kanao S, Kuhara S, et al. Whole-heart coronary MR angiography under a
alization of the coronary arteries with submillimeter in- single breath-hold: A comparative study with respiratory-gated acquisition using a multi-
plane spatial resolution. The utility of coronary MRA for element phased-array coil. Clin Radiol. 2011;66:1060–1063.
28. Doyle M, Scheidegger MB, de Graaf RG, et al. Coronary artery imaging in multiple 1-sec
assessing native coronary artery stenosis is still the focus of breath holds. Magn Reson Imaging. 1993;11:3–6.
intense clinical investigation. Several competing approaches 29. Liu YL, Riederer SJ, Rossman PJ, et al. A monitoring, feedback, and triggering system for
have been described and are currently being evaluated for reproducible breath-hold MR imaging. Magn Reson Med. 1993;30:507–511.
30. Wang Y, Grimm RC, Rossman PJ, et al. 3D coronary MR angiography in multiple breath-
the diagnosis of coronary artery disease in native coronary holds using a respiratory feedback monitor. Magn Reson Med. 1995;34:11–16.
arteries and coronary artery bypass grafts. Technical and 31. Wang Y, Christy PS, Korosec FR, et al. Coronary MRI with a respiratory feedback moni-
tor: The 2D imaging case. Magn Reson Med. 1995;33:116–121.
methodologic advances, in concert with a more experienced 32. Taylor AM, Jhooti P, Keegan J, et al. Magnetic resonance navigator echo diaphragm monitor-
application of knowledge, will make further improvements ing in patients with suspected diaphragm paralysis. J Magn Reson Imaging. 1999;9:69–74.
in visualizing the coronary arteries with MRI possible. 33. Holland AE, Goldfarb JW, Edelman RR. Diaphragmatic and cardiac motion during sus-
pended breathing: Preliminary experience and implications for breath-hold MR imaging.
Radiology. 1998;209:483–489.
34. McConnell MV, Khasgiwala VC, Savord BJ, et al. Comparison of respiratory suppression
References methods and navigator locations for MR coronary angiography. AJR Am J Roentgenol.
1997;168:1369–1375.
1. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics—2012 35. Danias PG, Stuber M, Botnar RM, et al. Navigator assessment of breath-hold dura-
update: A report from the American Heart Association. Circulation. 2012;125:e2–e220. tion: Impact of supplemental oxygen and hyperventilation. AJR Am J Roentgenol.
2. Budoff MJ, Georgiou D, Brody A, et al. Ultrafast computed tomography as a diagnostic 1998;171:395–397.
modality in the detection of coronary artery disease: A multicenter study. Circulation. 36. Li D, Paschal CB, Haacke EM, et al. Coronary arteries: Three-dimensional MR imaging
1996;93:898–904. with fat saturation and magnetization transfer contrast. Radiology. 1993;187:401–406.
3. Achenbach S. Current clinical applications of cardiac computed tomography. J Cardiovasc 37. Oshinski JN, Hofland L, Mukundan S Jr, et al. Two-dimensional coronary MR angiogra-
Transl Res. 2011;4:449–458. phy without breath holding. Radiology. 1996;201:737–743.
4. Achenbach S, Kondo T. Technical advances in cardiac CT. Cardiol Clin. 2012;30:1–8. 38. Ehman RL, Felmlee JP. Adaptive technique for high-definition MR imaging of moving
5. Hendel RC, Patel MR, Kramer CM, et al. ACCF/ACR/SCCT/SCMR/ASNC/NASCI/SCAI/ structures. Radiology. 1989;173:255–263.
SIR 2006 appropriateness criteria for cardiac computed tomography and cardiac magnetic 39. Post JC, van Rossum AC, Hofman MB, et al. Three-dimensional respiratory-gated MR
resonance imaging: A report of the American College of Cardiology Foundation Quality angiography of coronary arteries: Comparison with conventional coronary angiography.
Strategic Directions Committee Appropriateness Criteria Working Group, American College AJR Am J Roentgenol. 1996;166:1399–1404.
of Radiology, Society of Cardiovascular Computed Tomography, Society for Cardiovascular 40. Manke D, Nehrke K, Bornert P. Novel prospective respiratory motion correction
Magnetic Resonance, American Society of Nuclear Cardiology, North American Society for approach for free-breathing coronary MR angiography using a patient-adapted affine
Cardiac Imaging, Society for Cardiovascular Angiography and Interventions, and Society of motion model. Magn Reson Med. 2003;50:122–131.
Interventional Radiology. J Am Coll Cardiol. 2006;48:1475–1497. 41. Taylor AM, Jhooti P, Wiesmann F, et al. MR navigator-echo monitoring of temporal
6. Einstein AJ, Henzlova MJ, Rajagopalan S. Estimating risk of cancer associated with changes in diaphragm position: Implications for MR coronary angiography. J Magn
radiation exposure from 64-slice computed tomography coronary angiography. JAMA. Reson Imaging. 1997;7:629–636.
2007;298:317–323. 42. Sachs TS, Meyer CH, Hu BS, et al. Real-time motion detection in spiral MRI using navi-
7. Choi HS, Choi BW, Choe KO, et al. Pitfalls, artifacts, and remedies in multi-detector row gators. Magn Reson Med. 1994;32:639–645.
CT coronary angiography. Radiographics. 2004;24:787–800. 43. Stuber M, Botnar RM, Danias PG, et al. Submillimeter three-dimensional coronary MR
8. Lieberman JM, Botti RE, Nelson AD. Magnetic resonance imaging of the heart. Radiol angiography with real-time navigator correction: Comparison of navigator locations.
Clin North Am. 1984;22:847–858. Radiology. 1999;212:579–587.
9. Paulin S, von Schulthess GK, Fossel E, et al. MR imaging of the aortic root and proximal 44. Korin HW, Ehman RL, Riederer SJ, et al. Respiratory kinematics of the upper abdominal
coronary arteries. AJR Am J Roentgenol. 1987;148:665–670. organs: A quantitative study. Magn Reson Med. 1992;23:172–178.
10. Stuber M, Botnar RM, Fischer SE, et al. Preliminary report on in vivo coronary MRA at 45. Li D, Kaushikkar S, Haacke EM, et al. Coronary arteries: Three-dimensional MR imag-
3 Tesla in humans. Magn Reson Med. 2002;48:425–429. ing with retrospective respiratory gating. Radiology. 1996;201:857–863.
LWBK1209-ch20_p305-323.indd 321 17/05/13 5:11 PM

46. Hardy CJ, Cline HE. Broadband nuclear magnetic resonance pulses with two-dimen- 78. Yu J, Paetsch I, Schnackenburg B, et al. Use of 2D sensitivity encoding for slow-infu-
sional spatial selectivity. J Appl Phys. 1989;66:1513–1516. sion contrast-enhanced isotropic 3-T whole-heart coronary MR angiography. AJR Am J
47. Wang Y, Riederer SJ, Ehman RL. Respiratory motion of the heart: Kinematics and Roentgenol. 2011;197:374–382.
the implications for the spatial resolution in coronary imaging. Magn Reson Med. 79. Huber ME, Paetsch I, Schnackenburg B, et al. Performance of a new gadolinium-based
1995;33:713–719. intravascular contrast agent in free-breathing inversion-recovery 3D coronary MRA.
48. Keegan J, Gatehouse P, Yang GZ, et al. Coronary artery motion with the respiratory cycle Magn Reson Med. 2003;49:115–121.
during breath-holding and free-breathing: Implications for slice-followed coronary artery 80. Botnar RM, Stuber M, Kissinger KV, et al. Free-breathing 3D coronary MRA: The impact
imaging. Magn Reson Med. 2002;47:476–481. of “isotropic” image resolution. J Magn Reson Imaging. 2000;11:389–393.
49. Danias PG, McConnell MV, Khasgiwala VC, et al. Prospective navigator correction of 81. Kato S, Kitagawa K, Ishida N, et al. Assessment of coronary artery disease using mag-
image position for coronary MR angiography. Radiology. 1997;203:733–736. netic resonance coronary angiography: A national multicenter trial. J Am Coll Cardiol.
50. Stuber M, Botnar RM, Kissinger KV, et al. 3D real-time navigator corrected black-blood 2010;56:983–991.
coronary MRA. J Cardiovasc Magn Reson. 1999;1:340. 82. Nehrke K, Bornert P, Mazurkewitz P, et al. Free-breathing whole-heart coronary MR
51. Shea SM, Kroeker RM, Deshpande V, et al. Coronary artery imaging: 3D segmented angiography on a clinical scanner in four minutes. J Magn Reson Imaging. 2006;23:
k-space data acquisition with multiple breath-holds and real-time slab following. J Magn 752–756.
Reson Imaging. 2001;13:301–307. 83. Weber OM, Martin AJ, Higgins CB. Whole-heart steady-state free precession coronary
52. Huber ME, Oelhafen ME, Kozerke S, et al. Single breath-hold extended free-breathing artery magnetic resonance angiography. Magn Reson Med. 2003;50:1223–1228.
navigator-gated three- dimensional coronary MRA. J Magn Reson Imaging. 2002;15: 84. Stuber M, Botnar RM, Danias PG, et al. Double-oblique free-breathing high resolu-
210–214. tion three-dimensional coronary magnetic resonance angiography. J Am Coll Cardiol.
53. Weiger M, Bornert P, Proksa R, et al. Motion-adapted gating based on k-space weighting 1999;34:524–531.
for reduction of respiratory motion artifacts. Magn Reson Med. 1997;38:322–333. 85. Sodickson DK, Manning WJ. Simultaneous acquisition of spatial harmonics (SMASH):
54. Sachs TS, Meyer CH, Irarrazabal P, et al. The diminishing variance algorithm for real- Fast imaging with radiofrequency coil arrays. Magn Reson Med. 1997;38:591–603.
time reduction of motion artifacts in MRI. Magn Reson Med. 1995;34:412–422. 86. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: Sensitivity encoding for fast
55. Jhooti P, Gatehouse PD, Keegan J, et al. Phase ordering with automatic window selection MRI. Magn Reson Med. 1999;42:952–962.
(PAWS): A novel motion-resistant technique for 3D coronary imaging. Magn Reson Med. 87. Madore B, Glover GH, Pelc NJ. Unaliasing by fourier-encoding the overlaps using the
2000;43:470–480. temporal dimension (UNFOLD), applied to cardiac imaging and fMRI. Magn Reson
56. Huber ME, Hengesbach D, Botnar RM, et al. Motion artifact reduction and vessel Med. 1999;42:813–828.
enhancement for free-breathing navigator-gated coronary MRA using 3D k-space reor- 88. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially
dering. Magn Reson Med. 2001;45:645–652. parallel acquisitions (GRAPPA). Magn Reson Med. 2002;47:1202–1210.
57. Dewan M, Hager GD, Lorenz CH. Image-based coronary tracking and beat-to-beat 89. Hardy CJ, Darrow RD, Saranathan M, et al. Large field-of-view real-time MRI with a
motion compensation: Feasibility for improving coronary MR angiography. Magn Reson 32-channel system. Magn Reson Med. 2004;52:878–884.
Med. 2008;60:604–615. 90. Sodickson DK, Stuber M, Botnar RM, et al. Accelerated coronary MRA in volunteers
58. Keegan J, Gatehouse PD, Yang GZ, et al. Non-model-based correction of respiratory and patients using double-oblique 3D acquisitions with SMASH. Proc ISMRM. 1999;2:
motion using beat-to-beat 3D spiral fat-selective imaging. J Magn Reson Imaging. 1249.
2007;26:624–629. 91. Kaul MG, Stork A, Bansmann PM, et al. Evaluation of balanced steady-state free preces-
59. Stehning C, Bornert P, Nehrke K, et al. Free-breathing whole-heart coronary MRA with 3D sion (TrueFISP) and K-space segmented gradient echo sequences for 3D coronary MR
radial SSFP and self-navigated image reconstruction. Magn Reson Med. 2005;54:476–480. angiography with navigator gating at 3 Tesla. Rofo. 2004;176:1560–1565.
60. Lai P, Bi X, Jerecic R, et al. A respiratory self-gating technique with 3D-translation com- 92. Etienne A, Botnar RM, Van Muiswinkel AM, et al. “Soap-bubble” visualization and
pensation for free-breathing whole-heart coronary MRA. Magn Reson Med. 2009;62 quantitative analysis of 3D coronary magnetic resonance angiograms. Magn Reson Med.
:731–738. 2002;48:658–666.
61. Spuentrup E, Manning WJ, Botnar RM, et al. Impact of navigator timing on free-breath- 93. Oppelt A, Graumann R, Barfuss H, et al. FISP—a new fast MRI sequence. Electromedica.
ing submillimeter 3D coronary magnetic resonance angiography. Magn Reson Med. 1986;54:15–18.
2002;47:196–201. 94. Mansfield P. Real-time echo-planar imaging by NMR. Br Med Bull. 1984;40:187–190.
62. Stuber M, Botnar RM, Spuentrup E, et al. Three-dimensional high-resolution fast spin- 95. Meyer CH, Hu BS, Nishimura DG, et al. Fast spiral coronary artery imaging. Magn
echo coronary magnetic resonance angiography. Magn Reson Med. 2001;45:206–211. Reson Med. 1992;28:202–213.
63. Manning WJ, Li W, Boyle NG, et al. Fat-suppressed breath-hold magnetic resonance 96. Burstein D. MR imaging of coronary artery flow in isolated and in vivo hearts. J Magn
coronary angiography. Circulation. 1993;87:94–104. Reson Imaging. 1991;1:337–346.
64. Brittain JH, Hu BS, Wright GA, et al. Coronary angiography with magnetization-pre- 97. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance angiography for the
pared T2 contrast. Magn Reson Med. 1995;33:689–696. detection of coronary stenoses. N Engl J Med. 2001;345:1863–1869.
65. Botnar RM, Stuber M, Danias PG, et al. Improved coronary artery definition 98. Deshpande VS, Shea SM, Laub G, et al. 3D magnetization-prepared true-FISP: A new
with T2-weighted, free-breathing, three-dimensional coronary MRA. Circulation. technique for imaging coronary arteries. Magn Reson Med. 2001;46:494–502.
1999;99:3139–3148. 99. Weber OM, Pujadas S, Martin AJ, et al. Free-breathing, three-dimensional coronary
66. Li D, Dolan RP, Walovitch RC, et al. Three-dimensional MRI of coronary arteries using artery magnetic resonance angiography: Comparison of sequences. J Magn Reson
an intravascular contrast agent. Magn Reson Med. 1998;39:1014–1018. Imaging. 2004;20:395–402.
67. Stuber M, Botnar RM, Danias PG, et al. Contrast agent-enhanced, free-breathing, 100. Weber OM, Martin AJ, Higgins CB. Whole-heart coronary MR angiography using a
three-dimensional coronary magnetic resonance angiography. J Magn Reson Imaging. magnetization-prepared SSFP sequence. J Cardiovasc Magn Reson. 2003;5:23–24.
1999;10:790–799. 101. Hamdan A, Asbach P, Wellnhofer E, et al. A prospective study for comparison of MR
68. Stuber M, Botnar RM, Kissinger KV, et al. Free-breathing black-blood coronary MR and CT imaging for detection of coronary artery stenosis. JACC Cardiovasc Imaging.
angiography: Initial results. Radiology. 2001;219:278–283. 2011;4:50–61.
69. Meyer CH, Pauly JM, Macovski A, et al. Simultaneous spatial and spectral selective exci- 102. McKinnon GC. Ultrafast interleaved gradient-echo-planar imaging on a standard scan-
tation. Magn Reson Med. 1990;15:287–304. ner. Magn Reson Med. 1993;30:609–616.
70. Stanisz GJ, Odrobina EE, Pun J, et al. T1, T2 relaxation and magnetization transfer in 103. Botnar RM, Stuber M, Danias PG, et al. A fast 3D approach for coronary MRA. J Magn
tissue at 3T. Magn Reson Med. 2005;54:507–512. Reson Imaging. 1999;10:821–825.
71. Edelman RR, Chien D, Kim D. Fast selective black blood MR imaging. Radiology. 104. Bornert P, Aldefeld B, Nehrke K. Improved 3D spiral imaging for coronary MR angiogra-
1991;181:655–660. phy. Magn Reson Med. 2001;45:172–175.
72. Jara H, Yu BC, Caruthers SD, et al. Voxel sensitivity function description of flow-induced 105. Bornert P, Stuber M, Botnar RM, et al. Direct comparison of 3D spiral vs. Cartesian
signal loss in MR imaging: Implications for black-blood MR angiography with turbo gradient-echo coronary magnetic resonance angiography. Magn Reson Med. 2001;46:
spin-echo sequences. Magn Reson Med. 1999;41:575–590. 789–794.
73. Hofman MB, Henson RE, Kovacs SJ, et al. Blood pool agent strongly improves 3D mag- 106. Zheng J, Li D, Bae KT, et al. Three-dimensional gadolinium-enhanced coronary magnetic
netic resonance coronary angiography using an inversion pre-pulse. Magn Reson Med. resonance angiography: Initial experience. J Cardiovasc Magn Reson. 1999;1:33–41.
1999;41:360–367. 107. Green JD, Schirf BE, Omary RA, et al. Projection imaging of the right coronary artery
74. Stillman AE, Wilke N, Li D, et al. Ultrasmall superparamagnetic iron oxide to enhance with an intravenous injection of contrast agent. Magn Reson Med. 2004;52:699–703.
MRA of the renal and coronary arteries: Studies in human patients. J Comput Assist 108. Dirksen MS, Kaandorp TA, Lamb HJ, et al. Three-dimensional navigator coronary MRA
Tomogr. 1996;20:51–55. with the aid of a blood pool agent in pigs: Improved image quality with inclusion of the
75. Taylor AM, Panting JR, Keegan J, et al. Safety and preliminary findings with the intravas- contrast agent first-pass. J Magn Reson Imaging. 2003;18:502–506.
cular contrast agent NC100150 injection for MR coronary angiography. J Magn Reson 109. Taylor AM, Keegan J, Jhooti P, et al. A comparison between segmented k-space FLASH
Imaging. 1999;9:220–227. and interleaved spiral MR coronary angiography sequences. J Magn Reson Imaging.
76. Li D, Carr JC, Shea SM, et al. Coronary arteries: Magnetization-prepared contrast- 2000;11:394–400.
enhanced three-dimensional volume-targeted breath-hold MR angiography. Radiology. 110. Wang SJ, Hu BS, Macovski A, et al. Coronary angiography using fast selective inversion
2001;219:270–277. recovery. Magn Reson Med. 1991;18:417–423.
77. Bi X, Carr JC, Li D. Whole-heart coronary magnetic resonance angiography at 3 Tesla 111. Stuber M, Bornert P, Spuentrup E, et al. Selective three-dimensional visualization of
in 5 minutes with slow infusion of Gd-BOPTA, a high-relaxivity clinical contrast agent. the coronary arterial lumen using arterial spin tagging. Magn Reson Med. 2002;47:
Magn Reson Med. 2007;58:1–7. 322–329.
LWBK1209-ch20_p305-323.indd 322 17/05/13 5:11 PM

112. Evans AJ, Blinder RA, Herfkens RJ, et al. Effects of turbulence on signal intensity in 118. Weber OM, Schalla S, Martin AJ, et al. Interventional cardiac magnetic resonance imag-
gradient echo images. Invest Radiol. 1988;23:512–518. ing. Semin Roentgenol. 2003;38:352–357.
113. Klemm T, Duda S, Machann J, et al. MR imaging in the presence of vascular stents: A 119. Omary RA, Green JD, Schirf BE, et al. Real-time magnetic resonance imaging-guided
systematic assessment of artifacts for various stent orientations, sequence types, and field coronary catheterization in swine. Circulation. 2003;107:2656–2659.
strengths. J Magn Reson Imaging. 2000;12:606–615. 120. Green JD, Omary RA, Schirf BE, et al. Catheter-directed contrast-enhanced coronary
114. Lenhart M, Volk M, Manke C, et al. Stent appearance at contrast-enhanced MR angiog- MR angiography in swine using magnetization-prepared True-FISP. Magn Reson Med.
raphy: In vitro examination with 14 stents. Radiology. 2000;217:173–178. 2003;50:1317–1321.
115. Hug J, Nagel E, Bornstedt A, et al. Coronary arterial stents: Safety and artifacts during 121. Spuentrup E, Ruebben A, Schaeffter T, et al. Magnetic resonance-guided coronary artery
MR imaging. Radiology. 2000;216:781–787. stent placement in a swine model. Circulation. 2002;105:874–879.
116. Buecker A, Spuentrup E, Ruebben A, et al. Artifact-free in-stent lumen visualization by 122. van Elderen SG, Versluis MJ, Westenberg JJ, et al. Right coronary MR angiography at 7
standard magnetic resonance angiography using a new metallic magnetic resonance imag- T: A direct quantitative and qualitative comparison with 3 T in young healthy volunteers.
ing stent. Circulation. 2002;105:1772–1775. Radiology. 2010;257:254–259.
117. Buecker A, Spuentrup E, Ruebben A, et al. New metallic MR stents for artifact-free coro- 123. van Elderen SG, Versluis MJ, Webb AG, et al. Initial results on in vivo human coronary
nary MR angiography: Feasibility study in a swine model. Invest Radiol. 2004;39:250–253. MR angiography at 7 T. Magn Reson Med. 2009;62:1379–1384.
LWBK1209-ch20_p305-323.indd 323 17/05/13 5:11 PM

Evan Appelbaum
CHAPTER Thomas H. Hauser
21
Susan B. Yeon
René M. Botnar
Warren J. Manning
Coronary Artery Imaging—

Clinical Applications
■■ Magnetic Resonance Imaging of Normal noninvasive method to directly visualize the proximal/middle
Coronary Arteries native coronary vessels for the accurate identification/exclusion
of LM/multivessel CAD.
■■ Anomalous Coronary Artery Identification
Over the last decade, coronary magnetic resonance
■■ Coronary Artery Aneurysms/Kawasaki imaging (MRI) has evolved as a potential replacement for
Disease diagnostic x-ray coronary angiography among patients with
suspected anomalous CAD and coronary artery aneurysms,
■■ Coronary Magnetic Resonance Imaging for
and has reached sufficient maturity, such that it may obviate
Identification of Native Vessel Coronary the need for invasive x-ray angiography when performed at
Stenoses experienced centers. The technical aspects related to coro-
Comparative Studies versus Coronary CTA nary MRI acquisition have been reviewed in a prior chapter.
(MDCT) This chapter will review the published clinical data for
■■ Intracoronary Stents coronary MRI in the assessment of anomalous CAD, coro-
nary artery aneurysms, native coronary artery stenoses, and
■■ Magnetic Resonance Imaging for Coronary coronary artery bypass graft disease.
Artery Bypass Graft Assessment Since the mid-1980s, numerous investigators have con-
■■ Summary tributed to our current understanding of the clinical assess-
ment of coronary MRI in comparison with conventional
x-ray coronary angiography. The largest body of published
Despite decades of progress in both prevention and early diag-
clinical experience has used electrocardiogram (ECG)-
nosis, coronary artery disease (CAD) remains the leading cause
triggered targeted three-dimensional (3D) segmented k-space
of mortality for both men and women in the United States and
gradient-echo and whole-heart steady-state free precession
throughout the Western world (1). For over 40 years, invasive
(SSFP) methods, with increasing reports using whole-heart
x-ray coronary angiography has been the “gold standard” for
coronary MRI with MR contrast agents and 3-T methods.
the diagnosis of significant (≥50% diameter stenosis) CAD
with over a million diagnostic x-ray coronary angiograms per-
formed annually in the United States (1) and higher volumes
in Europe and Japan. Although numerous noninvasive tests Magnetic Resonance Imaging of
are available to help discriminate among those with and with- Normal Coronary Arteries
out significant angiographic disease, studies continue to dem-
onstrate that over 50% of patients referred for elective x-ray From a historical perspective, conventional ECG-triggered
coronary angiography are found to have no significant steno- spin-echo MRI was intermittently successful at imaging the
ses (2). Despite the absence of disease, these patients remain native coronary arteries (8,9), but it had relatively low spa-
subjected to the cost, inconvenience, and potential morbidity tial resolution (1.5 × 2 mm). It was the breath-hold, two-
of invasive x-ray angiography (3,4). In addition, data suggest dimensional (2D) segmented k-space gradient-echo approach
that in selected high-risk populations, the incidence of sub- described in humans by Edelman (10) that first offered a
clinical stroke associated with diagnostic cardiac catheteriza- robust approach for imaging the native coronary arteries.
tion may exceed 20% (5). Since surgical revascularization of As implemented across numerous vendor platforms, the
left main (LM) and multivessel proximal coronary disease has LM, left anterior descending (LAD), and right coronary artery
the greatest impact on patient mortality, and >90% of coro- (RCA) are visualized in the nearly all compliant subjects (Table
nary segments undergoing intervention fall within the proxi- 21.1) (11–16). Early reports had reduced (67% to 77%) suc-
mal/middle segments (6,7), it would be desirable to have a cess for imaging of the left circumflex coronary artery (LCX), a
324
LWBK1209-ch21_p324-341.indd 324 5/18/13 12:58 PM

Chapter 21 ■ Coronary Artery Imaging—Clinical Applications 325
uccessful Visualization of the Native Coronary Arteries Using 2D and 3D Segmented

S
Table 21.1
k-space Gradient-echo Coronary Magnetic Resonance Imaginga
Investigator Technique Resp Comp # of Sub RCA (%) LM (%) LAD (%) LCX (%)
Manning (12) 2D GRE BH 25 100 96 100 76

Pennell (11) 2D GRE BH 26 95 95 91 76
Duerinckx (13) 2D GRE BH 20 100 95 86 77
Sakuma (14) 2D GRE cine BH 18 100 100 100 67
Masui (15) 2D GRE BH 13 85 92 100 92
Davis (16) 2D GRE BH 33a 100 100 100 100
Li (17) 3D GRE Mult averages 14 100 100 86 93
Post (18) 3D GRE Retro Nav G 20 100 100 100 100
Wielopolski (19) 3D Seg EPI BH 32 100 100 100 100
Botnar (20) 3D GRE Pro Nav G/C 13 97 100 100 97
Weber (21) 3D SSFP Pro Nav G/C 12 100 100 100 100
a
Including 18 heart transplant recipients.
GRE, gradient echo; Resp Comp, respiratory compensation; BH, breath-hold; Sub, subjects; RCA, right coronary artery; LM, left main coronary
artery; LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery.
finding likely related to the use of an anterior surface coil (with of an anomalous vessel is sometimes only suspected after
reduced signal from the posteriorly directed LCX). Improved the invasive procedure, particularly in a case where there
(>95%) success has been reported with the now routine use of was unsuccessful engagement or visualization of a coronary
anterior and posterior thoracic phased array cardiac/thoracic artery. In addition, the declining routine use of a pulmonary
coils. Despite relatively limited spatial resolution, normal prox- artery catheter during x-ray coronary angiography has made
imal coronary artery diameter is similar to values obtained by characterization of the anterior versus posterior trajectory of
x-ray coronary angiography and pathology (22,23). Currently, the anomalous vessels more difficult to discern.
both targeted 3D segmented k-space gradient-echo and whole- Coronary MRI has several advantages in the diagnosis
heart SSFP coronary MRI methods have been the dominant of coronary anomalies. In addition to being noninvasive
imaging approaches with reported successful visualization of and not requiring ionizing radiation (an important consid-
all of the major vessels in nearly every subject (Table 21.1) eration among this group that is dominated by adolescents
(21,20). In addition to improved visualization of the origin of and younger adults) or iodinated contrast agents, coronary
the native coronary arteries, another distinct advantage of the MRI provides a definitive 3D “road map” of the mediastinal
3D approaches is increased contiguous visualization of length/ structures. With 3D coronary MRI, one can subsequently
distal segments, as compared with 2D methods. With 3D coro- acquire and/or reconstruct an image in arbitrary single and
nary MRI, similar success (as measured by length of coronary double oblique orientations (Fig. 21.1).
artery seen) is found among healthy adults and patients with Early reports of coronary MRI to visualize anomalous
angiographic CAD (17–19,24). coronary arteries included case report confirmation of x-ray
angiographic data. Subsequently, there have been at least
six published series (28–33) of patients who underwent a
Anomalous Coronary Artery blinded comparison of coronary MRI data with x-ray an-
Identification giography (Table 21.2). These early coronary MRI studies
often used a 2D breath-hold, ECG-triggered segmented k-
The ability of coronary MRI to reliably identify the major space gradient-echo approach (28–33). These 2D coronary
coronary arteries led to its early adoption for the identifi- MRI studies have uniformly reported excellent accuracy,
cation and characterization of anomalous CAD. Though including several studies in which coronary MRI was de-
uncommon (<1% of the general population) (25,26) and termined to be superior to x-ray angiography (31,32). Most
most often benign, congenital coronary anomalies in which centers now use 3D coronary MRI because of superior re-
the anomalous segment courses anterior to the aorta and construction capabilities with similar excellent results (30)
posterior to the pulmonary artery are a well-recognized cause (Table 21.2). As a result, clinical coronary MRI is now the
of myocardial ischemia and sudden cardiac death, especially preferred test for young patients in whom anomalous disease
among adolescents and young adults (27). These adverse is suspected or known anomalous disease needs to be further
events commonly occur during or immediately following clarified, or if the patient has another cardiac anomaly asso-
intense exercise and are thought to be related to compres- ciated with coronary anomalies (e.g., tetralogy of Fallot). An
sion of the anomalous segment or vessel kinking during peri- important technical aspect of coronary MRI for assessment
ods of high cardiac output, or coexistent eccentric stenoses of suspected anomalous CAD is the reduced spatial require-
(27). Projection x-ray coronary angiography had tradition- ments. Thus, more rapid coronary MRI acquisitions with
ally been the imaging modality for the clinical diagnosis and lower spatial resolution methods are generally sufficient to
characterization of these anomalies. However, the presence define the coronary artery origin and course.
LWBK1209-ch21_p324-341.indd 325 16/05/13 9:34 PM

Figure 21.1. Free-breathing 3D

coronary MRI using T2 prepulse
navigator gating with real-time motion
correction. (A) Transverse orientation
depicting a malignant-type anomalous
LAD originating from the RCA and
traversing between the aortic root and
the pulmonary artery. (B) Transverse
image in another patient with a malig-
nant-type anomalous origin of the RCA
from the left coronary cusp. Ao, aorta;
PA, pulmonary artery; LA, left atrium;
A B
RA, right atrium.
In a somewhat analogous fashion, 2D breath-hold coro- generalized vasculitis of unknown etiology, usually occurring
nary MRI has also been used to define the altered coronary in children younger than 5 years. Infants and children with this
artery orientation in the cardiac transplant population (16). syndrome may show evidence of myocarditis and/or pericar-
Among cardiac transplant recipients, coronary MRI has ditis, with nearly 20% developing coronary artery aneurysms.
documented a 25-degree anterior (clockwise) ostial rotation, These aneurysms are the source of both short- and long-term
likely explaining the more complex coronary engagement morbidity and mortality (46). Approximately half of the chil-
during x-ray angiography. dren with coronary aneurysms during the acute phase of the
disease will have angiographically normal-appearing vessels
1 or 2 years later (46,47). For afflicted young children, trans-
Coronary Artery Aneurysms/ thoracic echocardiography is usually adequate for diagnosing
Kawasaki Disease and following these aneurysms, but transthoracic echocar-
diography is often inadequate after adolescence and in obese
Though coronary artery aneurysms are relatively uncommon, children. These patients are therefore often referred for serial
data indicate an important role for coronary MRI for assess- x-ray coronary angiographic studies. Data from a series of
ment of this condition, especially given the common need for adolescents and young adults with coronary artery aneurysms
serial assessment. The vast majority of acquired coronary (Figs. 21.2 and 21.3) defined on x-ray angiography have con-
aneurysms in children and younger adults are a result of firmed the high accuracy of coronary MRI for both the iden-
mucocutaneous lymph node syndrome (Kawasaki disease), a tification and the characterization (diameter, length) of these
aneurysms (Fig. 21.4) (48,49). Though no longitudinal stud-
ies have been reported, it is likely that coronary aneurysms
nomalous Coronary
A can now be effectively followed with serial noncontrast coro-
Table 21.2
Magnetic Resonance Imaging nary MRI examinations, an approach particularly beneficial
for young patients for whom repeated exposure to ionizing
Correctly Classified radiation is often a concern. Good correlation between cor-
Investigator # Patients Anomalous Vessels onary MRI and x-ray coronary angiography has also been
McConnell (28) 15 14 (93%) reported for ectatic coronary arteries (distinct from Kawasaki
Post (31) 19 19 (100%)a disease) among adults (50).
Vliegen (32) 12 11 (92%)b
Taylor (29) 25 24 (96%)
Bunce (30) 26 26 (100%)c Coronary Magnetic Resonance
Razmi (33) 12 12 (100%) Imaging for Identification of
a
Including three patients originally misclassified by x-ray angiography.
Native Vessel Coronary Stenoses
b
Including five patients unable to be classified by x-ray angiography.
c
Including 11 patients unable to be classified by x-ray angiography. Although data support a broad clinical role for coronary
MRI in the assessment of suspected anomalous CAD and
LWBK1209-ch21_p324-341.indd 326 16/05/13 9:34 PM

ree-breathing 3D k-space Segmented Gradient-echo Coronary Magnetic Resonance

F
Table 21.3 Imaging Using Retrospective and Prospective Navigators for Identification of Focal
Greater or Equal to 50% Diameter Coronary Stenoses
For ≥50% Diameter
Investigator Stenosis # Subjects # Vessels Sensitivity Specificity
Retrospective Navigator Gating

Post (18) 20 21 (27%) 38% (0–57) 95% (85–100)
Muller (34) 35 83%a 94%a
Woodard (35) 10 10 (100%) 70%
Sandstede (36) 30 30 (100%) 81%b 89%b
van Geuns (37) 32 50% (50–55)c 91% (73–95)c
Huber (38) 40 20 (50%) 73% (25–100) 50% (25–82)
Sardanelli (39) 42 40% of segments 82% (57–100) 89% (72–100)
90% proximal 90% proximal
Prospective Navigators with Real-time Motion Correction
Bunce (40) 34 88% 72%
Moustapha (41) 25 92% 55%
90% (proximal) 92% (proximal)
Sommer (42) 112 74% 63%
88% (good qual.) 91% (good qual.)
Bogaert (43) 19 85–92% 50–83%
Plein (44) 10 75% 85%
Ozgun (45) 14 TFE 91% 57%
SSFP 76% 85%
a
Excluding five patients for “lack of cooperation” and 15 segments for being uninterpretable.
b
Based on 23 (77%) with high-quality scans.
c
Based on 74% of coronary artery segments analyzable by MRI.
A B
Figure 21.2. (A) Oblique 3D

free-breathing T2 prepulse coronary
MRI and (B) 3D dual inversion black
blood coronary MRI in a young adult
with RCA coronary artery aneurysms
(arrows) as a result of Kawasaki dis-
ease with (C) the corresponding x-ray
angiogram. Flow-related signal within
the aneurysms is homogeneous, with
no evidence of thrombosis. Ao, aorta.
(From: Greil GF, Stuber M, Botnar
RM, et al. Coronary magnetic reso-
nance angiography in adolescents and
young adults with Kawasaki disease.
Circulation. 2002;105(8):908–911, C
with permission.)
LWBK1209-ch21_p324-341.indd 327 16/05/13 9:34 PM

Figure 21.3. (A) Transverse 3D T2

prepulse coronary MRI of a subject
with a left coronary artery aneurysm
and (B) corresponding x-ray angio-
gram, demonstrating good correlation
of coronary MRI findings (black arrow,
LAD aneurysm; white arrow, LCX
A B
aneurysm).
coronary artery bypass graft patency, data are currently in sive CVMR have shown comparable accuracy to MDCT
evolution regarding clinical coronary MRI for routine iden- and single photon emission computed tomography (SPECT)
tification of coronary artery stenoses among patients pre- myocardial perfusion imaging (MPI) suggesting more rou-
senting with chest pain under consideration for x-ray angi- tine use of coronary MRI for the evaluation of CAD may be
ography, and no efficacy data have been reported regarding realized in the near future.
“screening coronary” MRI in the general or at-risk popula- K-space gradient-echo sequences depict flow in the cor-
tions. However, based on data published from multicenter onary lumen, with rapidly moving laminar blood flow ap-
studies (51,52), for populations in which the concern is LM pearing “bright” and areas of stagnant flow and/or focal
or multivessel disease coronary MRI approaches are appro- turbulence appearing “dark” because of local saturation
priate. More recently, multidetector computed tomography (stagnant flow) or dephasing (turbulence) (Fig. 21.5). Areas
(MDCT) has thrust the noninvasive coronary arterial assess- of focal stenoses produce varying severity of “signal void”
ment for CAD in to the limelight with rapid acquisition in the coronary MRI with the severity of the signal loss re-
of high-resolution images of the coronary arteries that has lated to the angiographic stenosis (Fig. 21.6) (53). However,
challenged the need for routine coronary x-ray angiography. gradient-echo coronary MRI may sometimes be misleading.
Coronary MRI has responded to the challenge with further If there is slow blood flow distal to a stenosis, there may be
improvements in speed and image quality through advances complete loss of signal in the segment distal to the lesion, de-
in hardware, software, and sequence design. Several recent spite the absence of a total occlusion. Similarly, because lumen
single and multicenter studies using high field strength (3 T), signal is insensitive to the direction of blood flow, a total occlu-
whole-heart acquisitions individually or part of a comprehen- sion with adequate retrograde (or antegrade collateral blood
12 28
y = 1.12x – 0.41 y = 1.02x + 1.11
R2 = 0.97 R2 = 0.85
p = NS 24 p = 0.03
10
20
8
MRA (mm)
MRA (mm)
16
6 Figure 21.4. Correlation of coro-
12 nary artery aneurysm length (A) and
diameter (B) as determined by coro-
4
8 nary x-ray angiography and coronary
MRI. (Adapted from: Greil GF, Stuber
2 4 M, Botnar RM, et al. Coronary mag-
netic resonance angiography in adoles-
0 cents and young adults with Kawasaki
0 0 2 4 6 8 10 12 0 4 8 12 16 20 24 28 disease. Circulation. 2002;105(8):
A X-Ray (mm) B X-Ray (mm) 908–911, with permission.)
LWBK1209-ch21_p324-341.indd 328 16/05/13 9:34 PM

A B C
Figure 21.5. Transverse (A) and oblique (B) 2D breath-hold coronary MRI. In panel A, a 45-year-old
woman with atypical chest pain demonstrates a signal void (arrows) in the proximal LAD. The more distal
LAD and diagonal vessel are also visualized, as well as the proximal LCX. Panel C demonstrates the corre-
sponding right anterior oblique (RAO) caudal x-ray angiogram (XRA), confirming the tight ostial LAD
stenosis (arrow) seen by MRI.
flow) to the distal segment may result in signal in the lumen resolution, but the technique has successfully demonstrated
distal to the occlusion. SSFP sequences are less dependent on proximal coronary stenoses in several clinical studies (Table
blood flow for signal intensity, but the overall depiction of 21.4) (13,53–56). When reported, the distance from the ves-
stenoses is similar. sel origin to the focal stenosis on coronary MRI correlates
Because of time constraints of a breath-hold, 2D breath- closely with x-ray angiography (Fig. 21.7) (53). However,
hold coronary MRI has relatively limited in-plane spatial there has been wide variation in reported sensitivity and spec-
ificity of this approach, differences that are likely the result of
technical issues/methodology—including the wide variation
100
A in patient selection, presence of arrhythmias, prevalence of
Mean
disease, and technical (MR vendor, echo time, receiver coils,
+/– SE timing of the acquisition, acquisition duration, breath-hold
90 89%
+/– SD maneuvers) issues, and the need for somewhat exhaust-
ing 20- to 40-second breath-holds to complete a study. To
80 date, no multicenter 2D coronary MRI study using a uni-
form hardware, software, and scanning protocol has been
reported, and the approach has now largely been supplanted
X-ray stenosis (%)
70 70% by 3D coronary MRI methods.
60
ree-breathing Targeted 3D
F
50 p < 0.0001 k-space Gradient-echo
Navigator Coronary
Table 21.4
Magnetic Resonance
40
Imaging: Multicenter
37%
Trial Results
30
Left Main/
p < 0.0005
Patients (%) 3VD (%)
20
Severe Partial Irregular wall Sensitivity 93 100
Degree of MR signal loss Specificity 58 85
Prevalence 42 15
Figure 21.6. Breath-hold 2D segmented k-space gradient-echo Positive predictive value 70 54
coronary MRI comparison of focal MR signal loss versus x-ray coro- Negative predictive value 81 100
nary artery diameter stenosis. Note the strong correlation between
the severity of signal loss by coronary MRI and the degree of x-ray Adapted from: Kim WY, Danias PG, Stuber M, et al. Coronary mag-
angiographic stenosis. (From: Pennell DJ, Bogren HG, Keegan J, et al. netic resonance angiography for the detection of coronary stenoses.
Assessment of coronary artery stenosis by magnetic resonance imag- N Engl J Med. 2001;345(26):1863–1869.
ing. Heart. 1996;75(2):127–133, with permission.)
LWBK1209-ch21_p324-341.indd 329 16/05/13 9:34 PM

60 retrospective navigators and prospective navigators with

real-time correction in combination with targeted 3D
Distance to arterial origin by (x-ray) (mm)
k-space segmented gradient echo (Table 21.3) (18,34–45).

50 Early studies using retrospective diaphragmatic navigators
used relatively prolonged acquisition times (up to 260 ms per
R–R interval) (18,34,35), whereas current reports have used
40
acquisition intervals of <120 milliseconds (36–39) or acquisi-
tion timing that is patient specific. These single-center reports
were encouraging with overall sensitivity and specificity of up
30
to 90% for proximal coronary disease (39).
Subsequently, single-center targeted 3D k-space seg-
20 mented gradient-echo studies using more sophisticated
prospective diaphragmatic navigators with real-time mo-
tion correction have shown improved results, especially for
10 the proximal coronary segments and in subjects with high
image-quality scans (Table 21.3) (18,34–45). An interna-
tional multicenter, free-breathing 3D targeted k-space seg-
0 mented coronary MRI study of 109 patients without prior
0 10 20 30 40 50 60
x-ray angiography, demonstrated high sensitivity (though
Distance to arterial origin by MRI (mm)
only modest specificity) and high negative predictive value of
Figure 21.7. Scatterplot comparing the distance from the coronary coronary MRI for the identification of coronary disease (as
origin to the stenosis as measured by x-ray and magnetic resonance cor- defined as ≥50% diameter stenosis by quantitative coronary
onary angiography. (From: Danias PG, McConnell MV, Khasgiwala angiography) (Table 21.4) (51). The sensitivity and negative
VC, et al. Prospective navigator correction of image position for predictive value were particularly high for the identifica-
coronary MR angiography. Radiology. 1997;203(3):733–736, with tion of LM and multivessel disease, demonstrating a clinical
permission.)
role for coronary MRI for this subset. Accordingly, we have
found coronary MRI to be especially valuable for patients
who present with a dilated cardiomyopathy/congestive heart
With the increasing availability of MR navigators, most car- failure in the absence of clinical infarction to determine the
diovascular magnetic resonance (CVMR) imaging centers have etiology (ischemic versus nonischemic) (57). For this group,
migrated to a free-breathing targeted 3D gradient-echo or “whole- more conventional noninvasive tests often have suboptimal
heart” SSFP coronary MRI for ease in patient acceptance (free diagnostic accuracy, with coronary MRI also being superior
breathing) and for improved signal-to-noise ratio (SNR) with to late gadolinium enhancement CVMR methods (57).
facilitated multiplanar reconstructions. As with 2D gradient- The widespread recognition of the benefits of SSFP cine
echo methods, a focal stenosis/turbulence appears as a signal MR for superior SNR and contrast-to-noise ratio (CNR),
void along the course of the vessel (Fig. 21.8). Data from sev- with decreased dependence on inflow of unsaturated pro-
eral single-center sites have now been published using both tons, has led to excitement regarding SSFP coronary MRI.
Figure 21.8. (A) Free-breathing

3D T2 prepulse coronary MRI with
navigator gating and real-time motion
correction and (B) corresponding
x-ray angiography (XRA) in a patient
with proximal (dashed arrow) and
mid-RCA stenoses (solid and dotted
arrows). (From: Stuber M, Botnar
RM, Danias PG, et al. Double-oblique
free-breathing high resolution three-
dimensional coronary magnetic reso-
nance angiography. J Am Coll Cardiol.
A B
1999;34(2):524–531, with permission.)
LWBK1209-ch21_p324-341.indd 330 16/05/13 9:34 PM

ree-breathing,
F
Electrocardiogram-triggered
Table 21.6 1.5 T “Whole-heart” 3D
SSFP Coronary Magnetic
Resonance Imaging:
Multicenter Trial Results
Patient
(n = 138) (%) Vessel (%)
Sensitivity 88 83
Specificity 72 90
Positive predictive value 71 67
Negative predictive value 88 96
Accuracy 79 89
Adapted from: Kato S, Kitagawa K, Ishida N, et al. Assessment

of coronary artery disease using magnetic resonance coronary
angiography. A national multicenter trial. J Am Coll Cardiol.
2010;56(12):983–991.
a multicenter study (seven sites) from Japan (52) compared

Figure 21.9. 3D reconstruction of a whole-heart SSFP coronary the diagnostic accuracy of whole-heart SSFP coronary MRI
MRI following computer-assisted image segmentation enables major to x-ray coronary angiography in 138 subjects suspected of
coronary vessels to be visualized. (Courtesy of Oliver Weber, PhD.) having CAD. The results suggest more widespread use of
whole-heart SSFP coronary MRA is feasible with good diag-
nostic accuracy across centers (Table 21.6).
Comparative data for noncontrast and contrast-enhanced
Despite inferior in-plane spatial resolution (typically 1.2 × coronary MRA in patients are sparse. Conventional extracel-
1.2 mm), single-center data using free-breathing, navigator lular gadolinium-based agents appear to offer minimal ben-
gated, and corrected whole-heart SSFP among patient series, efit. Considerable interest in intravascular agents has been
are particularly impressive for reconstruction (Fig. 21.9) and boosted by early reports demonstrating increased CNR,
suggest superior accuracy for this approach (Table 21.5) both using gadofosveset sodium (ABLAVAR, Lantheus
(58–61) with sensitivities of 80% to 94% and specificity ex- Medical Imaging, Inc., Billerica, MA) (63) and Gadomer-17
ceeding 90% (Fig. 21.10). An approach that uses the affine (Schering, Berlin, Germany) (64). In small series, overall
prospective navigator algorithm (compensates for motion in sensitivity and specificity of free-breathing navigator cor-
both the inferior–superior and anterior–posterior orienta- rected Gadomer-17 coronary MRI was 80% and 93%, re-
tions) offers theoretical advantages (59), but no direct com- spectively, and similar to that of noncontrast single-center
parison with a single diaphragmatic navigator is available. An studies (Fig. 21.11).
early comparative study of k-space segmented gradient echo High (3 T) field coronary MRI offers the theoretical ben-
versus SSFP did confirm the expected improvement of SNR efit of a doubling of SNR. In practice, the SNR gain is more
and CNR but no benefit with regard to accuracy for identifi- modest owing to lack of optimization of surface coils, shim-
cation of CAD (62). As already mentioned, single-center SSFP ming, and so forth. Initial comparative data regarding the
coronary MRI data have been quite favorable (58–61) and clinical use of 3-T coronary MRI have been reported. These
ree-breathing, Electrocardiogram-triggered “Whole-heart” 3D SSFP Coronary

F
Table 21.5 Magnetic Resonance Imaging for Identification of Focal Greater or Equal to 50%
Diameter Coronary Stenoses
For ≥50% Diameter Stenosis
Investigator # Subjects Prev CAD (%) Sensitivity (%) Specificity (%)
Ichikawa (58) 92 94 94
Jahnke (59) 32 50 79 91
Sakuma (60) 101 80 95
Sakuma (61) 131 82 90
LWBK1209-ch21_p324-341.indd 331 16/05/13 9:34 PM

A B C
Figure 21.10. Whole-heart coronary MR angiography (TR/TE = 4.6/2.3 milliseconds, navigator-gated

3D steady-state free precession sequence with fat saturation and T2 preparation) in a 54-year-old man with
LCX stenosis (arrow). (A) Curved multiplanar reformatted image of the LCX artery. (B) 3D volume rendering
image of the LCX artery. (C) Catheter angiography. (Courtesy of Hajime Sakuma, MD, Mie, Japan.)
suggest superior SNR but similar sensitivity and specificities diagnostic accuracy (>90%). Further study is warranted to
of 82% and 89% vs. 82% and 88%, respectively, for 1.5T see if the improved accuracy can be realized in a multicenter,
and 3 T coronary MRI for the detection of significant CAD “real-world” setting. In addition, higher field strengths (i.e.,
(Fig. 21.12) (65). 7 T) are being explored to further boost SNR and enhance
The addition of an exogenous MR contrast agent gado- overall study accuracy (66b).
benate dimeglumine (Gd-BOPTA, MultiHance, Bracco
Diagnostics, Inc., Germany) at high field strength (3 T) has
Comparative Studies versus
been studied in hopes of further increasing the diagnostic
Coronary CTA (MDCT)
accuracy of coronary MRA with additional improvements
in CNR and SNR. Yang et al (66) found that whole-heart Single-center studies have compared the performance of
coronary MRA using Gd-BOPTA at 3 T demonstrated high whole-heart coronary MRA with MDCT for the diagnosis
Figure 21.11. Free-breathing, navigator echo gated segmented 3D fast gradient echo coronary
MRA showing the RCA of a healthy volunteer. Images were taken with endogenous contrast (left)
using a T2 prep magnetization preparation scheme, and after administration of 0.075 mmol/kg of
B-22956 (right) using an Inversion Recovery preparation pulse. Note the much better delineation of
distal portions of the RCA after contrast. High vascular containment of this contrast agent allows for
an efficient nulling of myocardial signal, resulting in a 100% increase in vessel-myocardium contrast
after B-22956. (Reprinted with permission from: Paetsch I, Huber ME, Bornstedt A, et al. Improved
three-dimensional free-breathing coronary magnetic resonance angiography using gadocoletic acid
(B-22956) for intravascular contrast enhancement. J Magn Reson Imaging. 2004;20:288–293.)
LWBK1209-ch21_p324-341.indd 332 16/05/13 9:34 PM

A B C
D E F
Figure 21.12. Comparison of multiplanar reformatted coronary MRI at (A,D) 3 T and at (B,E) 1.5 T and
(C,F) corresponding x-ray angiogram in patients (A–C) with no hemodynamically significant stenosis of the
RCA, including atherosclerotic wall irregularities but no hemodynamically relevant coronary artery stenosis
and (D–F) in a patient with a proximal stenosis (arrow) of the LAD. At both field strengths there is good cor-
relation between coronary MRI and the x-ray angiogram. (Images courtesy of Torsten Sommer, MD, PhD.)
of CAD. Pouler et al (67) found that whole-heart coronary median of 2 years. Ten cardiac events (including five severe
MRA at 1.5 T compared favorably with 40/64 MDCT events) occurred in the 84 patients with CAD on coronary
with similarly high sensitivity and negative predictive value. MRI, whereas only one cardiac event (and no severe events)
(Table 21.7). Subsequently, a single-center 3 T whole-heart was observed in the 123 subjects without CAD on coronary
coronary MRI compared to 64-slice MDCT showing similar MRI. If confirmed in larger series, these data may support
accuracy (68) but without any major improvement in per- an expanded role of screening coronary MRI in high-risk
formance over 1.5 T (Table 21.7). populations with a negative result conveying an excellent
The superiority of coronary MRI over multidetector CT prognosis.
for CAD evaluation is evident in one particular group of pa- It is important to consider native vessel coronary MRI
tients: Those with highly calcified coronary in which MDCT assessment within the greater context of the comprehensive
has significantly reduced specificity and accuracy (69) (Table CVMR examination. Several noninvasive imaging modali-
21.7) (Fig. 21.13). We therefore suggest consideration of ties exist that may be used to evaluate patients suspected of
coronary MRI over MDCT for coronary anatomic evalua- having CAD. These include stress echocardiography, nuclear
tions in the following situations: (1) heavily calcified coro- stress MPI, and multidetector coronary CTA (MDCT). While
nary arteries (i.e., at least one plaque with a calcium score of each test offers advantages and disadvantages (vs. CVMR)
>100), (2) severe iodinated contrast allergy, or (3) concern for specific components of cardiac assessment, none offer the
for radiation exposure. broad comprehensive assessment available with CVMR.
In comparison with MDCT and other noninvasive tests, Recent studies that have compared the diagnostic accuracy
there are very limited prognostic data regarding the implica- of multiparametric CVMR (a combination of coronary MRI,
tions of a normal coronary MRI. Recently, Yoon et al. (70) late gadolinium enhancement, cine functional, and stress/rest
reported on the prognosis of 209 patients who underwent perfusion) with SPECT MPI. Most notably, the CE-MARC
whole-heart SSFP coronary MRI at 1.5 T and followed for a study (71) was a large, prospective, multicenter international
LWBK1209-ch21_p324-341.indd 333 16/05/13 9:34 PM

Table 21.7 Single-center Coronary MRI versus Multidetector Coronary CTA
40/64-slice MDCT versus 1.5-T Whole-heart 3D Coronary MRAa
MDCT (%) Coronary MRA (%)
Sensitivity 94 100
Specificity 88 72
Positive predictive value 70 50
Negative predictive value 98 100
Accuracy 90 78
64-slice MDCT versus 3-T Whole-heart 3D Coronary MRAb
MDCT (%) Coronary MRA (%) p Value
Sensitivity 90 87 0.16
Specificity 83 77 0.06
Positive predictive value 88 83 0.62
Negative predictive value 87 82 0.57
Accuracy 87 83 0.38
Impact of Calcified Plaque (Calcium Score > 100) on the Diagnostic Accuracy of 64-slice MDCT and 3D Free-breathing
Coronary MRAc
MDCT (%) Coronary MRA (%) p Value
Sensitivity 75 81 0.56
Specificity 48 75 0.002
AUC ROC 65 83 0.03
a
Adapted from: Pouleur AC, de Waroux JB, Kefer J, et al. Direct comparison of whole-heart navigator-gated magnetic resonance coronary angi-
ography and 40- and 64-slice multidetector computed tomography to detect the coronary artery stenosis in patients scheduled for conventional
coronary angiography. Circulation: Cardiovascular Imaging. 2008;1(2):114–121.
b
Adapted from: Hamdan A, Asbach P, Wellnhofer E, et al. A prospective study for comparison of MR and CT imaging for detection of coronary
artery stenosis. J Am Coll Cardiol Img. 2011;4(1):50–61.
c
Adapted from: Liu X, Zhao X, Francois CJ, et al. Comparison of free-breathing 3D coronary MR angiography and 64-MDCT angiography for
the detection of coronary stenoses in patients with high calcium scores. AJR: Am J Roentgenol. 2007; 189(6):1326–1332.
X-ray Angiography Coronary CTA Coronary MRI
A B C
Figure 21.13. Calcified coronary artery plaque and stenosis assessment by (A) x-ray angiography, (B) coro-
nary CTA (64-slice), and (C) coronary MRI (3DTFE). Note difficulty in visualizing the severe ostial RCA stenosis
(red arrow) by CTA with adjacent (bright) calcified plaque. The stenosis is better visualized by either x-ray angiog-
raphy or coronary MRI which are not susceptible to calcium blooming artifact. (Courtesy of David Maintz, MD.)
LWBK1209-ch21_p324-341.indd 334 16/05/13 9:34 PM

omparison of
C
Multiparametric CVMR
and SPECT Myocardial
Table 21.8 Perfusion Imaging for the
Diagnosis of CAD (Stenosis
>70%). Results of the
CE-MARC Study
Coronary
n = 752 MRA (%) MPI (%) p Value
Sensitivity 86.5 66.5 <0.0001

Specificity 83.4 82.6 0.916
PPV 77.2 71.4 0.061
NPV 90.5 79.1 <0.0001
Adapted from: Greenwood J, Maredia N, Younger JF, et al.

Cardiovascular magnetic resonance and single-photon emission
computed tomography for diagnosis of coronary artery disease
(CE-MARC): A prospective trial. Lancet. 2012;379(9814):453–460.
study designed to assess the diagnostic accuracy of CVMR as

compared to SPECT MPI for the diagnosis of CAD (as de-
fined by >70% stenosis on quantitative x-ray coronary angi-
ography). Seven hundred fifty-two subjects underwent multi-
parametric CVMR (cine functional, coronary MRI, stress/rest
perfusion, and late gadolinium enhancement imaging), phar-
macologic SPECT MPI (stress/rest) and coronary x-ray angiog-
raphy. CVMR had superior sensitivity and negative predictive
value as compared to SPECT MPI (Table 21.8). The individual
CVMR component data have not been published, but analysis
of the 55% of coronary MRI datasets deemed interpretable, Figure 21.14. Transverse, 2D breath-hold gradient-echo coronary
there was no overall impact on the accuracy of CVMR for MRI at the level of the LAD in a patient with a patent stent. Note
CAD detection. The reason for the surprisingly low success of the signal void (black marker) corresponding to the site of the stent.
coronary MRI in this population was not reported. (Courtesy of Christopher Kramer, MD.)
Intracoronary Stents indirect evidence of a patency by documentation of antegrade

flow.
Improvements in long-term patency rates for percutaneous
coronary interventions using conventional and drug-eluting
intracoronary stents have resulted in their widespread use in Magnetic Resonance Imaging
over 80% of the growing number of percutaneous coronary for Coronary Artery Bypass
artery revascularizations. Typically made from high-grade Graft Assessment
stainless steel, preliminary data demonstrate no short- or long-
term adverse events for patients who undergo “early” CVMR Until the advent of intracoronary stents over a decade
scanning after stent implantation (72,73), but the presence of ago, coronary artery bypass graft surgery was among the
stents does pose imaging problems. Although the attractive most common procedures performed in the United States.
force and local heating are negligible both at 1.5 (74–79) and 3 Unfortunately, early (<1 month) vein graft occlusion occurs
T (80), the local susceptibility effects lead to substantial signal in up to 15% of patients because of mechanical issues, and
voids/artifacts at the site of the stent (Fig. 21.14). The size of an accelerated atherosclerotic process often leads to late
the artifact is dependent on both the stent material (increased (5–10 years) stenoses/occlusion in the majority of grafts
with stainless steel and decreased with titanium stents) (81,82) (84,85).
and the MR sequence (larger with gradient-echo sequences). Compared to the native coronary arteries, reverse sa-
A novel MR-lucent stent has also been reported (83) (Fig. phenous vein and internal mammary artery grafts are rel-
21.15). The signal void/artifact precludes direct evaluation of atively easy to image (in the absence of adjacent vascular
intrastent and peristent coronary integrity, although assess- clips) because of their relatively stationary position during
ment of blood flow/direction proximal and distal to the stent the cardiac and respiratory cycles and their larger lumen.
using MR flow methods or spin-labeling methods may provide Furthermore, their predictable and less convoluted course
LWBK1209-ch21_p324-341.indd 335 16/05/13 9:34 PM

A B
Figure 21.15. MR-lucent stent. (A) Multidetector CT (MDCT) with Aachen Resonance prototype
MR stent in the RCA. (B) 3D coronary MRI in the same animal. Note the absence of local susceptibility
effects on the MR image. (Courtesy of Arno Buecker, MD.)
has allowed imaging of bypass grafts, even with conven- for spin-echo techniques and bright signal for gradient-
tional MR techniques. echo approaches). If signal consistent with flow is identified
With schematic knowledge of the origin and touchdown in the area of the graft lumen on two levels, it is likely to
site of each graft, conventional free-breathing ECG-gated be patent. If a patent lumen is seen at only one level (e.g.,
2D spin-echo (84–87) and 2D gradient-echo MR (88–91) for spin-echo techniques, a signal void is seen at only one
in the transverse plane both have been used to reliably as- level), a graft is considered “indeterminate.” If a patent graft
sess bypass graft patency (Fig. 21.16) (Table 21.9) (87–95). lumen is not seen at any level, the graft is considered oc-
Patency is generally determined by visualizing a patent graft cluded. Combining spin-echo and gradient-echo imaging in
lumen in at least two contiguous transverse levels along its the same patient does not appear to improve accuracy (89).
expected course (presence of flow appearing as signal void Both 3D noncontrast (92) and contrast-enhanced coronary
C
A B
Figure 21.16. (A) Coronal and (B) transverse schematic of the anatomic location of coronary artery
bypass grafts (CABG; RCA gr, LAD gr, LCX gr) originating from the aortic root (AA) and anastomosing with
the distal native coronary arteries with location of contiguous transverse slices. (C) Transverse ECG-gated con-
ventional spin-echo coronary MRI image demonstrating flow (white arrow) in an anatomic area corresponding
to the RCA vein graft indicating graft patency at that level. (From: Rubinstein RI, Askenase AD, Thickman
D, et al. Magnetic resonance imaging to evaluate patency of aortocoronary bypass grafts. Circulation.
1987;76(4): 786–791, with permission.)
LWBK1209-ch21_p324-341.indd 336 16/05/13 9:34 PM

ensitivity, Specificity, and Accuracy of Coronary Magnetic Resonance Imaging for

S
Table 21.9
Assessment of Coronary Artery Bypass Graft Patency
Investigator Technique # Grafts Patency (%) Sens (%) Spec (%) Accuracy (%)
White (88) 2D spin-echo 72 69 86 59 78

Rubinstein (96) 2D spin-echo 47 62 90 72 83
Jenkins (97) 2D spin-echo 41 63 89 73 83
Galjee (89) 2D spin-echo 98 74 98 85 89
White (88) 2D GRE 28 50 93 86 89
Aurigemma (90) 2D GRE 45 73 88 100 91
Galjee (89) 2D GRE 98 74 98 88 96
Engelmann (91) 2D GRE 55 100 (IMA) 100 100
66 (SVG) 92 85 89
Molinari (92) 3D GRE 51 76.5 91 97 96
Bunce (95) 3D SSFP 56 SVG 82 84 45 78
23 IMA 96
Wintersperger (94) CE-3D GRE 39 87 97 100 97
Vrachliotis (93) CE-3D GRE 45 67 93 97 95
SVG, saphenous vein graft; IMA, internal mammary artery; Sens, sensitivity; Spec, specificity.
MRI have also been described for the assessment of graft supporting struts or rings, and graft stents) (Figs. 21.14 and
patency (93,94), with slightly improved results. The accu- 21.17). The inability to identify severely diseased yet pat-
racy of ECG-gated SSFP sequences appears to be similar to ent grafts is also a hindrance to clinical use and acceptance.
that of spin-echo and gradient-echo approaches (Table 21.9) Langerak (100) has reported on the use of free-breathing
(95). Data suggest that the use of phase velocity mapping T2 prep submillimeter 3D coronary MRI for assessment of
for assessment of coronary artery bypass graft flow may be saphenous vein graft stenoses (Figs. 21.18 and 21.19). Very
superior to graft imaging (98,99). good agreement occurred between quantitative x-ray angi-
A practical limitation of coronary MRI bypass graft as- ography for assessment of both graft occlusion and graft
sessment is related to local signal loss/artifacts because of stenoses (Table 21.10). This group has also advocated as-
nearby metallic objects (hemostatic clips, ostial stainless steel sessment of rest and adenosine stress coronary artery flow
graft markers, sternal wires, coexistent prosthetic valves and assessment, using phase velocity MR techniques (98,99).
A B C
Figure 21.17. (A) Posterior-anterior (PA) chest x-ray in a patient with coronary artery bypass grafts.
Note the sternal wires (dashed arrow) as well as the coronary artery bypass graft markers (solid arrow).
(B) Transverse coronary MRI in the same patient. Note the large local artifacts (signal voids) related to the
sternal wires (dashed arrow) and bypass graft markers (solid arrows). The sizes of the artifacts are related to the
type of graft marker used. (C) Barium and tantalum markers (arrow) result in the smallest artifacts. The sizes of
the artifacts are also somewhat less with spin-echo/black blood MRI, as compared with gradient-echo imaging.
LWBK1209-ch21_p324-341.indd 337 16/05/13 9:34 PM

B1
B2
B3
A
B4
B5
B6
B7-x
Figure 21.18. (A) X-ray angiogram (XRA) of widely patent saphenous vein graft (SVG) to the obtuse
marginal branch (OM) of the left circumflex coronary artery. (B1–7) Individual slices of the 3D MRI in the
oblique coronal plane. (C) MPR of the 3D scan demonstrates the widely patent vein graft. (From: Langerak
SE, Vliegen HW, de Roos A, et al. Detection of vein graft disease using high-resolution magnetic resonance
angiography. Circulation. 2002;105(3):328–333, with permission.)
Summary
iagnostic Accuracy of
D
Coronary Magnetic Over the past two decades, coronary MRI has been trans-
Table 21.10 Resonance Imaging Flow formed from a laboratory curiosity to a clinically useful imag-
ing tool in selected populations—including the identification/
Assessment for Saphenous
characterization of anomalous coronary arteries, serial assess-
Vein Graft Disease
ment of children and young adults with coronary artery aneu-
Sensitivity (%) Specificity (%) rysms, and the assessment of coronary artery bypass graft
patency. Coronary MRI also appears to be clinically valuable
Graft occlusion 83 (36–100) 100 (92–100) for assessment of native vessel stenosis in selected patients,
Graft stenosis ≥50% 82 (57–96) 88 (72–97) especially those patients presenting with a dilated cardio-
Graft stenosis ≥70% 73 (39–94) 80 (64–91) myopathy/congestive heart failure with suspected left main/
multivessel CAD. A normal coronary MRI in this population
Adapted from: Langerak SE, Vliegen HW, de Roos A, et al. strongly suggests a nonischemic cardiomyopathy. Mounting
Detection of vein graft disease using high-resolution magnetic
evidence suggests that coronary MRI as part of a comprehen-
resonance angiography. Circulation. 2002;105(3):328–333, with
permission.
sive CVMR study to assess for CAD in a population with sus-
pected CAD may be a reasonable first-line test.
LWBK1209-ch21_p324-341.indd 338 16/05/13 9:34 PM

B1
B2
B3
A
B4
B5
B6
C B7-X
Figure 21.19. (A) X-ray angiogram (XRA) of saphenous vein graft (SVG) to the left anterior descend-
ing (LAD) coronary artery with a 56% proximal stenosis (*). (B1–7) Individual slices of the MRI obtained
in the oblique plane. (C) MPR of the 3D scan demonstrating the loss of graft lumen (tapering graft contour)
corresponding to the x-ray angiographic stenosis (arrow). (From: Langerak SE, Vliegen HW, de Roos A,
et al. Detection of vein graft disease using high-resolution magnetic resonance angiography. Circulation.
2002;105(3):328–333, with permission.)
References 13. Duerinckx AJ, Urman MK. Two-dimensional coronary MR angiography: Analysis of ini-
tial clinical results. Radiology. 1994;193(3):731–738.
14. Sakuma H, Caputo GR, Steffens JC, et al. Breath-hold MR cine angiography of coro-
1. Go AS, Hozaffarian D, Roger VL, et al. AHA Statistical Update. Heart disease and stroke nary arteries in healthy volunteers: Value of multiangle oblique imaging planes. Am J
statistics—2013 update. A report from the American Heart Association. Circulation. Roentgenol. 1994;163(3):533–537.
2013;127:e6–e245. 15. Masui T, Isoda H, Mochizuki T, et al. MR angiography of the coronary arteries. Radiat
2. Patel MR, Peterson ED, Dai D, et al. Low diagnostic yield of elective coronary angiogra- Med. 1995;13(1):47–50.
phy. N Eng J Med. 2010;362:886–895. 16. Davis SF, Kannam JP, Wielopolski P, et al. Magnetic resonance coronary angiography in
3. Johnson LW, Lozner EC, Johnson S, et al. Coronary arteriography 1984–1987: A report heart transplant recipients. J Heart Lung Transplant. 1996;15(6):580–586.
of the Registry of the Society for Cardiac Angiography and Interventions. I. Results and 17. Li D, Paschal CB, Haacke EM, et al. Coronary arteries: Three-dimensional MR imaging
complications. Cathet Cardiovasc Diagn. 1989;17(1):5–10. with fat saturation and magnetization transfer contrast.Radiology. 1993;187(2):401–406.
4. Davidson CJ, Mark DB, Pieper KS, et al. Thrombotic and cardiovascular complications 18. Post JC, van Rossum AC, Hofman MB, et al. Three-dimensional respiratory-gated MR
related to nonionic contrast media during cardiac catheterization: Analysis of 8,517 angiography of coronary arteries: Comparison with conventional coronary angiography.
patients. Am J Cardiol. 1990;65(22):1481–1484. AJR Am J Roentgenol. 1996;166(6):1399–1404.
5. Omran H, Schmidt H, Hackenbroch M, et al. Silent and apparent cerebral embolism 19. Wielopolski PA, van Geuns RJ, de Feyter PJ, et al. Breath-hold coronary MR angiography
after retrograde catheterisation of the aortic valve in valvular stenosis: A prospective, with volume-targeted imaging. Radiology. 1998;209(1):209–219.
randomised study. Lancet. 2003;361(9365):1241–1246. 20. Botnar RM, Stuber M, Danias PG, et al. Improved coronary artery definition with T2-weighted,
6. Kelle S, Hug J, Koehler U, et al. Potential intrinsic error of non-invasive coronary angiog- free-breathing, three-dimensional coronary MRA. Circulation. 1999;99(24):3139–3148.
raphy. J Cardiovasc Magn Reson. 2005;7(2):401–407. 21. Weber OM, Martin AJ, Higgins CB. Whole-heart steady-state free precession coronary
7. Rizzo MJ, Moynihan J, Ryan K, et al. Distance from the ostium to the culprit coronary lesions: artery magnetic resonance angiography. Magn Reson Med. 2003;50(6):1223–1228.
Implications for new noninvasive coronary imaging techniques. Circulation. 1997;96:I-306. 22. Botnar RM, Leiner T, Kissinger KV, et al. Improved motion compensation in coronary
8. Paulin S, von Schulthess GK, Fossel E, et al. MR imaging of the aortic root and proximal MRA [abstract]. Proc Int Soc Magn Reson Med. 2004;11:2555.
coronary arteries. Am J Roentgenol. 1987;148(4):665–670. 23. Scheidegger MB, Hess OM, Boesiger P. Validation of coronary artery MR angiogra-
9. Lieberman JM, Botti RE, Nelson AD. Magnetic resonance imaging of the heart. Radiol phy: Comparison of measured vessel diameters with quantitative contrast angiography
Clin North Am. 1984;22(4):847–858. proceed. Soc Magn Reson. 1994:497.
10. Edelman RR, Manning WJ, Burstein D, et al. Coronary arteries: Breath-hold MR angiog- 24. Stuber M, Botnar RM, Danias PG, et al. Double-oblique free-breathing high resolu-
raphy. Radiology. 1991;181(3):641–643. tion three-dimensional coronary magnetic resonance angiography. J Am Coll Cardiol.
11. Pennell DJ, Keegan J, Firmin DN, et al. Magnetic resonance imaging of coronary arteries: 1999;34(2):524–531.
Technique and preliminary results. Br Heart J. 1993;70(4):315–326. 25. Engel HJ, Torres C, Page HL Jr. Major variations in anatomical origin of the coronary
12. Manning WJ, Li W, Boyle NG, et al. Fat-suppressed breath-hold magnetic resonance coro- arteries: Angiographic observations in 4,250 patients without associated congenital heart
nary angiography. Circulation. 1993;87(1):94–104. disease. Cathet Cardiovasc Diagn. 1975;1(2):157–169.
LWBK1209-ch21_p324-341.indd 339 16/05/13 9:34 PM

26. Chaitman BR, Lesperance J, Saltiel J, et al. Clinical, angiographic, and hemodynamic 55. Post JC, van Rossum AC, Hofman MB, et al. Clinical utility of two-dimensional magnetic
findings in patients with anomalous origin of the coronary arteries. Circulation. resonance angiography in detecting coronary artery disease. Eur Heart J. 1997;18(3):
1976;53(1):122–131. 426–433.
27. Cheitlin MD, De Castro CM, McAllister HA. Sudden death as a complication of anoma- 56. Nitatori T, Yokoyama K, Hachiya J, et al. Comparison of 2D coronary MR angiography
lous left coronary origin from the anterior sinus of Valsalva, a not-so-minor congenital with conventional angiography—difference in imaging accuracy according to the severity
anomaly. Circulation. 1974;50(4):780–787. of stenosis. Asian Oceania J Radiol. 1998;3:15–19.
28. McConnell MV, Ganz P, Selwyn AP, et al. Identification of anomalous coronary arter- 57. Hauser TH, Yeon S, Appelbaum E, et al. Discrimination of ischemic vs. non-ischemic
ies and their anatomic course by magnetic resonance coronary angiography. Circulation. cardiomyopathy among patients with heart failure using combined coronary MRI and
1995;92(11):3158–3162. delayed enhancement MR. J Cardiovasc Magn Reson. 2005;7:94.
29. Taylor AM, Thorne SA, Rubens MB, et al. Coronary artery imaging in grown up congeni- 58. Ichikawa Y, Sakuma H, Makino K, et al. Diagnostic accuracy of whole heart coronary
tal heart disease: Complementary role of magnetic resonance and x-ray coronary angiog- magnetic resonance angiography for the detection of significant coronary stenoses in
raphy. Circulation. 2000;101(14):1670–1678. patients with suspected coronary artery disease. J Cardiovasc Magn Reson. 2005;7:60.
30. Bunce NH, Lorenz CH, Keegan J, et al. Coronary artery anomalies: Assessment with free- 59. Jahnke C, Paetsch I, Nehrke K, et al. Rapid and complete coronary arteries tree visualiza-
breathing three-dimensional coronary MR angiography. Radiology. 2003;227(1):201–208. tion with magnetic resonance imaging. J Cardiovasc Magn Reson. 2005;7:221.
31. Post JC, van Rossum AC, Bronzwaer JG, et al. Magnetic resonance angiography of 60. Sakuma H, Ichikawa Y, Suzawa N, et al. Assessment of coronary arteries with total study
anomalous coronary arteries. A new gold standard for delineating the proximal course? time of less than 30 minutes by using whole-heart coronary MR angiography. Radiology.
Circulation. 1995;92(11):3163–3171. 2005;237(1):316–321.
32. Vliegen HW, Doornbos J, de Roos A, et al. Value of fast gradient echo magnetic resonance 61. Sakuma H, Ichikawa Y, Chino S, et al. Detection of coronary artery stenosis with whole-heart
angiography as an adjunct to coronary arteriography in detecting and confirming the course coronary magnetic resonance angiography. J Am Coll Cardiol. 2006;48(10)1946–1950.
of clinically significant coronary artery anomalies.Am J Cardiol. 1997;79(6):773–776. 62. Maintz D, Aepfelbacher FC, Kissinger KV, et al. Coronary MR angiography:
33. Razmi RM, Meduri A, Chun W, et al. Coronary magnetic resonance angiography Comparison of quantitative and qualitative data from four techniques. Am J Roentgenol.
(CMRA): The gold standard for determining the proximal course of anomalous coronary 2004;182(2):515–521.
arteries [abstract]. J Am Coll Cardiol. 2001;37:380. 63. Stuber M, Botnar RM, Danias PG, et al. Contrast agent-enhanced, free-breathing,
34. Muller MF, Fleisch M, Kroeker R, et al. Proximal coronary artery stenosis: Three- three-dimensional coronary magnetic resonance angiography. J Magn Reson Imaging.
dimensional MRI with fat saturation and navigator echo. J Magn Reson Imaging. 1999;10(5):790–799.
1997;7(4):644–651. 64. Herborn CU, Schmidt M, Bruder O, et al. MR coronary angiography with SH L 643
35. Woodard PK, Li D, Haacke EM, et al. Detection of coronary stenoses on source and pro- A: Initial experience in patients with coronary artery disease. Radiology. 2004;233(2):
jection images using three-dimensional MR angiography with retrospective respiratory 567–573.
gating: Preliminary experience. Am J Roentgenol. 1998;170(4):883–888. 65. Sommer T, Hackenbroch M, Hofer U, et al. Coronary HR angiography at 3.0 T versus
36. Sandstede JJ, Pabst T, Beer M, et al. Three-dimensional MR coronary angiography using that at 1.5 T: initial results in patients suspected of having coronary artery disease.
the navigator technique compared with conventional coronary angiography. AJR Am J Radiology. 2005;234:718–725.
Roentgenol. 1999;172(1):135–139. 66. Yang Q, Li K, Liu X, et al. Contrast-enhanced whole-heart coronary magnetic resonance
37. van Geuns RJ, de Bruin HG, Rensing BJ, et al. Magnetic resonance imaging of the coro- angiography at 3.0-T. A comparative study with x-ray angiography in a single center. J
nary arteries: Clinical results from three dimensional evaluation of a respiratory gated Am Coll Cardiol. 2009;54(1):69–76.
technique. Heart. 1999;82(4):515–519. 66b. van Elderen SG, Versluis MJ, Webb AG, et al. Initial results on in vivo human coronary
38. Huber A, Nikolaou K, Gonschior P, et al. Navigator echo-based respiratory gating HR angiography at 7T. Magn Reson Med. 2009;62(6):1379–1384.
for three-dimensional MR coronary angiography: Results from healthy volunteers 67. Pouleur AC, de Waroux JB, Kefer J, et al. Direct comparison of whole-heart navigator-
and patients with proximal coronary artery stenoses. Am J Roentgenol. 1999;173(1): gated magnetic resonance coronary angiography and 40- and 64-slice multidetector com-
95–101. puted tomography to detect the coronary artery stenosis in patients scheduled for conven-
39. Sardanelli F, Molinari G, Zandrino F, et al. Three-dimensional, navigator-echo MR coro- tional coronary angiography. Circ Cardiovasc Imaging. 2008;1(2):114–121.
nary angiography in detecting stenoses of the major epicardial vessels, with conventional 68. Hamdan A, Asbach P, Wellnhofer E, et al. A prospective study for comparison of MR and
coronary angiography as the standard of reference.Radiology. 2000;214(3):808–814. CT imaging for detection of coronary artery stenosis. J Am Coll Cardiol Img. 2011;4(1)
40. Bunce N, Rahman S, Jhooti P, et al. The assessment of coronary artery disease by com- 50–61.
bined magnetic resonance coronary arteriography and perfusion [abstract]. J Cardiovasc 69. Liu X, Zhao X, Francois CJ, et al. Comparison of free breathing 3D coronary MR angiog-
Magn Reson. 2001;3:118. raphy and 64-MDCT angiography for the detection of coronary stenoses in patients with
41. Moustapha AI, Proctor M, Muthupillai R, et al. Coronary magnetic resonance angiog- high calcium scores. AJR Am J Roentgenol. 2007;189(6)1326–1332.
raphy using a free breathing, T2 weighted, three-dimensional gradient echo sequence 70. Yoon YE, Kitagawa K, Kato S, et al. Prognostic value of coronary magnetic resonance
with navigator respiratory and ECG gating can be used to detect coronary artery disease angiography for prediction of cardiac events in patients with suspected coronary artery
[abstract]. J Am Coll Cardiol. 2001;37:380. disease. J Am Coll Cardiol 2012;60(22)2316–2322.
42. Sommer T, Hackenbroch M, Hofer U, et al. Submillimeter 3D coronary MR angiography 71. Greenwood J, Maredia N, Younger JF, et al. Cardiovascular magnetic resonance and
with real-time navigator correction in 112 patients with suspected coronary artery disease single-photon emission computed tomography for diagnosis of coronary artery disease
[abstract]. J Cardiovasc Magn Reson. 2002;4:28. (CE-MARC): A prospective trial. Lancet. 2012;379(9814)453–460.
43. Bogaert J, Kuzo R, Dymarkowski S, et al. Coronary artery imaging with real-time navi- 72. Syed MA, Murphy M, Arai AE. Safety of magnetic resonance imaging in the first few
gator three-dimensional turbo-field-echo MR coronary angiography: Initial experience. days after primary angioplasty/stenting for acute myocardial infarction. Circulation.
Radiology. 2003;226(3):707–716. 2004;110:17.
44. Plein S, Ridgway JP, Jones TR, et al. Coronary artery disease: Assessment with a compre- 73. Patel MR, Honeycutt E. Clinical safety of cardiac MRI prior to discharge in patients with
hensive MR imaging protocol—initial results.Radiology. 2002;225(1):300–307. acute myocardial infarction and coronary artery stenting with both bare metal and drug-
45. Ozgun M, Hoffmeier A, Quante M, et al. Comparison of spoiled TFE and balanced TFE eluting stents. J Cardiovasc Magn Reson. 2005;7:50.
coronary MR angiography [abstract]. J Cardiovasc Magn Reson. 2004;6:268–269. 74. Strohm O, Kivelitz D, Gross W, et al. Safety of implantable coronary stents dur-
46. Akagi T, Rose V, Benson LN, et al. Outcome of coronary artery aneurysms after Kawasaki ing 1H-magnetic resonance imaging at 1.0 and 1.5 T. J Cardiovasc Magn Reson.
disease. J Pediatr. 1992;121(5 pt 1):689–694. 1999;1(3):239–245.
47. Kato H, Ichinose E, Yoshioka F, et al. Fate of coronary aneurysms in Kawasaki disease: 75. Kramer CM, Rogers WJ Jr, Pakstis DL. Absence of adverse outcomes after magnetic reso-
Serial coronary angiography and long-term follow-up study. Am J Cardiol. 1982;49(7): nance imaging early after stent placement for acute myocardial infarction: A preliminary
1758–1766. study. J Cardiovasc Magn Reson. 2000;2(4):257–261.
48. Greil GF, Stuber M, Botnar RM, et al. Coronary magnetic resonance angiography in ado- 76. Gerber TC, Fasseas P, Lennon RJ, et al. Clinical safety of magnetic resonance imaging
lescents and young adults with Kawasaki disease.Circulation. 2002;105(8):908–911. early after coronary artery stent placement. J Am Coll Cardiol. 2003;42(7):1295–1298.
49. Mavrogeni S, Papadopoulos G, Douskou M, et al. Magnetic resonance angiography 77. Hug J, Nagel E, Bornstedt A, et al. Coronary arterial stents: Safety and artifacts during
is equivalent to x-ray coronary angiography for the evaluation of coronary arteries in MR imaging. Radiology. 2000;216(3):781–787.
Kawasaki disease. J Am Coll Cardiol. 2004;43(4):649–652. 78. Scott NA, Pettigrew RI. Absence of movement of coronary stents after placement in a
50. Mavrogeni S, Manginas A, Papadakis E, et al. Correlation between magnetic resonance magnetic resonance imaging field. Am J Cardiol. 1994;73(12):900–901.
angiography (MRA) and quantitative coronary angiography (QCA) in ectatic coronary 79. Shellock FG, Shellock VJ. Metallic stents: Evaluation of MR imaging safety. Am J
vessels. J Cardiovasc Magn Reson. 2004;6(1):17–23. Roentgenol. 1999;173(3):543–547.
51. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance angiography for the 80. Schmiedel A. MR imaging at 3.0 tesla and coronary artery stents. Ex vivo evaluation of
detection of coronary stenoses. N Engl J Med. 2001;345(26):1863–1869. magnetic attraction forces and heating [abstract]. J Cardiovasc Magn Reson. 2003;5:17–18.
52. Kato S, Kitagawa K, Ishida N, et al. Assessment of coronary artery disease using mag- 81. Spuentrup E, Ruebben A, Schaeffter T, et al. Magnetic resonance—guided coronary artery
netic resonance coronary angiography. A national multicenter trial. J Am Coll Cardiol. stent placement in a swine model. Circulation. 2002;105(7):874–879.
2010;56(12)983–991. 82. Maintz D, Botnar RM, Fischbach R, et al. Coronary magnetic resonance angiogra-
53. Pennell DJ, Bogren HG, Keegan J, et al. Assessment of coronary artery stenosis by mag- phy for assessment of the stent lumen: A phantom study. J Cardiovasc Magn Reson.
netic resonance imaging. Heart. 1996;75(2):127–133. 2002;4(3):359–367.
54. Manning WJ, Li W, Edelman RR. A preliminary report comparing magnetic resonance 83. Spuentrup E, Ruebben A, Mahnken AH, et al. Artifact-free coronary MR angiography
coronary angiography with conventional angiography. N Engl J Med. 1993;328(12): and coronary vessel wall imaging in the presence of a new metallic coronary MRI stent.
828–832. Circulation. 2005;111:1019–1026.
LWBK1209-ch21_p324-341.indd 340 16/05/13 9:34 PM

84. Desai ND, Cohen EA, Naylor CD, et al. Randomized comparison of radial-artery and 93. Vrachliotis TG, Bis KG, Aliabadi D, et al. Contrast-enhanced breath-hold MR angi-
saphenous-vein coronary bypass grafts. N Engl J Med. 2004;351:2302–2309. ography for evaluating patency of coronary artery bypass grafts. Am J Roentgenol.
85. Fitzgibbon GM, Kafka HP, Leach AJ, et al. Coronary bypass graft fate and patient out- 1997;168(4):1073–1080.
come: Angiographic follow-up of 5,065 grafts related to survival and reoperation in 1,388 94. Wintersperger BJ, Engelmann MG, von Smekal A, et al. Patency of coronary bypass
patients during 25 years. J Am Coll Cardiol. 1996;28(3):616–626. grafts: Assessment with breath-hold contrast-enhanced MR angiography—value of a non-
86. van Geuns RJ, Wielopolski PA, de Bruin HG, et al. MR coronary angiography with breath- electrocardiographically triggered technique. Radiology. 1998;208(2):345–351.
hold targeted volumes: Preliminary clinical results. Radiology. 2000;217(1):270–277. 95. Bunce NH, Lorenz CH, John AS, et al. Coronary artery bypass graft patency: Assessment
87. White RD, Caputo GR, Mark AS, et al. Coronary artery bypass graft patency: Noninvasive with true fast imaging with steady-state precession versus gadolinium-enhanced MR angi-
evaluation with MR imaging. Radiology. 1987;164(3):681–686. ography. Radiology. 2003;227(2):440–446.
88. White RD, Pflugfelder PW, Lipton MJ, et al. Coronary artery bypass grafts: Evaluation of 96. Rubinstein RI, Askenase AD, Thickman D, et al. Magnetic resonance imaging to evaluate
patency with cine MR imaging. Am J Roentgenol. 1988;150(6):1271–1274. patency of aortocoronary bypass grafts. Circulation. 1987;76(4):786–791.
89. Galjee MA, van Rossum AC, Doesburg T, et al. Value of magnetic resonance imaging 97. Jenkins JP, Love HG, Foster CJ, et al. Detection of coronary artery bypass graft patency
in assessing patency and function of coronary artery bypass grafts. An angiographically as assessed by magnetic resonance imaging. Br J Radiol. 1988;61(721):2–4.
controlled study. Circulation. 1996;93(4):660–666. 98. Langerak SE, Kunz P, Vliegen HW, et al. MR flow mapping in coronary artery bypass
90. Aurigemma GP, Reichek N, Axel L, et al. Noninvasive determination of coronary artery bypass grafts: A validation study with Doppler flow measurements. Radiology. 2002;222(1):127–
graft patency by cine magnetic resonance imaging. Circulation. 1989;80(6):1595–1602. 135.
91. Engelmann MG, Knez A, von Smekal A, et al. Non-invasive coronary bypass graft imag- 99. Langerak SE, Vliegen HW, Jukema JW, et al. Value of magnetic resonance imaging for the
ing after multivessel revascularisation. Int J Cardiol. 2000;76(1):65–74. noninvasive detection of stenosis in coronary artery bypass grafts and recipient coronary
92. Molinari G, Sardanelli F, Zandrino F, et al. Value of navigator echo magnetic resonance arteries. Circulation. 2003;107(11):1502–1508.
angiography in detecting occlusion/patency of arterial and venous, single and sequential 100. Langerak SE, Vliegen HW, de Roos A, et al. Detection of vein graft disease using high-
coronary bypass grafts. Int J Card Imaging. 2000;16(3):149–160. resolution magnetic resonance angiography. Circulation. 2002;105(3):328–333.
LWBK1209-ch21_p324-341.indd 341 16/05/13 9:34 PM

CHAPTER
22 Hajime Sakuma
Charles B. Higgins
Coronary Blood Flow Measurements
■■ Magnetic Resonance Imaging Techniques for of the coronary flow velocity reserve were 94%, 95%, and
Measuring Coronary Blood Flow 94%, respectively, when stress thallium-201 single-photon
emission computed tomography was used as a gold stan-
■■ Magnetic Resonance Measurement of
dard (6). Another study demonstrated that a coronary flow
Coronary Flow Reserve in Patients with velocity reserve below 2 by the Doppler technique had a
Coronary Arterial Stenosis sensitivity of 92% and specificity of 82% for predicting the
■■ Magnetic Resonance Flow Measurement presence of a significant stenosis in the coronary artery on
in the Coronary Artery Bypass Graft selective coronary angiography (7). However, intracoronary
flow velocity measurement with a Doppler guidewire is inva-
■■ Magnetic Resonance Blood Flow sive and available only during cardiac catheterization.
Quantification in the Coronary Sinus Fast phase-contrast cine magnetic resonance imaging
■■ Summary (MRI) is an emerging application of MRI that can provide
noninvasive assessments of blood flow and flow reserve in
Selective coronary angiography has been used to evaluate human coronary arteries. Several studies have demonstrated
coronary artery disease. However, an assessment of the the usefulness of this technique in detecting coronary ar-
anatomic severity of a coronary stenosis does not ade- terial restenosis after percutaneous revascularization pro-
quately determine the functional significance of the lesion cedures and in assessing patency and stenosis in coronary
(1,2). Quantitative coronary angiography, which was artery bypass conduits. In addition, magnetic resonance
designed to minimize variability in interpretation, cannot (MR) measurement of the coronary sinus blood flow allows
reliably predict the physiologic significance of a stenosis of the noninvasive assessment of global myocardial blood flow.
intermediate severity (3). An assessment of the functional This chapter reviews the current status and potential clinical
significance of a stenosis is particularly important in lesions applications of MR measurements of blood flow and flow
with intermediate severity because the interpretation of reserve in the coronary artery during coronary artery bypass
such lesions significantly influences therapeutic decisions grafting (CABG).
in patients with coronary artery disease. The functional
significance of a coronary arterial stenosis can be evaluated
by measuring the coronary flow reserve, which is the ratio Magnetic Resonance Imaging
of maximal hyperemic coronary flow to the baseline coro- Techniques for Measuring
nary flow (4,5). In the presence of normal epicardial coro- Coronary Blood Flow
nary artery and normal myocardial microcirculation, the
administration of a vasodilator (e.g., dipyridamole and Phase-contrast cine MRI can provide noninvasive flow mea-
adenosine) induces approximately a three- to four-fold surements at multiple temporal phases in the cardiac cycle.
increase in coronary blood flow. In patients with signifi- The MR measurement of blood flow in large vessels—such
cant stenoses in the coronary arteries, however, compensa- as the aorta, pulmonary artery, and carotid artery—has been
tory dilatation takes place in the downstream microcircu- well established and validated (8). However, MR blood flow
lation to maintain myocardial blood flow. Thus, the ability quantification in the coronary artery has been very challeng-
to augment coronary blood flow during pharmacologic ing because the coronary artery is small (<3 to 4 mm) and
stress is attenuated in patients with significant coronary subject to both cardiac and respiratory motion. Several dif-
arterial stenosis. ferent approaches have been proposed in the past for quanti-
The evaluation of blood flow velocity and flow velocity fying blood flow in the coronary arteries with MRI, including
reserve with an intracoronary Doppler guidewire allows a a time-of-flight method (9), bolus tagging (10), and a phase-
functional assessment of the severity of a stenosis. A study in contrast method (11). A phase-contrast cine MRI has been
which an intracoronary Doppler guidewire was used showed most commonly used to measure blood flow in the human
that the sensitivity, specificity, and overall predictive accuracy coronary vessels in recent studies.
342
LWBK1209-ch22_p342-349.indd 342 16/05/13 9:33 PM

Chapter 22 ■ Coronary Blood Flow Measurements 343
A B
C D
Figure 22.1. Breath-hold phase-contrast cine magnetic resonance (MR) images of the left anterior descend-
ing artery (arrows) in a subject without significant stenosis in the left anterior descending artery. Magnitude
image (A) and phase-difference image (B) at rest and magnitude image (C) and phase-difference image (D)
after intravenous administration of dipyridamole. The increase of the coronary blood flow velocity after phar-
macologic stress is observed as a darkening of the signal on phase-difference image.
MR flow measurement in the coronary artery can be measurement was 4.2 ± 1.8 according to Grist et al. (16) and
achieved with breath-hold acquisition or respiratory-triggered 5 ± 2.6 according to Davis et al. (17).
acquisition. In 1993, Edelman et al. (11) demonstrated the Errors from several sources affect blood flow quantifica-
feasibility of breath-hold measurement of blood flow ve- tion with phase-contrast MR sequences (18–20). Currently
locity in the coronary artery at a single diastolic phase with the major limitations of breath-hold MR measurement of
the use of a segmented k-space, phase-contrast MR acquisi- coronary blood flow are suboptimal spatial resolution and
tion. Keegan et al. (12) then measured coronary blood flow temporal resolution. If the spatial resolution of MRI is com-
at multiple phases in the cardiac cycle within a single breath- promised in relation to the vessel size, partial volume aver-
hold time by using segmented k-space, phase-contrast cine aging of bright signal from flowing blood and attenuated
MRI. We measured diastolic peak velocities in the left ante- signal from static tissue around the vessel lumen result in
rior descending artery at rest and during dipyridamole stress an overestimation of blood flow volume on phase-difference
with this technique (13). The average coronary flow velocity images. In addition, the peak flow velocity in the coronary
reserve in eight healthy subjects was 3.14 ± 0.59 (Fig. 22.1). artery measured by phase-contrast MRI is substantially
Volumetric blood flow in the coronary arteries can be cal- lower than the peak velocity by intracoronary Doppler
culated in theory by integrating the product of the mean ve- guidewire because the highest velocity at the center and
locity and area of the vessel over the cardiac cycle (14–17). lower velocities at the peripheral areas of the vessel lumen
The normal coronary flow reserve with volumetric MR flow are averaged within MRI voxel (21). The lower velocities
LWBK1209-ch22_p342-349.indd 343 16/05/13 9:33 PM

that were obtained by phase-contrast MRI, compared with temporal resolution (30 milliseconds) and substantially re-
those obtained with intravascular Doppler guidewire, were duced underestimation of flow velocity that was inevitable
also observed in CABG (22). Insufficient temporal resolu- with a breath-hold FLASH sequence.
tion of the cardiac cycle is another source of the error. If
the data acquisition window in the cardiac cycle is insuf-
ficient, blurring of the vessel by in-plane motion results in Magnetic Resonance
inaccurate measurements of blood flow velocity and flow Measurement of Coronary
volume. Hofman et al. (23) reported that the duration of Flow Reserve in Patients with
the acquisition window for an accurate MR flow volume Coronary Arterial Stenosis
quantification should be less than 25 milliseconds for the
right coronary artery and 120 milliseconds for the left coro- Measurements of coronary flow reserve obtained with phase-
nary artery. These figures indicate that further improvement contrast cine MRI can be used to evaluate the functional
is necessary to obtain accurate flow volume quantification in significance of stenoses in the left main and left anterior
the right coronary artery. It should be noted that blood flow descending coronary arteries in patients (Fig. 22.2). To vali-
measured during breath-hold MR acquisition may be differ- date this approach, Shibata et al. (21) compared coronary
ent from physiologic blood flow during regular breathing flow velocity reserve measurements obtained by breath-hold
because breath-holding at inspiration causes an increase in phase-contrast cine MRI with measurements by intracoro-
intrathoracic pressure, which then reduces systemic venous nary Doppler guidewire in 19 patients with varying degrees
return to the heart. Our study showed that MR measure- of stenosis in the left anterior descending arteries. Although
ments of cardiac output are significantly depressed during the mean MR flow velocity in the coronary artery in the
breath-holding at deep inspiration (24). No significant dif- baseline state was significantly lower than that assessed by
ference was observed between breath-hold measurement Doppler guidewire (12.5 ± 4.9 cm/sec vs. 32.4 ± 12.1 cm/
at shallow inspiration and nonbreath-hold measurement, a sec, p < 0.001), a significant linear correlation was observed
finding that indicates the importance of using shallow inspi- between MR and Doppler measurements of the coronary
ration to obtain physiologic blood flow in breath-hold MR flow velocity reserve, with a correlation coefficient of 0.91.
sequences. In another study (29), breath-hold MR measurement of the
Phase-contrast cine MR sequences with respiratory- coronary flow velocity reserve in the proximal left anterior
triggering acquisition permit the nonbreath-hold assessment descending artery correlated well with myocardial perfusion
of coronary blood flow. Navigator techniques are based reserve in the anterior wall of the left ventricle (LV) mea-
on one-dimensional image acquisitions; these are used to sured by positron emission tomography (PET) and water
monitor the position of the diaphragm and accept image labeled with oxygen-15.
data only if the position of the diaphragm is within a cer- The noninvasive MR measurement of coronary flow re-
tain range. Since image data can be acquired for a prolonged serve has been shown to be useful in identifying the func-
scan time during free breathing, navigator technique makes tional significance of stenoses in the left anterior descending
it possible to acquire phase-contrast cine MR images ac- artery in patients with coronary artery disease. Hundley
quired with better spatial and temporal resolutions. Nagel et al. (30) demonstrated a statistically significant correlation
et al. (25) reported that a higher degree of temporal reso- between MR assessments of coronary flow reserve and the
lution and shorter acquisition window in the cardiac cycle
with navigator-corrected nonbreath-hold MRI increased the
accuracy of measurements of coronary blood flow velocity,
35
especially in the right coronary arteries.
The use of fast scan techniques can improve the accu- 30
During stress
racy of breath-hold MR flow quantification in the coronary
vessels (26). In comparison with respiration-triggered acqui- 25
sition, the breath-hold method requires less scan time; this
Velocity
20
(cm/s)
feature is critically important in studies in which pharmaco-

logic stress is used and makes it possible to assess blood flow 15
and flow reserve in multiple coronary arteries within a short
scan time. Langerak et al. (27) used turbo-field echo-planar 10 At rest
MRI to measure blood flow in CABG. In this approach,
5
breath-hold flow mapping with spatial and temporal reso-
lutions of 0.8 mm2 and 23 milliseconds, respectively, can 0
be achieved. These investigators showed that breath-hold 0 100 200 300 400 500 600 700 800
turbo-field echo-planar MRI provides fast and accurate flow Time after R wave trigger (ms)
measurements and allows motion-compensated flow quanti-
Figure 22.2. Blood flow velocity curves in the left anterior
fication in multiple CABGs during a short infusion of ade-
descending artery at rest and after intravenous administration of dipyr-
nosine. In a recent study by Keegan et al. (28), high temporal idamole in a patient with significant stenosis. Coronary flow velocity
resolution spiral sequences were developed. A free-breathing reserve can be calculated from flow velocity curves of the coronary
spiral sequence reduced MRI time by a factor of 10 when artery before and during pharmacologic stress. The coronary flow
compared with a free-breathing FLASH sequence. In addi- velocity reserve in this patient (1.9) was reduced because of the steno-
tion, a breath-hold spiral sequence can provide an excellent sis in the left anterior descending artery.
LWBK1209-ch22_p342-349.indd 344 16/05/13 9:33 PM

3.5 is a routine procedure and gold standard for assessing bypass

Patients without grafts, but this technique is invasive and entails some risks.
3.0
restenosis MRI techniques such as ECG-gated spin-echo MRI (34,35),
2.5 cine MRI (36,37), and three-dimensional contrast-enhanced
Flow velocity
LAD artery
MR angiography (38,39) have been shown to be useful for

reserve in
2.0
predicting graft patency. However, these MRI methods of
1.5 morphologic assessment do not differentiate between non-
occluded stenotic grafts and normal grafts (40). MR flow
1.0
Patients with measurement is an effective method of functionally assessing
0.5 restenosis a CABG, and it noninvasively detects stenosis and occlusion
in a graft.
0 MR blood flow quantification in saphenous vein grafts
1 2 3 4 5 6
was initially reported by Hoogendoorn et al. (41), who
Months after coronary interventions
used nonbreath-hold MRI. They reported that the average
Figure 22.3. Serial changes of the coronary flow velocity reserve blood flow volume in normal saphenous vein grafts was 71
after percutaneous balloon angioplasty and stent placement in the ± 17 mL/min, whereas the average blood flow volume in
left anterior descending arteries in patients without restenosis and the stenotic/occluded grafts was significantly reduced to 9
in patients with significant stenosis at follow-up selective coronary ± 8 mL/min (p < 0.001). In the study of Galjee et al. (42),
angiography. The coronary flow velocity reserve gradually decreased
adequate MR flow velocity profiles were obtained in 85%
after several months in the patients who demonstrated restenosis in the
of the angiographically patent grafts with nonbreath-hold
coronary artery.
phase-contrast cine MRI. Graft flow was characterized
by a biphasic pattern, with one peak during systole and a
severity of coronary arterial stenosis by quantitative coro- second peak during diastole. A recent study demonstrated
nary angiography. The sensitivity and specificity of MR cor- that MR flow measurement in vein grafts is useful in assess-
onary flow velocity reserve for identifying stenosis of 70% ing functional improvement after percutaneous intervention
or greater in the left main or left anterior descending arteries (43). The feasibility of MR blood flow measurements in an
were 100% and 83%, respectively. internal mammary artery graft was demonstrated with the
Restenosis of the coronary artery is a major clinical prob- use of nonbreath-hold phase-contrast cine MRI (44) and
lem in patients who undergo coronary interventions. Saito breath-hold phase-contrast cine MRI (45). Ishida et al. (46)
et al. (31) showed that restenosis after coronary balloon an- measured blood flow in internal mammary artery bypass
gioplasty and stent placement can be detected noninvasively grafts in 26 patients with a breath-hold phase-contrast cine
by monitoring serial changes of the coronary flow velocity re- MR sequence (Figs. 22.4 and 22.5). The ratio of diastolic to
serve using breath-hold phase-contrast MRI. In patients with- systolic peak velocity in patients with stenotic grafts (>70%
out restenosis at follow-up x-ray angiography after 6 months, diameter) on selective angiography was 0.61 ± 0.44, which
the flow velocity reserve was 1.97 ± 0.37 at 1 month and was significantly lower than that in patients without graft
2.29 ± 0.31 at 6 months. In contrast, the coronary flow ve- stenosis (1.88 ± 0.96, p < 0.01). In addition, the average
locity reserve showed a gradual decrease in patients with blood flow volume in stenotic grafts (16.9 ± 5.5 mL/min)
restenoses at follow-up x-ray angiography. The coronary was significantly lower than that in normal grafts (79.8 ±
flow velocity reserve was 2.27 ± 0.49 at 1 month and 1.52 38.2 mL/min, p < 0.01). The sensitivity and specificity of
± 0.15 at 6 months (Fig. 22.3). A study by Hundley et al. MR blood flow measurement in predicting significant ste-
(32) also indicated that phase-contrast cine MRI can be used nosis were 85.7% and 94.1%, respectively. These results
to detect restenosis noninvasively in patients with recurrent indicated that MR blood flow measurement is useful in
chest pain after percutaneous revascularization. A coronary predicting significant stenoses in internal mammary artery
flow reserve value of 2 or less was 100% sensitive and 82% grafts (Fig. 22.6).
specific for detecting a luminal diameter narrowing of 70% Recent studies demonstrated that MR measurement of
or greater, and it was 82% sensitive and 100% specific for the flow reserve is useful for identifying stenosis in the vein
detecting a luminal diameter narrowing of 50% or greater. grafts. Langerak et al. (27) measured blood flow in 20 nor-
In another study by Nagel et al. (33), coronary flow velocity mal grafts before and during adenosine stress. The mean by-
reserve in the artery distal to the stent could be determined in pass graft flow increased from 30.8 ± 13.5 mL/min to 76.7
29 of 38 patients (76%) after successful percutaneous coro- ± 36.5 mL/min (p < 0.05) to yield a flow reserve of 2.7.
nary intervention. By using a threshold of 1.2, a sensitivity Bedaux et al. (47) determined the value of MR flow and flow
of 83% and a specificity of 94% were achieved for detecting reserve measurements in detecting significant vein graft dis-
stenosis of 75% or greater. ease. An algorithm combining rest graft flow of 20 mL/min
or less and graft flow reserve of 2 or less had a sensitivity and
specificity of 78% and 80%, respectively. In a recent study
Magnetic Resonance Flow by Langerak et al. (48), MR flow velocity measurements
Measurement in the Coronary were obtained at rest and during stress in 69 patients (time
Artery Bypass Graft after CABG surgery 9.3 ± 5.4 years). MR flow reserve mea-
surement was successful in 80% of grafts. The flow velocity
In many circumstances, a CABG must be assessed postopera- reserve allowed an accurate differentiation between single
tively for patency or stenosis. Selective coronary angiography vein grafts with and without stenosis of 70% or greater as
LWBK1209-ch22_p342-349.indd 345 16/05/13 9:33 PM

A B A B
30 45
40
25 35
30
20 25
Velocity
Velocity
(cm/(s)
(cm/s)
20
15
15
10 10
5
5 Systolic peak < Diastolic peak 0
–5 Systolic peak > Diastolic peak
0 –10
0 100 200 300 400 500 600 700 0 100 200 300 400 500
C Time after ECG R wave (ms) C Time after ECG R wave (ms)
Figure 22.4. Magnitude (A) and phase-difference (B) MR images Figure 22.5. Magnitude (A) and phase-difference (B) MR images
acquired with phase-contrast cine MR imaging (MRI) in a patient acquired with phase-contrast cine MRI in a patient with significant
without significant stenosis in the internal mammary artery bypass stenosis in the internal mammary artery bypass conduit (distal anasto-
conduit (distal anastomosis without significant stenosis to the left mosis with significant stenosis to the left descending artery [arrow]).
descending artery [arrow]). Internal mammary artery graft was dem- Blood flow velocity curve in the graft with stenosis (C) showed pre-
onstrated as a small bright area on magnitude image (A), and flow in dominant flow during systolic phase.
the vessel was visually recognized on phase-difference image (B). Blood
flow velocity curve in the graft without stenosis (C) was characterized
by predominant flow during diastolic phase.
200 p < 0.01
reflected by the receiver operating characteristic (ROC) area

of 0.89 (optimal cutoff value of 1.43). In sequential vein
grafts, multivariate analysis of combined baseline and stress 160
velocity parameters was required to obtain a high diagnos-
tic accuracy (ROC area 0.88) because the velocity reserve
alone could not reliably differentiate between grafts with
flow volume
MR blood
and without stenosis (ROC area 0.70). Salm et al. (49) re-
(mL/min)
120
ported that the coronary flow velocity reserve and coronary
flow volume reserve had equivalent diagnostic accuracy in
the detection of significant stenosis in single vein grafts. The 79.8
80
mean difference between the coronary flow velocity reserve ±38.2
and coronary flow volume reserve was not significant, and
the area under the ROC curve was 0.96 for the flow velocity
reserve and 0.93 for the flow volume reserve in their study. 40
When Ishida et al. (46) evaluated pharmacologic flow re-
serve in the internal mammary artery grafts with dipyridam- 16.9
ole, the mean flow velocity reserve ratio in the grafts with ±5.5
significant stenoses (1.39 ± 1.46) was reduced in comparison 0
with that in normal grafts (2 ± 1.43, p < 0.01). However, this Patients without Patients with
study indicated that MR measurement of the vasodilator flow graft stenosis graft stenosis
reserve does not further increase the detection of graft stenosis Figure 22.6. Blood flow volume in the internal mammary artery
in patients early after operation because blood flow volume grafts in patients with and without significant stenosis on selective
and the ratio of diastolic to systolic velocity measured at rest angiography. MR blood flow volume <35 mL/min predicts significant
effectively discriminate between normal and stenotic grafts. graft stenosis (>70% diameter stenosis) with sensitivity and specificity
The limited diagnostic value of flow reserve measurements for of 85.7% and 94.1%, respectively.
LWBK1209-ch22_p342-349.indd 346 16/05/13 9:33 PM

detecting stenosis in internal mammary arterial grafts early LV myocardial mass is measured with cine MRI, both the total
after operation was mainly a consequence of the small di- myocardial blood flow and the average coronary blood flow
ameter and limited flow reserve ratios of the normal grafts. per gram of myocardial mass can be quantified (Fig. 22.7). To
Akasaka et al. (50) measured flow velocity in internal mam- validate this approach, several investigators compared MR
mary artery grafts by Doppler guidewire and showed that in- measurements of coronary sinus blood flow with the val-
ternal mammary artery grafts early after operation are char- ues determined by reference methods. Lund et al. (52) com-
acterized by a higher peak velocity at baseline to compensate pared blood flow in the coronary sinus measured by MRI
for their small diameter. The early postoperative flow reserve with that assessed by an ultrasonic volumetric flow meter
(1.8 ± 0.3) is significantly lower than in old grafts (2.6 ± 0.3) in dogs. Coronary sinus flow measured by phase-contrast
because of the high baseline flow velocity. Further studies are cine MRI and total coronary blood flow measured by flow-
required to determine the diagnostic value of pharmacologic meter showed a good correlation (r = 0.98, p < 0.001); the
flow reserve in internal mammary artery grafts. mean difference between total coronary blood flow by flow-
meter and coronary sinus flow by MRI was 3.1 ± 8.5 mL/
min. Mean blood flow per gram of myocardial mass was
Magnetic Resonance Blood 0.40 ± 0.09 mL/min/g by MRI and 0.44 ± 0.08 mL/min/g
Flow Quantification in by flowmeter (p = not significant). Schwitter et al. (53) mea-
the Coronary Sinus sured coronary sinus blood flow by phase-contrast cine MRI
and myocardial blood flow by PET and [13N] ammonia in
Blood flow in the coronary sinus reflects the global myocardial healthy volunteers. Coronary sinus flow divided by LV mass
blood flow because it represents approximately 96% of the correlated highly with the 13N-PET flow data (0.77 ± 0.19
total myocardial blood flow of the LV (51). If blood flow in the mL/min/g by MRI and 0.73 ± 0.15 mL/min/g by PET [r =
coronary sinus is measured with phase-contrast cine MRI and 0.95]). Koskenvuo et al. (54,55) compared breath-hold MR
A B
1
Volume flow (mL/sec)
0.8
0.6
0.4
0.2
–0.2
0 100 200 300 400 500 600 700 800
C Time after R wave (ms)
Figure 22.7. MR measurement of blood flow volume through the coronary sinus. Magnitude (A) and
phase-difference (B) images were acquired on the imaging plane that is perpendicular to the coronary sinus
using a phase-contrast cine MR sequence. Volume flow curve (C) can be generated by integrating products of
blood flow velocity and vessel area over the cardiac cycle.
LWBK1209-ch22_p342-349.indd 347 16/05/13 9:33 PM

measurements of coronary blood flow per gram of myocar- The MR measurement of blood flow through the coro-
dial mass with myocardial blood flow determined by PET nary sinus allows a noninvasive assessment of global cor-
and oxygen-15–labeled water, which is regarded as the most onary hemodynamics. Since the coronary sinus is larger
accurate approach for measuring myocardial blood flow. than the native coronary arteries, the errors in blood flow
Coronary sinus blood flow divided by myocardial mass was quantification associated with the limited spatial and tem-
0.65 ± 0.20 mL/min/g with breath-hold MRI, which showed poral resolution of current phase-contrast MR sequences
a good agreement with the PET measurement of myocardial are substantially smaller. The MR quantification of total
blood flow of 0.65 ± 0.20 mL/min/g. coronary blood flow and coronary blood flow per gram of
This integrated MR approach for measuring global myo- myocardium during stress and at rest seems to be an ideal
cardial blood under resting and hyperemic conditions has method for evaluating coronary hemodynamics in diffuse
been shown to be useful in evaluating diffuse myocardial myocardial diseases of the heart (57), such as hypertensive
diseases, such as hypertrophic cardiomyopathy and cardiac heart diseases and cardiomyopathies, and may be useful in
transplant. In a study in 29 patients with hypertrophic car- evaluating endothelial dysfunction of the coronary circula-
diomyopathy, the baseline myocardial blood flow was 0.74 tion, which develops in the earliest stage of atherosclerotic
± 0.23 mL/min/g in healthy subjects and 0.62 ± 0.27 mL/ disease and may precede obstruction of the epicardial coro-
min/g in patients with hypertrophic cardiomyopathy (p = nary arteries.
not significant) (56). After administration of dipyridamole,
the myocardial blood flow in patients with hypertrophic car-
diomyopathy increased to a level significantly less than that in References
healthy subjects (1.03 ± 0.40 mL/min/g vs. 2.14 ± 0.51 mL/ 1. White CW, Wright CB, Doty DB, et al. Does visual interpretation of the coronary
min/g, p < 0.01), so that the patients with hypertrophic car- angiogram predict the physiological importance of a coronary stenosis? N Engl J Med.
1984;310:819–824.
diomyopathy had a severely decreased flow reserve ratio 2. Marcus ML, Skorton DJ, Johnson MR, et al. Visual estimates of percent diameter coronary
(1.72 ± 0.49 vs. 3.01 ± 0.75, p <0.01). A significant inverse stenosis: A battered gold standard. J Am Coll Cardiol. 1988;11:882–885.
relationship was found between the coronary flow reserve 3. Vogel RA. Assessing stenosis significance by coronary angiography. Are the best variables
good enough? J Am Coll Cardiol. 1988;12:692–693.
ratio and LV mass (r = 0.46, p < 0.05). Schwitter et al. (53) 4. Could KL, Lipscomb K, Hamilton GW. Physiologic basis for assessing critical coronary
evaluated myocardial blood flow and flow reserve in recipi- stenosis: Instantaneous flow response and regional distribution during coronary hyperemia
ents of orthotopic heart transplants with phase-contrast as measures of coronary flow reserve. Am J Cardiol. 1974;33:87–94.
5. White CW. Clinical applications of Doppler coronary flow reserve measurements. Am J
MRI. The coronary flow reserve was reduced in patients Cardiol. 1993;71:10D–16D.
with transplanted hearts (2 ± 0.4) in comparison with that 6. Joye JD, Schulman DS, Lasorda D, et al. Intracoronary Doppler guide wire versus stress
single-photon emission computed tomographic thallium-201 imaging in assessment of
in control subjects (3.9 ± 1.4, p < 0.005). A moderate but intermediate coronary stenoses. J Am Coll Cardiol. 1994;24:940–947.
significant negative correlation was noted between myo- 7. Redberg RF, Sobol Y, Chou TM, et al. Adenosine induced coronary vasodilatation during
cardial blood flow after pharmacologic stress and LV mass transesophageal Doppler echocardiography: Rapid and safe measurement of coronary flow
reserve ratio can predict significant left anterior descending coronary stenosis. Circulation.
(r = 0.56, p < 0.005). 1995;92:190–196.
8. Szolar DH, Sakuma H, Higgins CB. Cardiovascular application of magnetic resonance flow
and velocity measurements. J Magn Reson Imaging. 1996;6:78–89.
Summary 9. Poncelet BP, Weisskoff RM, Wedeen WJ, et al. Time of flight quantification of coronary
flow with echo-planar MRI. Magn Reson Med. 1993;30:447–457.
10. Chao H, Burstein D. Multibolus stimulated echo imaging of coronary artery flow. J Magn
Recent advances in fast MRI sequences have considerably Reson Imaging. 1997;7:603–605.
11. Edelman RR, Manning WJ, Gervino E, et al. Flow velocity quantification in human coronary
extended the capabilities of cardiac MRI in the functional arteries with fast breath-hold MR angiography. J Magn Reson Imaging. 1993;3:699–703.
assessment of coronary blood flow and flow reserve. The 12. Keegan J, Firmin D, Gatehouse P, et al. The application of breath-hold phase velocity map-
ping techniques to the measurement of coronary artery blood flow velocity: Phantom data
MR measurement of coronary flow reserve can be used to and initial in vivo results. Magn Reson Med. 1994;31:526–536.
identify restenosis of the left main and left anterior descend- 13. Sakuma H, Blake LM, Amidon TM, et al. Noninvasive measurement of coronary flow
ing coronary arteries after percutaneous interventions in reserve in humans using breath-hold velocity encoded cine MR imaging.Radiology. 1996;
198:745–750.
clinical patients. Further refinements in MR pulse sequences, 14. Clarke GD, Eckels R, Chaney C, et al. Measurement of absolute epicardial coronary artery
including the use of phase-contrast echo-planar or spiral MR flow and flow reserve with breath-hold cine phase-contrast magnetic resonance imaging.
methods, may make it possible to quantify blood flow vol- Circulation. 1995;91:2627–2634.
15. Hundley WG, Lange RA, Clarke GD, et al. Assessment of coronary arterial flow and flow
ume and flow reserve more accurately in all major branches reserve in humans with magnetic resonance imaging. Circulation. 1996;93:1502–1508.
of the coronary arteries. 16. Grist TM, Polzin JA, Bianco JA, et al. Measurement of coronary blood flow and flow
reserve using magnetic resonance imaging. Cardiology. 1997;88:80–89.
The MR measurement of blood flow in CABG permits a 17. Davis CP, Liu P, Hauser M, et al. Coronary flow and coronary flow reserve measurements
functional evaluation of the graft and is useful in the nonin- in humans with breath-hold magnetic resonance phase contrast velocity mapping. Magn
vasive detection of graft stenosis. In most clinical patients Reson Med. 1997;37:537–544.
18. Polzin JA, Korosec FR, Wedding KL, et al. Effect of through-plane myocardial motion on
with coronary bypass conduits, however, it is necessary to phase-difference and complex-difference measurements of absolute coronary artery flow. J
know the status of the native coronary arteries or detect Magn Reson Imaging. 1996;6:113–123.
the presence of myocardial ischemia. Although the MR 19. Frayne R, Polzin JA, Mazaheri Y, et al. Effect of and correction for in-plane myocardial
motion on estimates of coronary volume flow rates. J Magn Reson Imaging. 1997;7:
flow measurement of flow in bypass grafts cannot by itself 815–828.
replace invasive coronary angiography, comprehensive 20. Hofman MBM, van Rossum AC, Sprenger M, et al. Assessment of flow in the right human
coronary artery by magnetic resonance phase contrast velocity measurement: Effects of
assessments of myocardial perfusion and blood flow in cardiac and respiratory motion. Magn Reson Med. 1996;35:521–531.
the coronary artery bypass conduits during stress and at 21. Shibata M, Sakuma H, Isaka N, et al. Assessment of coronary flow reserve with fast
rest within a single MR examination will be highly useful cine phase contrast magnetic resonance imaging: Comparison with the measurement by
Doppler guide wire. J Magn Reson Imaging. 1999;10:563–568.
after bypass graft surgery and may reduce the need for 22. Langerak SE, Kunz P, Vliegen HW, et al. MR flow mapping in coronary artery bypass grafts:
catheterization. A validation study with Doppler flow measurements. Radiology. 2002;222:127–135.
LWBK1209-ch22_p342-349.indd 348 16/05/13 9:33 PM

23. Hofman MB, Wickline SA, Lorenz CH. Quantification of in-plane motion of the coronary 41. Hoogendoorn LI, Pattynama PMT, Buis B, et al. Noninvasive evaluation of aortocoronary
arteries during the cardiac cycle: Implication for acquisition window duration for MR flow bypass grafts with magnetic resonance flow mapping. Am J Cardiol. 1995;75:845–848.
quantification. J Magn Reson Imaging. 1998;8:568–576. 42. Galjee MA, van Rossum AC, Doesburg T, et al. Value of magnetic resonance imaging in
24. Sakuma H, Kawada N, Kubo H, et al. Effect of breath holding on blood flow measurement assessing patency and function of coronary artery bypass grafts: An angiographically con-
using fast velocity encoded cine MRI. Magn Reson Med. 2001;45:346–348. trolled study. Circulation. 1996;93:660–666.
25. Nagel E, Bornstedt A, Hug J, et al. Noninvasive determination of coronary blood flow 43. Langerak SE, Vliegen HW, Jukema JW, et al. Vein graft function improvement after
velocity with magnetic resonance imaging: Comparison of breath-hold and navigator tech- percutaneous intervention: Evaluation with MR flow mapping. Radiology. 2003;228:
niques with intravascular ultrasound. Magn Reson Med. 1999;41:544–549. 834–841.
26. Sakuma H, Saeed M, Takeda K, et al. Quantification of coronary arterial volume flow rate 44. Debatin JF, Strong JA, Sostman HD, et al. MR characterization of blood flow in native and
using fast velocity encoded cine MR imaging. AJR Am J Roentgenol. 1997;168:1363–1367. grafted internal mammary arteries. J Magn Reson Imaging. 1993;3:443–450.
27. Langerak SE, Kunz P, Vliegen HW, et al. Improved MR flow mapping in coronary artery 45. Sakuma H, Globits S, O’Sullivan M, et al. Breath-hold MR measurements of blood flow
bypass grafts during adenosine-induced stress. Radiology. 2001;218:540–547. velocity in internal mammary arteries and coronary artery bypass grafts. J Magn Reson
28. Keegan J, Gatehouse PD, Mohiaddin RH, et al. Comparison of spiral and FLASH phase Imaging. 1996;6:219–222.
velocity mapping, with and without breath-holding, for the assessment of left and right 46. Ishida N, Sakuma H, Cruz BP, et al. MR flow measurement in the internal mammary artery
coronary artery blood flow velocity. J Magn Reson Imaging. 2004;19:40–49. to coronary artery bypass graft: A comparison with graft stenosis on X-ray angiography.
29. Sakuma H, Hoslenvuo JW, Niemi P, et al. Assessment of coronary flow reserve using fast Radiology. 2001;220:441–447.
velocity-encoded cine MR imaging: Validation study using positron emission tomography. 47. Bedaux WL, Hofman MB, Vyt SL, et al. Assessment of coronary artery bypass graft disease
AJR Am J Roentgenol. 2000;175:1029–1033. using cardiovascular magnetic resonance determination of flow reserve. J Am Coll Cardiol.
30. Hundley WG, Hamilton CA, Clarke GD, et al. Visualization and functional assessment of 2002;40:1848–1855.
proximal and middle left anterior descending coronary stenoses in humans with magnetic 48. Langerak SE, Vliegen HW, Jukema JW, et al. Value of magnetic resonance imaging for the
resonance imaging. Circulation. 1999;99:3248–3254. noninvasive detection of stenosis in coronary artery bypass grafts and recipient coronary
31. Saito Y, Sakuma H, Okinaka T, et al. Assessment of coronary flow reserve using cine phase arteries. Circulation. 2003;107:1502–1508.
contrast MRI for noninvasive detection of restenosis after angioplasty and stenting. J 49. Salm LP, Langerak SE, Vliegen HW, et al. Blood flow in coronary artery bypass vein grafts:
Cardiovasc Magn Reson. 2001;3(suppl 3):209–214. Volume versus velocity at cardiovascular MR imaging. Radiology. 2004;232:915–920.
32. Hundley WG, Hillis LD, Hamilton CA, et al. Assessment of coronary arterial restenosis 50. Akasaka T, Yoshikawa J, Yoshida K, et al. Flow capacity of internal mammary artery
with phase-contrast magnetic resonance imaging measurements of coronary flow reserve. grafts: Early restriction and later improvement assessed by Doppler guide wire; comparison
Circulation. 2000;101:2375–2381. with saphenous vein grafts. J Am Coll Cardiol. 1995;25:640–647.
33. Nagel E, Thouet T, Klein C, et al. Noninvasive determination of coronary blood flow veloc- 51. van Rossum AC, Visser FC, Hofman MBM, et al. Global left ventricular perfusion:
ity with cardiovascular magnetic resonance in patients after stent deployment. Circulation. Noninvasive measurement with cine MR imaging and phase velocity mapping of coronary
2003;107:1738–1743. venous outflow. Radiology. 1992;182:685–691.
34. White RD, Caputo GR, Mark AS, et al. Coronary artery bypass graft patency: Noninvasive 52. Lund GK, Wendland MF, Schimawaka A, et al. Coronary sinus flow measurement by
evaluation with MR imaging. Radiology. 1987;164:681–686. means of velocity-encoded cine MR imaging: Validation by using flow probes in dogs.
35. Rubinstein RI, Askenase AD, Thickman D, et al. Magnetic resonance imaging to evaluate Radiology. 2000;217:487–493.
patency of aortocoronary bypass grafts. Circulation. 1987;76:786–791. 53. Schwitter J, De Marco T, Kneifel S, et al. Magnetic resonance-based assessment of global
36. White RD, Pflugfelder PW, Lipton MJ, et al. Coronary artery bypass grafts: Evaluation of coronary flow and flow reserve and its relation to left ventricular function parameters: A
patency with cine MR imaging. AJR Am J Roentgenol. 1988;150:1271–1274. comparison with positron emission tomography. Circulation. 2000;101:2696–2702.
37. Aurigemma GP, Reichek N, Axel L, et al. Noninvasive determination of coronary artery 54. Koskenvuo JW, Sakuma H, Niemi P, et al. Global myocardial blood flow and global
bypass graft patency by cine magnetic resonance imaging. Circulation. 1989;80:1595– flow reserve measurements by MRI and PET are comparable. J Magn Reson Imaging.
1602. 2001;13:361–366.
38. Vrachliotis TG, Bis KG, Aliabadi D, et al. Contrast-enhanced breath-hold MR angiog- 55. Koskenvuo JW, Hartiala JJ, Knuuti J, et al. Assessing coronary sinus blood flow in patients
raphy for evaluating patency of coronary artery bypass grafts. AJR Am J Roentgenol. with coronary artery disease: A comparison of phase-contrast MR imaging with positron
1997;168:1073–1080. emission tomography. Am J Roentgenol. 2001;177:1161–1166.
39. Wintersperger BJ, Engelmann MG, von Smekal A, et al. Patency of coronary bypass grafts: 56. Kawada N, Sakuma H, Yamakado T, et al. Hypertrophic cardiomyopathy: MR measure-
Assessment with breath-hold contrast-enhanced MR angiography—value of a non-electro- ment of coronary blood flow and vasodilator flow reserve in patients and healthy subjects.
cardiographically triggered technique. Radiology. 1998;208:345–351. Radiology. 1999;211:129–135.
40. van Rossum AC, Bedaux WL, Hofman MB. Morphologic and functional evaluation of 57. von Schulthess GK, Schwitter J. Cardiac MR imaging: Facts and fiction. Radiology. 2001;
coronary artery bypass conduits. J Magn Reson Imaging. 1999;10:734–740. 218:326–328.
LWBK1209-ch22_p342-349.indd 349 16/05/13 9:33 PM

Liesbeth P. Salm
C h apt e r Jeroen J. Bax
23
Hildo J. Lamb
J. Wouter Jukema
Albert de Roos

and Computed Tomography of
Coronary Artery Bypass Grafts
■■ Introduction Cardiovascular Magnetic

■■ Cardiovascular Magnetic Resonance of Resonance of Coronary Artery
Coronary Artery Bypass Grafts Bypass Grafts
Anatomy Assessment: Angiography
Anatomy Assessment: Angiography
Functional Assessment: Flow Velocity
Functional Assessment: Left Ventricular Function, During the past decades, a considerable amount of effort
Contrast-enhancement, Viability has been invested to achieve noninvasive visualization of the
coronary arteries and bypass grafts with CMR. The larger
■■ Computed Tomography of Coronary Artery size, straight course, and immobility during the cardiac cycle
Bypass Grafts allowed the evaluation of graft patency even in the earliest
Patency of Bypass Grafts with CT studies, while assessment of the coronary arteries could not
Assessment of Graft Disease with Multidetector be achieved. In these initial investigations, two-dimensional
CT (2D) spin-echo and gradient-echo techniques were applied to
Preoperative Assessment with Multidetector acquire successive axial slices during repetitive breath-holds
CT (1–8). With the spin-echo technique the presence of blood
Multimodality Imaging flow is depicted as an absence of signal. If absence of sig-
nal at different levels of the bypass graft occurred, the graft
■■ Conclusion was considered to be patent. In contrast, flowing blood is
depicted as a bright signal during imaging with gradient MR
techniques. As shown in Table 23.1, both acquisition tech-
Introduction niques have been evaluated in several studies with conven-
tional angiography as the standard of reference, demonstrat-
Coronary artery bypass grafting (CABG) is a commonly per- ing sensitivities and specificities varying from 71% to 100%
formed surgical procedure for alleviation of symptoms and and 88% to 100%, respectively. Pooled analysis of these
prolonging survival for patients with ischemic heart disease. eight studies (with 180 patients and 381 grafts) revealed a
Subsequent bypass graft disease occurs frequently, requiring weighted mean sensitivity and specificity of 81% and 94%,
coronary angiography for diagnosis. Coronary angiogra- with inclusion of 95% of grafts. Despite these promising
phy is an invasive procedure that includes x-ray exposure, results in the distinction between patent and occluded grafts,
hospitalization, and a small risk of complications, including these 2D techniques were still limited by their low signal-to-
arrhythmias, coronary artery dissection, and cardiac death. noise ratio and low spatial resolution.
A noninvasive diagnostic approach for the assessment of Substantial progress in image quality was achieved by
bypass graft anatomy and function is of great benefit. This the development of three-dimensional (3D) imaging tech-
chapter reviews the research that has been performed in niques, allowing the acquisition of volume slabs with several
evaluating bypass grafts noninvasively using cardiovascu- thin slices, which resulted in a higher spatial resolution. To
lar magnetic resonance (CMR) and computed tomography improve patient comfort, navigator techniques have been
(CT). developed that permit real-time monitoring of diaphragm
350
LWBK1209-ch23_p350-364.indd 350 16/05/13 9:33 PM

Chapter 23 ■ Cardiovascular Magnetic Resonance and Computed Tomography of Coronary Artery Bypass Grafts 351
ssessment of Patency in Vein and Arterial Grafts by MR Angiography Compared with

A
Table 23.1
Invasive Angiography
MR Evaluable
Author Patients Grafts Graft Type Technique Grafts (%) Sensitivity (%) Specificity (%)
White et al. (4) 25 72 Vein 2D SE 90 72 91

Rubinstein et al. (6) 20 47 Vein 2D SE 100 72 90
Jenkins et al. (3) 22 45 Vein 2D SE 100 73 89
Frija et al. (2) 28 52 Vein and arterial 2D SE 100 71 97
White et al. (5) 10 28 Vein and arterial 2D GE 100 86 93
Aurigemma et al. (1) 20 45 Vein and arterial 2D GE 100 100 88
Vanninen et al. (7) 8 8 GEA 2D GE 100 100 100
Galjee et al. (8) 47 84 Vein 2D SE 92 84 98
2D GE 92 88 98
WEIGHTED MEAN — — — — 95 81 94
2D
Kessler et al. (9) 8 21 Vein and arterial 3D NAV 90 100 87
Engelmann et al. (10) 16 55 Vein and arterial 3D CE 100 85 95
Vrachliotis et al. (11) 15 45 Vein and arterial 3D CE 98 93 97
Wintersperger et al. 27 76 Vein and arterial 3D CE 100 81 95
(12)
Kalden et al. (13) 22 59 Vein and arterial 3D CE 100 93 93
Molinari et al. (14) 18 51 Vein and arterial 3D NAV 96 92 97
Langerak et al. (15) 38 56 Vein 3D NAV 100 83 98
Wittlinger et al. (16) 34 82 Vein and arterial 3D NAV 90 78 96
Bunce et al. (17) 34 79 Vein and arterial 3D CE 100 73 85
WEIGHTED MEAN — — — — 96 85 94
3D
SE, spin-echo; GE, gradient-echo; GEA, gastroepiploic artery; NAV, navigator; CE, contrast-enhanced.
motion and free-breathing during data acquisition. In addi- specificity of 82% and 88% was observed for the detection
tion, improved enhancement of blood/muscle contrast can of ≥50% luminal narrowing in vein grafts (15). The number
be expected by the administration of intravascular contrast of evaluable grafts was 50 of 56. In contrast, a discouraging
agents. An example of a typical MR angiography (MRA) sensitivity of 38% was reported (13). Evaluation of graft ste-
acquisition protocol with navigator respiratory gating is de- nosis remains difficult due to the presence of metallic clips,
picted in Figure 23.1. In Figure 23.2 an example of a patent, stents, or calcifications in the course of the graft, which may
nonstenosed vein graft as confirmed by conventional angiog- cause signal voids. Since the presence of graft stenosis rather
raphy is provided. Pooled analysis of nine studies using 3D than occlusion may be the cause of recurrent complaints
techniques, with over 200 patients included, revealed a slight and the remaining nongrafted coronary vasculature cannot
increase in weighted sensitivity from 81% to 85%, with no be evaluated in sufficient detail in a single MR session at
loss in specificity (9–17) (Table 23.1). No difference in di- this stage, the value of MRA to evaluate graft stenosis needs
agnostic accuracy for the evaluation of graft patency was to be further explored before this technique can be recom-
noted between arterial and vein grafts. The sensitivity and mended for routine use.
specificity of arterial grafts were 85% and 95% as compared
to 86% and 93% in vein grafts, respectively (10,11,13–17).
Functional Assessment: Flow Velocity
An initial study was performed to assess graft patency post-
operatively with MRA at 3 Tesla (T) (18). MRA confirmed Flow through a blood vessel may or may not be impaired
all grafts to be patent; however, no comparison with inva- by a stenosis, visualized by angiography (19,20). The assess-
sive angiography was made. An example of MR angiogram ment of blood flow through coronary arteries and bypass
at 3 T from this study is shown in Figure 23.3. These studies grafts has gained wide attention. To evaluate the hemody-
illustrate the potential of MRA to evaluate graft patency in namic impairment of a lesion, flow is measured at rest and
clinical routine. Its safe and noninvasive nature in combina- during pharmacologically induced stress with adenosine or
tion with the high specificity (∼94%) and negative predictive dipyridamole (21). By dividing the flow value during stress
value (∼96%) suggests that in patients presenting with recur- by the flow value at rest, the coronary flow reserve (CFR)
rent complaints after bypass grafting, MRA may function as is calculated (22,23). Flow-limiting stenoses cause a com-
a first-line investigation tool to rule out graft occlusion prior pensatory vasodilation at rest to maintain sufficient blood
to more invasive diagnostic procedures. flow to the myocardium. As a consequence, the blood vessel
In addition to assessment of patency, few attempts have cannot respond adequately to an increase in absolute flow
been made to evaluate graft stenosis. A sensitivity and by vasodilation during stress and CFR will be reduced. At
LWBK1209-ch23_p350-364.indd 351 16/05/13 9:33 PM

A C
Figure 23.1. Example of a typical MR angiography (MRA) acquisition protocol. A: Typical planscan for
coronary MRA. The red lines represent the axial imaging volume, the large green box is the volume used for
localized shimming, the rectangular green box is the position in the right hemidiaphragm of the respiratory
navigator, and the blue box is the position of a saturation band for suppression of image artifacts. B: Based on
axial scout images, the three-point planscan is used to select three points in space, one at the origo of the coro-
nary artery or bypass graft, one at the most distal point, and one in the middle of the first two points. From
this information, an imaging plane is automatically calculated in plane with the coronary artery or bypass
graft of interest. C: MRA of a patient with a bypass graft to the left coronary system (white arrow) and a vis-
ible native right coronary artery (black arrow). This imaging approach can be used clinically to assess bypass
graft patency. AO, aorta; LV, left ventricle.
LWBK1209-ch23_p350-364.indd 352 16/05/13 9:33 PM

Figure 23.2. Example of an MR angiogram of a vein graft (right panel) in comparison with a coronary
angiogram (left panel). A free-breathing, 3D, navigator-gated sequence was used for acquisition, here shown in
a multiplanar reformat reconstruction.
first, blood flow was determined by Doppler flow transduc- patients during catheterization. Blood flow correlated well
ers at open-chest procedures (19,22–26), limiting extensive with Doppler-derived velocity of blood flow both in vitro
use of CFR in clinical practice. When the diameter of intra- and in vivo (27–29). Invasive Doppler-derived CFVR has
vascular catheter-based Doppler ultrasonographic devices proven its potential in numerous clinical applications, such
could be reduced to 0.018 in., it became feasible to mea- as in identifying hemodynamic significant stenoses in native
sure the velocity of the blood flow and calculate the coro- coronary arteries and vein grafts (30,31), in the functional
nary flow velocity reserve (CFVR) for coronary arteries in assessment of stenoses of intermediate severity (32), in
the determination of the need for and the outcome after
coronary intervention (33–35), and in the prediction of
restenosis (36).
With the use of CMR, blood flow velocity can be mea-
sured using phase-contrast, velocity-encoded sequences
(37). To use such a sequence accurately, an imaging plane
is planned perpendicular to the target vessel. Throughout
the cardiac cycle, the acquisition yields anatomic modu-
lus images paired with phase images, in which every pixel
contains a different velocity value. Volume flow (mL/min)
can be obtained by calculating the integrated volumetric
flow rate of all pixels in the vessel lumen per heartbeat and
multiply it with the heart rate. To display the flow pattern,
a flow rate-versus-time graph is drawn. Alternatively, the
central peak velocity (cm/sec) can be obtained by selecting
several pixels in the vessel center. Figure 23.4 depicts a typi-
cal example of a CMR examination with velocity mapping
Figure 23.3. 3-T MR angiographic image (curved reconstruction) in a vein graft.
using an intravascular contrast agent for saphenous vein bypass graft
visualization. There is no evidence of stenosis. The bypass graft is pat- Early Magnetic Resonance Studies in Vein Grafts
ent from the proximal aortic anastomosis (asterisk) to the distal anas-
tomoses with the left coronary artery (arrow: Diagonal artery; open For vein grafts, feasibility to quantify flow and character-
arrow: Marginal artery). (Reproduced with permission from: Hoppe ize the flow pattern noninvasively by CMR was demon-
H, Reineke D, Rosskopf AB, et al. Morphological and functional strated (38,39). In an early study graft flow <20 mL/min
3-Tesla magnetic resonance imaging of saphenous vein coronary artery and a loss of the biphasic flow pattern, typical for bypass
bypass grafts. Interact Cardiovasc Thorac Surg. 2011;12:582–585.) grafts, would indicate a dysfunctional graft (40). In another
LWBK1209-ch23_p350-364.indd 353 16/05/13 9:33 PM

study adequate biphasic flow profiles could be obtained in 28 to 164 mL/min) was observed. Mean flow and peak
62 of 73 angiographically patent grafts (8). A significant velocity were lower in IMA grafts compared with native
difference in flow between single and sequential grafts to IMAs. In a different feasibility study, respiratory motion
three vascular regions was demonstrated. These early vein was compensated using a breath-hold, segmented k-space,
graft flow studies were limited by the use of a gradient-echo gradient-echo sequence to quantify velocity in native and
sequence with limited spatial resolution (1.9 × 1.2 × 5 mm3) grafted IMAs (42). Comparison with a free-breathing tech-
and no compensation for respiratory motion on 0.5- to nique in native IMAs demonstrated a higher peak veloc-
0.6-T magnets. ity using the breath-hold sequence because of elimination
of respiratory motion artifacts and averaging of veloci-
ties. This sequence allowed imaging time to decrease from
Early Magnetic Resonance Studies in Arterial Grafts
approximately 4 minutes at free breathing to a 20-second
Feasibility to quantify flow in native and grafted internal breath-hold. By means of a view-sharing reconstruction,
mammary arteries (IMAs) was demonstrated using a free- an effective temporal resolution of 64 milliseconds could
breathing gradient-echo sequence on a 1.5-T MR scanner be obtained, allowing an acquisition of 7 to 13 temporal
(41). A large intersubject variation of IMA graft flow (range: phases per cardiac cycle.
C D
A
Figure 23.4. Example of an MR flow velocity study. A: Panel A shows a sagittal scout scan, on which the
transversal survey scans at the level of the ascending aorta are planned. Panel B depicts the transversal survey
scans showing two vein grafts, one to the circumflex region (arrows) and one to the left anterior descending
artery (open arrows). Ao, aorta; PT, pulmonary trunk; SCV, supracaval vein. Panel C shows a coronal, oblique
survey scan of the two vein grafts. Two differently orientated survey scans are used to plan the flow velocity
scan. Panel D illustrates the flow velocity scan (modulus and phase image) in mid-diastole of the first graft at
rest, which is used to obtain volume flow. (Reproduced with permission from: Salm LP, Vliegen HW, Langerak
SE, et al. Evaluation of saphenous vein coronary artery bypass graft flow by cardiovascular magnetic reso-
nance. J Cardiovasc Magn Reson. 2005;7:631–637.)
LWBK1209-ch23_p350-364.indd 354 16/05/13 9:33 PM

between grafts with <50% and ≥50% stenosis was demon-

strated for MR-derived average and diastolic peak velocity
during stress (44).
The detection of ≥70% angiographic stenosis in IMA
grafts by breath-hold MR flow velocity at rest and during
stress was investigated (45). At CABG surgery, titanium
clips were used to avoid metal artifacts at MR imaging. A
significant difference between grafts with <70% and ≥70%
stenosis was demonstrated for baseline mean blood flow and
the diastolic-to-systolic velocity ratio. Threshold values of
35 mL/min for baseline mean flow and 1 for diastolic-to-
systolic velocity ratio were proposed to separate IMA grafts
with <70% and ≥70% stenosis with respective sensitivity
A and specificity of 86% and 94%, and 86% and 88%. CFR
did not differ significantly between IMA grafts with <70%
and ≥70% stenosis.
The value of MR flow in the prediction of vein graft dis-
ease was assessed (46). An algorithm was formulated combin-
ing baseline flow <20 mL/min or CFR <2 to detect grafts or
run-offs with a significant stenosis (≥50%) or a myocardial
infarction in the graft vascular territory, yielding a sensitiv-
ity of 78% with a specificity of 80%. The algorithm was de-
signed to exclude normal-functioning vein grafts from further
invasive examinations. A different approach for the detection
of stenotic vein and arterial grafts or recipient vessels by MR
with velocity mapping was also formulated, in which single
and sequential vein and arterial grafts were separately ana-
B C
lyzed (47). Sensitivity and specificity for detecting single vein
grafts with ≥70% stenosis were 96% and 92%, respectively.
For sequential vein grafts sensitivity and specificity were 94%
and 71%, respectively. A proposed cut-off point for separating
<70% and ≥70% stenosis in single vein grafts was 1.43 for
CFVR. Not enough stenoses in arterial grafts were present to
formulate an adequate model.
The detection of graft disease was further improved by
a combined approach with MRA and MR flow measure-
D
B ments, which showed a sensitivity and specificity of 95.2%
Figure 23.4. (continued) B: One of the survey images is depicted, and 96.8% for detecting significant stenoses (≥50%) in IMA
visualizing a vein graft in plane (Panel A). Panels B and C show the grafts and 100% and 97.8% in vein grafts (48).
modulus and phase images. The graft is pictured as a white, respec- These studies show that MR flow velocity assessment has
tively black spot in the center of the image. In every image in the potential to become a diagnostic tool in predicting patency
cardiac cycle 4 pixels in the center of the graft are selected to obtain and the presence or absence of significant stenosis in both
a velocity curve (Panel D; white dots: Resting curve). The MR flow vein and IMA grafts in daily practice. However, only if a
velocity acquisition is repeated during adenosine-induced stress to full examination of all grafts and native coronary arteries in
obtain the stress velocity curve (black dots). Ao, aorta; PA, pulmonary a single MR imaging session is feasible, then may this tech-
artery; SCV, supracaval vein.
nique reach its full value.
Applied Studies Using Magnetic Resonance

Flow Velocity in Bypass Grafts
Angiographically Controlled Magnetic Resonance
Studies in Vein and Arterial Grafts Some studies were conducted using MR flow velocity as a
tool to evaluate a certain procedure (surgical or percutane-
To evaluate the diagnostic value of MR flow velocity in ous intervention) or noninvasive modality. A comprehensive
bypass grafts, several angiographically controlled studies approach of contrast-enhanced MRA, MR flow quantifica-
were performed. With the use of breath-hold sequences, tion at rest, and MR cardiac function assessment was used
flow measurements at rest and during pharmacologically to evaluate the status of bypass grafts after CABG surgery
induced stress with determination of CFR in bypass grafts were (49,50). Patency of grafts was accurately assessed by this MR
achieved. By the use of a breath-hold turbo-field echo-planar approach.
imaging sequence, a significant increase at stress for flow and Significant improvements by MR flow after a percutane-
velocity parameters was observed and CFR (mean 2.7 ± 1.1 for ous intervention in vein grafts were observed for baseline
single grafts) could be calculated (43). A significant difference and stress mean velocity (51). Interestingly, CFVR did not
LWBK1209-ch23_p350-364.indd 355 16/05/13 9:33 PM

eference Values for Velocity

R Bypass Graft MR Flow Velocity Assessment at 3 T
and Flow Parameters in Initial studies have been performed evaluating bypass graft
Table 23.2
Nonstenotic Single and flow postoperatively at 3-T MR. Patency of all bypass
Sequential Vein Grafts grafts was confirmed by 3-T MR flow velocity imaging
(18,54). A better agreement of MR flow velocity assess-
Parameter Single Grafts Sequential Grafts ment with intraoperative Doppler flow measurements was
demonstrated for vein grafts than for IMA grafts, possibly
Velocity
due to the smaller luminal diameter of the IMAs or the
Baseline APV (cm/s) 8.6 ± 3.4 12.2 ± 5a
Stress APV (cm/s) 20.2 ± 9.5 27.2 ± 10.6a
different characteristics of the endothelium (54). These
Velocity reserve 2.4 ± 0.8 2.3 ± 0.7 preliminary results suggest that diagnostic accuracy may
Baseline SPV (cm/s) 14.7 ± 5.8 18.1 ± 5.7a further be improved with the use of 3-T MR. Future stud-
Baseline DPV (cm/s) 16.7 ± 5.9 22 ± 9.4a ies should further focus on developing a comprehensive
Stress SPV (cm/s) 24.9 ± 10.6 31.4 ± 14.5 CMR approach to the post-CABG surgery patient provid-
Stress DPV (cm/s) 35 ± 45.3 45.3 ± 17.2a ing both anatomical (MRA) and functional information
Baseline DSVR 1.3 ± 0.7 1.3 ± 0.6 (MR flow velocity mapping or perfusion with delayed
Stress DSVR 1.5 ± 0.4 1.5 ± 0.5 enhancement) in a single session and validate it against a
Flow gold standard.
Baseline flow (mL/min) 36.4 ± 17.7 58.7 ± 33.9a
Stress flow (mL/min) 89.5 ± 47.9 128.7 ± 57.2a
Flow reserve 2.9 ± 2.5 2.5 ± 0.9 Functional Assessment: Left Ventricular
Baseline SPF (mL/s) 1.3 ± 0.6 1.8 ± 0.8a Function, Contrast-Enhancement,
Baseline DPF (mL/s) 1.6 ± 0.6 2.2 ± 1.3 Viability
Stress SPF (mL/s) 2.4 ± 1.2 3.4 ± 1.9a
Stress DPF (mL/s) 3.5 ± 1.7 5 ± 2.6a Functional CMR assessment has been used as gold standard
Baseline DSFR 1.6 ± 1.3 1.3 ± 0.6 to validate other imaging techniques. In one study, global
Stress DSFR 1.6 ± 0.7 1.5 ± 0.5 and regional left ventricular (LV) function by 3D MR served
as the gold standard for the assessment of LV function by
Data are mean values ± SDs. Two different image analysis methods of 99m-Technetium sestamibi gated SPECT in patients post-
the same flow image, the 4-pixel and whole lumen area methods were
CABG surgery (55). After cardiac surgery, exaggerated sys-
used to obtain peak velocity and flow measurements, respectively.
a
Significantly different (p < 0.05) from measurement in single vein
tolic anteromedial translation of the entire heart within the
grafts. chest has been observed, enhancing difficulties in assessing
(Reproduced with permission from: Langerak SE, Vliegen HW, regional wall motion (especially in the septum). They dem-
Jukema JW, et al. Vein graft function improvement after percuta- onstrated that an automated software algorithm (56,57) for
neous intervention: evaluation with MR flow mapping. Radiology. the assessment of LV function by perfusion SPECT agreed
2003;228:834–841.) well with MR functional assessment in this particular group
of patients.
Several studies focused on evaluating myocardial injury
by CMR with gadolinium delayed enhancement (DE) dur-
significantly improve after an intervention as a result of an ing off-pump and on-pump CABG surgery. Conventional
equal recovery of both baseline and stress mean veloci- on-pump surgery uses a cardiopulmonary bypass and aor-
ties. Reference values confirmed a significant difference tic cross-clamping, which may lead to a systemic inflamma-
between single and sequential vein graft parameters, un- tory response syndrome with multiorgan dysfunction (58)
derscoring the need to evaluate these grafts separately and myocardial damage as a result of ischemia (59). Patients
(Table 23.2). Since an MR velocity map can be analyzed to were randomized to either off-pump or on-pump CABG
evaluate both flow and velocity, the diagnostic accuracy of surgery and were examined before and after surgery by DE-
both flow and velocity analysis approach was investigated CMR (60). Troponin I measurements were also obtained,
(52). A similar diagnostic accuracy was demonstrated for which correlated with mean mass of new myocardial hy-
the flow (92%) and velocity analysis (93%) in the detec- perenhancement. The authors found that off-pump CABG
tion of ≥70% stenosis. Velocity analysis seemed to be the resulted in a significantly better LV function early after sur-
preferred method, because it is less time consuming com- gery. However, it did not reduce the incidence or extent of
pared with flow analysis. irreversible myocardial injury. A 6-month follow-up CMR
The functional significance of a bypass graft stenosis was examination was then performed in the same patient group
evaluated by single-photon emission computed tomography to assess the diagnostic value in predicting viability and to
(SPECT) perfusion imaging, breath-hold MR velocity and cor- assess late regional wall motion recovery (61). A strong cor-
onary angiography (53). Agreement between SPECT perfusion relation between the transmural extent of hyperenhance-
imaging and MR velocity assessment was 80% (k = 0.61). All ment and regional function recovery at 6 months was dem-
grafts with a stenosis ≥50% and normal perfusion had normal onstrated, revealing DE-CMR as a powerful predictor of
CFVR, suggesting no hemodynamic significance of the ste- myocardial damage after CABG surgery. Also, right ven-
nosis. Reduced CFVR was observed in 33% of non-stenosed tricular function was determined with CMR and showed
grafts with abnormal perfusion, suggestive of microvascular impairment early after surgery and recovery after 6 months
disease. for both on-pump and off-pump CABG surgery patient
LWBK1209-ch23_p350-364.indd 356 16/05/13 9:33 PM

groups (62). In the same patient group the correlation of graft patency ranged from 45% to 95%. Ultrafast CT scan-
CABG-related myonecrosis with grafted vessel caliber was ning allowed acquisition of one axial image in 50 millisec-
studied (63). The authors found that new perioperative in- onds and repeating of scans after 8 milliseconds (79–81).
jury as detected by CMR occurs most often in the apical Total image acquisition time was 10 to 30 heartbeats, using
myocardium, suggesting inadequate distal myocardial pres- electrocardiograph (ECG) triggering. Sensitivity to detect
ervation as a main mechanism of injury. The occurrence of bypass graft patency using this technique was 93% to 96%
new myocardial necrosis after CABG was influenced by the with a specificity of 89% to 100%. With spiral CT scan-
caliber of the target vessel only in the posterior descending ning the complete heart could be scanned in one breath-
artery territory. hold acquisition of 24 to 30 seconds (10,82–84). Optimal
Correlation of elevated biochemical markers after CABG in-plane spatial resolution using this technique was 0.29
with the amount of perioperatively infarcted myocardium mm2. Reported diagnostic accuracy for the detection of
was confirmed in several studies (64,65). The prognostic graft patency was 96% for vein grafts and 88% for arterial
value of myocardial injury after coronary revascularization grafts (84).
assessed with DE-CMR and biochemical markers was investi- Thereafter, to use four detectors in a single spiral CT
gated (66). New LV hyperenhancement significantly predicted acquisition became feasible, allowing detection of graft pa-
adverse clinical outcome, whereas elevated biochemical mark- tency with very high accuracy: 98% to100% for vein grafts,
ers did not. and 97% to 100% for arterial grafts (85–91).
Furthermore, DE-CMR was used to evaluate a novel hy- Another approach in assessing bypass graft patency is to
brid technique of on-pump beating heart CABG (ONBEAT) estimate the graft blood flow. The contrast clearance curve of
and conventional on-pump CABG (ONSTOP) in patients a contrast bolus injection was evaluated by cine CT (92). This
with impaired ventricular function (67). The incidence of technique was tested in patients, showing a good agreement
new irreversible myocardial injury was significantly higher with observations at coronary angiography (k = 0.75) (93).
in ONBEAT than in ONSTOP patients. At 6-month follow-
up, only ONSTOP patients demonstrated an improvement
Assessment of Graft Disease with
in ventricular geometry.
Multidetector CT
Global LV function recovery may be predicted by DE-
CMR in patients with impaired ejection fraction undergoing In addition to assessment of bypass graft patency, the extent
CABG (68). The presence of ten or more viable plus nor- of graft disease may be evaluated by CT. With 4-detector row
mal segments, based upon the American Heart Association CT technology, adequate stenosis assessment of grafts was
16-segment format and defined as <50% transmural scar, prevented by motion artifacts, metal clip artifacts, or beam
predicted ≥3% improvement in LV ejection fraction with a hardening from calcium deposits (89–91). With 16-slice
sensitivity of 95% and specificity of 75%, suggesting DE- multidetector row spiral CT (MDCT) a scanning range of
CMR to be a simple and powerful tool to identify which the proximal IMA insertion to the heart apex became fea-
patients will benefit from surgical revascularization. sible. By using retrospective ECG gating and a segmental
Finally, MR perfusion imaging with gadolinium was vali- reconstruction algorithm, images with an in-plane spatial
dated for use in patients after CABG surgery (69). Future resolution of up to 0.35 mm2 could be acquired in a single
studies are warranted to determine the clinical value of this breath-hold (94,95). In comparison with coronary angiog-
technique in this group of patients. raphy, the 16-detector row CT technique was demonstrated
to detect bypass graft stenosis of ≥50% with a sensitivity of
75% to 100% and a specificity of 85% to 100% (96–104)
Computed Tomography of (Table 23.3). In Figure 23.5, a typical 16-detector row spiral
CT protocol is explained. The different image displays that
Coronary Artery Bypass Grafts
are available for evaluation of the cardiac CT examination
are shown in Figure 23.6.
Patency of Bypass Grafts with CT
Detection of bypass graft stenosis was improved by the
CT has been intensively investigated in its ability and value next-generation 64-slice MDCT technique. An increase in
to visualize bypass grafts noninvasively. Early studies were spatial and temporal resolution allowed accurate assessment
using conventional CT scanners to assess bypass graft of stenosis in recipient vessels and native coronary arteries
patency postoperatively in comparison with coronary angi- as well, compelling to a complete evaluation of the coronary
ography. Repeated axial images were acquired at multiple vasculature. Pooled analysis of 15 selected studies evaluat-
levels of the graft. If the contrast agent was able to visualize ing bypass grafts with 16- and 64-slice MDCT revealed a
the graft at two levels at least, a graft was scored as patent sensitivity of 94.4% and a specificity of 98% for the de-
(70,71). This technique required an intravenous bolus of 20 tection of significant stenosis (≥50%) (115). The reported
to 50 mL contrast agent per image and burdened the patient sensitivities and specificities of individual angiographically
with a large quantity of x-ray radiation. Then, CT scanners controlled studies are reported in Table 23.3. In recipient
could be programmed to acquire multiple axial images after vessels and nongrafted coronary arteries reported sensitiv-
one bolus injection of contrast agent (dynamic scanning), ity and specificity were 80% to 97%, and 86% to 97.3%
allowing faster acquisition of images (72–78). Scans could for the detection of significant stenosis by 64-slice MDCT
be repeated after a 1- to 4-second interval and per bolus (105,107,112,113,116).
injection four to six sequenced scans could be obtained. Preliminary results are available using 128-slice dual-
Diagnostic accuracy for this technique to detect bypass source MDCT in the evaluation of bypass graft patency (117).
LWBK1209-ch23_p350-364.indd 357 16/05/13 9:33 PM

ccuracy of MDCT in the Detection of Significant Stenoses (≥50% Luminal

A
Table 23.3
Narrowing) in Arterial and Vein Bypass Grafts, Controlled by Coronary Angiography
Number of Number Number of Evaluable

Author Patients of Grafts Detectors Grafts (%) Sensitivity (%) Specificity (%)
Schlosser et al. (100) 38 131 16 74 96 95

Martuscelli et al. (98) 96 278 16 88 90 100
Salm et al. (99) 25 67 16 91 100 95
Chiurlia et al. (97) 52 166 16 99.4 96 100
Anders et al. (96) 32 94 16 Observer 1: 78 80 85
Observer 2: 84 82 88
Yamamoto et al. (101) 42 101 16 99 100 93
Türkvatan et al. (102) 102 236 16 90.4 91.4 98.5
Vernhet-Kovacsik et al. (103) 19 29 16 86 75 100
Burgstahler et al. (104) 13 43 16 95 100 93
Malagutti et al. (105) 52 109 64 100 100 98
Pache et al. (106) 31 96 64 94 98 89
Dikkers et al. (107) 34 69 64 96 100 99
Ropers et al. (108) 50 138 64 100 100 94
Meyer et al. (109) 138 406 64 98 97 97
Jabara et al. (110) 50 147 64 87 95 100
Feuchtner et al. (111) 41 70 64 100 75 95
Nazeri et al. (112) 98 287 64 — 93 88
Romagnoli et al. (113) 78 213 64 99 Vein: 94.4 98.4
Arterial: 100 97.7
De Graaf et al. (114) 40 89 320 98 96 92
Due to high pitch and prospective ECG-synchronization it comprehensive noninvasive assessment, as is determined by
was feasible to reduce the administered radiation dose from several studies (99,121).
approximately 18.9 mSv at 64 slice to 2.3 mSv. However, CT technology is rapidly evolving with 320-detector row
to maintain a sufficient image quality at high heart rates re- technology being readily available. A full noninvasive com-
mains challenging. prehensive assessment of bypass grafts, recipient vessels,
The diagnostic performance of 320-slice MDCT was in- nongrafted arteries, and LV function is feasible with MDCT
vestigated in patients after CABG surgery, revealing a diag- and may detect significant coronary and graft disease with
nostic accuracy of 93%, 89%, and 80% in the detection high diagnostic accuracy. Future studies should focus to re-
of significant stenosis in grafts, recipient vessels, and non- duce the radiation burden and to establish this technique in
grafted coronary arteries, respectively (114). Figures 23.7 clinical practice.
and 23.8 show examples of 320-slice MDCT imaging from
this study.
Preoperative Assessment with
Retrospective ECG-gating allows reconstruction of CT
Multidetector CT
images at 0% to 90% during the cardiac cycle. The best
reconstruction interval to view bypass grafts anastomosed MDCT was used to preoperatively assess the surgical site
to the right coronary artery and branches was shown to before totally endoscopic CABG and the results were cor-
be at 50%, whereas grafts anastomosed to the left anterior related with findings at coronary angiography and dur-
descending or circumflex artery were best viewed at 60% ing surgery (122). MDCT provided extended information
to 70% of the cardiac cycle at four-channel CT (88). Sixty- about the coronary target site and is recommended to be
four–slice MDCT provided best image quality of various used as a planning tool before complex cardiac surgery,
segments and types of bypass grafts at 60% of the RR- such as endoscopic bypass grafting or minimally invasive
interval (118). At present, with 64-slice MDCT technol- direct CABG. In the same way MDCT was found to be
ogy prospective ECG-synchronization is feasible, resulting useful in the planning of redo CABG (123). Preoperative
in a considerable reduction in radiation dose (from ∼27 to mapping of patent grafts and other vital mediastinal
∼3.5 mSv) with a similar image quality and diagnostic ac- structures in relation to the sternum reduced the morbid-
curacy compared with a retrospective ECG-gating protocol ity through modification of the surgical approach. Also,
(119,120). The dose reduction was even more pronounced the availability of the right gastroepiploic artery was suc-
in female patients (120). cessfully investigated before CABG by MDCT (124). In
In addition to bypass graft CT angiography, MDCT al- a multicenter study 64-slice MDCT was used to identify
lows an accurate estimation of LV function (end-diastolic vol- target vessels and based on their findings the authors sug-
ume, end-systolic volume, and ejection fraction), providing a gest that MDCT may be used as a clinical alternative to
LWBK1209-ch23_p350-364.indd 358 16/05/13 9:33 PM

D
A B
Figure 23.5. Example of a typical 16-slice MDCT acquisition protocol. A: A typical coronary CT angi-
ography protocol starts with correct supine patient positioning, positioning of IV-line, placement of three
ECG leads and patient instruction concerning breath-holding. First, a scanogram for overview is acquired (A).
White lines indicate approximate cranial and caudal imaging margins to acquire a second noncontrast localizer
with prospective ECG triggering (30 to 40 slices, 3-mm slice thickness, 120 kV at 200 mA). B: Based on the
second localizer, the exact cranial and caudal margins are defined. These margins are depending on the clinical
setting, for example, for the CT angiography of an IMA graft a higher cranial slice is needed (B, arrow) than for
CT angiography of a vein graft (C, arrow). The actual contrast-enhanced, helical acquisition should be planned
from 1 cm above the defined cranial margin to 1 cm below the caudal margin, that is, the apex of the heart (D,
arrow). (continued)
LWBK1209-ch23_p350-364.indd 359 16/05/13 9:33 PM

F G
C
Figure 23.5. (continued) C: For optimal timing of the contrast arrival in the coronary arteries and bypass
grafts, automated bolus tracking is applied by placement of a region of interest in the ascending aorta (E). After
the signal intensity reaches a predefined threshold (see graph), the final CT angiography scan starts automatically.
Two levels are shown from the resulting CT angiography of a vein graft (F,G, arrows). The CT angiography was
performed with retrospective ECG gating (0.5-mm slice thickness, 120 kV at 250 mA, pitch and rotation time
depending on heart rate).
A B
Figure 23.6. For the evaluation of MDCT angiograms,

several imaging displays can be used. Panel A shows an
original axial slice. A curved multiplanar reconstruction
of a vein graft is depicted in Panel B. Panel C shows a
maximum intensity projection of an arterial graft. For an
overview of the coronary arteries and bypass grafts, 3D
C D
volume rendered reconstructions may be useful (Panel D).
LWBK1209-ch23_p350-364.indd 360 16/05/13 9:33 PM

Figure 23.7. Example of 320-slice

CT angiography of a 75-year-old man
with a history of CABG and percuta-
neous coronary intervention. Panel A
shows a 3D volume rendered recon-
struction of the heart. A single vein
graft is shown with an anastomosis
on the IM (arrowhead). The LAD is
functionally occluded (arrow). Panel
B shows a curved multiplanar recon-
struction, revealing a diseased native
IM, with a severely calcified lesion at
the origin (long arrow) and small ves-
sel caliber. In the LM is a patent stent
visible. In Panel C, a curved multipla-
nar reconstruction of the vein graft is
shown revealing proximal luminal nar-
rowing (long arrow). Two patent stents A B
are identified in the vein graft (short
arrows), without the presence of in-
stent restenosis. A patent anastomosis
with the IM is shown (arrowhead) with
good distal runoff and a non-significant
calcified lesion. Signs of an old myocar-
dial infarction are observed with calcifi-
cations in the apical and midventricular
anteroseptal myocardium (asterisk).
Graft and anastomosis (arrowhead)
patency as well as good distal runoff
in the IM were confirmed on invasive
coronary angiography (Panel D). IM,
intermediate branch; LAD, left anterior
descending artery; LM, left main artery.
(Reproduced with permission from: de
Graaf FR, van Velzen JE, Witkowska
AJ, et al. Diagnostic performance of
320-slice multidetector computed tomog-
raphy coronary angiography in patients
after coronary artery bypass grafting. C D
Eur Radiol. 2011;21:2285–2296.)
conventional angiography in the preoperative assessment PET to the pre-revascularization work-up was found to be
before cardiac surgery (125). cost-effective by diminishing the need for a diagnostic inva-
sive coronary angiogram (129). Further studies dedicated to
evaluate patients with bypass grafts using hybrid, multimo-
Multimodality Imaging
dality imaging are warranted.
Hybrid SPECT-CT and positron emission tomography
(PET)-CT scanners are now available. With the use of per-
fusion agents, preliminary studies suggest great potential in Conclusion
both anatomical and functional assessment of coronary artery
and bypass graft disease (126,127). In a single examination, CMR and CT have made rapid progress in the last three
evaluation of stress and rest myocardial perfusion with quan- decades in evaluating patency and significant stenoses in cor-
tification (PET), CT calcium scoring, and CT angiography is onary artery bypass grafts. Other applications of CMR and
feasible. One study investigated patients preoperatively with CT include pre- and postoperative assessment of the CABG
N-13 ammonia PET-CT and found that the results could procedure and evaluation of LV function. In the near future
influence planning for CABG and held important prognostic a full noninvasive approach including coronary anatomy
information compared with invasive coronary angiography and cardiac function may be offered to patients with bypass
alone (128). In this manner, the addition of N-13 ammonia grafts.
LWBK1209-ch23_p350-364.indd 361 16/05/13 9:33 PM

Figure 23.8. Example of 320-slice

CT angiography of an arterial graft
in a 74-year-old man who underwent
CABG 7 years ago. Panel A shows a 3D
volume rendered reconstruction of the
heart and bypass trajectory. An arterial
graft is visible with anastomoses to the
D1 and LAD. Panel B shows the curved
multiplanar reconstruction of a patent
graft without the presence of significant
graft stenosis. A patent anastomosis is
observed (arrowhead) with good run-
A B C off in the distal LAD. In Panel C, the
curved multiplanar reconstruction of
a heavily diseased LAD is shown with
severe proximal stenosis (arrow). In
Panel D, a curved multiplanar recon-
struction of the arterial graft and D1
are shown. The anastomosis of the graft
to the patent D1 is uninterpretable due
to a dense calcification. In Panels E and
F the concordant invasive coronary
angiograms are shown, confirming graft
and vessel patency. D1, first diagonal
branch; LAD, left anterior descend-
ing artery. (Reproduced with permis-
sion from: de Graaf FR, van Velzen
JE, Witkowska AJ, et al. Diagnostic
performance of 320-slice multidetec-
tor computed tomography coronary
angiography in patients after coronary
D E F artery bypass grafting. Eur Radiol.
2011;21:2285–2296.)
REFERENCES 13. Kalden P, Kreitner KF, Wittlinger T, et al. Assessment of coronary artery bypass grafts: Value of
different breath-hold MR imaging techniques. AJR Am J Roentgenol. 1999;172:1359–1364.
1. Aurigemma GP, Reichek N, Axel L, et al. Noninvasive determination of coronary artery 14. Molinari G, Sardanelli F, Zandrino F, et al. Value of navigator echo magnetic resonance
bypass graft patency by cine magnetic resonance imaging. Circulation. 1989;80:1595–1602. angiography in detecting occlusion/patency of arterial and venous, single and sequential
2. Frija G, Schouman Claeys E, Lacombe P, et al. A study of coronary artery bypass graft coronary bypass grafts. Int J Card Imaging. 2000;16:149–160.
patency using MR imaging. J Comput Assist Tomogr. 1989;13:226–232. 15. Langerak SE, Vliegen HW, de Roos A, et al. Detection of vein graft disease using high-
3. Jenkins JP, Love HG, Foster CJ, et al. Detection of coronary artery bypass graft patency resolution magnetic resonance angiography. Circulation. 2002;105:328–333.
as assessed by magnetic resonance imaging. Br J Radiol. 1988;61:2–4. 16. Wittlinger T, Voigtlander T, Kreitner KF, et al. Non-invasive magnetic resonance imaging
4. White RD, Caputo GR, Mark AS, et al. Coronary artery bypass graft patency: of coronary bypass grafts. Comparison of the haste- and navigator techniques with con-
Noninvasive evaluation with MR imaging. Radiology. 1987;164:681–686. ventional coronary angiography. Int J Cardiovasc Imaging. 2002;18:469–477.
5. White RD, Pflugfelder PW, Lipton MJ, et al. Coronary artery bypass grafts: Evaluation of 17. Bunce NH, Lorenz CH, John AS, et al. Coronary artery bypass graft patency: Assessment
patency with cine MR imaging. AJR Am J Roentgenol. 1988;150:1271–1274. with true fast imaging with steady-state precession versus gadolinium-enhanced MR angi-
6. Rubinstein RI, Askenase AD, Thickman D, et al. Magnetic resonance imaging to evaluate ography. Radiology. 2003;227:440–446.
patency of aortocoronary bypass grafts. Circulation. 1987;76:786–791. 18. Hoppe H, Reineke D, Rosskopf AB, et al. Morphological and functional 3-Tesla magnetic
7. Vanninen RL, Vainio PA, Manninen HI, et al. Gastroepiploic artery as an in situ coro- resonance imaging of saphenous vein coronary artery bypass grafts. Interact Cardiovasc
nary artery bypass graft: Evaluation of MRI and colour Doppler ultrasound in follow-up. Thorac Surg. 2011;12:582–585.
Scand J Thorac Cardiovasc Surg. 1995;29:7–10. 19. White CW, Wright CB, Doty DB, et al. Does visual interpretation of the coronary
8. Galjee MA, van Rossum AC, Doesburg T, et al. Value of magnetic resonance imaging arteriogram predict the physiologic importance of a coronary stenosis? N Engl J Med.
in assessing patency and function of coronary artery bypass grafts. An angiographically 1984;310:819–824.
controlled study. Circulation. 1996;93:660–666. 20. Topol EJ, Nissen SE. Our preoccupation with coronary luminology. The dissocia-
9. Kessler W, Achenbach S, Moshage W, et al. Usefulness of respiratory gated magnetic reso- tion between clinical and angiographic findings in ischemic heart disease. Circulation.
nance coronary angiography in assessing narrowings > or = 50% in diameter in native 1995;92:2333–2342.
coronary arteries and in aortocoronary bypass conduits. Am J Cardiol. 1997;80:989–993. 21. Uren NG, Melin JA, De Bruyne B, et al. Relation between myocardial blood flow and the
10. Engelmann MG, KnezA, von Smekal A, et al. Non-invasive coronary bypass graft imaging severity of coronary-artery stenosis. N Engl J Med. 1994;330:1782–1788.
after multivessel revascularisation. Int J Cardiol. 2000;76:65–74. 22. Gould KL, Lipscomb K, Hamilton GW. Physiologic basis for assessing critical coronary
11. Vrachliotis TG, Bis KG, Aliabadi D, et al. Contrast-enhanced breath-hold MR angiog- stenosis. Instantaneous flow response and regional distribution during coronary hyper-
raphy for evaluating patency of coronary artery bypass grafts. AJR Am J Roentgenol. emia as measures of coronary flow reserve. Am J Cardiol. 1974;33:87–94.
1997;168:1073–1080. 23. Gould KL, Kirkeeide RL, Buchi M. Coronary flow reserve as a physiologic measure of
12. Wintersperger BJ, Engelmann MG, von Smekal A, et al. Patency of coronary bypass stenosis severity. J Am Coll Cardiol. 1990;15:459–474.
grafts: Assessment with breath-hold contrast-enhanced MR angiography—value of a non- 24. Gould KL, Lipscomb K. Effects of coronary stenoses on coronary flow reserve and resis-
electrocardiographically triggered technique. Radiology. 1998;208:345–351. tance. Am J Cardiol. 1974;34:48–55.
LWBK1209-ch23_p350-364.indd 362 16/05/13 9:33 PM

25. Bittar N, Kroncke GM, Dacumos GC Jr., et al. Vein graft flow and reactive hyperemia in 57. Germano G, Erel J, Kiat H, et al. Quantitative LVEF and qualitative regional function
the human heart. J Thorac Cardiovasc Surg. 1972;64:855–860. from gated thallium-201 perfusion SPECT. J Nucl Med. 1997;38:749–754.
26. Wilson RF, White CW. Does coronary artery bypass surgery restore normal maximal 58. Ascione R, Lloyd CT, Underwood MJ, et al. Inflammatory response after coronary revascu-
coronary flow reserve? The effect of diffuse atherosclerosis and focal obstructive lesions. larization with or without cardiopulmonary bypass. Ann Thorac Surg. 2000;69:1198–1204.
Circulation. 1987;76:563–571. 59. Taggart DP. Biochemical assessment of myocardial injury after cardiac surgery: Effects of
27. Doucette JW, Corl PD, Payne HM, et al. Validation of a Doppler guide wire for intravas- a platelet activating factor antagonist, bilateral internal thoracic artery grafts, and coro-
cular measurement of coronary artery flow velocity. Circulation. 1992;85:1899–1911. nary endarterectomy. J Thorac Cardiovasc Surg. 2000;120:651–659.
28. Sudhir K, Hargrave VK, Johnson EL, et al. Measurement of volumetric coronary blood 60. Selvanayagam JB, Petersen SE, Francis JM, et al. Effects of off-pump versus on-pump
flow with a Doppler catheter: Validation in an animal model. Am Heart J. 1992;124: coronary surgery on reversible and irreversible myocardial injury: A randomized trial
870–875. using cardiovascular magnetic resonance imaging and biochemical markers. Circulation.
29. Labovitz AJ, Anthonis DM, Cravens TL, et al. Validation of volumetric flow measure- 2004;109:345–350.
ments by means of a Doppler-tipped coronary angioplasty guide wire. Am Heart J. 1993; 61. Selvanayagam JB, Kardos A, Francis JM, et al. Value of delayed-enhancement cardiovas-
126:1456–1461. cular magnetic resonance imaging in predicting myocardial viability after surgical revas-
30. Ofili EO, Labovitz AJ, Kern MJ. Coronary flow velocity dynamics in normal and diseased cularization. Circulation. 2004;110:1535–1541.
arteries. Am J Cardiol. 1993;71:3d–9d. 62. Pegg TJ, Selvanayagam JB, Karamitsos TD, et al. Effects of off-pump versus on-pump
31. White CW. Clinical applications of Doppler coronary flow reserve measurements. Am J coronary artery bypass grafting on early and late right ventricular function. Circulation.
Cardiol. 1993; 71:10d–16d. 2008;117:2202–2210.
32. Kern MJ, Donohue TJ, Aguirre FV, et al. Assessment of angiographically intermediate 63. Selvanayagam JB, Searle N, Neubauer S, et al. Correlation of coronary artery bypass
coronary artery stenosis using the Doppler flowire. Am J Cardiol. 1993;71:26d–33d. surgery-related myonecrosis with grafted vessel calibre: Insights from cardiovascular mag-
33. Segal J. Applications of coronary flow velocity during angioplasty and other coronary netic resonance imaging. Heart. 2006;92:1329–1330.
interventional procedures. Am J Cardiol. 1993;71:17d–25d. 64. Steuer J, Bjerner T, Duvernoy O, et al. Visualisation and quantification of peri-operative
34. Serruys PW, De Bruyne B, Carlier S, et al. Randomized comparison of primary stent- myocardial infarction after coronary artery bypass surgery with contrast-enhanced mag-
ing and provisional balloon angioplasty guided by flow velocity measurement. Doppler netic resonance imaging. Eur Heart J. 2004;25:1293–1299.
Endpoints Balloon Angioplasty Trial Europe (DEBATE) II Study Group. Circulation. 65. Pegg TJ, Maunsell Z, Karamitsos TD, et al. Utility of cardiac biomarkers for the diagnosis
2000;102:2930–2937. of type V myocardial infarction after coronary artery bypass grafting: Insights from serial
35. Nishida T, Di Mario C, Kern MJ, et al. Impact of final coronary flow velocity reserve on cardiac MRI. Heart. 2011;97:810–816.
late outcome following stent implantation. Eur Heart J. 2002;23:331–340. 66. Rahimi K, Banning AP, Cheng AS, et al. Prognostic value of coronary revascularisa-
36. Albertal M, Regar E, Van Langenhove G, et al. Value of coronary stenotic flow velocity tion-related myocardial injury: A cardiac magnetic resonance imaging study. Heart.
acceleration in prediction of angiographic restenosis following balloon angioplasty. Eur 2009;95:1937–1943.
Heart J. 2002;23:1849–1853. 67. Pegg TJ, Selvanayagam JB, Francis JM, et al. A randomized trial of on-pump beating
37. Lotz J, Meier C, Leppert A, et al. Cardiovascular flow measurement with phase-contrast heart and conventional cardioplegic arrest in coronary artery bypass surgery patients
MR imaging: Basic facts and implementation. Radiographics. 2002;22:651–671. with impaired left ventricular function using cardiac magnetic resonance imaging and
38. van Rossum AC, Galjee MA, Doesburg T, et al. The role of magnetic resonance in the biochemical markers. Circulation. 2008;118:2130–2138.
evaluation of functional results after CABG/PTCA. Int J Card Imaging. 1993;9(suppl 68. Pegg TJ, Selvanayagam JB, Jennifer J, et al. Prediction of global left ventricular functional
1):59–69. recovery in patients with heart failure undergoing surgical revascularisation, based on late
39. Galjee MA, van Rossum AC, Doesburg T, et al. Quantification of coronary artery gadolinium enhancement cardiovascular magnetic resonance. J Cardiovasc Magn Reson.
bypass graft flow by magnetic resonance phase velocity mapping. Magn Reson Imaging. 2010;12:56.
1996;14:485–493. 69. Arnold JR, Francis JM, Karamitsos TD, et al. Myocardial perfusion imaging after coro-
40. Hoogendoorn LI, Pattynama PM, Buis B, et al. Noninvasive evaluation of aortocoronary nary artery bypass surgery using cardiovascular magnetic resonance: A validation study.
bypass grafts with magnetic resonance flow mapping. Am J Cardiol. 1995;75:845–848. Circ Cardiovasc Imaging. 2011;4:312–318.
41. Debatin JF, Strong JA, Sostman HD, et al. MR characterization of blood flow in native 70. Albrechtsson U, Stahl E, Tylen U. Evaluation of coronary artery bypass graft patency with
and grafted internal mammary arteries. J Magn Reson Imaging. 1993;3:443–450. computed tomography. J Comput Assist Tomogr. 1981;5:822–826.
42. Sakuma H, Globits S, O’Sullivan M, et al. Breath-hold MR measurements of blood flow 71. Daniel WG, Dohring W, Stender HS, et al. Value and limitations of computed tomography
velocity in internal mammary arteries and coronary artery bypass grafts. J Magn Reson in assessing aortocoronary bypass graft patency. Circulation. 1983;67:983–987.
Imaging. 1996;6:219–222. 72. Brundage BH, Lipton MJ, Herfkens RJ, et al. Detection of patent coronary bypass grafts
43. Langerak SE, Kunz P, Vliegen HW, et al. Improved MR flow mapping in coronary artery by computed tomography. A preliminary report. Circulation. 1980;61:826–831.
bypass grafts during adenosine-induced stress. Radiology. 2001;218:540–547. 73. Guthaner DF, Brody WR, Ricci M, et al. The use of computed tomography in the diagno-
44. Langerak SE, Kunz P, Vliegen HW, et al. Magnetic resonance flow mapping in coronary sis of coronary artery bypass graft patency. Cardiovasc Intervent Radiol. 1980;3:3–8.
artery bypass grafts: A validation study with Doppler flow measurements. Radiology. 74. Godwin JD, Califf RM, Korobkin M, et al. Clinical value of coronary bypass graft evalu-
2002;222:127–135. ation with CT. AJR Am J Roentgenol. 1983;140:649–655.
45. Ishida N, Sakuma H, Cruz BP, et al. MR flow measurement in the internal mammary 75. Wilson PC, Gutierrez O, Moss A. Early evaluation of coronary artery bypass grafts: CT
artery-to-coronary artery bypass graft: Comparison with graft stenosis at radiographic or selective angiography. Eur J Radiol. 1984;4:22–27.
angiography. Radiology. 2001;220:441–447. 76. Foster CJ, Sekiya T, Brownlee WC, et al. Computed tomographic assessment of coronary
46. Bedaux WL, Hofman MB, Vyt SL, et al. Assessment of coronary artery bypass graft dis- artery bypass grafts. Br Heart J. 1984;52:24–29.
ease using cardiovascular magnetic resonance determination of flow reserve. J Am Coll 77. Muhlberger V, Knapp E, zur ND. Predictive value of computed tomographic determina-
Cardiol. 2002;40:1848–1855. tion of the patency rate of aortocoronary venous bypasses in relation to angiographic
47. Langerak SE, Vliegen HW, Jukema JW, et al. Value of magnetic resonance imaging for the results. Eur Heart J. 1990;11:380–388.
noninvasive detection of stenosis in coronary artery bypass grafts and recipient coronary 78. Kahl FR, Wolfman NT, Watts LE. Evaluation of aortocoronary bypass graft status by
arteries. Circulation. 2003;107:1502–1508. computed tomography. Am J Cardiol. 1981;48:304–310.
48. Stauder NI, Klumpp B, Stauder H, et al. Assessment of coronary artery bypass grafts by 79. Bateman TM, Gray RJ, Whiting JS, et al. Prospective evaluation of ultrafast cardiac
magnetic resonance imaging. Br J Radiol. 2007;80:975–983. computed tomography for determination of coronary bypass graft patency. Circulation.
49. Miller S, Scheule AM, Hahn U, et al. MR angiography and flow quantification of the 1987;75:1018–1024.
internal mammary artery graft after minimally invasive direct coronary artery bypass. 80. Stanford W, Rooholamini M, Rumberger J, et al. Evaluation of coronary bypass graft
AJR Am J Roentgenol. 1999;172:1365–1369. patency by ultrafast computed tomography. J Thorac Imaging. 1988;3:52–55.
50. Stauder NI, Scheule AM, Hahn U, et al. Perioperative monitoring of flow and patency 81. Stanford W, Brundage BH, MacMillan R, et al. Sensitivity and specificity of assessing
in native and grafted internal mammary arteries using a combined MR protocol. Br J coronary bypass graft patency with ultrafast computed tomography: Results of a multi-
Radiol. 2005;78:292–298. center study. J Am Coll Cardiol. 1988;12:1–7.
51. Langerak SE, Vliegen HW, Jukema JW, et al. Vein graft function improvement after percu- 82. Tello R, Costello P, Ecker C, et al. Spiral CT evaluation of coronary artery bypass graft
taneous intervention: Evaluation with MR flow mapping. Radiology. 2003;228:834–841. patency. J Comput Assist Tomogr. 1993;17:253–259.
52. Salm LP, Langerak SE, Vliegen HW, et al. Blood flow in coronary artery bypass vein grafts: 83. Ueyama K, Ohashi H, Tsutsumi Y, et al. Evaluation of coronary artery bypass grafts using
Volume versus velocity at cardiovascular MR imaging. Radiology. 2004;232:915–920. helical scan computed tomography. Catheter Cardiovasc Interv. 1999;46:322–326.
53. Salm LP, Bax JJ, Vliegen HW, et al. Functional significance of stenoses in coronary 84. Engelmann MG, von Smekal A, Knez A, et al. Accuracy of spiral computed tomography for
artery bypass grafts Evaluation by single-photon emission computed tomography perfu- identifying arterial and venous coronary graft patency. Am J Cardiol. 1997;80:569–574.
sion imaging, cardiovascular magnetic resonance, and angiography. J Am Coll Cardiol. 85. Yoo KJ, Choi D, Choi BW, et al. The comparison of the graft patency after coronary
2004;44:1877–1882. artery bypass grafting using coronary angiography and multi-slice computed tomography.
54. Reineke D, Ith M, Goeber V, et al. Comparison of arterial and venous coronary artery Eur J Cardiothorac Surg. 2003;24:86–91.
bypass flow measurements using 3-T magnetic resonance phase contrast imaging. Eur J 86. Ko YG, Choi DH, Jang YS, et al. Assessment of coronary artery bypass graft patency by
Radiol. 2012;81:e502–e506. multislice computed tomography. Yonsei Med J. 2003;44:438–444.
55. Tadamura E, Kudoh T, Motooka M, et al. Use of technetium-99 m sestamibi ECG-gated 87. Burgstahler C, Kuettner A, Kopp AF, et al. Non-invasive evaluation of coronary artery
single-photon emission tomography for the evaluation of left ventricular function follow- bypass grafts using multi-slice computed tomography: Initial clinical experience. Int J
ing coronary artery bypass graft: Comparison with three-dimensional magnetic resonance Cardiol. 2003;90:275–280.
imaging. Eur J Nucl Med. 1999;26:705–712. 88. Willmann JK, Weishaupt D, Kobza R, et al. Coronary artery bypass grafts: ECG-gated
56. Germano G, Kiat H, Kavanagh PB, et al. Automatic quantification of ejection fraction multi-detector row CT angiography—influence of image reconstruction interval on graft
from gated myocardial perfusion SPECT. J Nucl Med. 1995;36:2138–2147. visibility. Radiology. 2004;232:568–577.
LWBK1209-ch23_p350-364.indd 363 16/05/13 9:33 PM

89. Marano R, Storto ML, Maddestra N, et al. Non-invasive assessment of coronary artery 111. Feuchtner GM, Schachner T, Bonatti J, et al. Diagnostic performance of 64-slice com-
bypass graft with retrospectively ECG-gated four-row multi-detector spiral computed puted tomography in evaluation of coronary artery bypass grafts. AJR Am J Roentgenol.
tomography. Eur Radiol. 2004;14:1353–1362. 2007;189:574–580.
90. Nieman K, Pattynama PM, Rensing BJ, et al. Evaluation of patients after coronary artery 112. Nazeri I, Shahabi P, Tehrai M, et al. Assessment of patients after coronary artery bypass
bypass surgery: CT angiographic assessment of grafts and coronary arteries. Radiology. grafting using 64-slice computed tomography. Am J Cardiol. 2009;103:667–673.
2003;229:749–756. 113. Romagnoli A, Patrei A, Mancini A, et al. Diagnostic accuracy of 64-slice CT in eval-
91. Ropers D, Ulzheimer S, Wenkel E, et al. Investigation of aortocoronary artery bypass uating coronary artery bypass grafts and of the native coronary arteries. Radiol Med.
grafts by multislice spiral computed tomography with electrocardiographic-gated image 2010;115:1167–1178.
reconstruction. Am J Cardiol. 2001;88:792–795. 114. de Graaf FR, van Velzen JE, Witkowska AJ, et al. Diagnostic performance of 320-slice
92. Rumberger JA, Feiring AJ, Hiratzka LF, et al. Quantification of coronary artery bypass multidetector computed tomography coronary angiography in patients after coronary
flow reserve in dogs using cine-computed tomography. Circ Res. 1987;61:117–123. artery bypass grafting. Eur Radiol. 2011;21:2285–2296.
93. Tello R, Hartnell GG, Costello P, et al. Coronary artery bypass graft flow: Qualitative 115. Hamon M, Lepage O, Malagutti P, et al. Diagnostic performance of 16- and 64-section spi-
evaluation with cine single-detector row CT and comparison with findings at angiogra- ral CT for coronary artery bypass graft assessment: Meta-analysis. Radiology. 2008;247:
phy. Radiology. 2002;224:913–918. 679–686.
94. Dewey M, Lembcke A, Enzweiler C, et al. Isotropic half-millimeter angiography of 116. Laynez-Carnicero A, Estornell-Erill J, Trigo-Bautista A, et al. Non-invasive assessment
coronary artery bypass grafts with 16-slice computed tomography. Ann Thorac Surg. of coronary artery bypass grafts and native coronary arteries using 64-slice computed
2004;77:800–804. tomography: Comparison with invasive coronary angiography. Rev Esp Cardiol. 2010;63:
95. Gurevitch J, Gaspar T, Orlov B, et al. Noninvasive evaluation of arterial grafts with newly 161–169.
released multidetector computed tomography. Ann Thorac Surg. 2003;76:1523–1527. 117. Goetti R, Leschka S, Baumuller S, et al. Low dose high-pitch spiral acquisition 128-slice
96. Anders K, Baum U, Schmid M, et al. Coronary artery bypass graft (CABG) patency: dual-source computed tomography for the evaluation of coronary artery bypass graft
Assessment with high-resolution submillimeter 16-slice multidetector-row computed patency. Invest Radiol. 2010;45:324–330.
tomography (MDCT) versus coronary angiography. Eur J Radiol. 2006;57:336–344. 118. Desbiolles L, Leschka S, Plass A, et al. Evaluation of temporal windows for coronary
97. Chiurlia E, Menozzi M, Ratti C, et al. Follow-up of coronary artery bypass graft patency artery bypass graft imaging with 64-slice CT. Eur Radiol. 2007;17:2819–2828.
by multislice computed tomography. Am J Cardiol. 2005;95:1094–1097. 119. Andreini D, Pontone G, Mushtaq S, et al. Diagnostic performance of two types of low
98. Martuscelli E, Romagnoli A, D’Eliseo A, et al. Evaluation of venous and arterial conduit radiation exposure protocol for prospective ECG-triggering multidetector computed
patency by 16-slice spiral computed tomography. Circulation. 2004;110:3234–3238. tomography angiography in assessment of coronary artery bypass graft. Int J Cardiol.
99. Salm LP, Bax JJ, Jukema JW, et al. Comprehensive assessment of patients after coro- 2012;157;63–69.
nary bypass grafting by 16-detector row computed tomography. Am Heart J. 2005;150: 120. Esposito A, De Cobelli F, Colantoni C, et al. Gender influence on dose saving allowed by
775–781. prospective-triggered 64-slice multidetector computed tomography coronary angiography
100. Schlosser T, Konorza T, Hunold P, et al. Noninvasive visualization of coronary artery as compared with retrospective-gated mode. Int J Cardiol. 2012;158:253–259.
bypass grafts using 16-detector row computed tomography. J Am Coll Cardiol. 121. Spoeck A, Bonatti J, Friedrich GJ, et al. Evaluation of left ventricular function by 64-mul-
2004;44:1224–1229. tidetector computed tomography in patients undergoing totally endoscopic coronary
101. Yamamoto M, Kimura F, Niinami H, et al. Noninvasive assessment of off-pump coro- artery bypass grafting. Heart Surg Forum. 2008;11:E218–E224.
nary artery bypass surgery by 16-channel multidetector-row computed tomography. Ann 122. Herzog C, Dogan S, Diebold T, et al. Multi-detector row CT versus coronary angiogra-
Thorac Surg. 2006;81:820–827. phy: Preoperative evaluation before totally endoscopic coronary artery bypass grafting.
102. Türkvatan A, Biyikoglu SF, Buyukbayraktar FG, et al. Noninvasive evaluation of coro- Radiology. 2003;229:200–208.
nary artery bypass grafts and native coronary arteries: Is 16-slice multidetector CT useful? 123. Gasparovic H, Rybicki FJ, Millstine J, et al. Three-dimensional computed tomographic
Diagn Interv Radiol. 2009;15:43–50. imaging in planning the surgical approach for redo cardiac surgery after coronary revas-
103. Vernhet-Kovacsik H, Battistella P, Demaria R, et al. Early postoperative assessment of coro- cularization. Eur J Cardiothorac Surg. 2005;28:244–249.
nary artery bypass graft patency and anatomy: Value of contrast-enhanced 16-MDCT with 124. Lee JH, Chun EJ, Choi SI, et al. Prospective versus retrospective ECG-gated 64-detec-
retrospectively ECG-gated reconstructions. AJR Am J Roentgenol. 2006;186:S395–S400. tor coronary CT angiography for evaluation of coronary artery bypass graft patency:
104. Burgstahler C, Beck T, Kuettner A, et al. Non-invasive evaluation of coronary artery Comparison of image quality, radiation dose and diagnostic accuracy. Int J Cardiovasc
bypass grafts using 16-row multi-slice computed tomography with 188 ms temporal reso- Imaging. 2011;27:657–667.
lution. Int J Cardiol. 2006;106:244–249. 125. Simon AR, Baraki H, Weidemann J, et al. High-resolution 64-slice helical-computer-
105. Malagutti P, Nieman K, Meijboom WB, et al. Use of 64-slice CT in symptomatic patients assisted-tomographical-angiography as a diagnostic tool before CABG surgery: The dawn
after coronary bypass surgery: Evaluation of grafts and coronary arteries. Eur Heart J. of a new era? Eur J Cardiothorac Surg. 2007;32:896–901.
2007;28:1879–1885. 126. Hacker M, Jakobs T, Matthiesen F, et al. Combined functional and morphological imag-
106. Pache G, Saueressig U, Frydrychowicz A, et al. Initial experience with 64-slice cardiac CT: ing consisting of gated myocardial perfusion SPECT and 16-detector multislice spiral CT
Non-invasive visualization of coronary artery bypass grafts. Eur Heart J. 2006;27:976–980. angiography in the noninvasive evaluation of coronary artery disease: First experiences.
107. Dikkers R, Willems TP, Tio RA, et al. The benefit of 64-MDCT prior to invasive coro- Clin Imaging. 2007;31:313–320.
nary angiography in symptomatic post-CABG patients. Int J Cardiovasc Imaging. 2007; 127. Di Carli MF, Murthy VL. Cardiac PET/CT for the evaluation of known or suspected
23:369–377. coronary artery disease. Radiographics. 2011;31:1239–1254.
108. Ropers D, Pohle FK, Kuettner A, et al. Diagnostic accuracy of noninvasive coronary angi- 128. Plass A, Emmert MY, Gaemperli O, et al. The potential value of hybrid positron emis-
ography in patients after bypass surgery using 64-slice spiral computed tomography with sion tomography/dual-source computed tomography imaging in coronary bypass surgery.
330-ms gantry rotation. Circulation. 2006;114:2334–2341. Heart Surg Forum. 2011;14:E283–E290.
109. Meyer TS, Martinoff S, Hadamitzky M, et al. Improved noninvasive assessment of 129. Siegrist PT, Husmann L, Knabenhans M, et al. (13)N-ammonia myocardial perfusion
coronary artery bypass grafts with 64-slice computed tomographic angiography in an imaging with a PET/CT scanner: Impact on clinical decision making and cost-effective-
unselected patient population. J Am Coll Cardiol. 2007;49:946–950. ness. Eur J Nucl Med Mol Imaging. 2008;35:889–895.
110. Jabara R, Chronos N, Klein L, et al. Comparison of multidetector 64-slice computed
tomographic angiography to coronary angiography to assess the patency of coronary
artery bypass grafts. Am J Cardiol. 2007;99:1529–1534.
LWBK1209-ch23_p350-364.indd 364 16/05/13 9:33 PM

Chapter
William Stanford 24
CT Imaging of Coronary
Artery Calcification
■■ Introduction ■■ Conclusions
■■ Atherosclerotic Plaque Development ■■ Summary and Recommendations
■■ Historical Aspects in the Imaging of
Coronary Calcification Introduction
■■ Multidetector CT Scanners
Cardiovascular disease (CVD) affects millions of individu-
■■ Importance of Imaging Coronary als annually and is the leading cause of death in the United
Calcium States. In 2007 the American Heart Association (AHA) esti-
■■ Calcium Score Reporting mated there were 82,600,000 Americans who were suffer-
ing from some form of CVD and that 813,804 had died;
■■ Standardization of CT Scanners the economic loss was estimated at $286.6 billion (1). Of
■■ Reproducibility Issues those affected 16,300,000 were victims of angina and heart
attacks, and importantly 18% of these heart attack victims
■■ Radiation Dosage
were under age 65. Thus coronary artery disease (CAD) has
■■ Coronary Artery Calcium as an Indicator of a major impact on our economy and workforce.
Coronary Artery Stenosis Impacting these figures is the fact that CAD often remains
silent until a major catastrophic event occurs. This is ex-
■■ The Value of a Zero Calcium Score in
emplified by the fact that of the 286,000 individuals dying
Assessing Coronary Artery Stenosis from CAD in 2007, only 50% of men and 36% of women
■■ Coronary Artery Calcium Compared with had had previous symptoms. Because of this it becomes
other Imaging Modalities in the Assessment increasingly important to try to identify the individuals at
of Coronary Artery Stenosis risk before a major cardiovascular event occurs. Although
epidemiologic studies have identified numerous cardiac risk
■■ Coronary Artery Calcium as a Predictor of factors, traditional risk factors only predict approximately
Future Cardiac Events two-thirds of the patients who will eventually succumb to
■■ CAC as a Predictor of Events in Elderly heart disease (2). In fact, approximately one-third of indi-
viduals dying annually of heart disease have no identifiable
Patients
Framingham risk factors that would predict a future “hard”
■■ Calcium Score Progression cardiac event (3). Because of this considerable interest is oc-
■■ Effect of Statin Therapy on Coronary Artery curring in developing screening methodologies that can both
quantitatively measure the severity of CAD and accurately
Calcium Progression
predict future risk. Coronary artery calcification (CAC)
■■ Coronary Artery Calcification Scans as a has long been known to correlate with CAD severity and
Screening Tool Prior to Angiography multidetector computed tomography (MDCT) imaging has
■■ Coronary Artery Calcification in Emergency proven extremely sensitive in its detection. Since calcifica-
tion is a manifestation of the atherosclerotic disease process,
Room Patients with Chest Pain MDCT imaging has evolved as a major screening modality
■■ Current Recommendations from for individuals at risk for a coronary event. The fact that cal-
Professional Societies cification occurs early in the atherosclerotic process and that
365
LWBK1209-ch24_p365-375.indd 365 16/05/13 9:32 PM

MDCT screening of coronary calcification can often identify

an “at-risk population” and can do this early in the disease
process when the plaque begins to build within the wall of
the vessel and before the disease has advanced to cause ves-
sel stenoses or occlusion has been an important factor in the
development of imaging methodologies. Of these method-
ologies, CAC imaging has emerged as a prognostic indicator
of increased risk and numerous investigators have promoted
MDCT imaging of coronary calcium as an accurate, cost-
effective screening methodology for evaluating coronary
heart disease (CHD).
The indications for MDCT imaging of CAC have begun
to standardize over the past few years. Current applica-
tions have centered around: (1) The identification of indi-
viduals at risk for CAD, especially if their Agatston calcium
score is >400; (2) additional stratification of individuals at
intermediate risk per other stratification protocols such as
the Framingham Risk Score; (3) identification of individu-
als with extensive CAC where additional imaging such as
computed tomographic angiography (CTA) may be compro- Figure 24.1. Non-occlusive atherosclerotic plaque in the wall of
a coronary artery. The patient is at risk even though there is no lumi-
mised because of masking by the calcification; and (4) ruling
nal stenosis present. (Reprinted from: Davies MJ. Atlas of Coronary
out a cardiac etiology in patients presenting to emergency
Artery Disease. Philadelphia, PA: Lippincott-Raven; 1998, Figure
rooms (ERs) with chest pain and normal cardiac enzymes 2-29, with permission.)
and electrocardiograms (EKGs). In this chapter we explore
the use of coronary calcium imaging with emphasis on the
above applications.
angiographic narrowing at the site of plaque deposition (5).

Atherosclerotic Plaque Because of this myocardial perfusion and stress echocardiog-
Development raphy studies which rely on effects of occlusion are often
normal and thus traditional screening and risk-factor assess-
Coronary arteriosclerosis is a complex and unpredictable ments commonly underestimate an individual’s risk for sud-
process thought to be related to biochemical, genetic, and den cardiac death.
environmental factors. Since traditional risk factors, such as Unfortunately, these plaques are predisposed to spontane-
smoking, age, gender, obesity, hypertension, hyperlipidemia, ous rupture and thus type IV and V lesions are often referred
and diabetes mellitus are thought to be linked to overall to as “vulnerable plaques.” As such, vulnerable plaques are
risk, these assessments are important; however, they do a responsible for much of the morbidity and mortality from
relatively poor job in predicting individual risk for CHD. CAD. Why plaques rupture is unclear, but the process is
Regardless of the presence or absence of specific risk factors, likely multifactorial and related to biomechanical stresses
the development of arteriosclerosis remains an inevitable and localized plaque inflammation.
part of aging. Once a fibrous cap ruptures, the lipid core in the wall of
Atherosclerosis begins early in life. The earliest athero- the artery is exposed to flowing blood and an acute throm-
sclerotic lesion (type I plaque) is characterized by the ac- bogenic reaction frequently ensues. Repeated thrombosis
cumulation of lipid-laden macrophages (foam cells) within and subsequent recanalization may result in partial nar-
the intima of arterial walls. With further accumulations of rowing or complete occlusion, and if so, the narrowing is
intra- and extracellular lipids, type II (fatty streak) and type often manifested clinically by episodes of ischemia and/
III (pre-atheroma) lesions develop. These may be grossly vis- or infarction. If a severe stenosis is present at the site of
ible and often contain small foci of calcium on histologic plaque rupture, total vascular occlusion is often likely, and
examination. Although these lesions are hemodynamically this may account for many of the cardiac deaths seen an-
innocuous and potentially reversible, they represent lesions nually.
that can progress to the more clinically important type IV With repeated plaque rupture, type IV and V plaques be-
and V lesions (4). come transformed into type VI lesions, which are considered
Both type IV and type V lesions (atheroma) are well- histologically “complicated” plaques that are characterized
developed plaques characterized by intramural collections of by fibromuscular tissue deposition occurring from plaque
both cholesterol and phospholipids. These deposits are fre- repair and healing. With repeated injury and healing, type
quently covered by a thin, fibrous cap (fibroatheroma) and VI lesions often grow in size to eventually produce signifi-
often coexist without associated significant luminal narrow- cant arterial narrowing. Thus, type VI lesions have a greater
ing (Fig. 24.1). Therefore, the extent of the disease is often occurrence in patients with chronic and/or stable angina,
underestimated and undetected by conventional angiography. and because they produce narrowing, they are commonly
In fact, up to two-thirds of patients with acute myocardial detected by traditional imaging techniques that screen for
infarctions (MIs) or unstable angina often have only minimal the hemodynamic effects of stenoses (6).
LWBK1209-ch24_p365-375.indd 366 16/05/13 9:32 PM

Chapter 24 ■ CT Imaging of Coronary Artery Calcification 367
Historical Aspects in the Imaging ECG. Upon completion of the scan, the images can be recon-
of Coronary Calcification structed at a preselected phase of the cardiac cycle. In order
to avoid anatomical gaps in the data set, the pitch is set very
The association of calcium with coronary atherosclerosis low; thus patient radiation exposure is higher. Nevertheless,
dates back to before 1959 (7). Early studies using fluoros- the ability to reconstruct images during multiple phases of
copy showed that it was possible to identify calcific coronary the cardiac cycle and do so from the same high-resolution
plaques, but it was not until the development of the Electron data set can provide important information. However, the
Beam CT scanner in the early 1980s that the identification patient’s heart rate can be a major factor influencing image
of calcium became important in risk management protocols. quality, and if the heart rate exceeds 65 to 70 bpm, beta-
In 1983 Dr. Douglas Boyd, a researcher at University blockers are commonly administered to allow data collec-
of California, San Francisco, developed the Electron-beam tion during a single heartbeat. These advances have led to an
CT scanner (8). The unique design of this Electron-beam increased emphasis on CAC imaging.
Computed Tomography (EBCT) scanner had many advan-
tages in coronary artery imaging. Its excellent temporal reso-
lution produced images of the heart in near real time and did Importance of Imaging
so with excellent spatial resolution. This technology reduced Coronary Calcium
acquisition times to 50 to 100 milliseconds which, when
coupled with its spatial resolution, initially made EBCT the Strong correlations have been found between quantitative
new gold standard in CAC detection and quantification. measurements of coronary artery calcium (CAC) and mea-
With an EBCT 50- to 100-millisecond study the entire cor- surements of atherosclerotic plaque area and volume (11).
onary circulation could be scanned within a 20- to 30-second Supporting these observations was an article by Mautner
breath-hold. An additional advantage was that the effective et al. (12) who examined 1,298 segments from 50 heart
radiation dose was only approximately 0.7 mSv (9). specimens and observed that 93% of arteries with steno-
These scanners have now given way to newer-generation ses greater than 75% had CAC. Conversely, only 14% of
MDCT scanners. The introduction of multidetector-row arteries with stenoses less than 25% were associated with
scanners (4, 8, 16, 32, 64, and now 256 and 320 dual source calcium. However, in spite of these sensitivities calcium
scanners) have improved z-axis coverage per revolution and measurements derived from CT cannot predict site-specific
thus have decreased temporal resolutions to ±75 millisec- stenoses as shown in articles by Bormann et al. (13) and
onds (10). Sangiorgi et al. (14). Therefore, CAC measurements cannot
be used to predict site-specific locations nor the severity of
the stenoses.
Multidetector CT Scanners
MDCT scanners have detectors capable of generating up to Calcium Score Reporting
320 slice images with each gantry rotation, and gantry rota-
tion times have decreased from 1,000 to 280 milliseconds, Several CAC software scoring packages which enable opera-
and with segmented reconstruction and dual source tech- tors to quickly perform CAC quantitative measurements are
nologies, temporal resolutions of 75 milliseconds are now commercially available (Fig. 24.2). In these packages, coro-
possible. An added advantage of MDCT imaging is that the nary calcification thresholds are generally set at two to three
information generated can occur as a volumetric data set contiguous pixels at attenuations of +130 Hounsfield units.
and this permits reformations at various times in the cardiac The +130 Hounsfield unit threshold was chosen because it
cycle. MDCT images are commonly ECG gated, and this is approximately two standard deviations higher than the
further decreases motion unsharpness by allowing image attenuation of blood and has been found to correlate well
acquisition during the quieter phase of the cardiac contrac- with plaque calcium measurements (15).
tion. The latter is particularly important in coronary calcifi- Calcium scores are reported using the Agatston score
cation imaging. (15), volume score, and mass score.
Current multidetector CT scanners generate images by The Agatston score was the initial reporting score and is
both prospective and retrospective gating. In prospective used in much of the older literature. It used EBCT technol-
gating the scanner is activated only during the time needed ogy to identify lesions located over the course of the coro-
to acquire the image. Data collection is initiated from the nary artery. In calculating the score, a region of interest is
R-wave of the ECG and the operator can select the desired placed around each lesion and the area of the lesion is mul-
delay. This mode is frequently used in coronary artery cal- tiplied by a factor of 1 to 4 with 1 being a peak calcification
cium imaging because patient radiation dose can be kept to of 130 to 199 HU, 2 being a peak calcification of 200 to 299
a minimum. However, rapid or irregular heartbeats can af- HU, 3 being a peak calcification of 300 to 399 HU, and 4
fect both image quality and reproducibility. Current scan- being a peak calcification of >400 HU.
ners can be programmed to provide 0.5- to 0.6-mm slice The volume score reports the volume of the threshold cal-
thicknesses of the heart during one cardiac contraction and cifications.
do so with scanning times as rapid as 75 milliseconds (10). The mass score uses a calibration factor derived from the
Retrospective gating is also a commonly used operating simultaneous imaging of a phantom containing a known amount
mode. In this sequence helical scanning of the entire heart of calcium. The phantom is placed in the scanning field and a cali-
occurs along with the simultaneous recording of the patients’ bration factor determined. From it the calibration factor times
LWBK1209-ch24_p365-375.indd 367 16/05/13 9:32 PM

A B
Figure 24.2. Sixty-four–slice multidetector CT image showing calcium deposits in the left anterior
descending and circumflex coronary arteries (A). The calcifications show up as white areas in their respective
arteries. B: Same subject as in (A). The scoring program identifies each calcification (red area in LAD) and
prints out the Agatston, volume, and total scores.
the volume of a voxel times the number of voxels contain- serial CT scans on repeat imaging as well as between the
ing the threshold calcification times the mean CT Hounsfield various scanners. To assess for disease progression, CT mea-
number for each lesion equates to the mass score. The total surements must be accurate and reproducible. To explore
score is the sum of the individual scores. this issue, a number of studies have examined inter-scan
These values can then be compared to normalized popula- variability between EBCT and helical scanners and between
tion data bases, based on age- and gender-matched controls, the various helical scanners. These studies report a small
and can be used in assessing both disease severity and for inter-scan variability between EBCT and MDCT of approxi-
determining disease progression. The calcium volume score mately 15%; inter-reader variability of approximately 3%;
and calcium mass score are increasingly used in reporting and intra-reader variability of less than 1% (17,18).
calcium burden. Factors primarily responsible for the increased inter-scan
variability are slice misregistration resulting from cardiac
and respiratory motion, noncontiguous acquisition of image
Standardization of CT Scanners data (image gaps), image noise, changes in scanner calibra-
tion, cardiac arrhythmias/ECG triggering problems, and pa-
In order to assure calcium score reproducibility, it is impor- tient motion.
tant that CT scanner protocols be standardized so that scores The use of volumetric measurement (19) has helped re-
from one scanner can be compared with another. Toward that duce variability, as has the incorporation of a phantom as
end the Physics Task Group of the International Consortium part of the calcium mass examination. The volumetric and
on Standardization in Cardiac CT was formed (16). Using a calcium mass scores are increasingly being reported at many
phantom with inserts of calcium and water density material of the sites performing CAC screening.
embedded in an epoxy anthropomorphic body torso, scan- Reproducibility was recently studied in Multi-Ethnic
ning algorithms for all five commercially available scanners Study of Atherosclerosis (MESA) patients. Duplicate scan
were developed. The manufacturers were: Toshiba, Imatron, differences in the MESA study of 6,814 multiethnic individ-
General Electric, Phillips, and Siemens. These scanners were uals, scanned on three EBT and three MDCT scanners, were
calibrated against the phantom for temporal and spatial reso- 20.1% for the Agatston score, and 18.3% for the volume
lution and noise. To accommodate different patient sizes, cir- score (p < 0.01) (20). Current reported variability on 16-
cumferential rings were added. Using consortium-developed to 64-slice MDCT scanners range from 8% to 18% for the
scanner algorithms variations of 4% for Agatston scores, Agatston, volume, and mass scores. Thus in assessing CAC
7.9% for volume scores, and 4.9% for mass scores could be progression, this variability must be taken into account be-
achieved. The calculated calcium score was found to be within fore calcium progression can be assured.
±5 mgm of the actual calcium mass of the phantom (16).
Radiation Dosage
Reproducibility Issues
The effective radiation dosage is a measure of the total radia-
As mentioned, a major consideration in the utilization of cal- tion exposure to the patient. This is reported in millisieverts
cium scores is the variability of calcification scores between (mSv) and is frequently equated to months of background
LWBK1209-ch24_p365-375.indd 368 16/05/13 9:32 PM

radiation exposure. For an EBCT prospective calcium study, (28). These investigators evaluated the presence of CAC in
the effective dose approximates 1 mSv in males and 1.3 mSv detecting >50% obstructive plaque as compared to CTA and
in women (4 and 5.2 months of background radiation, conventional angiography in low-risk patients and found the
respectively) (21). following:
For a MDCT prospectively triggered coronary calcifica-
Ca Score 0 = 3.1% obstructive plaque
tion study, the effective dose approximates 1.5 mSv in males
Ca Score 1 - 200 = 15.1% obstructive plaque
and 1.8 mSv in women (6 and 7.2 months of background
Ca Score > 200 = 50% obstructive plaque
radiation). However, newer MDCT scanners can deliver pro-
spective coronary calcification scoring with radiation doses Their conclusions were that patients with low calcium
of <1 mSv (10). If the MDCT coronary calcium study were scores had a low presence of obstructive plaque.
retrospectively gated, the radiation exposure would increase A meta-analysis of MDCT calcification accuracy in detect-
to approximately 3 mSv in males and 3.6 mSv in women ing obstructive coronary plaque was that of Abdulla et al.
which would equate to 12 and 14.4 months of background (29). These investigators reported MDCT 64 calcium scores
radiation. Recent developments in modulation techniques in detecting significant obstructive CAD in 19 studies of
however can now ramp down the power during noncritical 1,120 patients with low CAC scores and 514 patients with
imaging times and this can decrease radiation exposure by high CAC scores. They found the following:
up to 80% (22).
With all imaging radiation dose should be as low as pos-
sible and since individuals may have multiple CT scans over CAC
their lifetime, total radiation doses may be problematic as Score Sensitivity Specificity PPV NPV Accuracy
shown by Kim et al. (23) who estimated if men 45 to 75 years
were to have a CAC scan every 5 years, the estimated can- <10 97.5 90 85 97 91
<100 97 88.5 86 91 89
cer risk using a median dose of 2.3 mSv was 42 cases per
>400 97.5 42 — — —
100,000 men and 62 cases per 100,000 women.
However, in spite of CT’s reported high calcium score

sensitivity, Bormann et al. (13) found, as stated previously,
Coronary Artery Calcium as that EBCT calcification scores were not predictive of signifi-
an Indicator of Coronary cant stenosis at the site of calcification and that no receiver
Artery Stenosis operator characteristic curve could be found that would sug-
gest a clinically useful calcification score as an indicator of
For many years researchers have recognized an association >70% site-specific stenosis. In a later study, Stanford et al.
between CAC and coronary artery stenosis, and a number (30) examined data from 150 patients undergoing EBCT
of reports have shown that as EBCT calcium scores increase, and coronary angiography from two institutions and found
the likelihood of having a significant stenosis (greater than only one patient with greater than 50% stenosis who had
50% luminal narrowing) increases. no coronary calcification. His conclusion was that the ab-
These studies were summarized by O’Rourke et al. (24) sence of calcific deposits seems to negate the presence of sig-
in a meta-analysis of 3,683 patients who underwent both nificant luminal stenosis in all but 5% to 7% of patients.
EBCT and coronary angiography. They found the presence However, in a recent report from the MESA trial, Rosen
of CAC had a weighted sensitivity for detecting significant et al. (31), in assessing 175 consecutive patients undergoing
stenoses of 80%, a weighted specifically of 39%, and a pre- coronary angiography to determine the extent of coronary
dictive accuracy of 59%. stenosis, found 16% of arteries with >75% stenosis had zero
Because of wide variations in calcium score reporting, Ca++ scores. Therefore, in using calcium scores to assess ob-
Rumberger et al. (25) attempted to optimize levels of cal- structive disease, these observations need to be taken into
cium that equated with the best sensitivity and specificity. account.
He suggested that a calcium score of 80 optimized specific- Since it is recognized that calcified plaques are often pres-
ity (85%) while still maintaining a reasonably good overall ent in non-obstructive lesions, the presence of calcium in as-
sensitivity of 84%. In addition, he found that calcium scores ymptomatic persons does not provide sufficient rationale for
between 167 and 370 correlated with a 90% specificity for revascularization, but rather suggests the need for more inten-
detecting stenosis with >50% narrowing, while calcium sive risk factor modification and possible further functional
scores >370 were associated with a 90% specificity for at assessment. The current thinking is that a positive calcium
least one luminal stenosis of 70% or more (26). score indicates atherosclerosis and not necessarily stenosis
The largest series reporting EBCT calcium as an indica- and that even though CT coronary artery calcium determi-
tor of stenosis was a multicenter study of 710 patients with nations have a high sensitivity and a high NPV for detect-
symptomatic CAD. This study included 456 men and 254 ing obstructive CAD, they have limited specificity. Still, CAC
women (mean age, 56 years) (27). The study reported a sen- reporting can assist the clinician in effectively “ruling out”
sitivity of 95% for the detection of significant stenosis and a angiographically significant lesions in symptomatic patients,
specificity of 44%. The positive predictive value was 72%, and a number of authors, as discussed later, have suggested
and the negative predictive value (NPV) was 84%. that CAC scoring be used as a filter prior to angiography and
Similar studies have been reported in the MDCT liter- that zero calcium scores may negate the necessity of undergo-
ature. A representative study was that of Venkatesh et al. ing invasive angiography and/or CTA.
LWBK1209-ch24_p365-375.indd 369 16/05/13 9:32 PM

The Value of a Zero Calcium Miranda et al. (37) also demonstrated a relationship
Score in Assessing Coronary between CAC and SPECT imaging. In 233 consecutive as-
Artery Stenosis ymptomatic patients who had CAC and SPECT imaging, no
patient with a CAC <100 had an abnormal SPECT, whereas
Possibly more important than a positive calcium score is the 4.1% with a moderate calcium score (101 to 400) had an
value of a zero score. Several investigators have addressed abnormal SPECT and 15% with a CAC >400 had an abnor-
this issue. One study by Becker et al. (32) used EBCT to mal SPECT. The best CAC cutoff for predicting an abnor-
study 1,347 symptomatic subjects with suspected CAD. He mal SPECT in this study was a score >400.
found an overall sensitivity of any calcium in predicting Several meta-analysis studies have also reviewed this issue.
stenosis was 99%; however, the specificity was only 32%. In a meta-analysis comparing exercise echocardiography and
Absolute scores ≥100 and/or >75th percentile were identi- MPS in diagnosing obstructive disease, Fleischmann et al.
fied as the cutoff levels that provided the highest sensitivities (38) found that exercise echocardiography (2,627 patients;
(86% to 89%) and the lowest false positive rates (20% to 69% men) had a sensitivity of 85% and a specificity of 77%
22%) for significant obstructive plaque. Becker concluded as compared to exercise SPECT (3,237 patients; 77% men)
that the absence of coronary calcium was a highly accurate that had a sensitivity of 87% and a specificity of 64%.
indicator for excluding significant coronary artery stenosis.
Coronary Artery Calcium
as a Predictor of Future
Coronary Artery Calcium Cardiac Events
Compared with Other Imaging
Modalities in the Assessment In recent years multiple studies have been published address-
of Coronary Artery Stenosis ing CAC as a predictor of cardiac risk. These data have been
reported as all-risk mortality, risk for any cardiac event
Compared with traditional, noninvasive testing, CAC quan- including revascularization and risk for hard events lim-
tification has been shown to be as good a predictor of occlu- ited to documented MIs and cardiac death. In all of these
sive CAD as ECG exercise testing and as good as myocardial instances a comparative baseline needed to be established.
perfusion scintigraphy (MPS). Supporting this was a report The most commonly used baseline is that of Hoff et al. (39)
by Kajinami et al. (33) who found a CAC sensitivity of 77% who determined age and gender distributions of coronary
and a specificity of 86% in symptomatic patients, compared calcium in 35,246 asymptomatic adult individuals. These
with 74% and 73% with ECG stress testing and 83% and data were developed using EBCT technology and reported
60% for thallium perfusion imaging in detecting significant as Agatston scores. Although similar studies are being col-
obstructive disease. lected for MDCT scanners, only 2,005 are currently included
Exercise stress EGG testing is a diagnostic test frequently in the Cleveland Clinic registry and they have not yet been
used in patients with suspected CAD. However, exercise ECG reported. Since the Hoff data are thought representative, this
testing often results in a substantial number of false negative is the data currently used in most commercial calcium scor-
results. Lamont et al. (34) studied 153 patients who under- ing algorithms for risk assessment comparisons.
went EBCT and coronary conventional angiography imaging Since CAC scores tend to vary, the absolute score is often
because of a positive treadmill stress test. In patients with a undependable. Subsequently risk guidelines were developed
CAC score of zero, the NPV was 93%. From this, the authors and many investigators now follow the guidelines proposed
concluded that the absence of CAC reliably identified patients by Rumberger et al. (40) which are as follows:
with a false-positive exercise ECG result. Raggi et al. (35) also
evaluated exercise ECG and found that in patients with low •• Ca++ scores 0 to 10 or <75% indicate low risk.
to intermediate pre-test probability of disease, a calcium score •• Ca++ scores 11 to 100 or >75% indicate intermediate risk
of zero can reliably exclude significant obstruction. and suggest the need for more aggressive therapy.
A number of studies have studied the relationship be- •• Ca++ scores >400 or >90% indicate high risk.
tween MPS and CAC in detecting significant coronary artery In 2003 Pletcher et al. (41) reported a meta-analysis of
stenosis. Berman et al. (36) reported a positive relationship studies between 1980 and 2003. Their analysis included
between stress-induced myocardial ischemia as seen on sin- 13,000 asymptomatic patients who were screened with
gle-photon emission computed tomography (SPECT) perfu- EBCT and followed for up to 3.6 years to evaluate the odds
sion studies versus EBCT CAC. He studied 1,195 patients ratios (OR) for hard coronary events. They found the OR
without known coronary disease and found the frequency for Agatston scores <100, 100 to 400, and >400 were 2.1,
of ischemic findings on SPECT imaging was <2% with CAC 4.2, and 7.2, and concluded that the EBCT Agatston scores
scores <100. SPECT ischemia increased progressively with were independent predictors of hard coronary event risk in
CAC >100 (p = <0.0001) and patients with CAC scores asymptomatic subjects.
>400 had a high likelihood of having myocardial ischemic In the same year, Shaw et al. (42) reported the relation-
changes. The authors concluded that ischemia as diagnosed ship of CAC to all-cause mortality in the largest series stud-
by SPECT is associated with a high likelihood of subclinical ied to date. The studies included 10,377 asymptomatic in-
atherosclerosis, but SPECT obstruction is rarely seen in pa- dividuals (40% women), who were followed for an average
tients with CAC scores <100 and that low CAC scores could of 5 years. They found that in both men and women, CAC
obviate the need for subsequent SPECT testing. was an independent predictor of death (p < 0.001) with
LWBK1209-ch24_p365-375.indd 370 16/05/13 9:32 PM

the risk increasing proportionally to the baseline calcium CAC imaging studies were done in 4,870 patients. The re-
scores: sults showed MACE events occurred in 146 (0.56%) asymp-
tomatic patients and in 702 (1.80%) symptomatic patients.
In 18 studies that included invasive coronary angiography
Calcium Score Relative Risk the presence of any CAC had a pooled sensitivity 98% and
an NPV 93% for detecting the presence of any significant
11–100 1.6
101–400 1.7
CAD. Thus on basis of review of over 64,000 patients, the
401–1,000 2.5 absence of CAC is associated with very low risk of future
>1,000 3 cardiovascular events in this low-risk group.
Thus, these large observational studies appear to show
that coronary calcium can provide independent incremental
Budoff and Gul (43) also reported mortality as compared
information in addition to traditional risk factor observa-
to calcium scores. Cumulative survival is shown in their
tions in the prediction of all-cause mortality.
Figure 5 (Fig. 24.3).
Other recent data using MDCT technology have repli-
cated these findings. An article by Jain et al. (44) evaluated
coronary calcium as a determinate of risk for Major Adverse
CAC as a Predictor of Events
Cardiac Events (MACE). Their study included 4,965 MESA in Elderly Patients
asymptomatic participants (mean age 62 years; 48% men)
who were followed for 5.8 years. They reported 297 CVD The above data also appear to hold true in older patients. In
events, of which 187 were CAD related. The CVD hazard the Rotterdam Coronary Calcification Study, 2,013 partici-
ratio for individuals with a CAC >0 was 2.3. Their conclu- pants (mean age 71 years) were evaluated (47).
sions were that the addition of CAC scoring significantly im- Compared to subjects in the lowest calcium score cat-
proved risk prediction. egory (0 to 100), the age-adjusted OR for MI in the highest
In a meta-analysis of articles between 1975 and 2007, score category was 7.7 for men and 6.7 for women. It was
Shareghi et al. (45) reviewed 35,765 asymptomatic pa- also reported that subjects were three times more likely to
tients including 16,406 (45%) who had zero CAC scores. have experienced a stroke with calcium scores >500 (OR
In these studies there were 48 hard events for an event rate 3.3) as compared to subjects in the lowest calcium score
of 0.027%. The pooled sensitivity for a cardiac event was category (<100), and adjusting the data to include carotid
98.1% and the pooled NPV was 99.9%. Their conclusions intimal thickness and cardiovascular risk did not alter the
were that a zero CAC is associated with a low cardiac event adjustment of these risk estimates. Only CAC was predictive
rate and that CAC scores can be used in counseling patients’ of MACE.
future cardiac events. Another study in elderly patients was by Raggi et al.
A second meta-analysis of studies between 1990 and (48). These investigators followed 35,388 patients (3,570
2008 with a follow-up of 51 months included 13 studies of patients >70 years; 50% women) for 5.8 years. The overall
25,903 asymptomatic individuals and 7 studies of 38,970 survival was 97.9%. The survival for <40-year-old men with
symptomatic individuals (46). Myocardial perfusion and Ca++ >400 compared to >80-year-old men with >400 score
was 88% versus 19%; for women it was 95% versus 44%
(p < 0.0001). Hence increased CAC scores were associated
with decreasing survival across all age deciles (p < 0.0001). In
this study 20,562 individuals had zero Ca++ score and the an-
Cumulative survival by coronary calcium score
nual mortality rates were 0.3% to 2.2% (p < 0.0001). Their
1.00 0 (n = 11,044) conclusion was that CAC can discriminate mortality risk even
1–10 (n = 3,567) in the elderly and CAC can be used to reclassify risk in elderly
11–100 (n = 5,032)
0.95 patients.
101–299 (n = 2,616)
Cumulative survival
Calcium Score Progression

0.90
300–399 (n = 561)
400–699 (n = 955)
0.85
Progression of CAC is generally calculated as a percent or
700–999 (n = 514) absolute change from the baseline score. Raggi et al. defined
0.80 a change >15% as true progression (49). In all studies of
progression the changes in scores must exceed scanner varia-
0.75 tions and reproducibility in order to be valid.
1,000 + (n = 964) Unfortunately, annual CAC progression rates may range
0.70 from 20% to 24% per year as per the Agatston and/or
0.0 2.0 4.0 6.0 8.0 10.0 12.0 the volume scores. Factors that may influence the rate of
Time to follow-up (years) change include the patient’s overall CAC score, gender, age,
Figure 24.3. Survival curves as compared to coronary calcium family history of premature CAD, ethnicity, diabetes, body
score. (Reprinted from: Budoff MJ, Gul KM. Expert review on coro- mass index, hypertension, and renal insufficiency. Most pa-
nary calcium. Vasc Health Risk Manag. 2008;4:315–324, Figure 5, tients will increase their CAC scores with time. However, a
with permission.) few (29% to 34%) may exhibit no change, especially if they
LWBK1209-ch24_p365-375.indd 371 16/05/13 9:32 PM

are at low Framingham risk, are women, or are individuals

with a zero score (50). Thus disease progression needs to be CAC
carefully evaluated if calcium scores are to be used in the Score Sensitivity Specificity PPV NPV
determinant. <400 100 91 74 100
>400 100 17 75 100
Effect of Statin Therapy on Their conclusion was that patients with CAC scores >400
Coronary Artery Calcium should not undergo CTA but undergo coronary angiography
Progression instead.
Finally, Abdulla et al. (29) examined CAC thresholds as a
A number of studies have evaluated changes in CAC scores gatekeeper prior to CTA imaging. They found CTA accura-
following treatment with statins. In representative stud- cies as compared to CAC scores to be the following:
ies untreated patients had an average CAC progression
of 36%, while those taking statins had changes averaging
13% (51). CAC
However, it must be recognized that other reports have Score Sensitivity Specificity PPV NPV Accuracy
failed to confirm these findings and therefore the ACCR/
<10 97.5 90 85 97 91
AHA consensus document paper of 2007 did not recom-
<100 97 88.5 86 91 89
mend CAC scoring as a means of following patients on >400 97.5 42 — — —
statin therapies (52).
They concluded CAC thresholds as a gatekeeper are arbi-
trary and do not necessarily warrant canceling CTA studies.
Coronary Artery Calcification
Scans as a Screening Tool Prior
to Angiography Coronary Artery Calcification in
Emergency Room Patients with
A more recent application of CAC imaging is in determin- Chest Pain
ing the influence of CAC in validating CTA observations in
coronary artery stenosis imaging. Ong et al. (53) studied 134 Numerous articles have addressed this issue. Representative
symptomatic patients undergoing CTA with MDCT 64 as articles by White et al. (56) have stated that in patients present-
compared to conventional angiography and found that mod- ing to the ER with chest pain with negative cardiac enzymes
erate to heavy CAC limits the accuracy of CTA in detecting and indeterminate ECGs that CAC can play an important
coronary artery stenosis. His data showed the following: role, and if there is no calcium present, the patients can be
safely discharged without risk of significant coronary events.
One study in which CAC was used in ER patients with
CAC
chest pain, negative enzymes and indeterminate ECGs was
Score Sensitivity Specificity PPV NPV
reported by Laudon et al. (57). These investigators per-
<142 85.4 98.1 76.7 99.2 formed CAC imaging in the emergency department in 104
>142 79.9 92.8 78.8 93.5 patients, and noted an NPV for CAD of 100% for a CAC
score of zero. In a similar study, McLaughlin et al. (58)
Confirming these conclusions was a report by Stolzmann reported an NPV of 98% in 134 patients in an ER setting. In
et al. (54). These investigators studied 100 patients involv- addition, Georgiou et al. (59) followed 198 patients present-
ing 1,462 coronary artery segments. The scanners used ing to the ER with chest pain and normal ECG and cardiac
were dual source scanners with prospective triggering. They enzymes. He found that patients with zero CAC scores can
found the following: be safely discharged given the extremely low rate of future
events (∼0.1% per year).
More recently, Rubinshtein et al. (60) assessed the
CAC severity of CAD using 64-MDCT in patients admitted to the
Score Sensitivity Specificity PPV NPV ER for investigation of acute coronary symptoms. In 668
consecutive patients, 231 had a low (<100) or zero CAC
<316 99 99 94 100 score; and of these, obstructive CAD was present in only
>316 98 99 94 99 (p ≤ 0.001) 9 of 125 patients (7%) with a zero score, and in only 18 of
106 (17%) with a low score (1 to 100). From these data,
Their conclusion was that CTA in patients with high CAC these investigators concluded a zero CAC score seemed to
was associated with an increased number of non-diagnostic be a good predictor for the exclusion of significant CAD.
segments, but overall accuracy remained high even in heavily However, a low CAC score is more controversial since a
calcified segments. number of studies have shown that the presence of non-
Diederichsen et al. (55) also reported on this issue. These calcified and potentially obstructive lesions is higher in pa-
investigators studied 109 patients undergoing CTA using tients with low calcium scores as compared to patients with
MDCT 64 scanning and found the following: a score of zero.
LWBK1209-ch24_p365-375.indd 372 16/05/13 9:32 PM

Current Recommendations from Conclusion

Professional Societies
A vast majority of heart attacks occur at the site of non-obstruc-
The 2003 European Cardiovascular Guidelines state, “The tive atherosclerotic plaques (5). Since exercise testing and/or
calcium score is an important parameter to detect asymp- cardiac imaging (nuclear scintigraphy/echocardiography) only
tomatic individuals at high risk for future CVD events, inde- diagnose high-grade coronary stenoses, they often fail to iden-
pendent of the traditional risk factors.” (61). tify a large number of asymptomatic patients at risk due to
In 2005 an AHA Consensus Statement addressing the im- non-obstructive coronary plaque. Framingham risk assessment
aging of coronary artery calcium in women stated, “Given models also miss a significant proportion of patients, and may
the evolving literature, current data indicate that CAD risk also inappropriately stratify intermediate- to low-risk patients.
stratification is possible. Specifically, low CAC scores are as- Expert Consensus Documents from the ACCR and AHA
sociated with a low adverse event risk, and high CAC scores have stated that in individuals in intermediate-risk catego-
are associated with a worse event-free survival.” (62). ries with multiple risk factors, high coronary calcium scores
In 2006 AHA published another scientific statement on CT (>400 or >75th percentile) indicate advanced coronary ath-
imaging of CAC (63). This statement concluded the following: erosclerosis and provide a rationale for intensification of
lipid-lowering therapy.
1. CT CAC can identify individuals at increased risk for MI
A negative test (zero score) makes the presence of ath-
or cardiac death.
erosclerotic plaque, including unstable or vulnerable plaque,
2. CT CAC is appropriate for further defining risk stratifica-
less likely and is consistent with a low risk (0.1% per year)
tion in individuals at intermediate risk via Framingham
for a cardiovascular event within the next 2 to 5 years.
scores.
3. CT CAC imaging is not recommended for individuals
classified as low risk. Summary and Recommendations
4. CT CAC imaging is not recommended for individuals
classified as high risk. The strengths and weaknesses of CT-derived measurements
5. CT CAC imaging is not recommended for following proof coronary calcium as a screening modality for CHD have
gression of coronary calcification. been reported in a number of review articles, and in posi-
6. CT CAC imaging is useful in evaluating luminal stenosis tion papers originating from AHA, American College of
in symptomatic patients. Cardiology, American College of Radiology, and other
Another professional society document is the ACCF/ prestigious societies. On the basis of these reports several
AHA Expert Consensus Document published in 2007 (64). assumptions concerning CAC seem valid and appear to
It concluded the following: have diagnostic implications. First, coronary calcification is
an unequivocal marker for atherosclerotic disease. Second,
1. CAC CT imaging has incremental value in further strati- CAC burdens correlate with greater overall severity of the
fying risk in intermediate-risk assessment algorithms such atherosclerotic process and with the probability of a ste-
as Framingham Risk Score. nosis somewhere within the coronary circulation. Third,
2. CAC can identify individuals at increased risk for MI and higher calcium scores, compared with age- and sex-matched
CVD death. controls, seem to signify a higher overall risk for a future
3. Asymptomatic individuals at low risk are not candidates cardiac event. Fourth, a negative CAC examination has a
for CAC scanning. high NPV for non-significant obstructive coronary disease,
4. Asymptomatic individuals at high risk are not candidates regardless of age or gender.
for CAC scanning. With these assumptions in mind, calcium scores are being
5. There is not enough evidence to reduce treatment in inter- used in (1) quantifying calcium burden to help assess risk in
mediate-risk patients with low calcium scores. asymptomatic individuals who have one or more traditional
6. At this time there is not enough evidence to compare CAC risk factors for CHD; (2) evaluating patients presenting with
imaging with alternative risk management approaches. equivocal symptoms of CHD; (3) helping assess opportune
7. Patients with atypical symptoms of cardiac disease may times for initiating therapies and instituting risk-factor mod-
benefit from CAC imaging. ifications; and (4) monitoring disease progression to measure
An anticipated study is the MESA study (65). The MESA the efficacy of therapeutic initiatives, particularly in patients
is a National Institutes of Health study which was initiated with high CAC scores.
in July 2000 to help establish the clinical utility of MDCT In an effort to bring clinical relevance to CAC screen-
and EBCT scanning of CAC as well as other modalities in ing, investigators have attempted to establish risk stratifica-
determining the role that these entities could play in assess- tion. These efforts have produced cut-points and thresholds,
ing CVD prevention. The study consists of a population- which seem clinically useful in optimizing the sensitivities
based sample of 6,814 men and women aged 45 to 84 years. and specificities of CAC measurements for detecting CAD.
All were asymptomatic for CVD. Measurement of CAC by Similarly, efforts at establishing risk strata based on calcium
EBCT or MDCT was done initially and repeated at 2.5 years. scores matched to age and gender are being used to pro-
The objective was to study the role of CAC as a predictor vide guidelines for determining individual risk for CHD.
of cardiac event risk. The study was to run 7 years. Events Empirically, those efforts can provide much-needed direc-
included acute MI, other CHD events, stroke, peripheral tion for clinicians who elect to use calcium scores to help
vascular disease events, and mortality. determine need and timeliness of therapeutic initiatives.
LWBK1209-ch24_p365-375.indd 373 16/05/13 9:32 PM

The use of calcium screening to assess risk seems most 21. Hunold P, Vogt FM, Schmermund A, et al. Radiation exposure during cardiac CT: Effective
dose at multi-detector CT and electron-beam CT. Radiology. 2003;226:145–152.
appropriate on a case-by-case basis. However, experts have 22. Jakobs TF, Becker CR, Ohnesorge B, et al. Multi-slice helical CT of the heart with retro-
been reluctant to recommend CAC scoring as a universal spective ECG gating: Reduction of radiation exposure by ECG-controlled tube current
modulation. Eur Radiol. 2002;12:1081–1086.
screening examination, primarily because of the low speci- 23. Kim KP, Einstein AJ, Berrington de Gonzalez A. Coronary artery calcification screening:
ficities in identifying occlusive CAD and the lack of stan- Estimated radiation dose and cancer risk. Arch Intern Med. 2009;169:1188–1194.
dardized threshold criteria for significance predicated by 24. O’Rourke RA, Brundage BH, Froelicher VF, et al. American College of Cardiology/
American Heart Association expert consensus document on electron-beam computed
both age and gender. tomography for the diagnosis and prognosis of coronary artery disease. Circulation. 2000;
Although outcome studies have shown that CAC quanti- 102:126–140.
fication by CT may have value as a predictor for CHD risk, 25. Rumberger JA, Behrenbeck T, Breen JF, et al. Electron beam computed tomography for
diagnosis of coronary artery disease: A cost analysis of various diagnostic testing pathways
the length of patient surveillance in most longitudinal stud- (abstract). Circulation. 1995;92:I–650.
ies has been relatively short and patient populations have 26. Rumberger JA, Brundage BH, Rader DJ, et al. Electron beam computed tomographic coro-
nary calcium scanning: A review and guidelines for use in asymptomatic persons. Mayo
been small. Furthermore, study design concerns, recruitment Clin Proc. 1999;74:243–252.
biases, and the use of inconsistent threshold criteria have re- 27. Budoff MJ, Georgiou D, Brody A, et al. Ultrafast computed tomography as a diagnos-
sulted in confusion as to the potential long-term benefits of tic modality in detection of coronary artery disease: A multicenter study. Circulation.
1996;93:808–904.
CAC screening. Until additional long-term outcome data are 28. Venkatesh V, Ellins ML, Yang S, et al. Incremental detection of coronary artery dis-
generated, consensus opinion is currently opposed to pop- ease by assessment of non-calcified plaque on coronary CT angiography. Clin Radiol.
ulation-based CT calcium screening programs. Currently, 2009;64:250–255.
29. Abdulla J, Pedersen JS, Budoff M, et al. Influence of coronary calcification on the diagnos-
CAC examinations are not indicated in patients with docu- tic accuracy of 64-slice computed tomography coronary angiography: A systematic review
mented CHD, or in younger, asymptomatic adults, because and meta-analysis. Int J Cardiovasc Imaging. 2011 (electronic pre-publication) Jun 12
(E pub ahead of print). DOI 10.1007/s10554-011.9902-6.
of the low prevalence of coronary disease in these popula- 30. Stanford W, Breen J, Thompson B, et al. Can the absence of coronary calcification in ultra-
tions. fast CT be used to rule out of (sic) nonsignificant coronary artery stenosis? (abstract). J Am
Coll Cardiol. 1992;19:189A.
31. Rosen BD, Fernandes V, McClelland RL, et al. Cardiovascular events with absent or
minimal coronary calcification: The Multi-Ethnic Study of Atherosclerosis (MESA). JACC
REFERENCES Cardiovasc Imaging. 2009;2(suppl 10):1175–1183.
32. Becker A, Leber A, White CW, et al. Multislice computed tomography for determination
1. American Heart Association. Heart Disease and Stroke Statistics—2011 Update. Dallas, of coronary artery disease in a symptomatic patient population. Int J Cardiovasc Imaging.
TX: American Heart Association; 2011. 2007;23:361–367.
2. Burke AP, Farb A, Tashko G, et al. Is calcification a marker of stable or unstable coronary 33. Kajinami K, Seki H, Takekoshi N, et al. Noninvasive prediction of coronary atherosclerosis
plaques? (abstract). Circulation. 1998;179(suppl I):I–655. by quantification of coronary calcification using electron beam computed tomography:
3. Taylor AJ, Burke AP, O’Malley PG, et al. A comparison of the Framingham Risk Index, Comparison with electrocardiographic and thallium exercise stress test results. J Am Coll
coronary artery calcification, and culprit plaque morphology in sudden cardiac death. Cardiol. 1995;26:1209–1221.
Circulation. 2000;101:1243–1248. 34. Lamont DH, Budoff MJ, Shavelle DM, et al. Coronary calcium scanning adds incremental
4. Stary HC, Chandler AB, Dinsmore RE, et al. A definition of advanced types of athero- value to patients with positive stress tests. Am Heart J. 2002;143:861–867
sclerotic lesions and histological classification of atherosclerosis: A report from the com- 35. Raggi P, Callister TQ, Cooil B, et al. Evaluation of chest pain in patients with low to
mittee on vascular lesions of the council on arteriosclerosis, American Heart Association. intermediate pretest probability of coronary artery disease by electron beam computed
Circulation. 1995;92:1355–1374. tomography. Am J Cardiol. 2000;85:283–288.
5. Little WC, Constantinescu M, Applegate RJ, et al. Can coronary angiography predict the 36. Berman DS, Wong ND, Gransar H, et al. Relationship between stress induced myocar-
site of a subsequent myocardial infarction in patients with mild-to-moderate coronary dial ischemia and atherosclerosis measured by coronary calcium tomography. J Am Coll
artery disease? Circulation. 1988;78:1157–1166. Cardiol. 2004;44:923–930.
6. Beckman JA, Ganz J, Creager MA, et al. Relationship of clinical presentation and calcifica- 37. Miranda RS, Schisterman EF, Gallagher AM, et al. The extent of coronary calcium by
tion of culprit coronary artery stenoses. Arterioscler Thromb Vasc Biol. 2001;21:1618–1622. electron beam computed tomography discriminates the likelihood of abnormal myocardial
7. Blankenhorn DH, Stern D. Calcification of the coronary arteries. Am J Roentgen Radium perfusion SPECT (abstract). Circulation. 2000;102(suppl II):II–543.
Ther Nucl Med. 1959;81:772–777. 38. Fleischmann KE, Hunink HG, Kuntz KM, et al. Exercise echocardiography or exercise
8. Boyd, DP, Lipton, MJ. Cardiac computed tomography. Proceedings of the IEEE. 1983; SPECT imaging? A meta-analysis of diagnostic test performance. JAMA. 1998;280:913–920.
3:298–397 39. Hoff JA, Chomka EV, Krainik JA, et al. Age and gender distributions of coronary artery
9. Morin RL, Gerber TC, McCullough CH. Radiation dose in computed tomography of the calcium detected by electron beam technology in 35,246 adults. Am J Cardiol. 2001;
heart. Circulation. 2003;107:917–922. 87:1335–1339.
10. Internet: www.Siemens.com: Flash Speed. Lowest Dose. New CT scanner from Siemens. 40. Rumberger JA, Brundage BH, Rader DJ, et al. Electron beam computed tomographic coro-
11. Rumberger JA, Simons DB, Fitzpatrick LA, et al. Coronary artery calcium area by electron- nary calcium scanning: A review and guidelines for use in asymptomatic persons. Mayo
beam computed tomography and coronary atherosclerotic plaque area: A histopathologic Clin Proc. 1999;74:243–252.
correlative study. Circulation. 1995;92:2147–2162. 41. Pletcher MJ, Tice JA, Pignone M, et al. Using the coronary artery calcium score to predict
12. Mautner GC, Mautner SL, Froehlich J, et al. Coronary artery calcification: Assessment with coronary heart disease events: A systematic review and meta-analysis. Arch Intern Med.
electron beam CT and histomorphometric correlation. Radiology. 1994;192:619–623. 2004;164:1285–1292.
13. Bormann JL, Stanford W, Stenberg RG, et al. Ultrafast tomographic detection of coronary 42. Shaw LJ, Raggi P, Schisterman E, et al. Prognostic value of cardiac risk factors and coro-
artery calcification as an indicator of stenosis. Am J Card Imaging. 1992;6:191–196. nary artery calcium screening for all-cause mortality. Radiology. 2003;28:826–833.
14. Sangiorgi G, Rumberger JA, Severson A, et al. Arterial calcification and not lumen steno- 43. Budoff MJ, Gul KM. Expert review on coronary calcium. Vasc Health Risk Manag.
sis is highly correlated with atherosclerotic plaque burden in humans: A histologic study 2008;4:315–324.
of 723 coronary artery segments using nondecalcifying methodology. J Am Coll Cardiol. 44. Jain A, McClelland RL, Polak JF, et al. Cardiovascular imaging for assessing cardiovas-
1998;31:126–133. cular risk in asymptomatic men versus women: The multi-ethnic study of atherosclerosis
15. Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of coronary artery calcium (MESA). Circ Cardiovasc Imaging. 2011;1:8–15.
using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827–832. 45. Shareghi S, Ahmadi N, Young E et al. Prognostic significance of zero coronary calcium
16. McCollough CH, Ulzheimer S, Halliburton SS, et al. Coronary artery calcium: A multi- scores on cardiac computed tomography. J Cardiovasc Comput Tomog. 2007;3:155–159.
institutional, multimanufacturer international standard for quantification at cardiac CT. 46. Sarwar A, Shaw LJ, Shapiro MD. et al. Diagnostic and prognostic value of absence of
Radiology. 2007;234:527–538. coronary artery calcification. JACC. 2009;6:675–688.
17. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium measurement: Agreement 47. Vliegenthart R, Oudkerk M, Song B, et al. Coronary calcification detected by electron-
of multidetector row and electron beam CT. AJR Am J Roentgenol. 2001;176:1295–1298. beam computed tomography and myocardial infarction: The Rotterdam Coronary
18. Knez A, Becker CR, Leber A, et al. A determination of coronary calcium with multi- Calcification Study. Eur Heart J. 2002;23:1596–1603.
slice spiral computed tomography: A comparative study with electron-beam CT. Int J 48. Raggi P, Gongora MC, Gopal A, et al. Coronary artery calcium to predict all-cause mortal-
Cardiovasc Imag. 2002;18:295–303. ity in elderly men and women. J Am Coll Cardiol. 2008;52:17–23.
19. Callister TQ, Cooil B, Raya SP, et al. Coronary artery disease: Improved reproducibility of 49. Raggi P, Cooil B, Ratti C, et al. Progression of coronary artery calcium and occurrence
calcium scoring with electron-beam CT volumetric method. Radiology. 1988;208:807–814. of myocardial infarction in patients with and without diabetes mellitus. Hypertension.
20. Detrano RC, Anderson M, Nelson J, et al. Coronary calcium measurements; effects of 2005;46:238–243.
CT scanner type and calcium measure on reproducibility-MESA study. Radiology. 2005; 50. Stanford W. Coronary artery calcium assessment. In: Ho V, Reddy GP, eds. Cardiovascular
236:477–484. Imaging. St Louis, MO: Saunders; 2010:32,457–465.
LWBK1209-ch24_p365-375.indd 374 16/05/13 9:32 PM

51. Schmermund A, Achenbach S, Budde T, et al. Effect of intensive versus standard lipid- 59. Georgiou D, Budoff MJ, Kaufer E, et al. Screening patients with chest pain in the emer-
lowering treatment with atorvastatin on the progression of calcified coronary atheroscle- gency department using electron beam tomography: A follow-up study. Am Coll Cardiol.
rosis over 12 months: A multicenter, randomized, double-blind trial. Circulation. 2006; 2001;38:105–110.
113:427–437. 60. Rubinshtein R, Gaspar T, Halon DA, et al. Prevalence and extent of obstructive coronary
52. ACCR/AHA 2007 Clinical expert consensus document on coronary artery calcium scoring artery disease in patients with zero or low calcium score undergoing 64-slice cardiac mul-
by computed tomography in global cardiovascular risk assessment and in evaluation of tidetector computed tomography for evaluation of a chest pain syndrome. Am J Cardiol.
patients with chest pain. J Am Coll Cardiol. 2007;49:378–402. 2007;99:472–475.
53. Ong TK, Chin SP, Liew CK, et al. Accuracy of 64-row multidetector computed tomography 61. De Backer G. European guidelines on cardiovascular disease prevention in clinical practice.
in detecting coronary artery disease in 134 symptomatic patients: Influence of calcification. Eur Heart J. 2003;24:1601–1610.
Am. Heart J. 2006;151:1323.e1–1323.e6 62. Mieres JH, Shaw LJ, Arai A, et al. The role of non-invasive testing in the clinical evaluation
54. Stolzmann P, Scheffel H, Leschka S, et al. Influence of calcifications on diagnostic accuracy of women with suspected coronary artery disease: American Heart Association consensus
of coronary CT angiography using prospective ECG triggering. AJR Am J Roentgenol. statement. Circulation. 2005;111:682–696.
2008;191:1684–1689. 63. Budoff MJ, Achenbach S, Blumenthal RS, et al. Assessment of coronary artery disease by
55. Diederichsen AC, Petersen H, Jensen LO, et al. Diagnostic value of cardiac 64-slice com- cardiac computed tomography: A scientific statement from the American Heart Association
puted tomography: Importance of coronary calcium. Scand Cardiovas J. 2009;43:337–344. Committee on Cardiovascular Imaging and Intervention, Council on Cardiovascular Radiology
56. White C, Read K, Kuo D. Assessment of chest pain in the emergency room. What is the role and Intervention, and Committee on Cardiac Imaging, Council on Clinical Cardiology.
of multidetector CT? Eur J Radiol. 2006;57:368–372. Circulation. 2006;114:1761–1791.
57. Laudon DA, Vukov LF, Breen JF, et al. Use of electron-beam computed tomography in 64. ACCF/AHA Expert Consensus Document: ACCF/AHA 2007 Clinical Expert Consensus
the evaluation of chest pain patients in the emergency department. Ann Emerg Med. Document on Coronary Artery Calcium Scoring by Computed Tomography in Global
1999;33:15–21. Cardiovascular Risk Assessment and in Evaluation of Patients With Chest Pain. Circulation.
58. McLaughlin VV, Balogh T, Rich S. Utility of electron beam computed tomography to 2007;115:402–426.
stratify patients presenting to the emergency room with chest pain. Am J Cardiol. 1999; 65. Bild DE, Detrano R, Peterson D, et al. Ethnic differences in coronary calcification: The
84:327–328, A328. Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2005;111:1313–1320.
LWBK1209-ch24_p365-375.indd 375 16/05/13 9:32 PM

Chapter
25 Marc Dewey
Lucia J.M. Kroft
Coronary Computed
Tomography Angiography
■■ ClinicalApplication of Coronary CT by adding functional testing such as myocardial stress CT

Angiography perfusion (CTP) to the armamentarium of cardiac CT (5–7).
Besides accurately ruling out coronary artery disease (Fig.
■■ Technical Implementation 25.1), cardiac CT is also extremely valuable for the com-
■■ Performing and Interpreting Coronary CT prehensive evaluation of the course of anomalous coronary
Angiography arteries (Figs. 25.2 and 25.3) and abnormal terminations of
coronary arteries, such as in the case of fistulae (Fig. 25.3).
■■ Remaining Issues
■■ Future Perspective: Myocardial CT
Perfusion Technical Implementation
Setting up coronary CT angiography in clinical practice
Clinical Application of Coronary requires two basic conditions.
CT Angiography First, there is the need of appropriate training of tech-
nicians to conduct the scan and of physicians to read the
Noninvasive coronary angiography using computed tomog- images. This sounds trivial but the learning curve for insti-
raphy (CT) results in high sensitivity for the detection of tutions and individuals is rather long with a minimum of
significant coronary artery stenosis that is far above 90%. about 6 and 12 months, respectively (8,9). Hands-on courses
Meta-analyses of patients with suspected coronary artery as well as longer-lasting fellowships with individual supervi-
disease have shown a pooled per-patient sensitivity of 97% sion provide the best available opportunity to prepare for
(with 95% confidence intervals of 96% to 98%) (1) (Table the high demands of cardiac CT (10). Second, the technical
25.1). As a result also the negative likelihood rate of coro- features of the CT system used should allow achieving diag-
nary CT angiography, that is, the odds that a negative result nostic image quality in almost every patient. Reassuringly,
is actually false, is very low (0.03 with confidence intervals on a worldwide perspective such technical requirements are
of 0.02 to 0.04) (1). This makes coronary CT angiography achieved in most sites (11).
the most accurate noninvasive imaging test for ruling out One of the basic technical requirements for cardiac CT is
coronary artery stenosis, and it is mainly recommended for the possibility to trigger the acquisition based on the ECG
this purpose in patients with low-to-intermediate pretest of the patient (Fig. 25.4). Prospectively ECG-triggered ac-
likelihood of disease (2,3) (Fig. 25.1). The pretest likelihood quisitions which are currently feasible on 64-row coronary
for coronary artery disease is determined beforehand by the CT angiography in patients with slow and stable heart rates
clinical presentation, risk factors, and diagnostic test results enable significant reductions in radiation exposure when
(such as exercise ECG) (4). compared with traditional retrospectively gated acquisi-
However, the moderate specificity of coronary CT angiog- tions. The slice collimation for coronary CT angiography
raphy in patients with suspected coronary artery disease of needs to be ≤0.625 mm, and for improvement of rendering
89% (95% confidence intervals of 86% to 92%) (1) makes the slice increment should be set to smaller values, such as
CT an imperfect test to confirm significant coronary artery 0.5 mm or less. With current standards, CT systems should
stenosis. Nevertheless, the probability that a diseased patient be equipped with at least 64 detector rows and should
has a positive coronary CT angiography is about nine times have a gantry rotation time of below 0.4 seconds (12–15).
higher than a positive result in a patient without coronary Nevertheless, in certain patients 64-row CT does not allow
disease (i.e., positive likelihood ratio). Therefore, further reliable delineation of coronary anatomy and ruling out of
refinements are needed to increase the specificity, for example, coronary artery disease. The main causes for non-diagnostic
376
LWBK1209-ch25_p376-390.indd 376 16/05/13 9:31 PM

Chapter 25 ■ Coronary Computed Tomography Angiography 377
Table 25.1 Diagnostic Accuracy of Coronary CT Angiography on the Per-Patient Level
Negative
Positive Likelihood
Mean Sensitivity Mean Specificity Area Under the Likelihood Ratio Ratio
Imaging Modality (%) (95% CI) (%) (95% CI) Curve (95% CI) (95% CI) (95% CI)
CT (89 studies) 97.2 (96.2, 98) 87.4 (84.5, 89.8) 0.98 (0.96, 0.99) 7.7 (6.2, 9.5) 0.03 (0.02, 0.04)
CT—Suspected coronary 97.6 (96.1, 98.5) 89.2 (86, 91.8) 0.98 (0.96, 0.99) 9.1 (7, 11.8) 0.03 (0.02, 0.04)
artery disease (45 studies)
CT—Suspected acute 95.6 (87.2, 98.6) 76.6 (50.9, 91.2) 0.94 (0.92, 0.96) 4.1 (2, 8.4) 0.06 (0.02, 0.19)
coronary syndrome
(7 studies)
CT—Covariate Analysis
CT: >16-row scanners 98.1 (97, 99) 89.4 (86, 92) — — —
CT: 12- to 16-row scanners 95.6 (94, 97) 84.7 (80, 89) — — —
p-value >16 vs. 12–16 p < 0.05 p = 0.07 — — —
CT: Heart rate <60 bpm 99 (98.1, 99.5) 85.8 (79.4, 90.5) — — —
CT: Heart rate ≥60 bpm 96.2 (94.7, 97.3) 87.7 (84.1, 90.5) — — —
p-value <60 bpm vs. p < 0.001 p = 0.55 — — —
≥60 bpm
With permission from: Schuetz GM, Zacharopoulou NM, Schlattmann P, et al. Meta-analysis: noninvasive coronary angiography using com-
puted tomography versus magnetic resonance imaging. Ann Intern Med. 2010;152:167–177.
coronary CT angiography are CT-dependent technical limi- of the entire cardiac volume within a single heartbeat (Fig.
tations in spatial and temporal resolution. Partial volume 25.6). Arrhythmia rejection performed prospectively by the
effect and beam hardening cause blooming artifacts espe- CT scanner is helpful to improve image quality while reduc-
cially in the presence of severe calcifications that hamper ing effective dose as compared to retrospective gating tech-
evaluation of the coronary artery lumen. Motion artifacts niques. Improving the coverage of coronary CT angiography
can be caused by postural motion (e.g., if the patient cannot and single heartbeat scanning can be done using either 320-
hold the breath), by high heart rates that result in too short row volume or fast dual-source CT (Fig. 25.7).
coronary rest-phase for motionless imaging, and by arrhyth-
mia (causing reconstruction artifacts, Fig. 25.5) (16,17). The
problems of arrhythmia can be overcome by fast coverage Performing and Interpreting
Coronary CT Angiography
Preparation for coronary CT angiography includes adminis-
tering oral and/or intravenous beta-blocker in patients with
heart rates exceeding 60 bpm. About 91% of cardiac CT sites
regularly use beta-blockade (11). Alternative medication (e.g.,
if channel blockers, calcium-antagonists and midazolam) may
be used in patients with contraindications for beta-blockers.
Slowing down the heart rate is important for avoiding motion
artifacts and for achieving diagnostic image quality. In addi-
tion, with slow and stable heart rates, radiation exposure can
be reduced, as acquisition can be limited to a short prospec-
tively triggered exposure interval during diastole.
Nitroglycerin is used by about 80% of cardiac CT sites
worldwide (11) and has the following advantages. First,
dilatation of the coronary arteries by nitroglycerin improves
the relative spatial resolution of the CT images. Second,
nitroglycerin is commonly used as an intracoronary injec-
tion during conventional coronary angiography and thus
comparison with CT results is improved when nitroglycerin
is also used for this test.
Figure 25.1. Three-dimensional display of an example of coronary Performing cardiac CT is optimally done in a systemati-
CT angiography in which coronary artery disease could be reliably cally unified way within the institution. Figure 25.8 gives an
ruled out. overview of the data acquisition procedure as typically done in
LWBK1209-ch25_p376-390.indd 377 16/05/13 9:31 PM

A B
Figure 25.2. Malignant-type coronary artery anomaly with the right coronary artery arising from the left
sinus of Valsalva (arrow, A,B). The anomalous right coronary artery may be compressed between the aorta
and pulmonary trunk on its interarterial way. This is a case with a low interarterial course of the anomalous
right coronary artery. Such low courses have been shown to lead less commonly than high courses to major
adverse cardiac events. LM, left main coronary artery.
64-row cardiac CT. Bolus tracking is most often used for ini- formed during the actual scan and thus has less variability. By
tiating the scan after contrast agent injection (Fig. 25.8) (11). using single-beat cardiac CT, the contrast agent volume can be
Alternatively, test bolus injection can be used to identify the further reduced in comparison to multi-beat 64-row CT. Further
optimal scan point but requires a slightly higher contrast agent reductions in contrast agent volume can be achieved using fast
volume (i.e., that of the test bolus itself). Bolus tracking is per- bolus tracking approaches available with single-beat imaging.
A B C
Figure 25.3. Left anterior descending coronary artery to pulmonary artery fistula and aberrant left circum-
flex coronary artery in a 74-year-old female patient. A,B: A small arteriovenous coronary artery fistula (arrow)
from the proximal left anterior descending coronary artery to the pulmonary artery (PA). C: The course of
an aberrant left circumflex artery (LCX) arising from the right sinus of Valsalva with a retro-aortic course
continuing in its normal location in the left atrioventricular groove. Arteriovenous coronary artery fistulas are
present in approximately 1/1,000 coronary angiographies. CT can be very useful and superior to invasive cor-
onary angiography in determining the anatomic relationships between the involved structures. Arteriovenous
fistulas bypass the myocardial tissue, which may cause symptoms (present in about half of the patients),
depending on the size of the fistula and the shunt volume.
LWBK1209-ch25_p376-390.indd 378 16/05/13 9:31 PM

A B
Figure 25.4. Effect of untriggered (A) versus ECG-triggered imaging (B) in a 52-year-old male patient with chronic atrial fibrillation. CT imag-
ing of the heart was performed before radiofrequency ablation therapy of the pulmonary veins. The CT images are used as roadmap for guiding
the procedure. This scan was repeated because the procedure had to be performed a second time. Before prospective triggering techniques became
available, we used an untriggered (non-ECG–synchronized or gated) protocol for pulmonary vein imaging (A). With untriggered protocols, the
pulmonary veins are already sharply displayed (arrows, A) because they hardly move in time. This is different from the ascending aorta (Ao) and
coronary arteries that move substantially throughout the cardiac cycle and are only displayed without motion on ECG-synchronized imaging (B).
Now that low-dose prospective ECG-triggering techniques have become available, we use these approaches for sharp imaging of the pulmonary
veins and the coronary arteries as well (arrowhead, B), at approximately the same radiation dose.
B C D
Figure 25.5. Issues with irregular heart rate in a 67-year-old female patient with suspected coronary artery disease who underwent 64-row CT that
resulted in helical step-artifacts. The patient had an irregular heart rate during scanning (red circle, A), resulting in step-artifact through at the level of the
aortic root (B–D). Some data are missing during image reconstruction because of a too short RR interval, whereas other data are projected twice because
of one too large RR interval (i.e., double projected left main coronary artery; arrows, D). The result is a poor-quality scan with nondiagnostic results.
LWBK1209-ch25_p376-390.indd 379 16/05/13 9:31 PM

A B
C D
Figure 25.6. Successful arrhythmia rejection with 320-row CT (arrow, A) in a 49-year-old female patient
with suspected coronary artery disease and irregular heart rate. Issues in patients with irregular heart rates
shown in Figure 25.5 can be overcome using single-beat scanning of the entire heart with prospective trigger-
ing and arrhythmia rejection that allows for sharp imaging of the coronary arteries (B–D). With this technique,
the scanner starts with prospective triggering at a desired point in the predefined cardiac cycle. If the heartbeat
is too short, the scanner automatically shuts down the exposure (arrow, A) and tries the next heartbeat, until
the RR cycle is long enough for image acquisition during diastole (arrowheads, A). In this patient, the RR
interval varied between 521 and 1198 milliseconds during scanning (i.e., heart rate varying between 50 and
115 bpm, A). During scanning, the third heartbeat was accepted. As multiple attempts were made during
image acquisition, the effective dose of 8 mSv was slightly higher than in patients with sinus rhythm. Image
quality of the coronary arteries was excellent in this patient with completely irregular heart rate. RCA, right
coronary artery; LAD, left anterior descending coronary artery; LCX, circumflex coronary artery.
LWBK1209-ch25_p376-390.indd 380 16/05/13 9:31 PM

Figure 25.7. Coverage of the entire heart in a single heartbeat is

important in coronary CT angiography. This coronal maximum-
intensity projection shows homogeneous attenuation along the left
ventricle and ascending aorta because scanning was done in a single
heartbeat. Fast single heartbeat imaging improves quality of coronary
artery imaging. Achieving an anatomical coverage along the z-axis
of the patient of at least 12 cm during the diastolic rest phase of the
cardiac cycle is required for single heartbeat CT. Two approaches are
available for single heartbeat imaging: Wide detector arrays with up to
320 detector rows covering up to 16 cm prospectively by volume imag-
ing, or second-generation dual-source CT using prospective high-pitch
spiral scanning. To compare, prior retrospective scanning approaches
using, for instance, 3.2-cm wide detector arrays (64-row CT) resulted in
overscanning, and thus higher effective dose, and needed about 8 to 10
heartbeats for covering the entire heart.
The flow and contrast agent amount is adjusted to the indi- right coronary artery (RCA) can occur (Fig. 25.10) and cov-
vidual patient weight to unify image quality. ering the entire cardiac cycle in patients with implantable
Important information before scanning relates to whether cardioverter-defibrillators may result in greater flexibility to
or not prior surgery has been performed. Coronary artery avoid artifacts.
bypass grafting should result in adequate anatomic cover- Reconstruction of images for coronary CT angiography is
age of the bypass and the coronary arteries (Fig. 25.9). In regularly done on minimal detector-row thickness (e.g., 0.625
case of cardiac pacemakers or implantable cardioverter- or 0.5 mm) that depends on the scanner. With reconstruction,
defibrillators, pronounced artifacts especially involving the overlapping slices by 0.5- to 0.25-mm increments are used
A B C
Figure 25.8. Overview of how to

perform CT of the heart using 64-row
scanners. The scan range (A) is defined
using the scanogram or is based on the
dimensions of the heart on noncontrast
calcium scoring. The start of the scan
range (B) is used to place a region of
interest in the descending aorta (C) that D E F
is used for initiation of the scan after
contrast agent injection. A mid-section
through the heart (D) can be used on
some scanners to identify the subse-
quent small cardiac reconstruction
field-of-view (E,F). A bolus tracking
scan (H) is performed at the same level
as the start level for the coronary scan
(G) and allows monitoring contrast
arrival in the previously defined region
of interest (I). When a certain threshold
of Hounsfield units has been reached,
for example, 180, scanning is initiated. G H I
LWBK1209-ch25_p376-390.indd 381 16/05/13 9:32 PM

Figure 25.9. One of the common pitfalls during cardiac CT is Figure 25.10. In patients who received pacemakers or an implant-
incomplete coverage (arrowheads) of the structures of interest such as able cardioverter-defibrillator as this is in this case (arrow pointing at
in this patient after coronary bypass grafting (which was not known). the defibrillator lead, asterisk pointing at the auricle lead), especially
As a result, the left internal mammary artery (LIMA) bypass graft to right coronary artery (RCA) assessment can be hampered by artifacts.
the left anterior descending coronary artery (LAD) as wells a venous In such patients CT may be improved by scanning the entire cardiac
(V) bypass graft were not completely covered. Such incidences can cycle allowing greater flexibility in image reconstruction with better
be avoided by proper information provided during the referral and chances to avoid artifacts.
informed consent process.
for optimal presentation of the dataset. The importance of mations and cross-sectional images are optimal for quantifi-
using thin slices for assessing the coronary arteries is shown cation of stenoses. However, automated approaches to gen-
in Figure 25.11. Small field-of-views used for cardiac as- erating curved multiplanar reformations along the coronary
sessment improve spatial resolution (Fig. 25.12) and should arteries are sometimes prone to tracking errors (Fig. 25.16).
always be used. Eighteen-centimeter field-of-views are most Therefore, findings on secondary reformations need to be
commonly used for this purpose (11). In contrast to other confirmed on the original axial or orthogonal slices. Stenoses
CT applications where full 360-degree rotation information that are formed on the basis of purely noncalcified plaques
is utilized, cardiac CT uses only half rotation information (Fig. 25.17) are ideal targets for automated probing tools that
(180 degrees) in order to improve temporal resolution (Fig. allow automated quantification of lesions. Automated quan-
25.12). Further improvements of temporal resolution, which tification of stenoses based on other types of plaques, for
are especially helpful in patients with higher heart rates, can example, those containing calcifications, sometimes leads to
be achieved with multisegment reconstruction that allows erroneous results (Fig. 25.18). Thus, visual confirmation of
splitting data used for reconstruction over multiple heart- the stenosis severity is key to avoiding under- or overestima-
beats (Fig. 25.13). Identifying the cardiac rest period for tion of coronary artery percent diameter stenosis. Important
coronary imaging is particularly important. With systematic advantages of CT over conventional angiography in patients
approaches, data may be reconstructed every 5% or 10% with chronic total occlusions are as follows: (1) The ability
over the entire cardiac cycle or parts of the cardiac cycle. to visualize the composition of the underlying plaque (calci-
Automated motion-mapping based on the CT raw data can fied vs. noncalcified), (2) tortuosity of the proximal vessel,
be used to identify the cardiac phase with the least motion, (3) location of the vessel segment distal from the occlusion,
and may be superior to systematic approaches (Fig. 25.14) and (4) the length of the occlusion (Fig. 25.19) (19). All of
(18). these parameters are important predictors of the success of
A systematic approach to reading coronary CT angiog- percutaneous revascularization of occlusions and CT may
raphy within a cardiac CT center is optimally utilized (Fig. also assist interventionalists to choose the optimal approach
25.15). Axial, sagittal, and coronal images are the main to revascularization.
source of information for detecting coronary stenoses (Fig. When reading coronary CT angiography, recognizing
25.15). Double-oblique reformations as well as double- image artifacts, especially motion artifacts, is of major im-
oblique thin-slice maximum-intensity projections are well portance that often requires high level of experience. Image
suited to confirm findings, whereas curved multiplanar refor- artifacts are the cause of false interpretations regarding
LWBK1209-ch25_p376-390.indd 382 16/05/13 9:32 PM

A B
Figure 25.11. The recon-

structed slice thickness greatly
influences the image quality
of cardiac CT. Images were
reconstructed based on raw
data obtained with 0.5-mm col-
limation using 3-mm (A), 2-mm
(B), 1-mm (C), and 0.5-mm
(D) slice thickness. Image qual-
ity for depiction of the main
coronary arteries but especially
side branches increases with
reductions in reconstructed slice
C D
thickness.
presence of coronary artery stenosis (17). Good image qual- viding the basis of high-quality interpretation within accept-
ity is the basis for correct diagnosis; valuing image quality able time.
includes information regarding the level of confidence of
diagnosis. During interpretation, artifacts should therefore
be recognized. Our rule of thumb: never trust a coronary Remaining Issues
artery that is not sharply defined. Blurred contours mean
motion artifacts. This may cause false-positive diagnosis One important remaining issue is the limited ability of CT to
because contrast within the lumen is not displayed well, as identify coronary in-stent restenosis. This is important since
motion smears out lumen and vessel wall, which may erro- also patients after coronary stenting would prefer CT over
neously suggest coronary artery stenosis. Also, blurring may MRI and conventional coronary angiography for future
cause false-negative diagnosis because the observer assumes imaging (20). Artifacts due to motion are pronounced in the
that it should probably be nothing in the blurred area and/ presence of metallic stent material and image quality needs
or misses the stenosis because it is not visualized. Whether to be excellent to enable ruling out (Fig. 25.20) or to detect
image quality is limited and confidence of diagnosis is re- in-stent restenosis (Fig. 25.21). Further improvements in
duced should be reported. spatial and temporal resolution or adding myocardial CTP
On average it takes 47 minutes for performing, reading, imaging may solve this issue.
and interpreting a cardiac CT examination. Especially, the The prognostic relevance of coronary plaque assessment
reading and interpretation time of on average 29 minutes per by CT has been shown in several single-center studies and ret-
case is surprisingly long (11) and highlights that adequate rospective multicenter registries (21–24). However, it would
preparation of patients is the key to success for cardiac CT. require prospective randomized trials before evidence-based
This is because good patient preparation improves image recommendations can be made regarding which changes in
quality which in turn reduces reading time. Appropriate patients management would improve clinical outcomes in
training in hands-on cardiac CT courses is essential for pro- patients with certain high-risk plaque characteristics.
LWBK1209-ch25_p376-390.indd 383 16/05/13 9:32 PM

A B C
D E F
Figure 25.12. Cardiac CT data should be reconstructed on small field-of-views for improved spatial resolu-
tion and using half-scan reconstruction for improved temporal resolution. With a full-field-of-view reconstruction
using full rotation data (A–C), both the spatial and temporal resolution is lower than with small-field-of-views
using half-scan reconstruction (D–F). Please note the sharp delineation of the right coronary artery on the small
field-of-views using half-scan reconstruction (D–F). All images in this figure are reconstructed from the same raw
data with 640 images of 0.5-mm slice thickness with a 0.25-mm slice increment (320-row volume technique).
A B
Figure 25.13. Further improvements of temporal resolution are feasible using multisegment (B) in com-
parison to half-scan reconstruction (A) that is most commonly used for cardiac CT (Fig. 25.12). Longer vessel
segments of the distal right coronary artery branches are visualized on maximum-intensity projections using
multisegment reconstruction (arrowheads, B) in comparison to half-scan reconstruction (asterisks, A).
LWBK1209-ch25_p376-390.indd 384 16/05/13 9:32 PM

A B
Figure 25.14. Identifying the cardiac rest phase for best imaging of the coronary arteries can be
improved using automated motion mapping approaches that are based on the raw data (B) in comparison
to reconstructions in increments over the cardiac cycle (in this case 75%, A). Using the 75% reconstruc-
tion (arrow, A) and other reconstructions in 5% increments (such as 70% and 80%, data not shown), a
coronary stenosis or motion artifact in the left anterior descending coronary artery was suspected. The best
diastolic phase that was identified using motion mapping showed smooth vessel contours and no significant
stenosis (arrow, B).
There is good evidence in patients with acute chest pain is weak (32,36). This means that next to anatomic evalua-
that CT reduces the length of in-hospital stay and may im- tion of stenosis, functional testing is necessary for diagnosing
prove the cost-effectiveness of diagnostic algorithms (25,26). ischemic heart disease. Myocardial stress CTP shows promise
Nevertheless, the clinical value of CT in terms of improved as a functional test in addition to coronary CT angiography.
health outcomes and reduction of major adverse clinical Preliminary studies have shown that CTP may provide incre-
events is difficult to prove. Cardiac CT indications are cur- mental value for the detection of hemodynamically signifi-
rently based on the expected ability to improve diagnosis and cant stenosis (5–7,37). Myocardial stress CTP may be espe-
treatment by indirect evidence from the diagnostic accuracy cially valuable in patients with reduced image quality such
of the test. In 2010 the updated appropriate use criteria for as in cases with severe calcified coronary sclerosis, motion
cardiac CT have been published (12). In this multi-societal artifacts, or after stent placement. In patients with stents,
paper, a variety of indications for use of cardiac CT in clini- adding myocardial stress CTP to coronary CT angiography
cal practice have been valued by a panel regarding the level has shown to improve the diagnostic accuracy for the detec-
of appropriateness as “appropriate” (CT is generally accept- tion of significant obstructive coronary artery disease (38).
able and reasonable), “uncertain” (CT may be acceptable Myocardial stress CTP is performed by the use of vaso-
and may be a reasonable approach), or “inappropriate” (CT dilators, for example, adenosine or dipyridamole. Severe
is not generally acceptable and not a reasonable approach). coronary artery stenosis compromises hyperemic flow that is
In addition to clinical expert recommendations, the decision visualized as perfusion defect. By performing myocardial
about appropriate indications of CT will be greatly influ- CTP at rest (these images may be used as the anatomic
enced by coverage decisions at national and/or local levels. images), it can be determined whether perfusion defects
are reversible or fixed. Myocardial CTP is acquired dur-
ing the first pass of a contrast agent. Myocardial stress
Future Perspective: CTP has several potential advantages as compared to the
Myocardial CT Perfusion other techniques: (1) Ability to combine imaging anatomy
(stenosis) and function (perfusion); (2) Potential cost sav-
One major limitation for anatomic imaging with coronary ings because both aspects can be imaged with a single
CT angiography, which also counts for invasive coronary scanner; (3) Fast scanning in rather short protocol (5); (4)
angiography, is the difficulty to predict the functional rel- Potential to better detect small areas of ischemia or infarc-
evance of a stenosis (27–35). Also, the relationship between tion as compared to nuclear medicine techniques (5), with
severity of stenosis and occurrence of complaints or ischemia ability to analyzed transmural differences in perfusion (39);
LWBK1209-ch25_p376-390.indd 385 16/05/13 9:32 PM

A B C
D E F
G H I
J K L
Figure 25.15. Systematic approach to reading for the presence of coronary artery stenoses. Axial images
(A–C) are the primary approach for visualization that can be followed by thin maximum-intensity projections
(below 5 mm, D–F). In this case a short high-grade (about 80%) diameter stenosis of the second marginal
branch (arrow, B,E) was found. The stenosis was due to a noncalcified plaque and vessel segments before
and after the short stenosis were normal (arrowhead, A,C,D,F). The stenosis (arrow) was also seen on a
double-oblique thin maximum-intensity projection (G), a three-dimensional reconstruction (H), and curved
multiplanar reformations (I,J). The latter were used for quantification of stenosis diameter on cross-sectional
images (I). Stenoses seen on secondary reformations such as curved multiplanar reformations should always
be confirmed on the axial slices. For correlation the stenosis is shown on conventional coronary angiography
(arrow, K,L).
LWBK1209-ch25_p376-390.indd 386 16/05/13 9:32 PM

A B
C D
Figure 25.16. Importance of checking automated probing of coronary tracks. With current post-
processing techniques, the coronary arteries are usually tracked automatically. Postprocessing errors by
loosing the coronary track may occur (arrow, A), which can result in a “pseudo-stenosis” in the curved
multiplanar reformation (arrow, B). After manual correction of the coronary track error (arrowhead, C),
the curved multiplanar reformation displays a normal continuity of the coronary artery (arrowhead, D).
A B C
Figure 25.17. Noncalcified plaque in the distal right coronary artery that results in an intermediate 60%
diameter stenosis (arrow, A) and is shown on a curved multiplanar reformation. For correlation the results of
conventional coronary angiography are shown before (arrow, B) and after percutaneous coronary intervention
(arrow, C).
LWBK1209-ch25_p376-390.indd 387 16/05/13 9:32 PM

A B
Figure 25.18. Visual assessment remains important for grading coronary artery stenosis, as automated
stenosis assessment may be difficult and depends on image quality. A: The coronary artery stenosis visually
assessed as being high-grade stenosis on the basis of a calcified and noncalcified plaque. B: The automated
stenosis grading that resulted in an area stenosis of 52% and a diameter stenosis of 51% (intermediate-grade).
With invasive coronary angiography a high-grade stenosis of 80% was found (not shown).
(5) Potential quantification of myocardial blood flow solved such as beam hardening and motion correction, lim-
(40). At present, a variety of techniques have been used in ited scan range, and the considerable additional radiation
preliminary CTP studies such as imaging at a single time and contrast dose. Development of robust myocardial CTP
point, depiction of image stacks in dynamic series, only with quantification of myocardial blood flow and further
rest imaging, only stress imaging, or using both stress and patient studies are needed to assess its value in clinical
rest. There are still many technical issues that need to be practice.
A B
Figure 25.19. Example of an occlusion (arrows, A) of the right coronary artery shown on a curved multi-
planar reformation. The length and the noncalcified plaque that constitutes the basis of the occlusion are easier
assessed on CT (A) than on conventional coronary angiography (B). A long length of the occlusion and calcifi-
cations on CT increase the chances for a percutaneous revascularization approach being unsuccessful.
LWBK1209-ch25_p376-390.indd 388 16/05/13 9:32 PM

A B C
Figure 25.20. Patent stent in the distal right coronary artery. Standard CT angiography window-level
presets and a standard cardiac CT kernel show the right coronary artery nicely, but stent lumen assessment is
prohibited (arrow, A). A dedicated stent kernel reconstruction with manual adaptation of the window-level
settings toward better assessment of the stent lumen (up to 1,500/300) allows ruling out in-stent restenosis
(arrow and inset, B). Conventional coronary angiography confirmed the patency of the stent (arrow, C).
A B
C D
Figure 25.21. In-stent restenosis in a left main coronary artery stent (arrow, A) that can only
be appreciated on cross-sectional display (B,C) that show a reduction of the luminal area to 1 mm2.
Conventional coronary angiography confirmed the presence of the stenosis (arrow, D).
LWBK1209-ch25_p376-390.indd 389 16/05/13 9:32 PM

References 22. Bamberg F, Sommer WH, Hoffmann V, et al. Meta-analysis and systematic review of
the long-term predictive value of assessment of coronary atherosclerosis by contrast-
enhanced coronary computed tomography angiography. J Am Coll Cardiol. 2011;57:
1. Schuetz GM, Zacharopoulou NM, Schlattmann P, et al. Meta-analysis: Noninvasive coro- 2426–2436.
nary angiography using computed tomography versus magnetic resonance imaging. Ann 23. Hadamitzky M, Distler R, Meyer T, et al. Prognostic value of coronary computed tomo-
Intern Med. 2010;152:167–177. graphic angiography in comparison with calcium scoring and clinical risk scores. Circ
2. Dewey M. Clinical indications. In: Dewey M, ed. Cardiac CT. Heidelberg: Springer; Cardiovasc Imaging. 2011;4:16–23.
2011:43–52. 24. Min JK, Dunning A, Lin FY, et al. Age- and sex-related differences in all-cause mortal-
3. Schlattmann P, Schuetz GM, Dewey M. Influence of coronary artery disease prevalence ity risk based on coronary computed tomography angiography findings results from the
on predictive values of coronary CT angiography: A meta-regression analysis. Eur Radiol. International Multicenter CONFIRM (Coronary CT Angiography Evaluation for Clinical
2011;21:1904–1913. Outcomes: An International Multicenter Registry) of 23854 patients without known coro-
4. Genders TS, Steyerberg EW, Alkadhi H, et al. A clinical prediction rule for the diagnosis of cor- nary artery disease. J Am Coll Cardiol. 2011;58:849–860.
onary artery disease: Validation, updating, and extension. Eur Heart J. 2011;32:1316–1330. 25. BlueCross BlueShield Association. Coronary Computed Tomographic Angiography in the
5. Blankstein R, Shturman LD, Rogers IS, et al. Adenosine-induced stress myocardial perfu- Evaluation of Patients with Acute Chest Pain. http://www.bcbs.com/blueresources/tec/
sion imaging using dual-source cardiac computed tomography. J Am Coll Cardiol. 2009;54: vols/26/26_09.pdf. In 2011.
1072–1084. 26. Goldstein JA, Chinnaiyan KM, Abidov A, et al. The CT-STAT (Coronary Computed
6. Bamberg F, Becker A, Schwarz F, et al. Detection of hemodynamically significant coronary Tomographic Angiography for Systematic Triage of Acute Chest Pain Patients to Treatment)
artery stenosis: Incremental diagnostic value of dynamic CT-based myocardial perfusion trial. J Am Coll Cardiol. 2011;58:1414–1422.
imaging. Radiology. 2011;260:689–698. 27. Paulin S. Assessing the severity of coronary lesions with angiography. N Engl J Med.
7. Ko SM, Choi JW, Song MG, et al. Myocardial perfusion imaging using adenosine-induced 1987;316:1405–1407.
stress dual-energy computed tomography of the heart: Comparison with cardiac magnetic 28. Topol EJ, Nissen SE. Our preoccupation with coronary luminology. The dissociation between
resonance imaging and conventional coronary angiography. Eur Radiol. 2011;21:26–35. clinical and angiographic findings in ischemic heart disease. Circulation. 1995;92:2333–
8. Chin S, Ong T, Chan W, et al. 64-row multi-detector computed tomography coronary 2342.
image from a centre with early experience: First illustration of learning curve. J Geriatr 29. Christou MA, Siontis GC, Katritsis DG, et al. Meta-analysis of fractional flow reserve
Cardiol. 2006;3:29–34. versus quantitative coronary angiography and noninvasive imaging for evaluation of myo-
9. Pugliese F, Hunink MG, Gruszczynska K, et al. Learning curve for coronary CT angiogra- cardial ischemia. Am J Cardiol. 2007;99:450–456.
phy: What constitutes sufficient training? Radiology. 2009;251:359–368. 30. Meijboom WB, van Mieghem CA, van Pelt N, et al. Comprehensive assessment of coro-
10. Zimmermann E, Germershausen C, Greupner J, et al. Improvement of skills and knowl- nary artery stenoses: Computed tomography coronary angiography versus conventional
edge by a hands-on cardiac CT course: Before and after evaluation with a validated ques- coronary angiography and correlation with fractional flow reserve in patients with stable
tionnaire and self-assessment. Rofo. 2010;182:589–593; epub. angina. J Am Coll Cardiol. 2008;52:636–643.
11. Maurer MH, Zimmermann E, Schlattmann P, et al. Indications, imaging technique, and read- 31. van Werkhoven JM, Schuijf JD, Jukema JW, et al. Comparison of non-invasive multi-slice
ing of cardiac computed tomography: Survey of clinical practice. Eur Radiol. 2012;22:59–72. computed tomography coronary angiography versus invasive coronary angiography and
12. Taylor AJ, Cerqueira M, Hodgson JM, et al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/ fractional flow reserve for the evaluation of men with known coronary artery disease. Am
SCAI/SCMR 2010 Appropriate use criteria for cardiac computed tomography: a report of J Cardiol. 2009;104:653–656.
the American College of Cardiology Foundation Appropriate Use Criteria Task Force, the 32. Hoffmann U, Bamberg F, Chae CU, et al. Coronary computed tomography angiography for
Society of Cardiovascular Computed Tomography, the American College of Radiology, the early triage of patients with acute chest pain: the ROMICAT (Rule Out Myocardial Infarction
American Heart Association, the American Society of Echocardiography, the American using Computer Assisted Tomography) trial. J Am Coll Cardiol. 2009;53:1642–1650.
Society of Nuclear Cardiology, the North American Society for Cardiovascular Imaging, 33. Tonino PA, De Bruyne B, Pijls NH, et al. Fractional flow reserve versus angiography for
the Society for Cardiovascular Angiography and Interventions, and the Society for guiding percutaneous coronary intervention. N Engl J Med. 2009;360:213–224.
Cardiovascular Magnetic Resonance. Circulation. 2010;122:e525–e555. 34. Tonino PA, Fearon WF, De Bruyne B, et al. Angiographic versus functional severity of
13. Mark DB, Berman DS, Budoff MJ, et al. ACCF/ACR/AHA/NASCI/SAIP/SCAI/SCCT 2010 coronary artery stenoses in the FAME study fractional flow reserve versus angiography in
expert consensus document on coronary computed tomographic angiography: A report multivessel evaluation. J Am Coll Cardiol. 2010;55:2816–2821.
of the American College of Cardiology Foundation Task Force on Expert Consensus 35. Pijls NH, Fearon WF, Tonino PA, et al. Fractional flow reserve versus angiography for
Documents. Circulation. 2010;121:2509–2543. guiding percutaneous coronary intervention in patients with multivessel coronary artery
14. Dewey M, Hamm B. CT coronary angiography: Examination technique, clinical results, disease: 2-year follow-up of the FAME (Fractional Flow Reserve Versus Angiography for
and outlook on future developments. Fortschr Röntgenstr. 2007;179:246–260. Multivessel Evaluation) study. J Am Coll Cardiol. 2010;56:177–184.
15. Dewey M, de Vries H, de Vries L, et al. The present and future of cardiac CT in research 36. Schuijf JD, Wijns W, Jukema JW, et al. Relationship between noninvasive coronary angi-
and clinical practice: Moderated discussion and scientific debate with representatives from ography with multi-slice computed tomography and myocardial perfusion imaging. J Am
the four main vendors. Rofo. 2010;182:313–321. Coll Cardiol. 2006;48:2508–2514.
16. Kroft LJ, Dewey M. Reading and reporting. In: Dewey M, ed. Cardiac CT. Heidelberg: 37. Feuchtner G, Goetti R, Plass A, et al. Adenosine stress high-pitch 128-slice dual-source
Springer; 2010;143–170. myocardial computed tomography perfusion for imaging of reversible myocardial isch-
17. Kroft LJ, de Roos A, Geleijns J. Artifacts in ECG-synchronized MDCT coronary angiogra- emia: Comparison with magnetic resonance imaging. Circ Cardiovasc Imaging. 2011;4:
phy. AJR Am J Roentgenol. 2007;189:581–591. 540–549.
18. Joemai RM, Geleijns J, Veldkamp WJ, et al. Automated cardiac phase selection with 38. Magalhaes TA, Cury RC, Pereira AC, et al. Additional value of dipyridamole stress myo-
64-MDCT coronary angiography. AJR Am J Roentgenol. 2008;191:1690–1697. cardial perfusion by 64-row computed tomography in patients with coronary stents.
19. Hoe J. CT coronary angiography of chronic total occlusions of the coronary arteries: How J Cardiovasc Comput Tomogr. 2011;5:449–458.
to recognize and evaluate and usefulness for planning percutaneous coronary interventions. 39. George RT, Arbab-Zadeh A, Miller JM, et al. Adenosine stress 64- and 256-row detector
Int J Cardiovasc Imaging. 2009;25(suppl 1):43–54. computed tomography angiography and perfusion imaging: A pilot study evaluating the
20. Schönenberger E, Schnapauff D, Teige F, et al. Patient acceptance of noninvasive and inva- transmural extent of perfusion abnormalities to predict atherosclerosis causing myocardial
sive coronary angiography. PLoS ONE. 2007;2:e246. ischemia. Circ Cardiovasc Imaging. 2009;2:174–182.
21. Motoyama S, Sarai M, Harigaya H, et al. Computed tomographic angiography character- 40. So A, Wisenberg G, Islam A, et al. Non-invasive assessment of functionally relevant coro-
istics of atherosclerotic plaques subsequently resulting in acute coronary syndrome. J Am nary artery stenoses with quantitative CT perfusion: Preliminary clinical experiences. Eur
Coll Cardiol. 2009;54:49–57. Radiol. 2012;22:39–50.
LWBK1209-ch25_p376-390.indd 390 16/05/13 9:32 PM

Chapter
26
Hildo J. Lamb
Lucia J.M. Kroft
Albert de Roos
CT Evaluation of Myocardial
Perfusion, Function, and
Late Enhancement
■■ Introduction outside the scope of coronary angiography. Some aspects of

the ischemic cascade cannot be evaluated with CVCT, for
■■ Myocardial Perfusion example, metabolic abnormalities cannot be derived from
Advantages of Stress CT MPI CT data, or will be discussed elsewhere, such as CT vessel
Patient Preparation wall imaging.
Static versus Dynamic Stress CT MPI
CT MPI Patterns Myocardial Perfusion
Calculation of Blood Flow
Accuracy of Stress CT MPI Investigations for After development of a significant coronary artery stenosis,
Detecting Ischemia diminished myocardial tissue perfusion is the next step in
Stress CT MPI Limitations the ischemic cascade. Noninvasive anatomic evaluation of
■■ Diastolic Heart Function the coronary arteries by CT angiography has been accepted
as imaging tool in patients with low to intermediate pre-test
■■ Systolic Heart Function probability for coronary artery disease (1). However, if a cor-
■■ Scar Tissue onary artery stenosis is found, the functional relevance of the
stenosis is difficult to predict, as the relationship between the
■■ Summary severity of stenosis and occurrence of complaints and/or isch-
emia is unclear. This mainly indicates that patients with ana-
tomically significant coronary artery stenosis may not have
Introduction ischemia, for example, because of development of sufficient
collateral circulation. This is the case not only for coronary
Cardiovascular CT (CVCT) for evaluation of coronary artery stenoses found on coronary CT angiography (2,3) but
artery disease has become a routine clinical application. also for invasive coronary angiography (4–7). Therefore, a
Once a patient has a clinical indication for coronary CT functional test is needed to assess the hemodynamic effect of
angiography, additional CT acquisitions can be considered. the anatomical stenosis found, which is essential in determin-
In principle, ECG-gated CT examinations allow acquisi- ing appropriate patient management (2–7).
tion of dynamic information such as myocardial function. Ischemic myocardium can be detected by stress myocar-
However, dynamic CT examinations require a substantially dial perfusion imaging (MPI) performed by using vasodila-
higher radiation dose, mainly because more than one cardiac tors such as adenosine or dipyridamole. Myocardial stress
time frame has to be imaged to be able to evaluate myocar- imaging identifies severe stenosis that reduces maximum
dial motion. On the other hand, sometimes ECG-triggered blood flow but not resting flow. Rest perfusion can be done
CT acquisitions are performed for other reasons, for exam- to determine whether perfusion defects are fixed, or partly
ple, in patients with high heart rate, where the optimal or completely reversible.
reconstruction phase is difficult to predict. In those patients Several imaging techniques are clinically available for
functional information is available as part of their standard stress MPI. Today, the nuclear medicine technique single-
CT examination and therefore needs to be evaluated. photon emission computed tomography (SPECT) is most
Figure 26.1 shows a schematic diagram of the ischemic widely applied to this purpose. Stress MPI using SPECT
cascade. This scheme is used to discuss CVCT applications has shown incremental value to coronary CT angiography
391
LWBK1209-ch26_p391-401.indd 391 5/18/13 1:00 PM

Figure 26.1. Schematic diagram of the ischemic cascade in relation to CT imaging goals. Cardiovascular
CT can be used to evaluate coronary anatomy, coronary stenosis, may be in the future myocardial perfusion,
in research setting diastolic heart function, in special clinical setting systolic function, and in research setting
may be delayed enhancement. See text for further explanation.
for the detection of coronary artery stenoses that are hemo- positron emission tomography cannot depict transmural dif-
dynamically significant (8). Also, magnetic resonance imaging ferences in myocardial perfusion, while ischemic myocardium
(MRI) has been shown a robust tool for stress MPI and is used is first manifested in the subendocardium. With MRI that has
in clinical practice. Currently, CT protocols are developed for improved spatial resolution, transmural differences can be
stress MPI as well. detected. Preliminary studies have shown that CT can also
depict transmural perfusion differences (13). CT may there-
fore improve the detection of small areas of ischemia or infarc-
Advantages of Stress CT MPI
tion (10); (4) CT potentially allows quantification of myocar-
Stress CT MPI has many potential advantages as compared dial blood flow (11); (5) Stress CT MPI can be particularly
with SPECT or MRI techniques. Major advantages are: (1) useful in cases of limited coronary CT angiography quality
Possibility for assessing morphology (stenosis) together with (11,14,15), such as in patients with severe calcifications, with
its hemodynamic functional effect (perfusion) in a single CT stents, in image quality-limiting motion artifacts, or in patients
examination (9–11); (2) CT has short scan times; the full CT with intermediate stenoses that are difficult to grade regard-
MPI protocol including stress and rest imaging can be per- ing hemodynamic consequences. Also, stress CT MPI may be
formed in 20 minutes (by using intravenous aminophylline combined with late enhancement imaging for assessing viabil-
and beta-blockers for rest scanning) (12). Stress and rest imag- ity (9,14); (6) Stress CT MPI may save costs in combination
ing including late enhancement have even been performed with coronary CT angiography because it can be performed
within 20 minutes (10); (3) CT has superior spatial resolu- during a single procedure as compared to multiple procedures
tion (10,11). Nuclear medicine techniques such as SPECT and with different imaging techniques.
LWBK1209-ch26_p391-401.indd 392 16/05/13 9:30 PM

Chapter 26 ■ CT Evaluation of Myocardial Perfusion, Function, and Late Enhancement 393
Figure 26.2. CT myocardial perfusion during adenosine stress (top row) in short-axis view at basal (red),
midventricular (blue), and apical (green) level (lower left image shows reconstructed slice levels in red, blue
and green respectively). Note subendocardial perfusion defect in the anteroseptal and anterior segments at
midventricular and apical level. CT (lower mid) and conventional (lower right) angiography show significant
LAD stenosis. (Courtesy of Bettencourt N, Rocha J, Ferreira N, et al. Incremental value of an integrated ade-
nosine stress-rest MDCT perfusion protocol for detection of obstructive coronary artery disease. J Cardiovasc
Comput Tomogr. 2011;5:392–405.)
Patient Preparation CT MPI is analogous to “static” SPECT MPI and dynamic

stress CT MPI is analogous to dynamic MRI MPI. Essential
Similar to SPECT or MRI MPI, a full stress CT MPI inves-
is that both static and dynamic CT MPI are performed dur-
tigation includes imaging at stress and rest. This requires
ing the first pass of contrast agent bolus passage through
two antecubital intravenous 18G to 20G cannulas, one for
the myocardial tissue. Differences in enhancement between
administering the stress agent and one for contrast agent.
normal and ischemic myocardium are maximum during the
Usually, vasodilator adenosine is administered, with injec-
upslope of myocardial bolus passage. During the downslope,
tion starting 3 to 4 minutes before scan start, until scan finish,
the differences become less and disappear (Fig. 26.3). This
with a dose of 0.14 mg/kg/min (140 mg/kg/min). Patients are
means that after bolus passage, normal and ischemic myo-
instructed to withdraw from coffee or tea for 24 hours before
cardium cannot be distinguished by enhancement differ-
examination. During the procedure, the patient is closely
ences anymore (16,17). Exact timing of the scan is therefore
monitored for heart rate, blood pressure, and ECG for isch-
of paramount importance.
emia. Beta-blockers and/or other medication may be used
With static CT MPI or “snapshot imaging,” only a single
depending on the protocol used.
image stack is acquired such as with coronary CT angiog-
raphy. Static imaging requires exact scan timing at a single
Static versus Dynamic Stress CT MPI
time point during the upslope or maximum of the contrast
agent bolus passage. Myocardial blood flow can then be
There are two basic techniques that can be applied for stress assessed qualitatively and semi-quantitatively. Static CT
CT MPI: Static imaging (Fig. 26.2) and dynamic imaging MPI can be performed by using helical or step-and-shoot
(Fig. 26.3). Roughly, regarding time resolution, static stress techniques.
LWBK1209-ch26_p391-401.indd 393 16/05/13 9:31 PM

700
600
LV
Attenuation density (HU)
500
400
300
Normal remote
200
100
Ischemic LAD
0
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70
Time (s)
Figure 26.3. Example of static versus dynamic CT myocardial perfusion imaging. The image
represents the tissue attenuation curves for the left ventricle (LV) and myocardial territories.
Differences in normal remote and ischemic myocardium (left anterior descending territory) are
maximum during upslope of myocardial bolus passage. For static imaging, a single image stack
is acquired at a single time point (e.g., at the vertical line) during the upslope or maximum. For
dynamic imaging, multiple image stacks are acquired at multiple time points during the upslope as
indicated with the time interval with two-pointed arrow at the x-axis. (Courtesy of Ho KT, Chua
KC, Klotz E, et al. Stress and rest dynamic myocardial perfusion imaging by evaluation of complete
time-attenuation curves with dual-source CT. JACC Cardiovasc Imaging. 2010;3:811–820; Mehra
VC, Ambrose M, Valdiviezo-Schlomp C, et al. CT-based myocardial perfusion imaging-practical
considerations: Acquisition, image analysis, interpretation, and challenges. J Cardiovasc Transl
Res. 2011;4:437–448.)
With dynamic CT MPI or “serial imaging,” multiple Calculation of Blood Flow

image stacks are acquired at multiple time points during the
With static CT MPI, blood flow or volume calculation is
upslope of contrast passage through the myocardial tissue as
limited to qualitative or semi-quantitative measures. Animal
to obtain a perfusion curve in time (Fig. 26.3). Next to qual-
studies have shown strong nonlinear correlation between
itatively and semi-quantitatively flow measures, dynamic
static CT MPI and microsphere myocardial blood flow (22).
imaging potentially allows absolute blood flow calculation
Also, by using static stress CT MPI in 40 patients, the trans-
(18). Dynamic CT MPI requires advanced CT scanners. At
mural perfusion ratio (TPR) has been measured by dividing
the moment, the technique most often used is shuttle-mode
the subendocardial attenuation density by the subepicardial
imaging with alternating table positions for covering the
attenuation density. It was found that TPR was statistical
left ventricle by moving to-and-fro between consecutive
significant inversely and linearly related to percent diameter
heart beats. This technique results in limited coverage of
stenosis, with a correlation of R = −0.63. In addition, abnor-
the left ventricle of 7.3 cm, as well as in contrast passage
mal TPR had sensitivity and specificity of 88% and 91% for
time delay within image stacks (19,16). Wide-volume scan-
detecting ischemia as compared to invasive coronary angi-
ners have the unique potential to overcome these problems
ography and SPECT MPI (13). It was found that because
by dynamic scanning with full coverage of the left ventricle
of the excellent spatial resolution, stress CT MPI allows the
without contrast time delay for each time point during first
detection of subendocardial ischemia.
pass perfusion (13).
With dynamic imaging, tissue attenuation curves derived
from serial imaging can be used for semi-quantitative or
absolute quantification of myocardial blood flow. Valdiviezo
CT MPI Patterns
et al. (18), have described three upslope evaluation techniques
The CT MPI patterns are basically the same as with other for dynamic imaging. The semi-quantitative method for cal-
perfusion techniques. In normal myocardium, no perfusion culating myocardial blood flow uses linear fitting applied
defects will be observed. A perfusion defect due to decreased to time-attenuation data by upslope analysis that requires
attenuation observed during stress but not at rest indicates only a portion of the contrast bolus. The myocardial tissue
reversible defect (or stress-induced ischemia). A perfusion attenuation curve is normalized by the left ventricular blood
defect observed at stress and rest indicates fixed defect (or pool attenuation curve (or, alternatively, by the left ventricle
myocardial infarct). In addition, late enhancement may be maximum). This method has shown high correlation with
used for assessment of viability (16,20,21). microspheres and seems most easy and practical to use.
LWBK1209-ch26_p391-401.indd 394 16/05/13 9:31 PM

Absolute measures of myocardial blood flow using model- well correlated to diminished fractional flow reserve of
based Patlak plot analysis and model-based convolution are ≤0.75, with sensitivity of 93% and specificity of 87%, where
more complex and require higher level of expertise. Patlak specificity is much improved as compared to coronary CT
plot analysis is also used in nuclear medicine MPI and uses angiography alone (19).
an upslope two-compartment model. Modeling with param- In the preliminary stress CT MPI studies performed until
eters arterial input function (blood pool) in time, myocardial now, and as has been shown in a study that performed both
attenuation in time, and the transfer constant of contrast static and dynamic stress CT MPI techniques, both techniques
to the extracellular space, is used for calculating myocar- show good detection of ischemic myocardium (21).
dial blood flow. For model-based deconvolution, that has
initially been developed for MRI, an attenuation-to-iodine Stress CT MPI Limitations
concentration curve is produced and used to create time–
iodine concentration curves for all tissue attenuation curves Stress CT MPI has just been started by feasibility studies and
as to measure absolute myocardial blood flow. The tech- there are still many limitations to overcome. At the moment
nique is difficult and requires high level of expertise (18). the studies performed in humans use a variety of techniques.
Still, absolute quantification methods may be limited by as- It is not clear yet what would be the most robust proto-
sumptions applied with modeling with regard to iodinated col. One discussion is what to do first: Stress first or rest
contrast and CT. first? Indication to do stress MPI first may be in high-risk
patients for detecting ischemia. Stress first has better sensi-
tivity to detect ischemia, however, with decreased sensitivity
Accuracy of Stress CT MPI Investigations
for detecting infarct. Rest imaging first may be advised in
for Detecting Ischemia
low-risk patients as stress perfusion imaging may be avoided
First feasibility studies already have recognized the potential if coronary artery disease can be ruled out after the rest
strength of stress CT MPI. Stress-induced myocardial per- scan. However, rest imaging only has decreased sensitivity
fusion defects could be identified with diagnostic accuracy for detecting ischemia (20).
and radiation doses comparable to SPECT, but now with Up until now, studies performed are single centre studies.
the great advantage of visualizing coronary artery stenoses Stress CT MPI studies that have been performed have high
as well (10). selection bias, for example, many patients have been selected
In a static stress CT MPI study that compared combined based on positive SPECT investigations (10,11,15,19,23,24).
coronary CT angiography and stress CT MPI, it was found There are several technical limitations that need to be over-
that the one-step combined CT protocol accurately predicted come such as beam-hardening artifacts simulating perfu-
coronary stenosis causing perfusion defects when compared to sion defects (10,11,15,18,24). Other problems are motion
combined invasive coronary angiography and SPECT with artifacts due to low temporal resolution (10,18,24). Current
a per-vessel sensitivity of 79% and specificity of 91% for limitations include limited scan range of 7.3 cm for dynamic
detecting perfusion defects (13). As compared to MRI, per- CT perfusion with 64-slice CTs (9,18,19). The latter may
vessel sensitivity of 95% and specificity of 96% has been be overcome by wide-volume scanners covering the full left
found for detecting reversible ischemia (23). Another static ventricle. Furthermore, radiation dose in stress CT MPI is of
stress CT MPI study found additional value of CT MPI for concern (18,19). The technique of stress CT MPI is associ-
identifying reversible nontransmural ischemia over coronary ated with an increase of radiation dose as compared to im-
CT angiography alone, and was compared with invasive an- aging of the coronary arteries only. Current patient studies
giography. Combined coronary CT angiography and stress that have included at least 30 patients have shown radiation
CT MPI protocol had per-vessel sensitivity comparable to doses for stress CT MPI protocols including rest imaging
quantitative invasive angiography (68% vs. 71%). However, that varied between approximately 3 and 22 mSv for static
specificity was much improved with combined CT protocol perfusion protocols (10,12,13,15,23–26), and 13 to 20 mSv
from 60% for coronary CT angiography only to 98% after for dynamic perfusion protocols (16,18).
including CT MPI information. To compare, the accuracy In these protocols, the full LV was covered in static stress
for detecting reversible ischemia was 76% for coronary CT CT MPI protocols, while with dynamic stress CT MPI pro-
angiography, 84% after including CT MPI, versus 66% for tocols the scan range was limited to only 7.3 cm of the left
quantitative invasive coronary angiography. Moreover, in ventricle. It should be noted that in these studies, effective
territories where CT coronary artery findings and perfu- radiation doses were calculated by multiplying the dose-
sion findings were concordant, the combined CT protocol length product with conventionally used conversion factors
was highly accurate in detection and exclusion of ischemia: K = 0.014 or K = 0.017 mSv/mGy/cm used for CT chest
Significant (≥50%) stenosis found on coronary CT angiog- imaging. However, recent investigations have shown that for
raphy in combination with CT perfusion defect was 98% accurate calculation of radiation dose for cardiac CT imag-
specific for detecting ischemia. Also, nonsignificant stenosis ing with multidetector CT scanners, a conversion factor of
(<50%) in combination with normal perfusion was 100% K = 0.03 should be applied. This means that the real effective
specific for exclusion of ischemia (24). radiation doses are actually double the doses than have been
Of dynamic stress CT MPI studies also, high degree of presented in the cardiac CT papers published (25,27).
correlation has been found with nuclear medicine studies, New scan techniques such as high-pitch scanning (the
with reported sensitivity of 83% and specificity of 78% for lowest presented effective dose of less than 3 mSv for a full
detecting diseased myocardium (16). Combined coronary stress CT MPI protocol has been achieved with high-pitch
CT angiography with stress CT MPI has also been shown 128-slice dual source CT technique) (23) and new CT data
LWBK1209-ch26_p391-401.indd 395 16/05/13 9:31 PM

reconstruction techniques such as iterative reconstruction and cardiac apex, and the length between these points was
(28) may help in lowering radiation dose for CT perfusion calculated for each phase of the 20 heart cycle phases and
protocols. Further research with development of robust plotted in a length-versus-time curve (Fig. 26.4). Changes in left
stress CT MPI protocols and multicenter studies with large ventricular length between two consecutive heart phases can
patient populations are needed to assess the value of stress be calculated to derive mitral septal tissue velocities (cm/sec).
CT MPI for use in clinical practice. The early peak mitral septal tissue velocity (cm/sec) repre-
sents the maximal tissue velocity (Ea) during early diastole.
CT results were compared to ultrasound measures of diastolic
Diastolic Heart Function heart function and showed good agreement and correlation.
Therefore, cardiac CT imaging is capable of determining
In the presence of myocardial ischemia based on signifi- transmitral velocity, mitral septal tissue velocity, and thereby
cant coronary artery stenosis cardiomyocyte metabolism is estimation of LV filling pressures (32). In addition, cardiac CT
down-regulated. Using CT it is yet impossible to evaluate and two-dimensional echocardiography were comparable for
energy status of myocardium. However, one of the first signs assessment of diastolic dysfunction (32). Accordingly, cardiac
of cardiometabolic disbalance is a change in diastolic heart CT may provide information on diastolic dysfunction when
function. Even in the presence of sustained normal global ECG-gated CT acquisitions are performed.
systolic left ventricular heart function, subtle changes in
diastolic heart function can be detected. In clinical setting,
cardiac ultrasound is the most used diagnostic approach. Systolic Heart Function
Classic echocardiographic evaluation of diastolic heart func-
tion comprises assessment of early and late filling proper- With advancing ischemic time or severity of coronary
ties, for example, using parameters such as E/A ratio and artery stenosis, systolic heart function becomes hampered.
deceleration time. More advanced ultrasound measures are Clinically, one of the most useful systolic function param-
based on tissue Doppler techniques for determination of eters is ejection fraction, an easily obtainable measurement
longitudinal mitral valve annular motion (Ea) in relation to from dynamic datasets. Prospectively triggered coronary CT
early filling properties allowing estimation of left ventricular angiography with short acquisition interval does not allow
filling pressures (E/Ea). In complicated cases MRI of dia- evaluation of systolic heart function, since there is no or
stolic heart function can be performed yielding similar func- only limited dynamic information acquired. Based on retro-
tional parameters as compared to ultrasound (29,30). Very spective ECG-gating, dynamic CT scans can be performed
limited reports are available on CT assessment of diastolic allowing reconstruction of 10 to 20 heart phases evenly dis-
heart function, mainly because regular CT examinations tributed over the cardiac cycle, resulting in a time resolution
are acquired without dynamic information. One interesting per cardiac frame of around 50 to 100 milliseconds. Desired
approach has tried to study diastolic heart function with CT time resolution for evaluation of systolic heart function is
using an indirect measurement based on time-enhancement below 40 milliseconds per heart frame (33); consequently,
curves measured in the ascending aorta during inspiratory CT evaluation of systolic heart function is hampered by
breath-holds (31). The upslope of this curve was presumed under-sampling of time. However, an estimation of, for
to correlate with diastolic heart function and was compared example, ejection fraction is possible, with restrictions of
to ultrasound-derived Ea. There was a significant correla- temporal resolution in mind. In the early days of cardiac
tion between CT-derived upslope of contrast arrival in the CT, functional evaluation of the heart was performed using
ascending aorta and ultrasound-derived Ea, whereas there 16-slice MDCT scanners, for example, the study by Raman
was no relation to systolic measures of left ventricular heart et al. (34), one of the first reports to show the potential
function (31). Although this is an interesting approach, a role of CT for evaluation of heart function. At that time,
more direct measurement of diastolic heart function is a substantial amount of patient examinations was inad-
desired. equate for wall-motion assessment due to motion artifacts
Retrospectively, ECG-gated CT scans are sometimes per- or poor contrast between myocardium and blood pool.
formed in patients with high heart rate, where the optimal Advancement of CT technology allowed 64-slice CT to
reconstruction phase is difficult to predict. Once dynamic evaluate heart function more accurately concerning ejection
information is available, more sophisticated direct evalua- fraction and other global measures of cardiac dimensions
tion of diastolic heart function with CT is possible. Based and function, whereas regional wall-motion analysis was
on a retrospectively ECG-gated CT acquisition, outlining of still very difficult (35). Dual-step prospective ECG-triggered
endocardial contours in reconstructed short-axis slices of 20 128-slice dual-source CT allowed evaluation of global heart
phases of the cardiac cycle, a volume-versus-time curve can function, combined with coronary CT angiography, with a
be calculated (32) (Fig. 26.4). Changes in LV volume between reasonable total radiation dose of about 4 mSv (36). The
two consecutive heart phases can be plotted against time (first Rule Out Myocardial Infarction using Computer-Assisted
derivative), thereby constructing a transmitral flow-versus- Tomography (ROMICAT) trial using 64-slice CT (Fig. 26.5)
time curve (Fig. 26.4). From this curve, early and late peak reported incremental diagnostic value of regional left ven-
transmitral flow (mL/sec) can be derived and E/A ratio can be tricular heart function determined at rest over coronary CT
calculated (32). Mitral septal tissue velocity can be obtained angiography for detection of acute coronary syndrome (37).
by measuring the changing distance between two anatomic Adding regional left ventricular function to information
markers of the left ventricle in long-axis reconstruction. derived from coronary CT angiography resulted in a 10%
Measurements were positioned at the mitral septal annulus increase in sensitivity to detect acute coronary syndrome
LWBK1209-ch26_p391-401.indd 396 16/05/13 9:31 PM

A Transmitral flow
140
120
LV volume (mL)
100
80
60
40
20 ESV EDV
0
0 200 400 600 800 1000 1200
Time (ms)
300
E
Transmitral flow (mL/s)

200 A
100
0
200 400 600 800 1000 1200
–100
Time (ms)
–200
–300
–400
B Mitral septal tissue velocity

105
100
LV length (mm)
95
90
85
0 200 400 600 800 1000 1200
Time (ms)
Mitral septal tissue velocity (cm/s)
6
Ea
4
0
200 400 600 800 1000 1200
–2
Time (ms)
–4
–6
Figure 26.4. CT assessment of diastolic heart function. A: Based on a retrospectively ECG-gated CT acqui-
sition, endocardial contours in reconstructed short-axis slices of 20 phases of the cardiac cycle are outlined
(A), from these measurements a volume-versus-time curve can be calculated (A, right top). Changes in LV vol-
ume between two consecutive heart phases can be plotted against time (first derivative), thereby constructing
a transmitral flow-versus-time curve (A, right bottom). From this curve, early and late peak transmitral flow
(mL/sec) can be derived and E/A ratio can be calculated. B: Mitral septal tissue velocity can be obtained by
measuring the changing distance between two anatomic markers of the left ventricle in long-axis reconstruc-
tion. Measurements were positioned at the mitral septal annulus and cardiac apex and the length between
these points was calculated for each phase of the 20 heart cycle phases (B, left) and plotted in a length-versus-
time curve (B, right top). Changes in left ventricular length between two consecutive heart phases can be cal-
culated to derive mitral septal tissue velocities (cm/sec) (B, bottom right). The early peak mitral septal tissue
velocity (cm/sec) represents the maximal tissue velocity (Ea) during early diastole. (Courtesy of Boogers MJ,
van Werkhoven JM, Schuijf JD, et al. Feasibility of diastolic function assessment with cardiac CT: Feasibility
study in comparison with tissue Doppler imaging. JACC Cardiovasc Imaging. 2011;4:246–256.)
LWBK1209-ch26_p391-401.indd 397 16/05/13 9:31 PM

A C E
B D F
Figure 26.5. Curved multiplanar reformation CT images of left anterior descending artery (A) and invasive
coronary angiography (B) showing nonobstructive (30%) coronary atherosclerosis. Midventricular short-axis
and two-chamber views obtained during end-systole (C,D) and end-diastole (E,F) demonstrating anterior
wall hypokinesis. (Courtesy of Seneviratne SK, Truong QA, Bamberg F, et al. Incremental diagnostic value of
regional left ventricular function over coronary assessment by cardiac computed tomography for the detec-
tion of acute coronary syndrome in patients with acute chest pain: From the ROMICAT trial. Circ Cardiovasc
Imaging 2010;3:375–383.)
by cardiac CT and significantly improved the overall accu- Scar Tissue

racy for extent of plaque and presence of stenosis (37). The
diagnostic accuracy of regional left ventricular function for The final stage of the ischemic cascade is myocardial infarc-
detection of acute coronary syndrome (ACS) has 89% sen- tion. Main reason for imaging in this irreversible phase of
sitivity and 86% specificity in patients with significant ste- ischemic heart disease is to assess the extent of myocardial
nosis and 60% sensitivity and 86% specificity in patients infarction and evaluate viability of the remaining tissue. In
with inconclusive coronary CT angiography (37). Therefore, the past decade, MRI has developed as gold-standard for
it was concluded that regional left ventricular function evaluation of myocardial scar tissue, based on late imaging
assessment at rest improves diagnostic accuracy for ACS in after administration of gadolinium contrast. The approach
patients with acute chest pain, especially in those with coro- is based on the larger distribution volume of contrast agent
nary artery disease and thus may be helpful to guide further due to myocyte damage in acute infarction and due to the
management in patients at intermediate risk for ACS (37). presence of fibrous tissue in chronic ischemic heart disease.
To keep total radiation dose below 5 mSv, assessment of left In normal myocardial tissue, contrast agent passes relatively
ventricular regional wall motion and ejection fraction can faster than in damaged or fibrous tissue. This difference in
be acquired using more advanced CT scanners, for example, contrast behavior can be exploited to differentiate between
low radiation dual-source CT (38), dual-step dual-source normal, viable, and scar tissue. To some extent, iodine con-
CT (39), or 320-row cardiac CT (40). trast has similar properties as compared to MRI contrast
LWBK1209-ch26_p391-401.indd 398 16/05/13 9:31 PM

50% 240
MRI - Signal intensity (AU)

y = 1.06x + 0.01
40% R2 = 0.93 200
Infarct size MRI

160
30%
120
20%
80
10% 40 ∗†
0% 0
0% 10% 20% 30% 40% 50% Infarcted LV blood Remote
Infarct size TTC pathology myocardium pool myocardium
50%
160
y = 1.03x
CT - Attenuation (HU)
Infarct size MSCT
40% R2 = 0.96
120
30%
80 ∗
20%
10% 40
0% 0
0% 10% 20% 30% 40% 50% Infarcted LV blood Remote
Infarct size TTC pathology myocardium pool myocardium
Figure 26.6. CT-delayed enhancement is shown in short-axis view in an animal model for myocardial
infarction as compared to MRI. Correlation between infarct size as determined postmortem by TTC pathol-
ogy is excellent for both CT and MRI (mid panel). Difference in contrast between remote myocardium and
infarcted myocardium is less for CT as compared to MRI (right panel). *p < 0.001 compared with infarcted
myocardium. †p < 0.001 compared with left ventricular (LV) blood pool. (Courtesy of Baks T, Cademartiri F,
Moelker AD, et al. Multislice computed tomography and magnetic resonance imaging for the assessment of
reperfused acute myocardial infarction. J Am Coll Cardiol 2006;48:144–152.)
agents. Main difference with MRI is that contrast resolution CT in patients with acute myocardial infarction was a reli-
between iodine and myocardium is much less for CT than able method for evaluating infarct size and transmurality of
contrast resolution between gadolinium-based agents and ischemic damage. Combination of early perfusion defects
myocardium for MRI. At the present time, MRI remains and late enhancement at CT following acute myocardial
first choice for evaluation of scar burden and viability, but infarction could predict segment myocardial dysfunction
CT is capable of late enhancement imaging too. One of the and myocardial functional recovery, with echocardiography
first reports was by Baks et al. (41), demonstrating that late as reference standard (44). Early and late CT-determined
enhancement CT imaging in a pig model can assess acute hypoenhancement showed good contrast-to-noise and cor-
reperfused myocardial infarction in good agreement with related well between CT and cardiac MRI (45). Dual-energy
in vivo delayed enhancement MRI and postmortem TTC late enhancement CT is also feasible for detection of reper-
pathology as standard of reference (Fig. 26.6). The relative fused chronic myocardial infarction (46). Although first
difference in CT attenuation value of infarcted myocar- results are reasonable, delayed enhancement CT requires a
dium compared to remote myocardium was about 191%, second acquisition 10 to 15 minutes after a coronary CT
whereas the relative MR signal intensity between infarcted angiogram, thereby reducing scan time efficiency and increas-
myocardium and remote myocardium was about 554% ing total radiation dose. A possible solution to reduce the
(41). Although promising, as noted, MRI results in much overall radiation dose is to perform late enhancement with
better contrast between infarcted and remote myocardium low-dose high-pitch 128-slice CT (47). Late enhancement CT
without the need of ionizing radiation. A similar study imaging in the high-pitch mode enables myocardial viability
in pigs and dogs published in the same period of time by assessment at a radiation dose as low as about 0.9 mSv, yield-
Lardo et al. (42) reported that the spatial extent of acute and ing good accuracy compared with MRI, although associated
healed myocardial infarction can be determined and quan- with a lower contrast-to-noise and higher noise level (47).
tified accurately with contrast-enhanced CT. They noted There are many technical factors and pitfalls to consider
that this feature, combined with existing high-resolution when performing and reading late enhancement CT exami-
coronary CT angiography, may have important implications nations, further refinement of contrast infusion protocol,
for the comprehensive assessment of cardiovascular disease imaging parameters, and their interaction with patient char-
(42). Later on, patient studies with late enhancement CT were acteristics are needed before viability imaging by CT can
performed immediately after invasive coronary angiography be of routine clinical use (48). Therefore, MRI will most
without iodine reinjection (43). In this study, late enhancement likely remain the imaging modality of choice for evaluation
LWBK1209-ch26_p391-401.indd 399 16/05/13 9:31 PM

Stress
Patient CTCA & Adenosine Delayed
perfusion & 10 min Discharge
preparation rest function infusion enhancement
function
Structure Function Perfusion Viability
Figure 26.7. Schematic diagram of possible comprehensive cardiac CT examination. (Courtesy of Williams
MC, Reid JH, McKillop G, et al. Cardiac and coronary CT comprehensive imaging approach in the assess-
ment of coronary heart disease. Heart. 2011;97:1198–1205.)
of myocardial infarct size and viability. In special circum- 6. Pijls NH, de Bruyne B, Peels K, et al. Measurement of fractional flow reserve to assess the
functional severity of coronary-artery stenoses. N Engl J Med. 1996;334:1703–1708.
stances late enhancement CT imaging can be considered 7. van Werkhoven JM, Schuijf JD, Jukema JW, et al. Comparison of non-invasive multi-slice
clinically, for example, in patients who cannot be studied by computed tomography coronary angiography versus invasive coronary angiography and
fractional flow reserve for the evaluation of men with known coronary artery disease. Am
MRI because of MR-incompatible devices. J Cardiol. 2009;104:653–656.
8. Rispler S, Keidar Z, Ghersin E, et al. Integrated single-photon emission computed tomography
and computed tomography coronary angiography for the assessment of hemodynamically
Summary significant coronary artery lesions. J Am Coll Cardiol. 2007;49:1059–1067.

9. Bastarrika G, Ramos-Duran L, Rosenblum MA, et al. Adenosine-stress dynamic myocar-
dial CT perfusion imaging: Initial clinical experience. Invest Radiol. 2010;45:306–313.
CT seems capable of imaging most aspects of the ischemic 10. Blankstein R, Shturman LD, Rogers IS, et al. Adenosine-induced stress myocardial perfu-
sion imaging using dual-source cardiac computed tomography. J Am Coll Cardiol. 2009;
cascade at the cost of a reasonable effective radiation dose 54:1072–1084.
below 5 mSv. Cardiac CT can potentially serve as a “one- 11. George RT, Ichihara T, Lima JA, et al. A method for reconstructing the arterial input
function during helical CT: Implications for myocardial perfusion distribution imaging.
stop–shop” for evaluation of ischemic heart disease (49). Radiology. 2010;255:396–404.
Such a comprehensive imaging strategy (Fig. 26.7) may 12. Cury RC, Magalhaes TA, Borges AC, et al. Dipyridamole stress and rest myocardial perfu-
improve diagnosis, reduce over-investigation, decrease cost, sion by 64-detector row computed tomography in patients with suspected coronary artery
disease. Am J Cardiol. 2010;106:310–315.
allow targeted treatment, and provide firmer reassurance for 13. George RT, Arbab-Zadeh A, Miller JM, et al. Adenosine stress 64- and 256-row detector
patients (50). Cardiovascular imaging specialists should be computed tomography angiography and perfusion imaging: A pilot study evaluating the
familiar with CT protocol design and comprehensive inter- transmural extent of perfusion abnormalities to predict atherosclerosis causing myocardial
ischemia. Circ Cardiovasc Imaging. 2009;2:174–182.
pretation of cardiac CT (49). At current state of technol- 14. Kurata A, Mochizuki T, Koyama Y, et al. Myocardial perfusion imaging using adenosine
ogy it seems preferable to only perform more advanced CT triphosphate stress multi-slice spiral computed tomography: Alternative to stress myocar-
dial perfusion scintigraphy. Circ J. 2005;69:550–557.
imaging options if alternatives are not possible. 15. Rocha-Filho JA, Blankstein R, Shturman LD, et al. Incremental value of adenosine-induced
stress myocardial perfusion imaging with dual-source CT at cardiac CT angiography.
Radiology. 2010;254:410–419.
16. Ho KT, Chua KC, Klotz E, et al. Stress and rest dynamic myocardial perfusion imaging
References by evaluation of complete time-attenuation curves with dual-source CT. JACC Cardiovasc
Imaging. 2010;3:811–820.
1. Taylor AJ, Cerqueira M, Hodgson JM, et al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/ 17. Mehra VC, Ambrose M, Valdiviezo-Schlomp C, et al. CT-based myocardial perfusion
SCAI/SCMR 2010 appropriate use criteria for cardiac computed tomography. A report of imaging-practical considerations: Acquisition, image analysis, interpretation, and chal-
the American College of Cardiology Foundation Appropriate Use Criteria Task Force, the lenges. J Cardiovasc Transl Res. 2011;4:437–448.
Society of Cardiovascular Computed Tomography, the American College of Radiology, the 18. Valdiviezo C, Ambrose M, Mehra V, et al. Quantitative and qualitative analysis and inter-
American Heart Association, the American Society of Echocardiography, the American pretation of CT perfusion imaging. J Nucl Cardiol. 2010;17:1091–1100.
Society of Nuclear Cardiology, the North American Society for Cardiovascular Imaging, 19. Bamberg F, Becker A, Schwarz F, et al. Detection of hemodynamically significant coronary
the Society for Cardiovascular Angiography and Interventions, and the Society for artery stenosis: Incremental diagnostic value of dynamic CT-based myocardial perfusion
Cardiovascular Magnetic Resonance. J Am Coll Cardiol. 2010;56:1864–1894. imaging. Radiology. 2011;260:689–698.
2. Hoffmann U, Bamberg F, Chae CU, et al. Coronary computed tomography angiography for 20. Techasith T, Cury RC. Stress myocardial CT perfusion: An update and future perspective.
early triage of patients with acute chest pain: The ROMICAT (Rule Out Myocardial Infarction JACC Cardiovasc Imaging. 2011;4:905–916.
using Computer Assisted Tomography) trial. J Am Coll Cardiol. 2009;53:1642–1650. 21. Weininger M, Schoepf UJ, Ramachandra A, et al. Adenosine-stress dynamic real-time myo-
3. Schuijf JD, Wijns W, Jukema JW, et al. Relationship between noninvasive coronary angiog- cardial perfusion CT and adenosine-stress first-pass dual-energy myocardial perfusion CT
raphy with multi-slice computed tomography and myocardial perfusion imaging. J Am Coll for the assessment of acute chest pain: Initial results. Eur J Radiol. 2012;81:3703–3710.
Cardiol. 2006;48:2508–2514. 22. George RT, Silva C, Cordeiro MA, et al. Multidetector computed tomography myocardial
4. Christou MA, Siontis GC, Katritsis DG, et al. Meta-analysis of fractional flow reserve perfusion imaging during adenosine stress. J Am Coll Cardiol. 2006;48:153–160.
versus quantitative coronary angiography and noninvasive imaging for evaluation of myo- 23. Feuchtner G, Goetti R, Plass A, et al. Adenosine stress high-pitch 128-slice dual-source
cardial ischemia. Am J Cardiol. 2007;99:450–456. myocardial computed tomography perfusion for imaging of reversible myocardial ischemia:
5. Meijboom WB, Van Mieghem CA, van Pelt N, et al. Comprehensive assessment of coro- Comparison with magnetic resonance imaging. Circ Cardiovasc Imaging. 2011;4:540–549.
nary artery stenoses: Computed tomography coronary angiography versus conventional 24. Ko BS, Cameron JD, Meredith IT, et al. Computed tomography stress myocardial perfu-
coronary angiography and correlation with fractional flow reserve in patients with stable sion imaging in patients considered for revascularization: A comparison with fractional
angina. J Am Coll Cardiol. 2008;52:636–643. flow reserve. Eur Heart J. 2012;33:67–77.
LWBK1209-ch26_p391-401.indd 400 16/05/13 9:31 PM

25. Geleijns J, Joemai RM, Dewey M, et al. Radiation exposure to patients in a multicenter detection of acute coronary syndrome in patients with acute chest pain: From the ROMICAT
coronary angiography trial (CORE 64). AJR Am J Roentgenol. 2011;196:1126–1132. trial. Circ Cardiovasc Imaging. 2010;3:375–383.
26. Tamarappoo BK, Dey D, Nakazato R, et al. Comparison of the extent and severity of myo- 38. Nakazato R, Tamarappoo BK, Smith TW, et al. Assessment of left ventricular regional wall
cardial perfusion defects measured by CT coronary angiography and SPECT myocardial motion and ejection fraction with low-radiation dose helical dual-source CT: Comparison
perfusion imaging. JACC Cardiovasc Imaging. 2010;3:1010–1019. to two-dimensional echocardiography. J Cardiovasc Comput Tomogr. 2011;5:149–157.
27. Gosling O, Loader R, Venables P, et al. Cardiac CT: Are we underestimating the dose? A 39. Takx RA, Moscariello A, Schoepf UJ, et al. Quantification of left and right ventricular
radiation dose study utilizing the 2007 ICRP tissue weighting factors and a cardiac specific function and myocardial mass: Comparison of low-radiation dose 2nd generation dual-
scan volume. Clin Radiol. 2010;65:1013–1017. source CT and cardiac MRI. Eur J Radiol. 2012;81:e598–e604.
28. Gosling O, Loader R, Venables P, et al. A comparison of radiation doses between state- 40. Chen CM, Liu YC, Chen CC, et al. Radiation dose exposure of patients undergoing 320-row
of-the-art multislice CT coronary angiography with iterative reconstruction, multislice CT cardiac CT for assessing coronary angiography and global left ventricular function. Int J
coronary angiography with standard filtered back-projection and invasive diagnostic coro- Cardiovasc Imaging. 2012;28(suppl 1):1–5.
nary angiography. Heart. 2010;96:922–926. 41. Baks T, Cademartiri F, Moelker AD, et al. Multislice computed tomography and magnetic
29. Paelinck BP, Lamb HJ, Bax JJ, et al. Assessment of diastolic function by cardiovascular resonance imaging for the assessment of reperfused acute myocardial infarction. J Am Coll
magnetic resonance. Am Heart J. 2002;144:198–205. Cardiol. 2006;48:144–152.
30. Paelinck BP, de Roos A, Bax JJ, et al. Feasibility of tissue magnetic resonance imaging: A 42. Lardo AC, Cordeiro MA, Silva C, et al. Contrast-enhanced multidetector computed
pilot study in comparison with tissue Doppler imaging and invasive measurement. J Am tomography viability imaging after myocardial infarction: Characterization of myocyte
Coll Cardiol. 2005;45:1109–1116. death, microvascular obstruction, and chronic scar. Circulation. 2006;113:394–404.
31. Nakahara T, Jinzaki M, Fukuda N, et al. Estimation of the left ventricular diastolic func- 43. Habis M, Capderou A, Sigal-Cinqualbre A, et al. Comparison of delayed enhancement
tion with cardiac MDCT: Correlation of the slope of the time-enhancement-curve with the patterns on multislice computed tomography immediately after coronary angiography
mitral annulus diastolic velocity. Eur J Radiol. 2012;81:234–238. and cardiac magnetic resonance imaging in acute myocardial infarction. Heart. 2009;95:
32. Boogers MJ, van Werkhoven JM, Schuijf JD, et al. Feasibility of diastolic function assess- 624–629.
ment with cardiac CT: Feasibility study in comparison with tissue Doppler imaging. JACC 44. Lessick J, Dragu R, Mutlak D, et al. Is functional improvement after myocardial infarction
Cardiovasc Imaging. 2011;4:246–256. predicted with myocardial enhancement patterns at multidetector CT? Radiology. 2007;
33. Pattynama PM, Lamb HJ, Van der Velde EA, et al. Left ventricular measurements with cine 244:736–744.
and spin-echo MR imaging: A study of reproducibility with variance component analysis. 45. Nieman K, Shapiro MD, Ferencik M, et al. Reperfused myocardial infarction: Contrast-
Radiology. 1993;187:261–268. enhanced 64-Section CT in comparison to MR imaging. Radiology. 2008;247:49–56.
34. Raman SV, Shah M, McCarthy B, et al. Multi-detector row cardiac computed tomography 46. Deseive S, Bauer RW, Lehmann R, et al. Dual-energy computed tomography for the detec-
accurately quantifies right and left ventricular size and function compared with cardiac tion of late enhancement in reperfused chronic infarction: A comparison to magnetic reso-
magnetic resonance. Am Heart J. 2006;151:736–744. nance imaging and histopathology in a porcine model. Invest Radiol. 2011;46:450–456.
35. Annuar BR, Liew CK, Chin SP, et al. Assessment of global and regional left ventricular 47. Goetti R, Feuchtner G, Stolzmann P, et al. Delayed enhancement imaging of myocardial
function using 64-slice multislice computed tomography and 2D echocardiography: A viability: Low-dose high-pitch CT versus MRI. Eur Radiol. 2011;21:2091–2099.
comparison with cardiac magnetic resonance. Eur J Radiol. 2008;65:112–119. 48. Mendoza DD, Joshi SB, Weissman G, et al. Viability imaging by cardiac computed tomog-
36. Feuchtner G, Goetti R, Plass A, et al. Dual-step prospective ECG-triggered 128-slice dual- raphy. J Cardiovasc Comput Tomogr. 2010;4:83–91.
source CT for evaluation of coronary arteries and cardiac function without heart rate 49. Tsai IC, Lee WL, Tsao CR, et al. Comprehensive evaluation of ischemic heart disease using
control: A technical note. Eur Radiol. 2010;20:2092–2099. MDCT. AJR Am J Roentgenol. 2008;191:64–72.
37. Seneviratne SK, Truong QA, Bamberg F, et al. Incremental diagnostic value of regional left 50. Williams MC, Reid JH, McKillop G, et al. Cardiac and coronary CT comprehensive imag-
ventricular function over coronary assessment by cardiac computed tomography for the ing approach in the assessment of coronary heart disease. Heart. 2011;97:1198–1205.
LWBK1209-ch26_p391-401.indd 401 16/05/13 9:31 PM

Chapter David Rosenbaum
27
Venkatesh Mani
Antoine Millon
Zahi A. Fayad
Atherosclerotic Plaque Imaging: MRI
■■ Atherosclerotic Plaques matrix (including collagen), proteoglycans, and fibronectin

elastic fibers; (b) lipids such as crystalline cholesterol, cho-
■■ Mri Techniques
lesteryl esters, and phospholipids; and (c) inflammatory cells
Magnetic Resonance Angiography such as monocyte-derived macrophages, T-lymphocytes, and
Plaque Morphometrics smooth muscle cells (1).
Plaque Characterization The occurrence of these components in varying propor-
Dynamic Contrast-Enhanced MRI tions in different plaques gives rise to a wide and heteroge-
Targeted Molecular Contrast Agents Imaging neous spectrum of lesions (2–4). Plaques are mainly developed
Imaging of Plaque Biomechanics within the arterial intima but the rearrangement of the wall
can also affect the media and the adventitia with, for example,
■■ Applications
development of neovascularization (5,6). Plaque formation is
Measurement of Plaque Size a very long process that initiates early in life. The natural his-
Determination of Plaque Composition tory of plaques is growth over dozens of years under the influ-
In Vivo Magnetic Resonance Imaging Studies on ence of local and systemic accelerating factors such as smok-
Carotid Artery Plaque ing, dyslipidemia, diabetes, hypertension.
In Vivo Magnetic Resonance Imaging Studies on There are two main types of complications of atheroscle-
Aortic Plaques rotic plaques:
In Vivo Magnetic Resonance Imaging Studies on Lumen stenosis
Coronary Artery Plaque The arterial narrowing leads to distal chronic ischemia
In Vivo Magnetic Resonance Imaging Studies on which first occurs in situations when an increase in O2
Peripheral Arteries intake is needed. This is the mechanism of chronic mani-
In Vivo Monitoring of Therapy with Magnetic festations of cardiovascular disease such as lower limb
Resonance Imaging claudication and angina pectoris. It has been demon-
strated that symptoms only occur after a >60% stenosis
■■ Limitations
is reached (7,8).
■■ Future Challenges
Sudden rupture and acute events
■■ Conclusion Plaque rupture or erosion leads to a sudden thrombus for-
■■ Acknowledgments mation and the complete vessel obstruction with possible
distal embolization of thrombotic material. This is the
mechanism of acute events such as stroke, transient isch-
emic attack (TIA), or acute coronary syndromes (ACS)
Atherosclerotic Plaques (9). Plaque rupture is due to plaque destabilization which
has been closely related to its composition, especially
Atherosclerotic plaques are manifestations of both arterial
with the presence of a lipid core and intraplaque hemor-
aging and atherosclerotic process. They occur in the aorta,
rhage (10,11).
carotid, and coronary and peripheral arteries, and they are
the primary cause of heart disease and stroke thus accounting Therefore, within all plaques, a subcategory of vulnerable/
for 50% of all deaths in Western societies. The initial mecha- high-risk plaques has been defined but the natural history of
nism of plaque formation is endothelial dysfunction due to plaques differs between the localizations of atherosclerosis
systemic (hypertension, diabetes) and/or local (shear stress, leading to different types of vulnerable/high-risk plaques.
turbulences) conditions. It explains the preferential localiza- Prone to rupture atherosclerotic coronary artery plaques,
tion of plaques in bifurcations. so-called “vulnerable” plaques, tend to have a thin (∼65 to 150
The main components of atherosclerotic plaque are: (a) mm) fibrous cap (FC) and a large lipid core. ACS often results
Fibrous elements such as connective tissue, extracellular from the rupture of modestly stenotic plaques. Different lesion
402
LWBK1209-ch27_p402-419.indd 402 17/05/13 5:11 PM

Chapter 27 ■ Atherosclerotic Plaque Imaging: MRI 403
types depend in part on stage or phase of atherosclerosis •• EVOLUTION over time with monitoring of potential
progression. The coronary “vulnerable” types IV and Va therapeutic effects of prescribed medications.
lesions and the “complicated” type VI lesion are the most
relevant to ACS according to the criteria of the American
Heart Association Committee on Vascular Lesions. Those MRI Techniques
types IV and Va lesions, although not necessarily stenotic at
angiography, frequently have increased density of resident In the past, the degree of arterial stenosis was considered the
macrophages that release proteolytic enzymes such as me- main determinant factor in determining the need to treat a
talloproteinases (12) that degrade the FC and increase the patient. This notion has, however, in recent years evolved to
risk of plaque rupture. Type IV lesions consist of extracel- include several other factors that are considered potentially
lular lipid intermixed with fibrous tissue that is covered by important markers for future cerebrovascular events. Some
a FC, whereas type Va lesions possess a predominantly ex- of these factors include plaque composition, size of the FC,
tracellular lipid core covered by a thin FC. The lipid core is intraplaque hemorrhage, plaque ulceration and plaque loca-
made highly thrombogenic by the presence of tissue factor tion, and the size of a lipid-rich necrotic core. This has been
(13). The type VI lesion is formed after disruption of type facilitated primarily by the development of dark blood MR
IV or Va lesions and results in ACS, consists of an occlu- imaging techniques that offer high-resolution direct visual-
sive thrombus. The majority of atherosclerotic lesions in ization of plaque components.
the coronary arteries are minimally stenosed (14) but these
relatively nonstenotic plaques with large lipid cores are actu-
ally vulnerable and at high risk for rupture and thrombosis.
The plaque caps are often thinnest at the shoulder regions, The most commonly clinically used MR imaging for visual-
where macrophages (15) and mast cells (16) accumulate, izing atherosclerosis is the magnetic resonance angiography
and disruption is a frequent occurrence (12). In contrast, the (MRA) approach. This is achieved using both noncontrast-
most severely stenotic plaques at angiography, which have a enhanced imaging (time-of-flight (TOF) angiography) or
high content of smooth muscle cells and collagen, and little contrast-enhanced angiography using Gd-chelates. MRA
lipid, are less susceptible to rupture. does not require any sophisticated postprocessing approach
In contrast to vulnerable plaques in the coronary arter- to get rid of calcium-related artifacts, reducing image inter-
ies, high-risk plaques in the carotid arteries are more often pretation time with a simple and fast luminographic display
severely stenotic. The term high-risk, rather than the clas- of the arterial anatomy, superimposable to that of DSA.
sic term vulnerable, should be preferred because vulnerable
implies the presence of a lipid-rich core. Indeed, high-risk
Plaque Morphometrics
carotid plaques are not necessarily lipid rich but rather het-
erogeneous, and are very rich in fibrous tissue. These plaques The direct visualization of atherosclerotic plaques should
often become symptomatic due to an intramural hematoma enhance our understanding of the natural history of this
or dissection that likely develops secondary to the impact of disease. Several invasive imaging techniques, such as x-ray
blood during systole against a resistant area of stenosis (17). angiography, intravascular ultrasonography, angioscopy,
Since the carotid arteries are superficial and not subject to and optical coherence tomography, in addition to nonin-
significant motion, they are much easier to image than the vasive imaging techniques, such as surface B-mode ultraso-
coronary arteries, and therefore, they are the most studied nography and ultrafast computed tomography, can be used
arteries with MRI (18). In the thoracic aorta, atherosclerosis to assess atherosclerotic vessels. Most of the standard tech-
is a significant marker of associated coronary disease (19,20). niques identify the luminal diameter or stenosis, wall thick-
In fact, parameters such as aortic wall thickness, luminal ir- ness, and plaque volume. However, none of these imaging
regularities, and plaque composition are strong predictors methods can completely characterize the composition of
of future vascular events (21). These plaques correspond to the atherosclerotic plaque, and therefore, have inadequate
noncalcified, frequently lipid-laden plaques (American Heart sensitivity and specificity to identify vulnerable or high-risk
Association types IV and Va) as described in the coronary arter- plaques (22).
ies. They have been found to be associated with a significantly High-resolution MR has emerged as the potential leading
increased risk for all vascular events (stroke, myocardial noninvasive imaging modality for characterizing atheroscle-
infarction, peripheral embolism, and cardiovascular death) rotic plaque in vivo. With improvements in imaging tech-
in patients who had noncalcified aortic plaques with a thick- nology, the ability of magnetic resonance (MR) to delineate
ness of more than 4 mm in the French Aortic Plaque Study these vessels has significantly improved (23–29). MR dif-
(FAPS) (17,21). ferentiates plaque components on the basis of biophysical
In summary, the useful information that could be pro- and biochemical parameters, such as chemical composition
vided by atherosclerotic plaque imaging with MRI are: and concentration, water content, physical state, molecular
motion, and diffusion. MR provides a method of imaging
Anatomic imaging
without the need for ionizing radiation that can be repeated
•• Plaque geometry: Size, localization, and percent of
sequentially over time. Dedicated pulse sequences have be-
stenosis.
come more readily available for time-efficient multislice
•• Plaque internal composition
imaging. Over time, a diverse array of black blood MRI
Functional imaging techniques (30,31) has developed into practical tools for the
•• Assessment of plaque internal biologic mechanisms arterial imaging and evaluation of atherosclerosis.
LWBK1209-ch27_p402-419.indd 403 17/05/13 5:11 PM

In black blood MRI, images are commonly acquired using blood turbo spin-echo (TSE) sequence approach currently
a rapid acquisition with relaxation enhancement (RARE) se- being investigated is the sampling perfection with application-
quence with double inversion recovery (DIR) pulses (30) to optimized contrast using different flip angle evolution
provide good contrast between the lumen and vessel wall. (SPACE) approach (37). Other 3D approaches include the
DIR modules consist of two 180-degree radiofrequency (RF) use of a steady-state free precession (SSFP) approach com-
pulses. Magnetization in the whole volume is inverted by bined with a diffusion preparation for nulling blood signal
the first nonselective RF pulse, while magnetization in the (35) and the MSDE approach with a TSE acquisition kernel.
slab of interest is subsequently restored by the second selec- New automated segmentation approaches for plaques have
tive RF pulse. The application of these two preparatory RF also been proposed in different vascular territories such as
pulses causes spins outside the slab of interest to be inverted, carotids (38) and femoral arteries (39).
with no effect on spins within the slab. Image acquisition be-
gins following a time delay (TI) necessary for magnetization
Plaque Characterization
of inverted blood flowing into the imaging slice to reach null
point (30). DIR techniques have been modified for multislice The main advantage of MRI over other noninvasive imaging
black blood imaging (23,24,27,32). The use of black blood modalities is its ability to characterize plaque components and
MRI to directly image atherosclerotic plaques provides the help determine constituents that may be vulnerable to rup-
unique opportunity of measuring plaque and wall changes ture. The characterization of atherosclerotic plaque by MR is
due to atherosclerotic disease with high accuracy, taking based on the signal intensities and morphologic appearance
into account the intrinsic variations of the diseased arterial of plaque on T1-weighted, proton-density–weighted, and T2-
wall. One technique for multislice black blood imaging is the weighted images, as previously validated (18,40,41). The
rapid extended coverage (REX) (24) method. This method fibrous tissues of plaque, consisting of extracellular matrix
allows for imaging up to 20 times faster than the conven- elaborated by smooth muscle cells, are associated with a
tionally used single-slice, black blood DIR technique. More short T1. This T1-shortening (increased signal intensity on
recently, these two-dimensional (2D) approaches have given T1-weighted images) is the consequence of specific interactions
way to three-dimensional (3D) dark blood imaging tech- between protein and water (42). Plaque lipids consist pri-
niques. It is an increasingly common imaging approach em- marily of unesterified cholesterol and cholesteryl esters
ployed in multiple vascular beds including the carotid (33), and are associated with a short T2 (41). The short T2
aorta (34) and femoral (35) arteries in an attempt to evalu- (decreased signal intensity of T2-weighted images) of the
ate systemic disease burden instead of focusing on specific lipid components is in part a consequence of the micellar
lesions. In such instances, 3D black blood approaches can structure of lipoproteins, their denaturation by oxidation,
subsequently be used as a measure to evaluate changes in or an exchange between cholesteryl esters and water mol-
burden in response to treatment and/or evaluate progression ecules (both from the fatty chain or the cholesterol ring),
or regression of disease over time (36). Due to a higher SNR with a further interchange between free and bound water
and decreased susceptibility of volume averaging artifacts, (43,44) (Fig. 27.1). Perivascular fat, mainly composed of
3D approaches are often preferred to the conventional 2D triglycerides, differs in appearance on MR images from
approaches. The ability to obtain isotropic voxels thereby atherosclerotic plaque lipids (45). The signal intensities
enabling multiplanar reformatting of images is yet another of the calcified regions of plaque, which consist primar-
benefit of 3D approaches. The most promising 3D black ily of calcium hydroxyapatite, are low on MR images
Figure 27.1. Multicontrast MRI (PD-, T1-weighted, and T2-weighted) showing a lipid core in the internal
carotid of a patient with atherosclerosis. Multicontrast MRI has emerged as a tool for noninvasive in vivo
determination of plaque components.
LWBK1209-ch27_p402-419.indd 404 17/05/13 5:11 PM

because of their low proton density and diffusion-mediated MRI has been used to identify intraplaque hemorrhage (67)
susceptibility effects (46). and also differentiate intraplaque and juxtaluminal hemor-
The MR appearance and evolution of thrombus or hem- rhage/thrombus in advanced human carotid atherosclerosis
orrhage in the central nervous system (47), pelvis (48), and (68,69). The use of gadolinium-based contrast agents has
aorta (49,50) has been studied. The identification of hem- enabled improved characterization of carotid plaque by not
orrhage with MRI depends on the structure of hemoglobin only improving visualization of the arterial lumen but also
and its oxidation state. At various stages, blood clots con- by differential contrast enhancement of various plaque com-
tain different products (i.e., oxyhemoglobin, deoxyhemoglo- ponents. For example, Yuan et al. (70,71) demonstrated an
bin, methemoglobin, hemosiderin/ferritin). Each of these has 80% enhancement of fibrous tissue while there was only
a set of specific MR relaxation properties (T1 and T2) that a 29% enhancement of the necrotic core when comparing
produce different signal intensities (51–55). Blood shortens postcontrast-enhanced with noncontrast-enhanced high-
the T1 and T2 of water. Presumably, T1-shortening is caused resolution MRI. MRI has also been used to study differences
by the formation of methemoglobin (which is paramagnetic) in carotid plaque composition between different ethnic and
from hemoglobin, whereas T2 shortening is caused by mag- racial groups (72).
netic susceptibility. In this regard, ferritin- or hemosiderin-
rich mature thrombus is associated with marked signal loss
Dynamic Contrast-Enhanced MRI
on T2-weighted images.
As mentioned previously, most in vivo studies of MR im- The formation of neoangiogenic vessels has been associated
aging and characterization of plaque have been performed with the hypoxic conditions in the thickened intimal layer
in a multicontrast approach (i.e., T1-weighting, proton- (>500 mm), and recent observations suggest that neovascular-
density–weighting, and T2-weighting) with high-resolution ization may promote plaque inflammation and progression
black blood spin-echo and fast spin-echo (FSE) MR se- already at an early stage of plaque formation by facilitating
quences. As the name implies, the signal from flowing blood extravasation of inflammatory cells. Plaque growth has also
is rendered black by the use of preparatory pulses (e.g., RF been intimately associated with intraplaque hemorrhage of
spatial saturation or inversion recovery pulses) for better the fragile neovasculature. Dynamic contrast enhancement
visualization of the adjacent vessel wall. Hatsukami et al. (DCE) MRI is a well-validated technique examining the
(56) introduced the use of bright blood imaging (i.e., 3D vascularity of tumors. In the past few years, this technique
fast TOF imaging) to visualize FC thickness and morpho- has been adapted for use in atherosclerotic plaque neovas-
logic integrity. This sequence provides enhancement of the cularization in both animal models (73–75) and in humans
signal from flowing blood and a mixture of T1-weighting (76). DCE-MRI uses repeated MRI acquisitions of the same
and proton-density contrast–weighting, which highlights the imaging slice/volume serially with high temporal resolution
FC. Multicontrast MRI of human carotid plaque had high after the injection of a Gd-based contrast agent. DCE-MRI
sensitivity and specificity for identification of unstable FCs has demonstrated the ability to differentiate fibrous tissue
and may enable prospective identification of high-risk ca- (enhancement ∼80%) from lipid core (enhancement ∼30%)
rotid plaques (57). by virtue of the neovessel density in fibrous tissue. It has
In addition, MR direct thrombus imaging using a T1- also been shown to correlate well with neovessel density on
weighted turbo field-echo sequence, which improves detec- histology (77,78) (Fig. 27.2). Both model-based and model-
tion of methemoglobin, enabled accurate identification of free approaches to analyzing DCE images have been pro-
complicated carotid lesions with thrombus in symptomatic posed and implemented. Appropriate modeling of the signal
patients that underwent imaging and subsequent endarter- intensity time course in the plaque provides information on
ectomy (58,59). contrast agent plaque uptake kinetics and estimations of the
Measurement of carotid wall volume and maximum area amount of neovascularization and vessel permeability (79).
using 3D contrast-enhanced MR has also been performed
(60). MRA and high-resolution black blood imaging of the
Targeted Molecular Contrast
vessel wall can be combined. MRA demonstrates the severity
Agents Imaging
of stenotic lesions and their spatial distribution, whereas the
high-resolution black blood wall characterization technique The ability to target specific molecules of plaques may greatly
may show the composition of plaques and facilitate risk enhance detection and characterization of atherosclerotic
stratification and selection of treatment modality. Spatial and atherothrombotic lesions using MRI (80). Contrast
resolution has recently been improved (≤250 mm) with the agents linked to antibodies (81–84) or peptides (85,86) that
design of new phased-array coils (61,62) tailored for carotid target specific plaque components or molecules that local-
arteries and new imaging sequences, such as long echo train ize to specific regions of atherosclerotic plaque (25,87–89)
FSE imaging with “velocity-selective” flow suppression and are examples of strategies that have been employed to image
DIR preparatory pulses (black blood imaging) (60,61,63). atherosclerosis with MRI.
Following surgery, the carotid plaque specimens were The ability to target mononuclear cells such as monocytes,
again imaged ex vivo with high-resolution MRI and revealed macrophages, and foam cells is an attractive means of iden-
good correlation between in vivo and ex vivo measurements tifying atherosclerosis since these cells have been shown to
of maximum wall area, minimum lumen area, and wall vol- play a pivotal role in the progression of atherosclerosis to
ume. High-resolution MRI also had good sensitivity and symptomatic disease (90,91). While current research is un-
specificity for determining the type of carotid lesion based on derway to target macrophages with gadolinium-based con-
the American Heart Association classification system (64–66). trast agents that target the macrophage scavenger receptor,
LWBK1209-ch27_p402-419.indd 405 17/05/13 5:11 PM

A B C
D E F
Figure 27.2. DCE-MRI. A–F: Single-slice axial T1-weighted MRI scans, with color-encoded overlay of
the contrast signal at 1 minute of the abdominal aorta. A,D: Baseline control and baseline PIO groups.
B,E: One-month control and PIO groups. C,F: Three-month control and PIO groups. Warm colors (orange
to red) indicate higher AUC values and cold colors (green to blue) indicate lower AUC values. Bar graph
summarizes the changes in the AUC at BL, 1M, and after 3M for the control and PIO groups. Values are
represented as mean ±SD. Area under the curve remained relatively unchanged after 3 months in control
group but regressed in treatment groups after 3 months. (From Vucic E, Dickson SD, Calcagno C, et al.
Pioglitazone modulates vascular inflammation in atherosclerotic rabbits noninvasive assessment with
FDG-PET-CT and dynamic contrast-enhanced MR imaging. JACC Cardiovasc Imaging. 2011;4:1100–
1109.)
macrophages have only been imaged with iron oxide com- easily be replaced by linking an MR contrast agent to the
pounds (USPIOs, SPIOs) that are removed from the circu- monoclonal antibody to target the molecule of interest.
lation by macrophages and other cells of the reticuloendo- However, the identification of molecules found only in ath-
thelial system (87–89). In a study by Kooi et al. (88) on 11 erosclerotic plaque will ultimately enable improved detec-
symptomatic patients scheduled for carotid endarterectomy, tion of atherosclerotic plaque assuming that the target is
75% of ruptured or rupture-prone lesions demonstrated up- expressed in adequate quantity for detection by molecular
take of USPIOs compared with only 7% of stable lesions MRI.
and a decrease in signal intensity of 24% on T2*. Gadofluorine M is a lipophilic, macrocyclic (1,528 Da),
Neovascularization has been shown to play an impor- water-soluble, gadolinium chelate complex (Gd-DO3A-
tant role in atherosclerosis and the integrin αvβ3 has been derivative) with a perfluorinated side chain. Sirol et al. (25)
targeted to identify regions in the vessel wall undergoing and Barkhausen et al. (101) demonstrated that Gadofluorine
neovascularization (82,83). Winter et al. (84,92) recently M enhanced aortic wall imaging in Watanabe heritable hy-
demonstrated in a rabbit model of atherosclerosis that re- perlipidemic rabbits but did not enhance the aorta of control
gions of neovascularization in plaque had a 47% increase rabbits. Gadofluorine M increased signal intensity by 164%
in signal intensity following treatment with αvβ3-targeted at 1-hour postcontrast and increased signal intensity 207% at
nanoparticles. 24-hour postcontrast.
Additional targets of interest for imaging of atheroscle- The ability to identify components of thrombus with
rotic plaque with molecular specific MR contrast agents are molecular MRI may enable enhanced detection and char-
oxidized low-density lipoprotein (oxLDL), tissue factor, en- acterization of both luminal thrombus and components
dothelial integrins, matrix metalloproteinases, and extracel- of thrombus organized in an old atherothrombotic lesion.
lular matrix proteins such as tenascin-C. Therefore, selection of targets pivotal in the coagulation
While targeted nuclear imaging through the use of anti- cascade, such as fibrin, factor XIII, integrins on the surface
bodies specific to oxLDL has been demonstrated promises of platelets, and tissue factor, is necessary to identify the
at detecting atherosclerosis (93), estimating plaque volume areas of old or active thrombus formation. The attach-
(94), and following progression/regression of atheroscle- ment of MR contrast agents to both monoclonal antibod-
rosis (95). Recently, studies evaluating oxLDL as a target ies or peptide ligands that specifically bind components of
for molecular MRI of atherosclerosis have been performed thrombus has been performed in animal models (102–104)
(96,97). In addition, molecules such as tissue factor (85) and and humans (92,105). In addition, thrombus resulting
endothelial integrins (98–100) such as E-selectin, P-selectin, from plaque rupture has been identified using fibrin-spe-
intracellular adhesion molecule-1, or vascular cell adhesion cific MR contrast agents in a rabbit carotid crush injury
molecule-1 have been targeted with antibodies linked to model (106,107). In the 25 arterial thrombi induced by
echogenic contrast agents. carotid crush injury, Botnar et al. (106) demonstrated a
While these agents utilize echocardiography or nuclear sensitivity and specificity of 100% for in vivo thrombus
imaging, the echogenic or nuclear contrast agent could detection using MRI. Sirol et al. (107) investigated a similar
LWBK1209-ch27_p402-419.indd 406 17/05/13 5:11 PM

Figure 27.3. MRI scans of the left

common carotid artery are shown,
before and after USPIO infusion at
baseline and after 12 weeks of statin
treatment. Pre-USPIO imaging remains
very similar in both time points, indi-
cating that USPIOs from the baseline
imaging session were not present in the
plaque at 12 weeks. Signal enhance-
ment (indicated by the blue arrowheads
on the image) can be seen after 12 weeks
of treatment, with no evidence of signal
voids. The yellow arrowhead indicates
signal void due to USPIO accumula-
tion. The red arrowhead indicates that
USPIO was not retained in the plaque,
as the signal void at 12 weeks is not
appreciable anymore. (Modified from
Tang TY, Howarth SP, Miller SR, et
al. The ATHEROMA (Atorvastatin
Therapy: Effects on Reduction of
Macrophage Activity) Study. Evaluation
using ultrasmall superparamagnetic
iron oxide-enhanced magnetic reso-
nance imaging in carotid disease. J Am
Coll Cardiol. 2009;53:2039–2050.)
fibrin-specific MR contrast agent in 12 guinea pigs demon-

Imaging of Plaque Biomechanics
strating that thrombus signal intensity was increased by
over fourfold after intravascular delivery of contrast agent The overarching goal of plaque imaging is to identify plaque
and thrombus was detected in 100% of animals postcon- features that predict risk of clinical events. Biomechanical
trast compared with 42% identification of thrombus pre- features of vulnerable plaques can be exploited to help strat-
contrast (107). ify this plaque risk. Therefore, a single morphologic feature
Inflammatory activity in atherosclerotic plaque is also may not truly represent the total vulnerability of the plaque,
identified as a strong risk factor for plaque rupture and pro- but it may be a combination of all the biomechanical fea-
gression and may be used as an imaging target to evaluate tures of these vulnerable features (108). Several studies have
interventional therapy. While inflammatory activity is typi- linked biomechanical forces to the pathogenesis of athero-
cally assessed in vivo using 18-FDG-PET, some approaches sclerotic plaque implying that biomechanical analyses may
using MRI have also been proposed. Inflammatory ac- reveal a predisposition for vulnerable plaque development.
tivity can be evaluated at the (sub) cellular level using in Thus plaque biomechanics play a vital role in determining
vivo molecular MRI. Various MRI contrast agents, such hallmarks of vulnerability (109).
as ultra-small particles of iron oxide, micelles, liposomes, Plaque mechanics using MRI are typically performed
low–molecular-weight Gd-chelates, and perfluorocarbon using computational models derived from vessel wall mor-
emulsions, have been used for in vivo visualization of vari- phology. Computational fluid dynamics (CFD) provides
ous inflammation-related targets, such as macrophages, an elegant method to derive geometry and flow boundary
oxLDL, endothelial cell expression, plaque neovascula- conditions from MR images (110). By modeling blood flow
ture, MMPs, apoptosis, and activated platelets/thrombus using patient- and vessel-specific geometry and blood flow
(Fig. 27.3). conditions consists of many steps. Some of these steps include
LWBK1209-ch27_p402-419.indd 407 17/05/13 5:11 PM

image segmentation, geometric reconstruction, mesh genera- a slice thickness of 3-mm proton-density–weighted and
tion, determination of boundary conditions, and numerical T2-weighted images were obtained. A significant correlation
analysis (111). Alternatively, MRI can also directly charac- between MRI and histology was observed in the analysis
terize arterial flow patterns, as well as measure blood flow of lipid-rich and fibrous areas. In two separate serial stud-
velocity using phase-contrast MRI (PC-MRI). The velocity ies (129,130), MRI demonstrated significant regression of
information provided by PC-MRI can be used to compute aortic plaque in vivo in atherosclerotic rabbits that under-
wall shear stress maps along different arterial segments went cholesterol lowering. After aortic balloon injury and
(112). MRI techniques also exist to directly measure tissue feeding of a high-cholesterol diet, one group of rabbits was
strain. Circumferential strain can be measured by displace- continued on the atherogenic diet (atherosclerosis progres-
ment encoding with stimulated echoes (DENSE) agreed sion), and another group was placed on a normal chow,
and has shown correlations with CINE measurements of low-cholesterol diet (atherosclerosis regression). A signif-
changes in lumen circumference in normal carotid arteries. icant reduction in vessel wall area and mean wall thick-
Such strain imaging approaches, however, may not have the ness was observed in the dietary regression group, and
required resolution to measure radial strain, another impor- an increase was seen in the dietary progression group. A
tant strain component (113). significant reduction in the lipid-rich component of plaques
was also observed in the dietary regression group, and an
increase in the dietary progression group. A small, nonsignif-
Applications icant increase in the fibrous plaque components was noted
in the dietary regression group, but there was a significant
Measurement of Plaque Size decrease in the fibrous composition of lesions in the progres-
Plaque size and stenosis percentage assessment are key sion group. A significant correlation was found between MR
answers to many clinical problems such as the determination and histopathology for atherosclerotic burden and plaque
of symptoms’ origin or whether or not to perform surgery composition (129). Using a 9.4-T (89-mm) bore system our
on a patient with severe carotid stenosis. team studied atherosclerosis in alive animals (124). The
In humans, MRI validation initially came from the com- achievable in-plane spatial resolution with MR microscopy
parison of histology pieces from excised abdominal aortic was approximately 50 mm to 97 mm, and the slice thickness
aneurysm (114), but most MRI validation studies were per- was 500 mm. Using transgenic apolipoprotein E-knockout
formed on carotid plaque of patients who later underwent mice, we showed an excellent agreement between MR micros-
carotid endarterectomy. Carotid lesions were imaged in vivo, copy and histologic findings for aortic plaque size, shape, and
then excised en bloc, and the histology specimens rescanned characteristics. This study was extended to follow the rapid
with the same parameters to finally perform a paired com- progression of atherosclerosis in animals with lesions of vary-
parison (18,71,115–119). Throughout their extensive vali- ing severity (131).
dation work, Yuan et al. showed small intra- and interob-
server variability with intraclass correlation coefficients In Humans
ranging from 0.9 to 0.98 in the measurement of the vessel
wall area of atherosclerotic lesions, for example (120). Corti MRI can also assess plaque composition in humans (132,133).
et al. (121) reported similar reproducibility data in repeated Initial studies on the application of MRI to characterize plaque
carotid and aortic MRIs. When averaging the vessel wall ex vivo focused on lipid assessment with nuclear MR spectros-
area values of five continuous images, the image-specific copy and chemical shift imaging (134–141). Unfortunately,
error was 2.5 mm2 for the carotid and 4.5 mm2 for the aorta the concentration of lipid in plaque is very low in compari-
corresponding to an error of 3.5% and 2.6%, respectively. son with that of water, and these techniques suffered from
Based on these data, it is admitted that MRI could accurately poor signal-to-noise ratio (SNR) (41,134,138). MRI ability
measure changes in plaque size over 5% in the aorta and 7% to identify plaque components was then assessed in stud-
in the carotid. Another key validation study performed by ies in which imaging of lesions was performed in patients
Fayad et al. (122) showed a strong correlation between mea- scheduled for carotid endarterectomy. Several teams dem-
surement of plaque thickness and composition of thoracic onstrated that MR techniques were able to identify in vivo
aortas as imaged by MRI or transesophageal echography lipid components (134,141), fibrous cap, and intraplaque
(TEE). hemorrhage (18,135). Later, Cai and Chung (142) went
further deep into the classification of plaques being able to
separate high- and low-risk lesion types using multicontrast
Determination of Plaque Composition imaging with different contrast-weighted images (TOF, and
In Animals T1-weighted, proton-density–weighted, and T2-weighted) on
the carotid arteries. Thus, they were able to show good agree-
MRI has been used to study plaque composition in several ment with the American Heart Association classification.
animal models, including rabbits (123), mice (124), pigs
(125), rats (126), and nonhuman primates (127). In rabbits
In Vivo Magnetic Resonance Imaging
with atheroma induced in the thoracic and abdominal aorta
Studies on Carotid Artery Plaque
by a combination of atherogenic diet and double-balloon
denudation, we validated the ability of MRI to quantify lipid- As mentioned above, in vivo images of advanced lesions in
rich and fibrous components of lesions (128). FSE sequences carotid arteries have been mostly obtained from patients
were obtained with an in-plane resolution of 0.35 mm, and referred for endarterectomy. The carotid arteries, with a
LWBK1209-ch27_p402-419.indd 408 17/05/13 5:11 PM

Figure 27.4. Two-dimensional cross-sectional images of the neck showing atherosclerosis in the carotid
arteries obtained using the REX (multislice dark blood spin-echo) sequence. Complex lesions consisting of a
lipid-rich necrotic core with fibrous tissue are observed in the right carotid bulb (arrows).
superficial location and relative absence of motion, present provides enhancement of the signal from flowing blood and
less of a technical challenge for imaging than do vessels such a mixture of T1-weighting and proton-density contrast–
as the aorta and coronary arteries (Fig. 27.4). Short T2 com- weighting, which highlights the FC. Multicontrast MRI of
ponents were quantified in vivo before surgery, and values human carotid plaque had high sensitivity and specificity
were correlated with values obtained in vitro after surgery for identification of unstable FCs and may enable prospec-
(18). Some of the MR studies of carotid arterial plaques tive identification of high-risk carotid plaques (57). In ad-
include imaging and characterization of normal and patho- dition, MR direct thrombus imaging using a T1-weighted
logic arterial walls (18), quantification of plaque size (116), turbo field-echo sequence, which improves detection of
and detection of FC “integrity” (56). methemoglobin, enabled accurate identification of compli-
Most in vivo studies of MR imaging and characterization of cated carotid lesions with thrombus in symptomatic patients
plaque have been performed in a multicontrast approach (i.e., who underwent imaging and subsequent endarterectomy
T1-weighting, proton-density–weighting, and T2-weighting) (59,143). Measurement of carotid wall volume and maxi-
with high-resolution black blood spin-echo and FSE MR mum area using 3D contrast-enhanced MR has also been
sequences. As the name implies and as explained in the previ- performed (60,61).
ous chapter, the signal from flowing blood is rendered black MRA and high-resolution black blood imaging of the ves-
by the use of preparatory pulses (e.g., RF spatial saturation or sel wall can be combined. MRA demonstrates the severity of
inversion recovery pulses) for better visualization of the ad- stenotic lesions and their spatial distribution, whereas the
jacent vessel wall. Hatsukami et al. (56) introduced the use high-resolution black blood wall characterization technique
of bright blood imaging (i.e., 3D fast TOF imaging) to visu- may show the composition of plaques and facilitate risk
alize FC thickness and morphologic integrity. This sequence stratification and selection of treatment modality. Spatial
LWBK1209-ch27_p402-419.indd 409 17/05/13 5:12 PM

resolution has been improved recently (≤250 mm) with the ization as shown in MRI was associated with prone to rup-
design of new phased-array coils (61) tailored for carotid ture plaques (155–157).
imaging (144) and new imaging sequences, such as long Three prospective population studies have performed MRI
echo train FSE imaging with “velocity-selective” flow sup- in primary prevention to relate clinical features and MRI
pression and DIR preparatory pulses (black blood imaging) findings in vivo. In a Multi-ethnic Study of Atherosclerosis
(61). Luo et al. (60) performed high-resolution black blood (MESA) substudy in subjects with the highest intima–media
MRI in vivo on 37 patients with carotid artery disease who thickness, the presence of a lipid-rich necrotic core was cor-
then underwent carotid endarterectomy. Following surgery, related to total cholesterol (158).
the carotid plaque specimens were again imaged ex vivo In the Atherosclerosis Risk in Community Study (ARIC),
with high-resolution MR and revealed good correlation be- the presence of a thin FC was related to platelet CD40, ge-
tween in vivo and ex vivo measurements of maximum wall netic variations within the MMP2 promoter region, and to
area, minimum lumen area, and wall volume. homozygous state for the Leu33Pro allele (159,160).
High-resolution MRI also had good sensitivity and spec- Recently, van den Bouwhuijsen et al. (161) published
ificity for determining the type of carotid lesion based on the a key study on the clinical determinants of the carotid
American Heart Association classification system (64–66). plaque composition. Participants of the population-based
MRI has been used in advanced human carotid atheroscle- Rotterdam Study with carotid wall thickening >2.5 mm un-
rosis (145) to identify intraplaque hemorrhage (146) and derwent MRI and cardiovascular risk factors assessment.
differentiate intraplaque and juxtaluminal hemorrhage/ MRI was performed on 1,006 subjects of 70 ± 10 years.
thrombus. It was found that lipid-rich necrotic core was correlated to
The use of gadolinium-based contrast agents has enabled hypercholesterolemia and that calcifications were associ-
improved characterization of carotid plaque by not only ated with hypertension and intraplaque hemorrhage with
improving visualization of the arterial lumen but also by smoking, hypertension. Male gender was also a strong de-
differential contrast enhancement of various plaque compo- terminant of plaque size as well as presence of intraplaque
nents. For example, Yuan et al. (71) demonstrated an 80% hemorrhage. Interestingly, diabetes was not associated with
enhancement of fibrous tissue while there was only a 29% a higher risk of any of the plaque component studied in the
enhancement of the necrotic core when comparing postcon- Rotterdam Study whereas in another much smaller study by
trast-enhanced with noncontrast-enhanced high-resolution Esposito et al. (162) on patients with much more advanced
MRI. MRI has also been used to study differences in ca- lesions diabetes was associated with the presence of vulner-
rotid plaque composition between different ethnic and racial able plaque components. These results are consistent with
groups (72). another study by Caballero et al. (163), which showed the
Several studies have tried successfully to link anatomic superiority of MRI over classic Doppler ultrasound for sub-
features of the carotid plaques with events or medical his- clinical atherosclerosis screening. In 36 patients with famil-
tory of patients as seen in Table 27.1. ial hypercholesterolemia MRI detected lipid-rich plaques in
In carotid arteries, it has been proved that the plaque the aorta in 94% of the patients versus 14% in the carotid
growth may be partially explained to intraplaque hemor- as seen in ultrasound.
rhages due to development of vasa vasorum. The intraplaque Recently, in a young-to-middle–aged population, MRI
inflammation process is at the source of this neovasculariza- examinations were able to detect lipid cores in 17% of the
tion. Moreover, Sluimer et al. have shown that neovascular- plaques that were found in those asymptomatic subjects
with a significant association with their blood cholesterol
level and no other traditional risk factor (164).
In an another recent study, carotid plaque stress analy-
linical and Morphologic
C sis from in vivo MRI using coupled fluid/structure interac-
Table 27.1 Findings in Carotid MRI tion was applied to 12 symptomatic and 8 asymptomatic
Studies patients, and followed by a detailed biomechanics analysis,
and morphologic feature study. This very interesting work
MRI Findings Clinical Feature showed (165) that local maximum stress values within the
FC region were found to be significantly higher in symptom-
Fibrous cap thinning or rupture History of recent TIA or stroke atic patients than that in asymptomatic patients. Although
(147)
there was no significant difference in lipid core size between
Intraplaque hemorrhage (148) Cerebrovascular events
the two groups, a larger lipid core and a significantly thin-
(ex vivo)
Intraplaque hemorrhage, Symptomatic carotid stenosis ner FC were found in the group of symptomatic patients.
fibrous cap disruption Plaques with a higher stenosis were also more likely to have
(147,149,150) (in vivo) extreme stress conditions upstream of plaque throat.
High-intensity signal related to Ipsilateral symptoms
thrombus presence on one
side (151)
Lipid-rich necrotic core Degree of stenosis (152) Studies on Aortic Plaques
Type VI lesion Ipsilateral stroke or TIA (153) In vivo MR characterization of atherosclerotic plaque in the
High-intensity signals in carotid Future coronary events in
human aorta has been reported (122,166), (Fig. 27.5).
plaques on T1-weighted patients with previous
The principal challenges associated with MRI of the tho-
coronary artery disease (154)
racic aorta are to obtain sufficient sensitivity for submillimeter
LWBK1209-ch27_p402-419.indd 410 17/05/13 5:12 PM

Figure 27.5. Two-dimensional cross-sectional multislice dark blood imaging of the abdominal aorta of a
patient with atherosclerosis. Eccentric plaques with spotty calcifications are seen (arrows).
imaging and to exclude artifacts caused by respiratory disease and cardiovascular risk factors were associated with
motion and blood flow. Summers et al. (166) used MRI aortic atherosclerosis.
to show that the wall thickness of the ascending aorta is Recent technical improvements with increased magnetic
increased in patients with homozygous familial hypercho- field strength, new dedicated sequences to suppress the
lesterolemia but only conventional T1-weighted spin-echo signal of flowing blood with double inversion preparation
images were obtained, and therefore, plaque composition pulses, and navigator techniques to correct respiratory mo-
was not analyzed. tion have increased MRI performances on the aortic bed.
Fayad et al. (122) assessed the composition of plaque Beyond arterial wall, some studies have focused on the
in the thoracic aorta with T1-weighted, proton-density– relation between wall shear stress as evaluated by MRI and
weighted, and T2-weighted images. Rapid high-resolution the localization of plaques showing that in aortic regions
imaging was performed with an FSE sequence in conjunc- with low shear stress, such as the inner aortic (169) curva-
tion with “velocity-selective” flow suppression preparatory ture, more atherosclerotic lesions were found. Recent clini-
pulses. The results of matched cross-sectional aortic imaging cal studies have shown the interest of MRI of thoracic aorta
with MR and TEE correlated strongly for plaque composi- to improve the detection of causes of strokes (170,171) or
tion and mean maximum plaque thickness. to link Lp(a) levels and abdominal aortic atherosclerosis
In asymptomatic subjects from the Framingham Heart (172).
Study (FHS), MR showed that the prevalence and burden
(i.e., plaque volume/aortic volume) of aortic atherosclero-
sis increased significantly with age and was greater in the
Studies on Coronary Artery Plaque
abdominal aorta compared with the thoracic aorta (167).
It was also found that long-term measures of risk factors Noninvasive imaging of coronary artery plaques would rep-
and the FHS coronary risk score are strongly associated with resent the “graal” of noninvasive plaque imaging as it has
asymptomatic aortic atherosclerosis as detected by MRI. In been widely demonstrated the powerful predictive power
patients referred for coronary angiography, Taniguchi et of coronary plaque presence and types. Preliminary studies
al. (168) also showed that the presence of coronary artery were made in the pig (40,173) before high-resolution black
LWBK1209-ch27_p402-419.indd 411 17/05/13 5:12 PM

blood MRI of both normal and atherosclerotic human coro-

nary arteries were performed. It has showed that the difficul-
ties encountered in imaging the coronary wall are caused by
a combination of cardiac and respiratory motion artifacts in
addition to the tortuous course, small size, and location of
the coronary arteries.
Fayad et al. (63) were the first to demonstrate the feasi-
bility of coronary plaque imaging in humans in vivo. The
first image acquisitions during breath-holding to minimize
respiratory motion were then improved 58/60 with real-
time navigator combined with slice position correction and
black blood FSE method to alleviate the need for breath-
holding. New improvements are still needed to allow reli-
able and precise MR characterization of coronary plaques
(174–177). MRI-based atherosclerotic plaque models with
fluid-structure interactions have been described to perform
flow and stress analysis and identify possible mechanical
and morphologic indices for accurate plaque vulnerability in
human coronary assessment (178).
Recently, Kawasaki et al. (179) have shown the possi-
bility to characterize coronary hyperintense plaque using
noncontrast T1-weighted imaging in cardiac MR correlated
to ultrasound attenuation and low-computed tomography
density.

Studies on Peripheral Arteries
MR techniques have been used for a long time to assess ath-
erosclerotic burden in peripheral artery disease (Fig. 27.6).
Contrast-enhanced MR angiography (CEMRA) has emerged
as an interesting tool in patients with contraindication to
iodine contrast agent injection during CT. The images are
produced from a series of overlapping body segments. The
peripheral arteries are typically divided into aortoiliac, fem-
oropopliteal, and calf (or below knee) stations, and each is
imaged separately during first passage of an intravenously
administered gadolinium-based contrast agent (GBCA). The
issue is the delayed transit time of the contrast agent during
its course in the limb, especially in patients with severe PAD Figure 27.6. Three-dimensional multiplanar reconstructions
where diagnostic accuracy of 96% to 100% in the pelvis of carotids, aorta, and femoral arteries obtained using the SPACE
and thigh, but only 43% in the calf (180). To overcome this sequence. All images were acquired in less than 30 minutes. Three-
dimensional dark blood imaging is an efficient method to scan multiple
problem, “hybrid” MRA techniques have evolved, with real-
vascular territories for atherosclerotic disease.
time tracking of the contrast bolus to allow accurate image
acquisition in the arterial phase, to minimize venous contam-
ination and to increase diagnostic accuracy (181,182). The
latest trials of very rapid dynamic 3D CEMRA using of view angiographic techniques, namely TOF and phase-contrast
sharing/keyhole techniques and in combination with parallel MR angiography, several novel techniques have since been
acceleration, report sensitivity and specificity of >90% for developed, including electrocardiograph (ECG)-gated FSE,
the detection of significant stenosis or occlusion (183,184). SSFP, and arterial spin labeling (186).
Majority of CEMRA data have been obtained in patients Concerning the assessment of plaque itself, MR tech-
with claudication, so its accuracy in patients with critical niques have been able to identify plaques in “angiographi-
lower limb ischemia is less clearly established. However, in a cally normal” femoral arteries with up to 76% stenosis as
study of 30 patients with critical limb ischemia who under- measured by cross-sectional area. Post balloon angioplasty
went CEMRA with high-resolution imaging of the calf ves- changes in lumen diameter and lesion size have been also
sels, diagnostic accuracy was similar to DSA (185). reported as well as plaque fissuration and dissection. Those
Unenhanced MR angiographic techniques are increas- alterations may explain post procedure complications in
ingly used to avoid severe complications caused by con- patients with post dilatation satisfactory considered angio-
trast agents such as iodinated contrast material-induced graphic images. In vivo cross-sectional analysis of arterial
nephropathy and gadolinium-induced nephrogenic systemic wall remodeling by MRI after endovascular brachytherapy
fibrosis. Alongside the two conventional unenhanced MR (187–189) has also been reported.
LWBK1209-ch27_p402-419.indd 412 17/05/13 5:12 PM

Recently, we described a 3D diffusion-prepared seg- decrease in the vessel wall area but no change in the lumi-
mented SSFP sequence for evaluating atherosclerotic plaque nal area was noted at 12 months, confirming the findings of
burden in inguinal and thigh segments of the femoral artery previous experimental studies (130,191). A longer follow-up
and comparing the results obtained with 2D-TSE techniques showed a continued reduction in the arterial wall area and
(35). Interobserver reproducibility for the 3D plaque burden even a small, but significant, increase in the arterial lumen at
measures was excellent. Results indicated a good correlation 24 months (197).
between 3D-DP-SSFP and 2D-TSE techniques for the ingui- Gottlieb et al. (198,199) confirmed that plaque regres-
nal segment but poorer correlation for the thigh segment. sion was strongly associated with LDL-C reduction when
measuring thoracic aorta atherosclerotic plaques total sur-
face by transesophageal MRI in 27 patients before and after
In Vivo Monitoring of Therapy with 6 months of therapy by simvastatin 20 to 80 mg daily. In
Magnetic Resonance Imaging fact, no difference was detected between the randomized
In Animals doses (simvastatin 20 vs. 80 mg). In another study, 43 pa-
tients with moderate hypercholesterolemia (LDL-C between
In a serial MRI study, hepatic hydroxymethylglutaryl coen- 100 and 250 mg/dL) and <80% carotid stenosis assessed
zyme A (HMG CoA) reductase inhibitors (statins) and a by duplex ultrasound were randomized to receive either
novel class of agents, the acylcholesterol acyltransferase a low (5 mg) or high (40/80 mg) dose of rosuvastatin for
(ACAT) inhibitors, showed beneficial effects in Watanabe 24 months. It resulted in a significant parallel reduction in
rabbits (190). In the rabbit model of aortic atherosclerosis, LDL-C levels and in lipid-rich necrotic core size whereas the
our team also showed that MRI could be used to document overall plaque burden remained unchanged. More recently,
arterial remodeling (191). High-resolution MRI and MR the ATHEROMA study demonstrated that aggressive lipid
microscopy might allow convenient and noninvasive quan- lowering with atorvastatin 80 mg daily over a 3-month pe-
titative assessment of serial changes in atherosclerosis in dif- riod was associated with a significant reduction in plaque
ferent animal models of disease progression and regression inflammation measured using ultra-small superparamag-
(192). Moreover, after treating hyperlipidemic rabbits with netic iron oxide particles (200). Using the same study de-
aortic atherosclerosis with statins, selective peroxisomal sign, Yonemura et al. (201) investigated the effects of 20
proliferator-activated receptor-gamma (PPAR-gamma), or versus 5 mg atorvastatin on thoracic and abdominal aor-
a combination of both, Corti et al. (193) demonstrated tic plaques in 40 hypercholesterolemic Japanese patients
that PPAR-gamma alone induced a significant regression who were randomized to receive either dose. They reported
in plaque but the largest regression was seen with PPAR- that 1-year 20 mg atorvastatin treatment induced marked
gamma and statin combination therapy. LDL-C reduction along with thoracic aortic plaques regres-
We have recently showed the good interscan and excel- sion, whereas it only retarded plaque progression in the ab-
lent intra- and interobserver reproducibility of black blood dominal aorta. In another study, Zhao et al. (202) exam-
contrast-enhanced MRI, suggesting that DCE-MRI could be ined carotid plaque composition by MRI quantitatively in
used in preclinical settings as a read-out for novel therapeu- eight patients with combined hyperlipidemia who had been
tic interventions for atherosclerosis (73). treated intensively with lovastatin, niacin, and colestipol for
10 years in comparison with an untreated, nonrandomized
In Humans control group of eight patients. In the control group, the
atherosclerotic lipid core comprised 17% of carotid intima–
The first point to be noted before mentioning therapeutical media cross-sectional area whereas only 1% of the cross-
interventions is that atherosclerosis develops naturally over sectional area belonged to the lipid core in the treated group.
time. Carotid and femoral arteries of 30 patients at high car- Beneficial effect of statin has also recently been evaluated in
diovascular risk have been studied during 2 years by MRI. It abdominal and thoracic aorta by MRI. It has been reported
showed that atherosclerotic burden increased significantly in that regression of aortic plaque was found on MRI after
both carotid and femoral arteries. However, carotid plaque 2 years of 20-mg atorvastatin (203), and more recently after
progression was associated with negative remodeling, 1 year of combination of statin with bisphosphonate (204)
whereas the increase in femoral plaque burden was compen- in the thoracic and the abdominal level. Abdominal aortic
sated by positive remodeling (194). plaques seem to be more resistant to statins than thoracic
Then, it is now clear that MRI has become a powerful plaques maybe because of the higher frequency of calcifica-
tool to quantify in vivo the degree of plaque regression fol- tion in the abdominal plaque. Lipid-lowering therapies may
lowing administration of different pharmacologic treatment also have effects on neovascularization. Dong et al. (79)
(195,196). We demonstrated that MR could be used in vivo with DCE-MRI showed decreased penetration of contrast
to measure the effect of lipid-lowering therapy (statins) in agent within carotid plaques after 1 year of intensive lipid-
asymptomatic untreated patients with hypercholesterolemia lowering treatment.
and carotid and aortic atherosclerosis (121). Atherosclerotic Taken together, these findings provide evidence that lipid-
plaques were visualized and measured with MR at different lowering therapy by statin may have a beneficial effect on
times after the initiation of lipid-lowering therapy. Significant plaque volume and composition, by reducing lipid-rich core
regression of atherosclerotic lesions was observed. Despite and local inflammation as assessed by noninvasive MRI.
the early and expected hypolipidemic effect of the statins, a HDL-cholesterol has demonstrated to be protective for
minimum of 12 months was needed to observe changes in the atherosclerosis in prospective studies. In the imaging field,
vessel wall. In fact, no changes were detected at 6 months. A MRI was able to detect accelerated atherogenesis in carotids
LWBK1209-ch27_p402-419.indd 413 17/05/13 5:12 PM

of patients with low HDL due to lecithin cholesterol acyl- TOF and weighted sequences or contrast-enhanced MRA
transferase gene mutations (205). Therefore, another way mask images. More standardized definition is needed as
to intervene on cholesterol to modify cardiovascular risk well as standardized method for quantitative assessment of
is to raise HDL-C. The current most promising treatments plaque dimension and plaque composition.
are CETP inhibitors. One of them, dalcetrapib, is currently CEMRI is limited by the risk of nephrogenic systemic
evaluated in randomized clinical trials and imaging studies fibrosis in patient with renal insufficiency. Eventually, an-
(206). In the dal-PLAQUE multicenter study, 130 patients other potential limitation of MRI is the limited sensitivity of
with, or with high risk of, coronary heart disease were ran- the technique in the detection of low concentration of tar-
domly assigned to dalcetrapib or placebo for 24 months. geted contrast agent, and therefore, much higher doses are
MRI-derived change in total vessel area was reduced in pa- needed as compared to the nuclear imaging techniques with
tients given dalcetrapib compared with those given placebo a probably higher toxicity.
after 24 months; absolute change from baseline relative to
placebo was −4.01 mm (2) (90% CI −7.23 to −0.80; nomi-
nal p = 0.04) (207,208). Future Challenges
Beyond the scope of cardiovascular drugs per se, the next
vast field of explorations will probably lie in chronic in- Noninvasive imaging of atherosclerosis by MRI is now fac-
flammatory diseases. Diseases like psoriasis, systemic lupus ing many challenges:
(209), rheumatoid arthritis (210), or even HIV (211) are
all associated with increased cardiovascular risk (212) and •• To improve techniques, magnets and antennas to improve
heavy atherosclerotic burden. These diseases have cardio- spatial and temporal resolution, and therefore, be able to
logic and arterial localizations and how new imaging mo- image rapid-moving vessels such as the coronaries.
dalities can be useful to their evaluation or treatment is now •• To develop new tools for functional imaging thus being
under investigation (213,214). Several authors have pointed able to approach in vivo plaque physiology like HDL
the potential role of MRI as a tool to monitor atherosclero- micelles (see Figure 27.7).
sis and arterial wall function during the course of these dis- •• To put in place standardized sequences for multicon-
eases or to assess the cardiovascular effects of their specific trast imaging and for each arterial bed and to develop
treatments (215,216). validated automated operator-independent software for
quantitative assessment of plaque dimension and com-
position.
Limitations •• To spread MRI into hospitals with MRI physicists able to
modify these imaging parameters and manage these formi-
MRI is limited by several clinical contraindications such dable imaging tool.
as presence of pacemaker or metallic implants and claus- •• To limit the high costs associated with this technique.
trophobia. Acquisition time for atherosclerosis MRI can be •• And finally to clarify the clinical implications of MRI
long and sometimes poorly tolerated. findings and translate them into therapeutical decisions.
Image quality may be poor because of motion artifact due
to patient movement, swallowing, or artery wall pulsation.
Despite technical improvement and the use of higher mag- Conclusion
netic fields, the main technical limitation of atherosclerotic
plaque imaging by MRI remains its temporal and spatial During the past 10 years, MRI as emerged as one of the (if
resolution, especially when it comes to image the coronary not “the”) most powerful tool to assess atherosclerosis in
circulation. vivo.
Identification of the different plaque components re- It is noninvasive and does not require contrast agent in-
quired analyzing multiple MRI sequences simultaneously. jection. It does not expose to radiation which becomes a real
However, standardized protocol is lacking and limits ath- issue in conventional cardiology with the large development
erosclerosis imaging by MRI in clinical practice. Another of CT. It now has adequate spatial resolution for large ves-
obstacle for the implementation of MRI as a standard tool sel assessment. It is reproducible and enables serial imaging
in the evaluation of atherosclerotic plaques is the interob- over time. The drawbacks are limited temporal resolution
server variability in the interpretation of the different im- which hampered coronary arteries imaging and the lack of
aging features. The reproducibility of MRI for identifica- generally accepted image acquisition protocols. MRI has
tion of atherosclerotic plaque features is variable depending now been validated in clinical practice for vessel wall dimen-
on the study and on the plaque component observed. The sions’ assessment and for qualitative assessment of plaque
intra- and interobserver agreement is good to excellent for composition. The exponential development of new imaging
calcifications and hemorrhage and moderate to excellent for tools will certainly break those barriers in the coming years.
FC identification. This can be explained by different crite- Allowing both anatomical and functional imaging, MRI
ria used to determine the same feature. Several sequences is playing a key role in research and in routine clinic when it
and combinations have been proposed for studying the FC will be widely available and standardized.
or the intraplaque hemorrhage. FC can be assessed using MRI enables us to better understand atherosclerosis
3D TOF, 3D multiple overlapping thin-slab acquisition (3D mechanisms by reaching the very heart of the plaques with
MOTSA) or T1-weighted images before and after gadolin- all their components as well as inflammation, neovascular-
ium. Intraplaque hemorrhage (IPH) can be assessed using ization, and wall stress.
LWBK1209-ch27_p402-419.indd 414 17/05/13 5:12 PM

Figure 27.7. The figure shown is a schematic depiction of a discoidal high-density lipoprotein agent with
gadolinium–lipid incorporated in the phospholipid coating. Representative MRI scans of the aorta of apolipo-
protein E-knockout mice are shown preinjection of a high-density lipoprotein agent (top panel) and 24-hour
postinjection (bottom panel). (Panels are modified from Cormode DP, Chandrasekar R, Delshad A, et al.
Comparison of synthetic high-density lipoprotein (HDL) contrast agents for MR imaging of atherosclerosis.
Bioconjug Chem. 2009; 20(5):937–943.)
In preventive cardiology, especially, where identifying References

prone to rupture plaques will help individual risk stratifi-
1. Libby P. Molecular bases of the acute coronary syndromes. Circulation. 1995;91:2844–2850.
cation and management. The BioImage Study within the 2. Fuster V, Fayad ZA, Badimon JJ. Acute coronary syndromes: Biology. Lancet. 1999;353
High-risk Plaque Initiative (217) is a joint research and de- (suppl 2):SII5–SII9.
velopment effort designed to evaluate these plaque imaging 3. Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation. 1995;92:657–671.
4. Fuster V, Lewis A. Conner Memorial Lecture. Mechanisms leading to myocardial infarc-
driven strategies for risk stratification and to determine its tion: Insights from studies of vascular biology. Circulation. 1994;90:2126–2146.
cost-effectiveness. 5. Moreno PR, Purushothaman KR, Fuster V, et al. Intimomedial interface damage and
adventitial inflammation is increased beneath disrupted atherosclerosis in the aorta:
The application of MRI opens up entirely new areas Implications for plaque vulnerability. Circulation. 2002;105:2504–2511.
for the diagnosis, prevention, and treatment (e.g., lipid- 6. Ware JA. Too many vessels? Not enough? The wrong kind? The VEGF debate continues.
lowering drug regimens) of atherosclerosis in all arterial Nat Med. 2001;7:403–404.
7. Ambrose JA, Tannenbaum MA, Alexopoulos D, et al. Angiographic progression of coro-
locations. nary artery disease and the development of myocardial infarction. J Am Coll Cardiol. 1988;
12:56–62.
8. Little WC, Constantinescu M, Applegate RJ, et al. Can coronary angiography predict the
site of a subsequent myocardial infarction in patients with mild-to-moderate coronary
Acknowledgments artery disease? Circulation. 1988;78:1157–1166.
9. Carr S, Farb A, Pearce WH, et al. Atherosclerotic plaque rupture in symptomatic carotid
artery stenosis. J Vasc Surg. 1996;23:755–765; discussion 765–766.
This work was supported in part by NIH/NHLBI R01 10. Underhill HR, Yuan C, Yarnykh VL, et al. Predictors of surface disruption with MR imag-
ing in asymptomatic carotid artery stenosis. AJNR Am J Neuroradiol. 2010;31:487–493.
HL071021, R01 HL078667; NIH/NBIB R01 EB009638; 11. Takaya N, Yuan C, Chu B, et al. Presence of intraplaque hemorrhage stimulates progres-
NIH/NHLBI Program of Excellence in Nanotechnology sion of carotid atherosclerotic plaques: A high-resolution magnetic resonance imaging
(PEN) Award, Contract #HHSN268201000045C; and study. Circulation. 2005;111:2768–2775.
12. Richardson PD, Davies MJ, Born GV. Influence of plaque configuration and stress distri-
NIH/NCRR CTSA UL1RR029887 (Imaging Core), and by bution on fissuring of coronary atherosclerotic plaques. Lancet. 1989;2:941–944.
funds from the Mount Sinai Cardiovascular Institute and 13. Toschi V, Gallo R, Lettino M, et al. Tissue factor modulates the thrombogenicity of
the Department of Radiology. David Rosenbaum’s work human atherosclerotic plaques. Circulation. 1997;95:594–599.
14. Nakagomi A, Celermajer DS, Lumley T, et al. Angiographic severity of coronary narrowing
was partially supported by the Fédération Francaise de is a surrogate marker for the extent of coronary atherosclerosis. Am J Cardiol. 1996;78:
Cardiologie, Paris, France. 516–519.
LWBK1209-ch27_p402-419.indd 415 17/05/13 5:12 PM

15. Moreno PR, Falk E, Palacios IF, et al. Macrophage infiltration in acute coronary syn- 51. Totterman S, Francis CW, Foster TH, et al. Diagnosis of femoropopliteal venous thrombo-
dromes. Implications for plaque rupture. Circulation. 1994;90:775–778. sis with MR imaging: A comparison of four MR pulse sequences. AJR Am J Roentgenol.
16. Kaartinen M, Penttila A, Kovanen PT. Accumulation of activated mast cells in the shoul- 1990;154:175–178.
der region of human coronary atheroma, the predilection site of atheromatous rupture. 52. Erdman WA, Jayson HT, Redman HC, et al. Deep venous thrombosis of extremities: Role
Circulation. 1994;90:1669–1678. of MR imaging in the diagnosis. Radiology. 1990;174:425–431.
17. Glagov S, Zarins C, Giddens DP, et al. Hemodynamics and atherosclerosis. Insights and per- 53. Bryant RG, Marill K, Blackmore C, et al. Magnetic relaxation in blood and blood clots.
spectives gained from studies of human arteries. Arch Pathol Lab Med. 1988;112:1018–1031. Magn Reson Med. 1990;13:133–144.
18. Toussaint JF, LaMuraglia GM, Southern JF, et al. Magnetic resonance images lipid, 54. Bass JC, Hedlund LW, Sostman HD. MR imaging of experimental and clinical thrombi at
fibrous, calcified, hemorrhagic, and thrombotic components of human atherosclerosis in 1.5 T. Magn Reson Imaging. 1990;8:631–635.
vivo. Circulation. 1996;94:932–938. 55. Rapoport S, Sostman HD, Pope C, et al. Venous clots: Evaluation with MR imaging.
19. Solberg LA, Strong JP. Risk factors and atherosclerotic lesions. A review of autopsy stud- Radiology. 1987;162:527–530.
ies. Arteriosclerosis. 1983;3:187–198. 56. Hatsukami TS, Ross R, Polissar NL, et al. Visualization of fibrous cap thickness and
20. Fazio GP, Redberg RF, Winslow T, et al. Transesophageal echocardiographically detected rupture in human atherosclerotic carotid plaque in vivo with high-resolution magnetic
atherosclerotic aortic plaque is a marker for coronary artery disease. J Am Coll Cardiol. resonance imaging. Circulation. 2000;102:959–964.
1993;21:144–150. 57. Mitsumori LM, Hatsukami TS, Ferguson MS, et al. In vivo accuracy of multisequence
21. Cohen A, Tzourio C, Bertrand B, et al. Aortic plaque morphology and vascular events: MR imaging for identifying unstable fibrous caps in advanced human carotid plaques. J
A follow-up study in patients with ischemic stroke. FAPS investigators. French Study of Magn Reson Imaging. 2003;17:410–420.
Aortic Plaques in Stroke. Circulation. 1997;96:3838–3841. 58. Moody AR. Magnetic resonance direct thrombus imaging. J Thromb Haemost. 2003;1:
22. Fayad ZA, Fuster V. The human high-risk plaque and its detection by magnetic resonance 1403–1409.
imaging. Am J Cardiol. 2001;88:42E–45E. 59. Cappendijk VC, Cleutjens KB, Heeneman S, et al. In vivo detection of hemorrhage in
23. Song HK, Wright AC, Wolf RL, et al. Multislice double inversion pulse sequence for human atherosclerotic plaques with magnetic resonance imaging. J Magn Reson Imaging.
efficient black-blood MRI. Magn Reson Med. 2002;47:616–620. 2004;20:105–110.
24. Mani V, Itskovich VV, Szimtenings M, et al. Rapid extended coverage simultaneous mul- 60. Luo Y, Polissar N, Han C, et al. Accuracy and uniqueness of three in vivo measurements
tisection black-blood vessel wall MR imaging. Radiology. 2004;232:281–288. of atherosclerotic carotid plaque morphology with black blood MRI. Magn Reson Med.
25. Sirol M, Itskovich VV, Mani V, et al. Lipid-rich atherosclerotic plaques detected by gado- 2003;50:75–82.
fluorine-enhanced in vivo magnetic resonance imaging. Circulation. 2004;109:2890–2896. 61. Fayad ZA, Connick TJ, Axel L. An improved quadrature or phased-array coil for MR
26. Itskovich VV, Mani V, Mizsei G, et al. Parallel and nonparallel simultaneous multislice cardiac imaging. Magn Reson Med. 1995;34:186–193.
black-blood double inversion recovery techniques for vessel wall imaging. J Magn Reson 62. Hayes JK, Collette DJ, Peters JL, et al. Monitoring body-core temperature from the tra-
Imaging. 2004;19:459–467. chea: Comparison between pulmonary artery, tympanic, esophageal, and rectal tempera-
27. Yarnykh VL, Yuan C. Multislice double inversion-recovery black-blood imaging with tures. J Clin Monit. 1996;12:261–269.
simultaneous slice reinversion. J Magn Reson Imaging. 2003;17:478–483. 63. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human coronary artery
28. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially lumen and wall imaging using black-blood magnetic resonance imaging. Circulation.
parallel acquisitions (GRAPPA). Magn Reson Med. 2002;47:1202–1210. 2000;102:506–510.
29. Vignaux OB, Augui J, Coste J, et al. Comparison of single-shot fast spin-echo and con- 64. Cai JM, Hatsukami TS, Ferguson MS, et al. Classification of human carotid atherosclerotic
ventional spin-echo sequences for MR imaging of the heart: Initial experience. Radiology. lesions with in vivo multicontrast magnetic resonance imaging. Circulation. 2002;106:
2001;219:545–550. 1368–1373.
30. Simonetti OP, Finn JP, White RD, et al. “Black blood” T2-weighted inversion-recovery 65. Chu B, Hatsukami TS, Polissar NL, et al. Determination of carotid artery atherosclerotic
MR imaging of the heart. Radiology. 1996;199:49–57. lesion type and distribution in hypercholesterolemic patients with moderate carotid steno-
31. Nishimura K, Hida S, Nishio Y, et al. The validity of magnetic resonance imaging (MRI) sis using noninvasive magnetic resonance imaging. Stroke. 2004;35:2444–2448.
in the staging of bladder cancer: Comparison with computed tomography (CT) and trans- 66. Chu B, Kampschulte A, Ferguson MS, et al. Hemorrhage in the atherosclerotic carotid
urethral ultrasonography (US). Jpn J Clin Oncol. 1988;18:217–226. plaque: A high-resolution MRI study. Stroke. 2004;35:1079–1084.
32. Parker DL, Goodrich KC, Masiker M, et al. Improved efficiency in double-inversion fast 67. Zhang L, Zalewski A, Liu Y, et al. Diabetes-induced oxidative stress and low-grade
spin-echo imaging. Magn Reson Med. 2002;47:1017–1021. inflammation in porcine coronary arteries. Circulation. 2003;108:472–478.
33. Koktzoglou I, Chung YC, Mani V, et al. Multislice dark-blood carotid artery wall imag- 68. Saam T, Cai JM, Cai YQ, et al. Carotid plaque composition differs between ethno-racial
ing: A 1.5 T and 3.0 T comparison. J Magn Reson Imaging. 2006;23:699–705. groups: An MRI pilot study comparing mainland Chinese and American Caucasian
34. Koktzoglou I, Li D. Diffusion-prepared segmented steady-state free precession: patients. Arterioscler Thromb Vasc Biol. 2005;25:611–616.
Application to 3D black-blood cardiovascular magnetic resonance of the thoracic aorta 69. Saam T, Kerwin WS, Chu B, et al. Sample size calculation for clinical trials using magnetic
and carotid artery walls. J Cardiovasc Magn Reson. 2007;9:33–42. resonance imaging for the quantitative assessment of carotid atherosclerosis. J Cardiovasc
35. Hayashi K, Mani V, Nemade A, et al. Comparison of 3D-diffusion-prepared segmented Magn Reson. 2005;7:799–808.
steady-state free precession and 2D fast spin echo imaging of femoral artery atherosclero- 70. Yuan YV, Kitts DD. Dietary fat source and cholesterol interactions alter plasma lipids and
sis. J Cardiovasc Imaging. 2010;26:309–321. tissue susceptibility to oxidation in spontaneously hypertensive (SHR) and normotensive
36. Mihai G, Chung YC, Merchant A, et al. T1-weighted-SPACE dark blood whole body Wistar Kyoto (WKY) rats. Mol Cell Biochem. 2002;232:33–47.
magnetic resonance angiography (DB-WBMRA): Initial experience. Magn Reson Imaging. 71. Yuan C, Kerwin WS, Ferguson MS, et al. Contrast-enhanced high resolution MRI for ath-
2010;31:502–509. erosclerotic carotid artery tissue characterization. J Magn Reson Imaging. 2002;15:62–67.
37. Mihai G, Winner MW, Raman SV, et al. Assessment of carotid stenosis using three- 72. Kampschulte A, Ferguson MS, Kerwin WS, et al. Differentiation of intraplaque versus
dimensional T2-weighted dark blood imaging: Initial experience. J Magn Reson Imaging. juxtaluminal hemorrhage/thrombus in advanced human carotid atherosclerotic lesions by
2012;35:449–455. in vivo magnetic resonance imaging. Circulation. 2004;110:3239–3244.
38. Liu W, Balu N, Sun J, et al. Segmentation of carotid plaque using multicontrast 3D gradi- 73. Calcagno C, Vucic E, Mani V, et al. Reproducibility of black blood dynamic contrast-
ent echo MRI. J Magn Reson Imaging. 2012;35:812–819. enhanced magnetic resonance imaging in aortic plaques of atherosclerotic rabbits. J Magn
39. Chiu B, Sun J, Zhao X, et al. Fast plaque burden assessment of the femoral artery using Reson Imaging. 2010;32:191–198.
3D black-blood MRI and automated segmentation. Med Phy. 2011;38:5370–5384. 74. Calcagno C, Cornily JC, Hyafil F, et al. Detection of neovessels in atherosclerotic plaques
40. Worthley SG, Helft G, Fuster V, et al. High resolution ex vivo magnetic resonance imaging of rabbits using dynamic contrast enhanced MRI and 18F-FDG PET. Arterioscler Thromb
of in situ coronary and aortic atherosclerotic plaque in a porcine model. Atherosclerosis. Vasc Biol. 2008;28:1311–1317.
2000;150:321–329. 75. Cornily JC, Hyafil F, Calcagno C, et al. Evaluation of neovessels in atherosclerotic plaques
41. Toussaint JF, Southern JF, Fuster V, et al. T2-weighted contrast for NMR characterization of rabbits using an albumin-binding intravascular contrast agent and MRI. J Magn Reson
of human atherosclerosis. Arterioscler Thromb Vasc Biol. 1995;15:1533–1542. Imaging. 2008;27:1406–1411.
42. Edzes HT, Samulski ET. Cross relaxation and spin diffusion in the proton NMR or 76. Calcagno C, Mani V, Ramachandran S, et al. Dynamic contrast enhanced (DCE) mag-
hydrated collagen. Nature. 1977;265:521–523. netic resonance imaging (MRI) of atherosclerotic plaque angiogenesis. Angiogenesis.
43. Witztum JL, Steinberg D. Role of oxidized low density lipoprotein in atherogenesis. J Clin 2010;13:87–99.
Invest. 1991;88:1785–1792. 77. Lobatto ME, Calcagno C, Metselaar JM, et al. Imaging the efficacy of anti-inflammatory
44. Fisel CR, Ackerman JL, Buxton RB, et al. MR contrast due to microscopically heteroge- liposomes in a rabbit model of atherosclerosis by non-invasive imaging. Methods
neous magnetic susceptibility: Numerical simulations and applications to cerebral physi- Enzymol. 2012;508:211–228.
ology. Magn Reson Med. 1991;17:336–347. 78. Vucic E, Dickson SD, Calcagno C, et al. Pioglitazone modulates vascular inflammation in
45. Rapp JH, Connor WE, Lin DS, et al. Lipids of human atherosclerotic plaques and xantho- atherosclerotic rabbits noninvasive assessment with FDG-PET-CT and dynamic contrast-
mas: Clues to the mechanism of plaque progression. J Lipid Res. 1983;24:1329–1335. enhanced MR imaging. JACC Cardiovasc Imaging. 2011;4:1100–1109.
46. Kucharczyk W, Henkelman RM. Visibility of calcium on MR and CT: Can MR show 79. Dong L, Kerwin WS, Chen H, et al. Carotid artery atherosclerosis: Effect of intensive lipid
calcium that CT cannot? AJNR Am J Neuroradiol. 1994;15:1145–1148. therapy on the vasa vasorum—evaluation by using dynamic contrast-enhanced MR imag-
47. Bradley WG Jr. MR appearance of hemorrhage in the brain. Radiology. 1993;189:15–26. ing. Radiology. 2011;260:224–231.
48. Yamashita Y, Hatanaka Y, Torashima M, et al. Magnetic resonance characteristics of 80. Lipinski MJ, Fuster V, Fisher EA, et al. Technology insight: Targeting of biological mol-
intrapelvic haematomas. Br J Radiol. 1995;68:979–985. ecules for evaluation of high-risk atherosclerotic plaques with magnetic resonance imag-
49. Bluemke DA. Definitive diagnosis of intramural hematoma of the thoracic aorta with MR ing. Nat Clin Pract Cardiovasc Med. 2004;1:48–55.
imaging. Radiology. 1997;204:319–321. 81. Kang HW, Josephson L, Petrovsky A, et al. Magnetic resonance imaging of inducible
50. Murray JG, Manisali M, Flamm SD, et al. Intramural hematoma of the thoracic aorta: E-selectin expression in human endothelial cell culture. Bioconjug Chem. 2002;13:
MR image findings and their prognostic implications. Radiology. 1997;204:349–355. 122–127.
LWBK1209-ch27_p402-419.indd 416 17/05/13 5:12 PM

82. Kerwin W, Hooker A, Spilker M, et al. Quantitative magnetic resonance imaging analysis 114. Kramer CM, Cerilli LA, Hagspiel K, et al. Magnetic resonance imaging identifies the fibrous
of neovasculature volume in carotid atherosclerotic plaque. Circulation. 2003;107:851–856. cap in atherosclerotic abdominal aortic aneurysm. Circulation. 2004;109:1016–1021.
83. Anderson SA, Rader RK, Westlin WF, et al. Magnetic resonance contrast enhance- 115. Yuan C, Mitsumori LM, Ferguson MS, et al. In vivo accuracy of multispectral magnetic
ment of neovasculature with alpha(v)beta(3)-targeted nanoparticles. Magn Reson Med. resonance imaging for identifying lipid-rich necrotic cores and intraplaque hemorrhage in
2000;44:433–439. advanced human carotid plaques. Circulation. 2001;104:2051–2056.
84. Winter PM, Morawski AM, Caruthers SD, et al. Molecular imaging of angiogen- 116. Yuan C, Beach KW, Smith LH Jr, et al. Measurement of atherosclerotic carotid plaque size in
esis in early-stage atherosclerosis with alpha(v)beta3-integrin-targeted nanoparticles. vivo using high resolution magnetic resonance imaging. Circulation. 1998;98:2666–2671.
Circulation. 2003;108:2270–2274. 117. Toussaint D, Massager N, Mathy L, et al. [Primary malignant lymphoma of the spine:
85. Botnar RM, Perez AS, Witte S, et al. In vivo molecular imaging of acute and subacute Clinical aspects and therapeutic advances]. Acta Orthop Belg. 1999;65:23–32.
thrombosis using a fibrin-binding magnetic resonance imaging contrast agent. Circulation. 118. Yuan C, Mitsumori LM, Beach KW, et al. Carotid atherosclerotic plaque: Noninvasive MR
2004;109:2023–2029. characterization and identification of vulnerable lesions. Radiology. 2001;221:285–299.
86. Johansson LO, Bjornerud A, Ahlstrom HK, et al. A targeted contrast agent for mag- 119. Yuan C, Zhao XQ, Hatsukami TS. Quantitative evaluation of carotid atherosclerotic
netic resonance imaging of thrombus: Implications of spatial resolution. J Magn Reson plaques by magnetic resonance imaging. Curr Atheroscler Rep. 2002;4:351–357.
Imaging. 2001;13:615–618. 120. Wang D, Shen W, Tian Y, et al. [Protective effect of active components extracted from
87. Ruehm SG, Corot C, Vogt P, et al. Magnetic resonance imaging of atherosclerotic plaque radix Astragali on human erythrocyte membrane damages caused by reactive oxygen spe-
with ultrasmall superparamagnetic particles of iron oxide in hyperlipidemic rabbits. cies]. Zhongguo Zhong Yao Za Zhi. 1996;21:746–748, 763.
Circulation. 2001;103:415–422. 121. Corti R, Fayad ZA, Fuster V, et al. Effects of lipid-lowering by simvastatin on human
88. Kooi ME, Cappendijk VC, Cleutjens KB, et al. Accumulation of ultrasmall superparamag- atherosclerotic lesions: A longitudinal study by high-resolution, noninvasive magnetic
netic particles of iron oxide in human atherosclerotic plaques can be detected by in vivo resonance imaging. Circulation. 2001;104:249–252.
magnetic resonance imaging. Circulation. 2003;107:2453–2458. 122. Fayad ZA, Nahar T, Fallon JT, et al. In vivo magnetic resonance evaluation of atheroscle-
89. Trivedi RA, U-King-Im JM, Graves MJ, et al. MRI-derived measurements of fibrous-cap rotic plaques in the human thoracic aorta: A comparison with transesophageal echocar-
and lipid-core thickness: The potential for identifying vulnerable carotid plaques in vivo. diography. Circulation. 2000;101:2503–2509.
Neuroradiology. 2004;46:738–743. 123. Skinner MP, Yuan C, Mitsumori L, et al. Serial magnetic resonance imaging of experimen-
90. Falk E. Plaque rupture with severe pre-existing stenosis precipitating coronary thrombo- tal atherosclerosis detects lesion fine structure, progression and complications in vivo. Nat
sis. Characteristics of coronary atherosclerotic plaques underlying fatal occlusive thrombi. Med. 1995;1:69–73.
Br Heart J. 1983;50:127–134. 124. Fayad ZA, Fallon JT, Shinnar M, et al. Noninvasive in vivo high-resolution magnetic
91. Ross R. Atherosclerosis—an inflammatory disease. N Engl J Med. 1999;340:115–126. resonance imaging of atherosclerotic lesions in genetically engineered mice. Circulation.
92. Winter PM, Caruthers SD, Yu X, et al. Improved molecular imaging contrast agent for 1998;98:1541–1517.
detection of human thrombus. Magn Reson Med. 2003;50:411–416. 125. Lin W, Abendschein DR, Haacke EM. Contrast-enhanced magnetic resonance angiogra-
93. Tsimikas S, Palinski W, Halpern SE, et al. Radiolabeled MDA2, an oxidation-specific, phy of carotid arterial wall in pigs. J Magn Reson Imaging. 1997;7:183–190.
monoclonal antibody, identifies native atherosclerotic lesions in vivo. J Nucl Cardiol. 1999; 126. Chandra NC, Spiro MJ, Spiro RG. Identification of a glycoprotein from rat liver mito-
6:41–53. chondrial inner membrane and demonstration of its origin in the endoplasmic reticulum.
94. Tsimikas S. Noninvasive imaging of oxidized low-density lipoprotein in atherosclerotic J Biol Chem. 1998;273:19715–19721.
plaques with tagged oxidation-specific antibodies. Am J Cardiol. 2002;90:22L–27L. 127. Kaneko CR. Eye movement deficits after ibotenic acid lesions of the nucleus prepositus
95. Tsimikas S, Shortal BP, Witztum JL, et al. In vivo uptake of radiolabeled MDA2, an oxidation- hypoglossi in monkeys. I. Saccades and fixation. J Neurophysiol. 1997;78:1753–1768.
specific monoclonal antibody, provides an accurate measure of atherosclerotic lesions rich 128. Helft G, Worthley SG, Fuster V, et al. Atherosclerotic aortic component quantification by
in oxidized LDL and is highly sensitive to their regression. Arterioscler Thromb Vasc Biol. noninvasive magnetic resonance imaging: An in vivo study in rabbits. J Am Coll Cardiol.
2000;20:689–697. 2001;37:1149–1154.
96. Briley-Saebo KC, Nguyen TH, Saeboe AM, et al. In vivo detection of oxidation-specific 129. Helft G, Worthley SG, Fuster V, et al. Progression and regression of atherosclerotic lesions:
epitopes in atherosclerotic lesions using biocompatible manganese molecular magnetic Monitoring with serial noninvasive magnetic resonance imaging. Circulation. 2002;105:
imaging probes. J Am Coll Cardiol. 2012;59:616–626. 993–998.
97. Briley-Saebo KC, Cho YS, Shaw PX, et al. Targeted iron oxide particles for in vivo mag- 130. McConnell MV, Aikawa M, Maier SE, et al. MRI of rabbit atherosclerosis in response to
netic resonance detection of atherosclerotic lesions with antibodies directed to oxidation- dietary cholesterol lowering. Arterioscler Thromb Vasc Biol. 1999;19:1956–1959.
specific epitopes. J Am Coll Cardiol. 2011;57:337–347. 131. Choudhury RP, Aguinaldo JG, Rong JX, et al. Atherosclerotic lesions in genetically modi-
98. Lindner JR, Song J, Christiansen J, et al. Ultrasound assessment of inflammation and renal fied mice quantified in vivo by non-invasive high-resolution magnetic resonance micros-
tissue injury with microbubbles targeted to P-selectin. Circulation. 2001;104:2107–2112. copy. Atherosclerosis. 2002;162:315–321.
99. Villanueva FS, Jankowski RJ, Klibanov S, et al. Microbubbles targeted to intercellular 132. Degnan AJ, Young VE, Gillard JH. Advances in noninvasive imaging for evaluating
adhesion molecule-1 bind to activated coronary artery endothelial cells. Circulation. clinical risk and guiding therapy in carotid atherosclerosis. Expert Rev Cardiovasc Ther.
1998;98:1–5. 2012;10:37–53.
100. Hamilton AJ, Huang SL, Warnick D, et al. Intravascular ultrasound molecular imaging of 133. Corti R, Fuster V. Imaging of atherosclerosis: Magnetic resonance imaging. Eur Heart J.
atheroma components in vivo. J Am Coll Cardiol. 2004;43:453–460. 2011;32:1709–1719b.
101. Barkhausen J, Ebert W, Heyer C, et al. Detection of atherosclerotic plaque with 134. Maynor CH, Charles HC, Herfkens RJ, et al. Chemical shift imaging of atherosclerosis at
Gadofluorine-enhanced magnetic resonance imaging. Circulation. 2003;108:605–609. 7.0 Tesla. Invest Radiol. 1989;24:52–60.
102. Flacke S, Fischer S, Scott MJ, et al. Novel MRI contrast agent for molecular imaging of 135. Toussaint JF, Southern JF, Fuster V, et al. 13C-NMR spectroscopy of human atheroscle-
fibrin: Implications for detecting vulnerable plaques. Circulation. 2001;104:1280–1285. rotic lesions. Relation between fatty acid saturation, cholesteryl ester content, and luminal
103. Schmitz SA, Taupitz M, Wagner S, et al. Magnetic resonance imaging of atherosclerotic obstruction. Arterioscler Thromb. 1994;14:1951–1957.
plaques using superparamagnetic iron oxide particles. J Magn Reson Imaging. 2001;14: 136. Altbach MI, Mattingly MA, Brown MF, et al. Magnetic resonance imaging of lipid depos-
355–361. its in human atheroma via a stimulated-echo diffusion-weighted technique. Magn Reson
104. Johnstone MT, Botnar RM, Perez AS, et al. In vivo magnetic resonance imaging of experi- Med. 1991;20:319–326.
mental thrombosis in a rabbit model. Arterioscler Thromb Vasc Biol. 2001;21:1556–1560. 137. Gold GE, Pauly JM, Glover GH, et al. Characterization of atherosclerosis with a 1.5-T
105. Yu X, Song SK, Chen J, et al. High-resolution MRI characterization of human thrombus imaging system. J Magn Reson Imaging. 1993;3:399–407.
using a novel fibrin-targeted paramagnetic nanoparticle contrast agent. Magn Reson Med. 138. Vinitski S, Consigny PM, Shapiro MJ, et al. Magnetic resonance chemical shift imaging
2000;44:867–872. and spectroscopy of atherosclerotic plaque. Invest Radiol. 1991;26:703–714.
106. Botnar RM, Buecker A, Wiethoff AJ, et al. In vivo magnetic resonance imaging of cor- 139. Mohiaddin RH, Firmin DN, Underwood SR, et al. Chemical shift magnetic resonance
onary thrombosis using a fibrin-binding molecular magnetic resonance contrast agent. imaging of human atheroma. Br Heart J. 1989;62:81–89.
Circulation. 2004;110:1463–1466. 140. Pearlman JD, Zajicek J, Merickel MB, et al. High-resolution 1H NMR spectral signature
107. Sirol M, Aguinaldo JG, Graham PB, et al. Fibrin-targeted contrast agent for improvement from human atheroma. Magn Reson Med. 1988;7:262–279.
of in vivo acute thrombus detection with magnetic resonance imaging. Atherosclerosis. 141. Soila K, Nummi P, Ekfors T, et al. Proton relaxation times in arterial wall and atheroma-
2005;182:79–85. tous lesions in man. Invest Radiol. 1986;21:411–415.
108. European Carotid Surgery Trialists Collaborative Group. Randomised trial of endarter- 142. Cai W, Chung AC. Multi-resolution vessel segmentation using normalized cuts in retinal
ectomy for recently symptomatic carotid stenosis: Final results of the MRC European images. Med Image Comput Comput Assist Interv. 2006;9:928–936.
Carotid Surgery Trial (ECST). Lancet. 1998;351:1379–1387. 143. Moody AR, Murphy RE, Morgan PS, et al. Characterization of complicated carotid
109. Kullo IJ, Edwards WD, Schwartz RS. Vulnerable plaque: Pathobiology and clinical impli- plaque with magnetic resonance direct thrombus imaging in patients with cerebral isch-
cations. Ann Intern Med. 1998;129:1050–1060. emia. Circulation. 2003;107:3047–3052.
110. Zheng J, Abendschein DR, Okamoto RJ, et al. MRI-based biomechanical imaging: Initial 144. Hayes CE, Mathis CM, Yuan C. Surface coil phased arrays for high-resolution imaging of
study on early plaque progression and vessel remodeling. Magn Reson Imaging. 2009; the carotid arteries. J Magn Reson Imaging. 1996;6:109–112.
27:1309–1318. 145. Saam T, Ferguson MS, Yarnykh VL, et al. Quantitative evaluation of carotid plaque com-
111. Seeger JM, Barratt E, Lawson GA, et al. The relationship between carotid plaque compo- position by in vivo MRI. Arterioscler Thromb Vasc Biol. 2005;25:234–239.
sition, plaque morphology, and neurologic symptoms. J Surg Res. 1995;58:330–336. 146. Zhang S, Cai J, Luo Y, et al. Measurement of carotid wall volume and maximum area
112. von Ingersleben G, Schmiedl UP, Hatsukami TS, et al. Characterization of atherosclerotic with contrast-enhanced 3D MR imaging: Initial observations. Radiology. 2003;228:
plaques at the carotid bifurcation: Correlation of high-resolution MR imaging with histo- 200–205.
logic analysis—preliminary study. Radiographics. 1997;17:1417–1423. 147. Yuan C, Zhang SX, Polissar NL, et al. Identification of fibrous cap rupture with magnetic
113. Kerwin WS, Canton G. Advanced techniques for MRI of atherosclerotic plaque. Top resonance imaging is highly associated with recent transient ischemic attack or stroke.
Magn Reson Imaging. 2009;20:217–225. Circulation. 2002;105:181–185.
LWBK1209-ch27_p402-419.indd 417 17/05/13 5:12 PM

148. Imparato AM, Riles TS, Gorstein F. The carotid bifurcation plaque: Pathologic findings 178. Tang D, Yang C, Kobayashi S, et al. 3D MRI-based anisotropic FSI models with cyclic
associated with cerebral ischemia. Stroke. 1979;10:238–245. bending for human coronary atherosclerotic plaque mechanical analysis. J Biomech Eng.
149. Moody AR, Allder S, Lennox G, et al. Direct magnetic resonance imaging of carotid 2009;131:061010.
artery thrombus in acute stroke. Lancet. 1999;353:122–123. 179. Kawasaki T, Koga S, Koga N, et al. Characterization of hyperintense plaque with noncon-
150. Takaya N, Yuan C, Chu B, et al. Association between carotid plaque characteristics and trast T(1)-weighted cardiac magnetic resonance coronary plaque imaging: Comparison
subsequent ischemic cerebrovascular events: A prospective assessment with MRI—initial with multislice computed tomography and intravascular ultrasound. JACC Cardiovasc
results. Stroke. 2006;37:818–823. Imaging. 2009;2:720–728.
151. Murphy RE, Moody AR, Morgan PS, et al. Prevalence of complicated carotid atheroma as 180. Wang Y, Winchester PA, Khilnani NM, et al. Contrast-enhanced peripheral MR angiogra-
detected by magnetic resonance direct thrombus imaging in patients with suspected carotid phy from the abdominal aorta to the pedal arteries: Combined dynamic two-dimensional
artery stenosis and previous acute cerebral ischemia. Circulation. 2003;107:3053–3058. and bolus-chase three-dimensional acquisitions. Invest Radiol. 2001;36:170–177.
152. Saam T, Underhill HR, Chu B, et al. Prevalence of American Heart Association type VI 181. Binkert CA, Baker PD, Petersen BD, et al. Peripheral vascular disease: Blinded study of
carotid atherosclerotic lesions identified by magnetic resonance imaging for different levels dedicated calf MR angiography versus standard bolus-chase MR angiography and film
of stenosis as measured by duplex ultrasound. J Am Coll Cardiol. 2008;51:1014–1021. hard-copy angiography. Radiology. 2004;232:860–866.
153. Parmar JP, Rogers WJ, Mugler JP 3rd, et al. Magnetic resonance imaging of carotid ath- 182. Morasch MD, Collins J, Pereles FS, et al. Lower extremity stepping-table magnetic reso-
erosclerotic plaque in clinically suspected acute transient ischemic attack and acute isch- nance angiography with multilevel contrast timing and segmented contrast infusion. J
emic stroke. Circulation. 2010;122:2031–2038. Vasc Surg. 2003;37:62–71.
154. Noguchi T, Yamada N, Higashi M, et al. High-intensity signals in carotid plaques on 183. Pereles FS, Collins JD, Carr JC, et al. Accuracy of stepping-table lower extremity MR
T1-weighted magnetic resonance imaging predict coronary events in patients with coro- angiography with dual-level bolus timing and separate calf acquisition: Hybrid peripheral
nary artery disease. J Am Coll Cardiol. 2011;58:416–422. MR angiography. Radiology. 2006;240:283–290.
155. Underhill HR, Yuan C, Yarnykh VL, et al. Arterial remodeling in [corrected] subclinical 184. Deutschmann HA, Schoellnast H, Portugaller HR, et al. Routine use of three-dimensional
carotid artery disease. JACC Cardiovasc Imaging. 2009;2:1381–1389. contrast-enhanced moving-table MR angiography in patients with peripheral arterial
156. Sluimer JC, Kolodgie FD, Bijnens AP, et al. Thin-walled microvessels in human coro- occlusive disease: Comparison with selective digital subtraction angiography. Cardiovasc
nary atherosclerotic plaques show incomplete endothelial junctions relevance of com- Intervent Radiol. 2006;29:762–770.
promised structural integrity for intraplaque microvascular leakage. J Am Coll Cardiol. 185. Owen AR, Robertson IR, Annamalai G, et al. Critical lower-limb ischemia: The diagnos-
2009;53:1517–1527. tic performance of dual-phase injection MR angiography (including high-resolution dis-
157. Kerwin WS, Oikawa M, Yuan C, et al. MR imaging of adventitial vasa vasorum in carotid tal imaging) compared with digital subtraction angiography. J Vasc Interv Radiol. 2009;
atherosclerosis. Magn Reson Med. 2008;59:507–514. 20:165–172.
158. Cai J, Hatsukami TS, Ferguson MS, et al. In vivo quantitative measurement of intact 186. Morita S, Masukawa A, Suzuki K, et al. Unenhanced MR angiography: Techniques and clini-
fibrous cap and lipid-rich necrotic core size in atherosclerotic carotid plaque: Comparison cal applications in patients with chronic kidney disease. Radiographics. 2011;31:E13–E33.
of high-resolution, contrast-enhanced magnetic resonance imaging and histology. 187. Corti R, Wyttenbach R, Alerci M, et al. Images in cardiovascular medicine. Effect of per-
Circulation. 2005;112:3437–3444. cutaneous transluminal angioplasty on severely stenotic femoral lesions: In vivo demon-
159. Wasserman BA, Astor BC, Sharrett AR, et al. MRI measurements of carotid plaque in the stration by noninvasive magnetic resonance imaging. Circulation. 2002;106:1570–1571.
atherosclerosis risk in communities (ARIC) study: Methods, reliability and descriptive 188. Coulden RA, Moss H, Graves MJ, et al. High resolution magnetic resonance imaging of
statistics. J Magn Reson Imaging. 2010;31:406–415. atherosclerosis and the response to balloon angioplasty. Heart. 2000;83:188–191.
160. Wagenknecht L, Wasserman B, Chambless L, et al. Correlates of carotid plaque presence 189. Wyttenbach R, Gallino A, Alerci M, et al. Effects of percutaneous transluminal angio-
and composition as measured by MRI: The Atherosclerosis Risk in Communities Study. plasty and endovascular brachytherapy on vascular remodeling of human femoropopliteal
Circ Cardiovasc Imaging. 2009;2:314–322. artery by noninvasive magnetic resonance imaging. Circulation. 2004;110:1156–1161.
161. van den Bouwhuijsen QJ, Vernooij MW, Hofman A, et al. Determinants of magnetic 190. Worthley SG, Roque M, Helft G, et al. Rapid oral endotracheal intubation with a fibre-
resonance imaging detected carotid plaque components: The Rotterdam Study. Eur Heart optic scope in rabbits: A simple and reliable technique. Lab Anim. 2000;34:199–201.
J. 2012;33:221–229. 191. Worthley SG, Helft G, Fuster V, et al. Serial in vivo MRI documents arterial remodeling in
162. Esposito L, Saam T, Heider P, et al. MRI plaque imaging reveals high-risk carotid plaques experimental atherosclerosis. Circulation. 2000;101:586–589.
especially in diabetic patients irrespective of the degree of stenosis. BMC Med Imaging. 192. Pohost GM, Fuisz AR. From the microscope to the clinic: MR assessment of atheroscle-
2010;10:27. rotic plaque. Circulation. 1998;98:1477–1478.
163. Caballero P, Alonso R, Rosado P, et al. Detection of subclinical atherosclerosis in 193. Corti R, Osende JI, Fallon JT, et al. The selective peroxisomal proliferator-activated
familial hypercholesterolemia using non-invasive imaging modalities. Atherosclerosis. receptor-gamma agonist has an additive effect on plaque regression in combination with
2012;222:468–472. simvastatin in experimental atherosclerosis: In vivo study by high-resolution magnetic
164. Du J, Wasserman BA, Tong W, et al. Cholesterol is associated with the presence of a lipid resonance imaging. J Am Coll Cardiol. 2004;43:464–473.
core in carotid plaque of asymptomatic, young-to-middle-aged African Americans with 194. Bianda N, Di Valentino M, Periat D, et al. Progression of human carotid and femoral
and without HIV infection and cocaine use residing in inner-city Baltimore, MD., USA. atherosclerosis: A prospective follow-up study by magnetic resonance vessel wall imaging.
Cerebrovasc Dis. 2012;33:295–301. Eur Heart J. 2012;33:230–237.
165. Gao H, Long Q, Kumar Das S, et al. Study of carotid arterial plaque stress for symptom- 195. Joshi FR, Lindsay AC, Obaid DR, et al. Non-invasive imaging of atherosclerosis. Eur
atic and asymptomatic patients. J Biomech. 2011;44:2551–2557. Heart J Cardiovasc Imaging. 2012;13:205–218.
166. Summers RM, Andrasko-Bourgeois J, Feuerstein IM, et al. Evaluation of the aortic 196. Underhill HR, Yuan C. Carotid MRI: A tool for monitoring individual response to cardio-
root by MRI: Insights from patients with homozygous familial hypercholesterolemia. vascular therapy? Expert Rev Cardiovasc Ther. 2011;9:63–80.
Circulation. 1998;98:509–518. 197. Corti R, Fuster V, Fayad ZA, et al. Lipid lowering by simvastatin induces regression of
167. Jaffer FA, O’Donnell CJ, Larson MG, et al. Age and sex distribution of subclinical aortic human atherosclerotic lesions: Two years’ follow-up by high-resolution noninvasive mag-
atherosclerosis: A magnetic resonance imaging examination of the Framingham Heart netic resonance imaging. Circulation. 2002;106:2884–2887.
Study. Arterioscler Thromb Vasc Biol. 2002;22:849–854. 198. Gottlieb I, Agarwal S, Gautam S, et al. Aortic plaque regression as determined by mag-
168. Taniguchi H, Momiyama Y, Fayad ZA, et al. In vivo magnetic resonance evaluation of netic resonance imaging with high-dose and low-dose statin therapy. J Cardiovasc Med
associations between aortic atherosclerosis and both risk factors and coronary artery dis- (Hagerstown). 2008;9:700–706.
ease in patients referred for coronary angiography. Am Heart J. 2004;148:137–143. 199. Corti R, Fuster V, Fayad ZA, et al. Effects of aggressive versus conventional lipid-
169. Frydrychowicz A, Stalder AF, Russe MF, et al. Three-dimensional analysis of segmen- lowering therapy by simvastatin on human atherosclerotic lesions: A prospective, ran-
tal wall shear stress in the aorta by flow-sensitive four-dimensional-MRI. J Magn Reson domized, double-blind trial with high-resolution magnetic resonance imaging. J Am Coll
Imaging. 2009;30:77–84. Cardiol. 2005;46:106–112.
170. Markl M, Dudler P, Fydrychowicz A, et al. Optimized 3D bright blood MRI of aortic 200. Tang TY, Howarth SP, Miller SR, et al. The ATHEROMA (Atorvastatin Therapy: Effects
plaque at 3 T. Magn Reson Imaging. 2008;26:330–336. on Reduction of Macrophage Activity) Study. Evaluation using ultrasmall superpara-
171. Harloff A, Simon J, Brendecke S, et al. Complex plaques in the proximal descending magnetic iron oxide-enhanced magnetic resonance imaging in carotid disease. J Am Coll
aorta: An underestimated embolic source of stroke. Stroke. 2010;41:1145–1150. Cardiol. 2009;53:2039–2050.
172. Momiyama Y, Ohmori R, Fayad ZA, et al. Associations between serum lipoprotein(a) levels 201. Yonemura A, Momiyama Y, Fayad ZA, et al. Effect of lipid-lowering therapy with ator-
and the severity of coronary and aortic atherosclerosis. Atherosclerosis. 2012;41:1145–1150. vastatin on atherosclerotic aortic plaques detected by noninvasive magnetic resonance
173. Worthley SG, Helft G, Fuster V, et al. Noninvasive in vivo magnetic resonance imag- imaging. J Am Coll Cardiol. 2005;45:733–742.
ing of experimental coronary artery lesions in a porcine model. Circulation. 2000;101: 202. Zhao XQ, Yuan C, Hatsukami TS, et al. Effects of prolonged intensive lipid-lowering
2956–2961. therapy on the characteristics of carotid atherosclerotic plaques in vivo by MRI: A case-
174. Stuber M, Botnar RM, Danias PG, et al. Double-oblique free-breathing high resolution control study. Arterioscler Thromb Vasc Biol. 2001;21:1623–1629.
three-dimensional coronary magnetic resonance angiography. J Am Coll Cardiol. 1999; 203. Yonemura A, Momiyama Y, Fayad ZA, et al. Effect of lipid-lowering therapy with ator-
34:524–531. vastatin on atherosclerotic aortic plaques: A 2-year follow-up by noninvasive MRI. Eur J
175. Stuber M, Botnar RM, Kissinger KV, et al. Free-breathing black-blood coronary MR angi- Cardiovasc Prev Rehabil. 2009;16:222–228.
ography: Initial results. Radiology. 2001;219:278–283. 204. Kawahara T, Nishikawa M, Furusawa T, et al. Effect of atorvastatin and etidronate com-
176. Botnar RM, Stuber M, Kissinger KV, et al. Noninvasive coronary vessel wall and plaque bination therapy on regression of aortic atherosclerotic plaques evaluated by magnetic
imaging with magnetic resonance imaging. Circulation. 2000;102:2582–2587. resonance imaging. J Atheroscl Thromb. 2011;18:384–395.
177. Botnar RM, Stuber M, Lamerichs R, et al. Initial experiences with in vivo right coronary 205. Duivenvoorden R, Holleboom AG, van den Bogaard B, et al. Carriers of lecithin choles-
artery human MR vessel wall imaging at 3 tesla. J Cardiovasc Magn Reson. 2003;5: terol acyltransferase gene mutations have accelerated atherogenesis as assessed by carotid
589–594. 3.0-T magnetic resonance imaging [corrected]. J Am Coll Cardiol. 2011;58:2481–2487.
LWBK1209-ch27_p402-419.indd 418 17/05/13 5:12 PM

206. Duivenvoorden R, Fayad ZA. Utility of atherosclerosis imaging in the evaluation of high- 212. Roifman I, Beck PL, Anderson TJ, et al. Chronic inflammatory diseases and cardiovascu-
density lipoprotein-raising therapies. Curr Atheroscl Rep. 2011;13:277–284. lar risk: A systematic review. Can J Cardiol. 2011;27:174–182.
207. Fayad ZA, Mani V, Woodward M, et al. Rationale and design of dal-PLAQUE: A 213. Gharib AM, Abd-Elmoniem KZ, Pettigrew RI, et al. Noninvasive coronary imaging for
study assessing efficacy and safety of dalcetrapib on progression or regression of atherosclerosis in human immunodeficiency virus infection. Curr Probl Diagn Radiol.
atherosclerosis using magnetic resonance imaging and 18F-fluorodeoxyglucose posi- 2011;40:262–267.
tron emission tomography/computed tomography. Am Heart J. 2011;162:214–221. 214. Kerekes G, Soltesz P, Nurmohamed MT, et al. Validated methods for assessment of sub-
e2. clinical atherosclerosis in rheumatology. Nat Rev Rheumatol. 2012;8:224–234.
208. Fayad ZA, Mani V, Woodward M, et al. Safety and efficacy of dalcetrapib on athero- 215. Croca SC, Rahman A. Imaging assessment of cardiovascular disease in systemic lupus
sclerotic disease using novel non-invasive multimodality imaging (dal-PLAQUE): A ran- erythematosus. Clin Dev Immunol. 2012;2012:694143.
domised clinical trial. Lancet. 2011;378:1547–1559. 216. Furer V, Fayad ZA, Mani V, et al. Noninvasive cardiovascular imaging in rheumatoid
209. McMahon M, Hahn BH, Skaggs BJ. Systemic lupus erythematosus and cardiovascular arthritis: Current modalities and the emerging role of magnetic resonance and positron
disease: Prediction and potential for therapeutic intervention. Expert Rev Clin Immunol. emission tomography imaging. Semin Arthritis Rheu. 2012;41:676–688.
2011;7:227–241. 217. Falk E, Sillesen H, Muntendam P, et al. The high-risk plaque initiative: Primary preven-
210. Kozera L, Andrews J, Morgan AW. Cardiovascular risk and rheumatoid arthritis—the tion of atherothrombotic events in the asymptomatic population. Curr Atheroscl Rep.
next step: Differentiating true soluble biomarkers of cardiovascular risk from surrogate 2011;13:359–366.
measures of inflammation. Rheumatology. 2011;50:1944–1954. 218. Cormode DP, Chandrasekar R, Delshad A, et al. Comparison of synthetic high density
211. Giannarelli C, Klein RS, Badimon JJ. Cardiovascular implications of HIV-induced dyslip- lipoprotein (HDL) contrast agents for MR imaging of atherosclerosis. Bioconjug Chem.
idemia. Atherosclerosis. 2011;219:384–389. 2009;20(5):937–943.
LWBK1209-ch27_p402-419.indd 419 17/05/13 5:12 PM

Chapter
28 Hans-Christoph Becker
Plaque Characterization—
Computed Tomography
■■ Development of Coronary Atherosclerosis phenomenon is called “positive remodeling” and explains

why such plaques may not be seen by cardiac catheter (6)
■■ Estimation of Cardiac Event Risk
(Fig. 28.1). Nonfatal plaque rupture or erosion at end stage
■■ Computed Tomography Technique may heal, organize, and subsequently calcify. Consecutively,
fibro-calcified lesions may reduce vessel lumen diameter
■■ Coronary Calcium Screening
(negative remodeling), with consecutive reduction of blood
■■ Plaque Imaging by Computed Tomographic flow and myocardial ischemia. In clinical terms develop-
Angiography ment of fibro-calcified plaques may lead to stable and
chronic angina.
■■ Clinical Value of Coronary Calcium
■■ Computed Tomography Plaque Analysis
■■ Plaque Imaging in Symptomatic Patients Estimation of Cardiac Event Risk
■■ Future Perspective In many patients, unheralded myocardial infarction (MI)
associated with a mortality of approximately 20% is the
first sign of coronary artery disease (CAD). The risk of an
Development of Coronary event strongly depends on risk factors such as hypertension,
Atherosclerosis hypercholesterolemia, smoking habits, family history, age,
and gender. Based on these risk factors, algorithms such as
Coronary atherosclerosis begins as early as the first decade the Framingham (7), PROCAM (8), and SCORE (9) provide
of life with endothelial dysfunction, proliferation of smooth an estimation of the midterm (10-year) risk for an individual
muscle cells, and deposition of fatty streaks in the coronary subject to experience a cardiac event. According to most of
artery wall (1). At the later stage of this still clinically silent the international guidelines, patients with a midterm risk of
disease, these lesions may further accumulate cholesterol less than 10% are considered to be at low risk and usually
within the intima and media coronary artery wall layer, with require advice for a healthy lifestyle but no specific therapy.
a fibrous cap separating the lipid pool from the coronary Patients with a midterm risk of more than 20% are consid-
artery lumen (2). Inflammatory processes with invasion of ered to be at high risk and, therefore, may be termed patients
macrophages and activation of matrix metalloproteases may with a CAD equivalent. Similar to patients with established
cause consecutive weakening of the fibrous cap (3). CAD, these asymptomatic patients may require intensive
Such vulnerable plaques can rupture when exposed to intervention for risk reduction, such as lifestyle changes and
shear stress, and the thrombogenic lipid material may enter lifetime medical treatment.
the bloodstream. In the most unfortunate event, thrombus Approximately 40% of the population is considered to
progression turns the vulnerable plaques into culprit lesions have a moderate (10% to 20%) midterm risk of CAD. All of
that occlude the coronary vessels, leading to myocardial isch- the currently available risk stratification schemes suffer from
emia, ventricular fibrillation, and death (4). Alternatively, the lack of accuracy to correctly determine the risk, and
studies show that plaque erosions with consecutive thrombus uncertainty exists as to how to treat patients identified as
formation constitute approximately 40% of cases of sudden intermediate risk. Providing information to reassure and/or
coronary death (5). Plaque erosions are more commonly seen treat intermediate-risk patients is a high priority. Currently,
in young women and men of 50 years of age, and are associ- besides testing for myocardial ischemia (e.g., by treadmill
ated with smoking, especially in premenopausal women. testing), assessment of the atherosclerotic plaque burden
In the initial stadium of atherosclerosis, the coronary ves- provides valid information for further risk stratification in
sel widens at the location of the atherosclerotic plaques. The patient cohorts (10).
420
LWBK1209-ch28_p420-430.indd 420 17/05/13 5:13 PM

Chapter 28 ■ Plaque Characterization—Computed Tomography 421
Figure 28.1. Non-calcified lesion (left arrow) in the proximal left anterior descending coronary artery. The
lesion is difficult to assess in the coronary angiogram (right arrow) because of positive remodeling.
Tests for myocardial ischemia, such as electrocardio- the entire coronary artery tree and to quantify the athero-
graphic (ECG) stress testing, are better suited to diagnose sclerotic plaque burden noninvasively (12).
patients with ischemic CAD than to assess clinically silent
atherosclerosis in the coronary arteries. Determination of
the intima–media thickness (IMT) and ankle bracelet index Computed Tomography
(ABI) by ultrasound and Doppler focus on the assessment of Technique
the atherosclerotic plaque burden in the carotid and periph-
eral arteries, respectively. However, only computed tomog- Initially, coronary CT was performed with electron beam CT
raphy (CT) and magnetic resonance imaging (MRI) have the (EBCT) since lack of any moving items allowed for exposure
ability to noninvasively assess the extent of the atheroscle- times as short as 100 milliseconds. Scan acquisition in EBCT
rotic plaque burden in the coronary arteries. MRI is superior is triggered by the ECG signal to the mid-diastole phase of
to CT in terms of soft-tissue differentiation (11). However, the cardiac cycle to further reduce cardiac motion artifacts
CT is currently superior to MRI in terms of spatial and (Table 28.1). The initial intent of EBCT was to measure
temporal resolution to image the small and constantly myocardial perfusion. However, EBCT can also detect coro-
moving structures such as the coronary arteries. Therefore, nary calcifications as a surrogate marker for coronary ath-
CT is the only reliable and practicable tool to investigate erosclerosis (13). Further attempts have also been made to
Table 28.1 Acquisition Parameters for Computed Tomography
EBCT 4-MDCT 64-MDCT Dual-source–CT
Spatial resolution 3 mm 3 mm 0.4a mm 0.4a mm

Temporal resolution 100 ms 250 ms 165b ms 75 ms
Scan time 20 s 20 s 10 s 200 ms
Contrast for CTA — — 80 mL ≅ 5 mL/s 40 mL ≅ 5 mL/s
Radiation exposurec CCS: 1 mSv CCS: 2 mSv CCS: 2 mSv CCS: 1 mSv
CTA: n.a. CTA: n.a. CTA: 10 mSv CTA: 1 mSv
a
With dual z-focus on x-ray tube.
b
Depending on x-ray tube.
c
May substantially be reduced in MDCT by applying ECG pulsing and a low-dose protocol.
EBCT, electron beam computed tomography; MDCT, multidetector-row CT; CTA, CT angiography; CCS, coronary calcium scanning.
LWBK1209-ch28_p420-430.indd 421 17/05/13 5:13 PM

use this scanner for CT angiography (CTA) and plaque imag- After the reconstruction, the image data are analyzed and
ing of the coronary arteries. However, low spatial resolution postprocessed by a dedicated workstation. Upon identifica-
(3 mm) and high image noise of EBCT limits the investigation of the specific lesions, the workstation automatically
tion to the most proximal part of the coronary arteries (14). displays the quantities of coronary calcium as the Agatston
Conventional mechanical CT scanners with an x-ray tube score, volume equivalent, and absolute mass (Fig. 28.2).
and detector ring rotating around the patient have also in- According to the method proposed by Agatston, any dense
creased in speed within recent years. The scan mode dedicated structure in the CT image greater than 130 HU in density
to coronary multidetector-row (MD) CT is called retrospec- is identified as calcification. The area of this calcification is
tive ECG gating (15). With this technique, the spiral CT scan multiplied by a factor that depends on the peak density of
is acquired with a small pitch (table feed per gantry rotation), the lesion. A factor between 1 and 4 is used for a peak den-
and image reconstruction is performed in the slow motion sity of 130 to 199 HU, 200 to 299 HU, 300 to 399 HU, and
diastolic phase of the cardiac cycle. Coronary calcium quanti- greater than 400 HU, respectively. The sum of all lesions in
ties measured with the first clinical available MDCT scanner all four coronary vessels (left main, left anterior descending,
with four detector rows and temporal resolution of 250 mil- circumflex, and right coronary artery) corresponds to the
liseconds are well comparable to EBCT (16,17). total Agatston score (13). This algorithm requires specific
EBCT image quality and nonoverlapping slice thickness of
3 mm. Therefore, the Agatston quantification method is of
Coronary Calcium Screening limited value for MDCT.
Originally, the first investigations with the EBCT allowed
A four detector-row CT with 500-millisecond gantry rota- for acquisition of half of the entire heart because no more
tion time is the minimal requirement for a coronary calcium than 20 slices could be acquired at once. With this approach,
measurement. Coronary calcium screening can be performed the limited reproducibility became obvious (20), and a num-
by any MDCT without contrast media and with 3-mm slices. ber of authors have suggested algorithms to improve the re-
Overlapping slice reconstruction improves reproducibility producibility. To allow for a reproducible measurement of
of the coronary calcium quantification (17). Retrospective coronary calcium for follow-up investigations, it was neces-
ECG gating is superior in acquiring the entire volume with- sary to introduce the calcium volume equivalent (19). With
out any gaps and reconstructing the images with overlapping isotropic interpolation, overlapping slice acquisition can be
increment. An overlapping slice reconstruction may help to simulated by a workstation, and the reproducibility of the
improve the reproducibility (18). Depending on the MDCT quantification can be improved. However, real overlapping
scanner, the scan time may range from 5 to 20 seconds, and slice acquisition is always superior to isotropic interpolation
the entire investigation is completed within 5 minutes. In (21) and should be used with MDCT. The volume score, as
total, 40 to 80 slices are generated with one investigation. well as the Agatston score, is limited since the quantity of
As a fundamental requirement for screening, the radia- the calcium volume depends on the image quality and the
tion exposure for coronary calcium scanning needs to be re- values cannot easily be transferred from EBCT to MDCT.
duced to a minimum (approximately 1 mSv). In particular, In general, it is possible to quantify the absolute amount
when assessing coronary atherosclerosis in asymptomatic of coronary calcium by using a standard calibration phan-
patients, redundant radiation should be avoided. On the tom in CT (22). For this standardization, the calibration
base of the ECG signal, the x-ray tube current is switched to phantom must be scanned with those parameters that will
its nominal value during the diastolic phase and is reduced be used later for patient investigation. As the mass of the
significantly during the systolic phase of the heartbeat. This calcium particles in the phantom are known, it is possible
technique is called prospective ECG tube current modula- to measure the volume and density of the calcium from the
tion or ECG pulsing, and is most effective in patients with CT image and then to calculate the calibration factor for
low heart rates. If the heart rate is lower than 60 bpm, the absolute quantification.
radiation exposure will be reduced by approximately 50% In patient investigations, however, the accuracy of the
(19). measurement differs from patient to patient, depending on
Alternatively, prospective ECG triggering is a radiation size. X-ray absorption is different in thin patients as com-
dose efficient scan mode for coronary calcium screening, if pared to obese patients, where radiation exposure will lead
exposure is dedicated to the time point of the least cardiac to a beam hardening, resulting in a different density of the
motion only. Newest generation dual source CT allows for same amount of calcium. Therefore, the calibration phantom
ECG triggered high pitch scan acquisition. This scan mode needs to be scanned with “fat rings” that simulate different
has proved to be the most efficient in terms of radiation ex- patient sizes and deliver appropriate calibration factors
posure. However, in higher heart rate image quality may be (Fig. 28.3). As such, a topogram is used to determine the
degraded by heart motion and small calcifications may then patient’s chest diameter in order to select the appropriate cali-
be missed with this technique. bration factor. In practice, the calibration phantom is mea-
One of the other prerequisites is constant kVp of the sured without fat rings, with one fat ring, and with two fat
x-ray tube. The photon energy of the x-ray beam has pivotal rings. For example, in patients with chest diameters measuring
impact on the density of coronary calcium in the CT image. smaller than 30 cm, 30 to 38 cm, and larger than 38 cm, cali-
Most modern CT scanners allow for 120 kVp tube voltage bration factors with no fat ring, one fat ring, and two fat rings
that is closest to the scan protocol with the EBCT. will be used, respectively, on each patient.
Even the smallest calcifications will become visible by the Recently, the International Consortium on Standardization
reconstruction with a no-edge enhancing soft-tissue kernel. in Cardiac CT tried to establish a standardized algorithm
LWBK1209-ch28_p420-430.indd 422 17/05/13 5:13 PM

Figure 28.2. Coronary screening images are loaded into a dedicated workstation for quantification of cor-
onary calcium. The software calculates the Agatston score, volume equivalent, and mass of coronary calcium
in all of the four major coronary artery segments.
for coronary calcium measurements for all CT vendors. This cardiac catheter may provide dynamic information about
consortium also tried to establish a database to collect data blood flow not assessable by CT.
in a standardized fashion in order to provide reference val- Homogenous and constant contrast in the coronary ar-
ues for the absolute mass of coronary calcium (22). tery lumen is necessary to ensure reliable measurements of
coronary artery plaque densities (24). Coronary artery en-
hancement depends mainly on contrast media density and
Plaque Imaging by Computed injection rate (25). The use of a dual-head injector for the
Tomographic Angiography sequential injection of contrast media and saline is manda-
tory to keep the contrast bolus compact. In MDCT scanners
CTA for plaque imaging and detection of coronary artery with shorter scan times, a saline chaser bolus also results in a
stenoses requires the highest temporal and spatial resolution wash-out of contrast media from the right ventricle, helping
as provided by CT. Most advanced CT scanners have either to reduce artifacts caused by the influx of undiluted contrast
a wide detector (320 detector row CT) or two gantries in media from the superior vena cava (26). The amount of con-
one unit. Both these dedicated CT scanners have the advan- trast media necessary for a CT angiogram of the coronary
tage that they are able to acquire the entire heart within one arteries with fast scan acquisition may be reduced down to
single heartbeat. Temporal resolution in these CT scanners 40 mL.
may be between 150 and 75 milliseconds and the spatial High quality CT images for plaque imaging may only be
resolution with thinnest collimation and rapidly moving achieved in patients with low heart rates, (e.g., <70 bpm)
z-focus on the x-ray tube, has improved from 1 to 0.4 mm (27). To reduce cardiac motion artifacts, patients com-
(23). Therefore, the spatial resolution in CT is only 50% monly need to be prepared with beta-blocker, aiming for
below coronary angiography (0.2 mm). On the other side, a heart rate of 60 bpm for patients. When considering
LWBK1209-ch28_p420-430.indd 423 17/05/13 5:13 PM

Figure 28.3. Calibration phantom and “fat rings” for calibrated measurement of coronary calcium. The
calcium inlets contain known mass of calcium hydroxyapatite. Scanning with parameters used for patient
investigation allows calculation of the calibration factor. The calibration factor needs to be adjusted to differ-
ent body sizes to correct for beam hardening.
beta-blocker administration, contraindications (bronchial atrium, blood mixed with contrast media from the supe-
asthma, AV block, severe congestive heart failure, aortic rior vena cava is prevented from entering the right atrium.
stenosis, etc.) (28) must be ruled out and informed consent Furthermore, as a reflex, the increased blood volume in the
must be obtained from the patient. When the patient’s heart right atrium leads to a decrease in heart rate. The Valsalva
rate is significantly higher than 60 bpm, 50 to 200 mg of maneuver can last several seconds under breath-hold condi-
Metoprololtartraat can be administered orally 30 to 90 min- tions. After release, high-dense contrast media may enter the
utes prior to the investigation. Alternatively, 5 to 20 mg of right atrium, and the heart rate recovers to its original or
Metoprololtartraat divided into four doses can be admin- even slightly higher frequency (29). As a result, CTA may
istered intravenously (28) immediately before scanning. not be homogeneously enhanced and image quality may
Monitoring of vital functions, heart rate, and blood pres- suffer from rapid changes of the heart rate. Most advanced
sure is essential during this approach. The positive effects of scanners with single heartbeat acquisition are feasible to
beta-blockers on coronary CTA are fourfold: Better patient scan acquisition without breath-hold commands.
compliance, less radiation exposure, less cardiac motion ar- After scan acquisition has been completed, reconstruc-
tifacts, and higher vascular enhancement. tion of the axial slices begins with a careful analysis of the
Patients should be instructed not to press when tak- ECG trace recorded with the helical scan. The image recon-
ing deep breath in order to avoid the Valsalva maneuver. struction interval is best seen placed in front of the P-wave
During the Valsalva maneuver, the intra-abdominal pressure of the ECG that corresponds to the mid-diastole interval,
increases, with blood from the inferior vena cava entering and the individual point-of-time adjustment for the recon-
the right atrium. With an increased pressure in the right struction seems to improve image quality (27). The lower
LWBK1209-ch28_p420-430.indd 424 17/05/13 5:13 PM

the heart rate, the easier it is to find the best interval for all coronary calcium is an independent predictor for coronary
three major branches of the coronary artery tree. Conversely events regardless of the country the persons were screened
in patients that had been scanned with higher heart rate, the (36). The likelihood to establish coronary calcium scanning
best time point for image reconstruction may be the end- as a broadly applied screening technique certainly depends
systole phase of the cardiac cycle right behind the T-wave in on the cost benefit ratio of such programs.
the ECG trace. The progression of coronary calcium in patients with hy-
Para-cardiac findings are frequently observed in CTA percholesteremia may depend on the intensity of the statin
studies of the heart and should be reported. These findings therapy. In asymptomatic hypercholesteremic patients with-
include lymph node enlargement, pulmonary nodule esoph- out therapy, statins, and more than 120 mg/dL cholesterol,
ageal hernias, or even tumors (30). These incidental findings the annual progression rate is 52 ± 36%, 25 ± 22%, and −7
should trigger an additional reconstruction with a larger ± 23%, respectively (37). However, a regression of coronary
field of view, or a more dedicated (CT) investigation might calcium appears very unlikely from the pathophysiologic
be recommended. point of view. The reproducibility of the measurement of
coronary calcium by CT is in the range of 10% and, there-
fore, a regression below this value may be very difficult to
Clinical Value of determine. Furthermore, it remains to be proved that the
Coronary Calcium progression of coronary calcium results in an increased risk
for a cardiac event.
Coronary calcium is a specific marker for coronary ath-
erosclerosis. Initially, such calcifications were detected by
fluoroscopy or conventional chest x-ray. However, EBCT Computed Tomography
provides a more sensitive means to detect coronary calci- Plaque Analysis
fications than fluoroscopy. In a cohort of 584 patients,
coronary calcium could be detected in 52% and 90% of The entire extent of coronary atherosclerosis with calcified
patients by fluoroscopy and EBCT, respectively. However, and non-calcified plaques may become visible by the admin-
only 109 patients within this entire cohort had proven CAD; istration of contrast media. Plaques in heart specimens with
therefore, detection of coronary calcium by EBCT is a poor low-density plaques (40 HU) and high-density plaques (90
method to discriminate between patients with and without HU) may consist predominantly of lipid and fibrous tis-
CAD (13). sue (37), respectively. It was recently reported that micro-
Arad et al. were the first to report on the attempt to pre- CT with ultrahigh spatial resolution is able to distinguish
dict cardiac events with coronary calcium as detected by between different plaque components such as lipid, fibrin,
EBCT. In their cohort of 1,173 patients, they observed 26 and calcium by the different Hounsfield units. In the early
soft events (percutaneous transluminal coronary angioplasty stages of atherosclerosis, the proliferation of smooth muscle
and bypass grafting) and hard events (MI and death) in a pe- cells initially increased the Hounsfield units of atheroscle-
riod of 19 months. If the Agatston score was above 160, the rotic plaques (38).
odds ratio for an event was 20 to 35.4. Raggi et al. (31) used The current gold standard to detect coronary athero-
age-specific and gender-specific percentiles derived from sclerosis in vivo is intravascular ultrasound (IVUS). Studies
nearly 10,000 patients to identify those at increased risk for comparing IVUS with MDCT have shown a good correla-
an event. Seventy percent of patients with an unheralded MI tion between the echogeneity and CT density of coronary
(n = 172) were above the 75th percentile with their calcium atherosclerotic lesions (39). Schroeder et al. (40) were able
score as compared to an asymptomatic cohort (n = 632). to demonstrate that plaques with low, intermediate, and high
Current strategies provide two different values for the risk echogeneity may correspond to plaques in MDCT with a den-
estimation—one by the conventional risk assessment and sity of 14 ± 26 HU, 91 ± 21 HU, and 419 ± 194 HU, respec-
another by the amount of coronary calcium. It was recently tively, as compared to IVUS (Fig. 28.4). The sensitivity and
hypothesized that the combined use of the Framingham risk specificity for CT to detect calcified and non-calcified coro-
assessment and the calcium measurement are superior to the nary atheroscleroses are between 78% and 94%, respectively.
use of the Framingham risk assessment alone (32). In the However, the sensitivity to detect non-calcified plaques in a
Framingham risk algorithm, higher age becomes the pre- lesion-by-lesion comparison between CTA and IVUS is only
dominant factor above all others, an assumption that does 53% (41). Most likely, plaque erosions are more difficult to
not fit all patients. Therefore, Grundy has proposed an al- detect by MDCT than vulnerable plaque with a larger lipid
ternative method whereby the age score in Framingham is core.
replaced by a scheme where the coronary calcium percentiles As mentioned previously, large lipid components are
are taken into account. If the amount of coronary calcium found predominantly in vulnerable plaques that are prone
falls between the 25th and 75th percentile, the Framingham to rupture with consecutive thrombosis and occlusion of
risk score remains unchanged. If the amount of calcium is the coronary artery. Routine investigation of asymptomatic
below the 25th or above the 75th percentile, the score is patients by coronary CTA is currently not indicated due to
the same as for patients approximately 10 years younger or the high radiation burden and contrast media administra-
older, respectively (33). tion necessary for this investigation. Furthermore, the pro-
Age-, race-, and sex-dependent reference values have respective value for the detection of non-calcified lesions has
cently been published on the base of large population based not been demonstrated to date and needs to be further inves-
screening studies (34,35). Both studies had shown that tigated in clinical studies.
LWBK1209-ch28_p420-430.indd 425 17/05/13 5:13 PM

Figure 28.4. Coronary CT angi-

ography compared to intravascular
ultrasound. The low-dense plaques in
CT compare well with low echogenic
lesions in ultrasound. Calcified plaques
in CT appear as bright objects with
echo shadowing in ultrasound.
Plaque Imaging in that in symptomatic patients the HR ratio for cardiac events
Symptomatic Patients in patients with plaques and coronary artery stenosis is 4.51
and 10.74, respectively (46).
The assessment of atherosclerotic plaque burden in symp- Therefore, particularly low-dense non-calcified plaques
tomatic patients may be of clinical interest. In a recent study, in the coronary arteries should be taken serious as a pre-
Leber et al. (42) reported that non-calcified lesions were dictor for cardiac event. However, there are currently no
found predominantly in patients with acute MI, whereas cal- guidelines on the treatment of such finding in CT. Moreover,
cified lesions were found more often in patients with chronic these results may not allow for any conclusion in asymp-
stable angina. In patients with acute coronary syndromes, tomatic persons in whom these findings may be detected
non-calcified lesions in the coronary artery can correspond incidentally.
to an intra-coronary thrombus (43) (Fig. 28.5).
We observed a plaque in a patient with unstable angina
who developed an MI in due course with a culprit lesion at Future Perspective
the location of the former plaque. The vulnerable plaque de-
tected prior to its rupture had a dark center surrounded by a According to the large prospective cohort s tudies such
bright rim most likely corresponding to a lipid pool and a fi- as the PACC (47), RECALL (48), and MESA (17) trial
brous cap (Fig. 28.6). Matsumoto et al. followed a cohort of coronary calcium has an independent predictive value
810 non-obstructive CAD over a period of around 3 years. for cardiac events. Coronary calcium percentile rankings
Patients with plaques less than 65 HU in density tented to in combination with conventional risk assessments as dis-
have three times more often acute coronary events than pa- cussed in this chapter may be incorporated into future pre-
tients without (44). Min et al. tried to establish the modified vention guidelines. However, coronary calcium screening
DUKE index to semi-quantify the amount of atherosclerotic programs are likely not cost efficient and are therefore not
plaque burden in the coronary arteries. They found that pa- established so far in any healthcare system anywhere in
tients who had moderate to severe plaques in particular in the world.
the left main stem showed high event rates within a period If the ability of coronary CTA to detect vulnerable plaques
of about 1.5 years (45). A recent meta-analysis confirmed in patients with acute coronary syndrome holds true, new
LWBK1209-ch28_p420-430.indd 426 17/05/13 5:13 PM

Figure 28.5. Patient with acute coronary syndrome and thrombus plaque formation in the proximal left
anterior descending coronary artery. Coronary angiography confirmed the finding and immediate intervention
was performed.
Figure 28.6. Vulnerable plaque with low-dense lipid core and bright fibrous cap as detected by CT
(left arrow). At time of the CT examination, the plaque was invisible by coronary angiography (right, arrow up).
Three weeks later, the patient developed a myocardial infarction with a culprit lesion in the proximal
left anterior descending coronary artery (left, arrow down).
LWBK1209-ch28_p420-430.indd 427 17/05/13 5:13 PM

References
1. Stary HC, Chandler AB, Glagov S, et al. A definition of initial, fatty streak, and intermediate
lesions of atherosclerosis. A report from the Committee on Vascular Lesions of the Council
on Arteriosclerosis, American Heart Association. Circulation. 1994;89:2462–2478.
2. Stary HC, Chandler AB, Dinsmore RE, et al. A definition of advanced types of athero-
sclerotic lesions and a histological classification of atherosclerosis. A report from the
Committee on Vascular Lesions of the Council on Arteriosclerosis, American Heart
Association. Circulation. 1995;92:1355–1374.
3. Pasterkamp G, Falk E, Woutman H, et al. Techniques characterizing the coronary athero-
sclerotic plaque: Influence on clinical decision making? J Am Coll Cardiol. 2000;36:13–21.
4. Virmani R, Kolodgie FD, Burke AP, et al. Lessons from sudden coronary death: A com-
prehensive morphological classification scheme for atherosclerotic lesions. Arterioscler
Thromb Vasc Biol. 2000;20:1262–1275.
5. Farb A, Burke AP, Tang AL, et al. Coronary plaque erosion without rupture into a lipid
core. A frequent cause of coronary thrombosis in sudden coronary death. Circulation.
1996;93:1354–1363.
6. Glagov S, Weisenberg E, Zarins CK, et al. Compensatory enlargement of human athero-
sclerotic coronary arteries. N Engl J Med. 1987;316:1371–1375.
7. Wilson PW, D’Agostino RB, Levy D, et al. Prediction of coronary heart disease using risk
factor categories. Circulation. 1998;97:1837–1847.
8. Assmann G, Cullen P, Schulte H. Simple scoring scheme for calculating the risk of acute
coronary events based on the 10-year follow-up of the prospective cardiovascular Munster
(PROCAM) study. Circulation. 2002;105:310–315.
9. Conroy RM, Pyorala K, Fitzgerald AP, et al. Estimation of ten-year risk of fatal cardiovas-
cular disease in Europe: The SCORE project. Eur Heart J. 2003;24:987–1003.
10. Greenland P, Abrams J, Aurigemma GP, et al. Prevention Conference V: Beyond secondary
prevention: Identifying the high-risk patient for primary prevention: Noninvasive tests of
atherosclerotic burden: Writing Group III. Circulation. 2000;101:E16–E22.
11. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human coronary artery lumen and
wall imaging using black-blood magnetic resonance imaging. Circulation. 2000;102:506–510.
12. Nikolaou K, Poon M, Sirol M, et al. Complementary results of computed tomography
Figure 28.7. A high resolution gated CT scan may be used in the and magnetic resonance imaging of the heart and coronary arteries: A review and future
future for a complete workup of patients with acute chest pain. This outlook. Cardiol Clin. 2003;21:639–655.
protocol allows simultaneous detection of pulmonary embolism, aortic 13. Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of coronary artery calcium
dissection, and coronary artery occlusion. using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827–832.
14. Schmermund A, Rensing BJ, Sheedy PF, et al. Intravenous electron-beam computed tomo-
graphic coronary angiography for segmental analysis of coronary artery stenoses. J Am
Coll Cardiol. 1998;31:1547–1554.
15. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial and temporal resolution
for electrocardiogram-gated multislice spiral CT of the heart. J Comput Assist Tomogr.
strategies need to be considered for appropriate treatment 2001;25:907–923.
of these patients. The noninvasive detection of vulnerable 16. Becker CR, Jakobs TF, Aydemir S, et al. Helical and single-slice conventional CT ver-
sus electron beam CT for the quantification of coronary artery calcification. AJR Am
plaques will probably justify intensive medical treatment J Roentgenol. 2000;174:543–547.
and follow-up. Furthermore, CTA may soon serve as a tool 17. Carr JJ, Nelson JC, Wong ND, et al. Calcified coronary artery plaque measurement with
for a complete diagnostic workup in patients with atypical cardiac CT in population-based studies: Standardized protocol of Multi-Ethnic Study
of Atherosclerosis (MESA) and Coronary Artery Risk Development in Young Adults
chest pain. A single CTA investigation will allow practitio- (CARDIA) study. Radiology. 2005;234:35–43.
ners to diagnose pulmonary emboli, aortic dissection, or 18. Ohnesorge B, Flohr T, Fischbach R, et al. Reproducibility of coronary calcium quantifica-
tion in repeat examinations with retrospectively ECG-gated multisection spiral CT. Eur
coronary thrombus (Fig. 28.7). Since radiation exposure is Radiol. 2002;12:1532–1540.
reduced and the amount of contrast media is limited, coro- 19. Callister TQ, Cooil B, Raya SP, et al. Coronary artery disease: Improved reproducibility
nary CTA is indeed considered as a screening tool. Coronary of calcium scoring with an electron-beam CT volumetric method. Radiology. 1998;208:
807–814.
CTA studies determined a strong predictive value of coro- 20. Hernigou A, Challande P, Boudeville JC, et al. Reproducibility of coronary calcification
nary plaque and stenosis for cardiac events. However, these detection with electron-beam computed tomography. Eur Radiol. 1996;6:210–216.
observations had only been reported in symptomatic per- 21. Ulzheimer S, Kalender WA. Assessment of calcium scoring performance in cardiac com-
puted tomography. Eur Radiol. 2003;13:484–497.
sons. It is yet unclear, if this observation can also be trans- 22. McCollough CH, Ulzheimer S, Halliburton SS, et al. Coronary artery calcium: A multi-
ferred to asymptomatic persons. Furthermore, the clinical institutional, multimanufacturer international standard for quantification at cardiac CT.
Radiology. 2007;243:527–538.
consequence needs to be determined before broad applica- 23. Flohr T, Stierstorfer K, Raupach R, et al. Performance evaluation of a 64-slice CT system
tion is recommended. with z-flying focal spot. RoFo. 2004;176:1803–1810.
Coronary CTA already is a valid pretest before cardiac 24. Schroeder S, Flohr T, Kopp AF, et al. Accuracy of density measurements within plaques
located in artificial coronary arteries by X-ray multislice CT: Results of a phantom study.
catheter for the triage of patients for conservative, interven- J Comput Assist Tomogr. 2001;25:900–906.
tional, or surgical therapy, and may reduce the use of inva- 25. Becker CR, Hong C, Knez A, et al. Optimal contrast application for cardiac 4-detector-row
sive diagnostic procedures in preselected patients in whom computed tomography. Invest Radiol. 2003;38:690–694.
26. Cademartiri F, Luccichenti G, Marano R, et al. Use of saline chaser in the intravenous
coronary interventions are required. The superior spatial administration of contrast material in non-invasive coronary angiography with 16-row
resolution of MDCT allows visualization of the morphology multislice Computed Tomography. Radiol Med. 2004;107:497–505.
27. Hong C, Becker CR, Huber A, et al. ECG-gated reconstructed multi-detector row CT coro-
of atherosclerotic plaques; however, it lacks any functional nary angiography: Effect of varying trigger delay on image quality. Radiology. 2001;220:
information. Positron emission tomography (PET) CT is a 712–717.
new modality that combines the morphology of CT with the 28. Ryan TJ, Anderson JL, Antman EM, et al. ACC/AHA guidelines for the management of
patients with acute myocardial infarction: Executive summary. A report of the American
functional analysis of PET. For the detection of vulnerable College of Cardiology/American Heart Association Task Force on Practice Guidelines
plaques, PET/CT may provide useful information in terms (Committee on Management of Acute Myocardial Infarction). Circulation. 1996;94:2341–
of linking non-calcified lesions with the uptake of F-18- 2350.
29. Mao SS, Oudiz RJ, Bakhsheshi H, et al. Variation of heart rate and electrocardiograph
Fluorodesoxy-glucose (FDG), which indicates inflammation trigger interval during ultrafast computed tomography. Am J Card Imaging. 1996;10:
(49,50). 239–243.
LWBK1209-ch28_p420-430.indd 428 17/05/13 5:13 PM

30. Horton KM, Post WS, Blumenthal RS, et al. Prevalence of significant noncardiac findings 41. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary artery stenoses by contrast-
on electron-beam computed tomography coronary artery calcium screening examinations. enhanced, retrospectively electrocardiographically-gated, multislice spiral computed
Circulation. 2002;106:532–534. tomography. Circulation. 2001;103:2535–2538.
31. Raggi P, Callister TQ, Cooil B, et al. Identification of patients at increased risk of first 42. Leber AW, Knez A, White CW, et al. Composition of coronary atherosclerotic plaques
unheralded acute myocardial infarction by electron-beam computed tomography. in patients with acute myocardial infarction and stable angina pectoris determined by
Circulation. 2000;101:850–855. contrast-enhanced multislice computed tomography. Am J Cardiol. 2003;91:714–718.
32. Greenland P, LaBree L, Azen SP, et al. Coronary artery calcium score combined with 43. Becker CR, Knez A, Ohnesorge B, et al. Imaging of noncalcified coronary plaques using
Framingham score for risk prediction in asymptomatic individuals. JAMA. 2004;291: helical CT with retrospective ECG gating. AJR Am J Roentgenol. 2000;175:423–424.
210–215. 44. Matsumoto N, Sato Y, Yoda S, et al. Prognostic value of non-obstructive CT low-dense coro-
33. Grundy SM. Coronary plaque as a replacement for age as a risk factor in global risk assess- nary artery plaques detected by multislice computed tomography. Circ J. 2007;71:1898–1903.
ment. Am J Card. 2001;88:8E–11E. 45. Min JK, Shaw LJ, Devereux RB, et al. Prognostic value of multidetector coronary computed
34. Detrano R, Guerci AD, Carr JJ, et al. Coronary calcium as a predictor of coronary events tomographic angiography for prediction of all-cause mortality. J Am Coll Cardiol. 2007;
in four racial or ethnic groups. N Engl J Med. 2008;358:1336–1345. 50:1161–1170.
35. Schmermund A, Mohlenkamp S, Berenbein S, et al. Population-based assessment of subclin- 46. Bamberg F, Sommer WH, Hoffmann V, et al. Meta-analysis and systematic review of the
ical coronary atherosclerosis using electron-beam computed tomography. Atherosclerosis. long-term predictive value of assessment of coronary atherosclerosis by contrast-enhanced
2006;185:177–182. coronary computed tomography angiography. J Am College Cardiol. 2011;57:2426–2436.
36. Schmermund A, Lehmann N, Bielak LF, et al. Comparison of subclinical coronary ath- 47. O’Malley PG, Taylor AJ, Gibbons RV, et al. Rationale and design of the Prospective Army
erosclerosis and risk factors in unselected populations in Germany and US-America. Coronary Calcium (PACC) Study: Utility of electron beam computed tomography as a
Atherosclerosis. 2007;195:e207–e216. screening test for coronary artery disease and as an intervention for risk factor modification
37. Callister TQ, Raggi P, Cooil B, et al. Effect of HMG-CoA reductase inhibitors on coro- among young, asymptomatic, active-duty United States Army Personnel. Am Heart J. 1999;
nary artery disease as assessed by electron-beam computed tomography. N Engl J Med. 137:932–941.
1998;339:1972–1928. 48. Schmermund A, Mohlenkamp S, Stang A, et al. Assessment of clinically silent atheroscle-
38. Langheinrich AC, Bohle RM, Greschus S, et al. Atherosclerotic lesions at micro CT: rotic disease and established and novel risk factors for predicting myocardial infarction and
Feasibility for analysis of coronary artery wall in autopsy specimens. Radiology. 2004; cardiac death in healthy middle-aged subjects: Rationale and design of the Heinz Nixdorf
231:675–681. RECALL Study. Risk Factors, Evaluation of Coronary Calcium and Lifestyle. Am Heart
39. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection and evaluation of ath- J. 2002;144:212–218.
erosclerotic coronary plaques with multislice computed tomography. J Am Coll Cardiol. 49. Rudd JH, Warburton EA, Fryer TD, et al. Imaging atherosclerotic plaque inflammation
2001;37:1430–1435. with [18F]-fluorodeoxyglucose positron emission tomography. Circulation. 2002;105:
40. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection of coronary lesions by 2708–2711.
multislice computed tomography: Results of the New Age pilot trial. Catheter Cardiovasc 50. Tatsumi M, Cohade C, Nakamoto Y, et al. Fluorodeoxyglucose uptake in the aortic wall at
Interv. 2001;53:352–358. PET/CT: Possible finding for active atherosclerosis. Radiology. 2003;229:831–837.
LWBK1209-ch28_p420-430.indd 429 17/05/13 5:13 PM

LWBK1209-ch28_p420-430.indd 430 17/05/13 5:13 PM
PART
4
Congenital Heart Disease
LWBK1209-ch29_p431-470.indd 431 16/05/13 9:27 PM

LWBK1209-ch29_p431-470.indd 432 16/05/13 9:27 PM
Chapter
Charles B. Higgins
Karen G. Ordovas 29
Congenital Heart Disease:
Magnetic Resonance Evaluation
of Morphology and Function
■■ Major Clinical Indications ■■ Additional Cyanotic Lesions

■■ Techniques Ebstein Malformation
Anatomy Hypoplastic Left Heart Syndrome
Function Total Anomalous Pulmonary Venous Connection
Volume and Velocity of Blood Flow ■■ Situs Abnormalities
■■ Acyanotic Lesions ■■ Postoperative Evaluation
Left-to-right Shunts Monitoring of Ventricular Function and
Atrioventricular Septal Defect/Atrioventricular Pulmonary Regurgitation in Postoperative
Canal (Endocardial Cushion Defect) Tetralogy of Fallot
Coarctation of the Aorta ■■ Adult Congenital Heart Disease
Arch Anomalies
Pulmonary Arterial Anomalies The overall objectives of imaging in congenital heart dis-
Pulmonary Valvular Stenosis ease (CHD) are the precise delineation of cardiovascular
Congenital Aortic Stenosis anatomy and the quantitative assessment of function. The
Coronary Artery Anomalies evaluation of CHD was one of the first applications of car-
diac magnetic resonance imaging (MRI) and continues to be
■■ Cyanotic Lesions
one of its most important indications. MRI has significant
Tetralogy of Fallot advantages over other modalities, including echocardiogra-
Pulmonary Atresia phy and angiography, for the definitive assessment of con-
Pulmonary Atresia with Intact Ventricular genital cardiovascular anomalies. MRI does not require the
Septum use of a contrast agent or ionizing radiation. The absence of
Abnormalities of Ventriculoarterial Connections ionizing radiation is a major advantage of MRI in children,
Truncus Arteriosus who in the past have been exposed to large doses of radia-
tion during cine angiography for initial diagnosis and post-
■■ Analysis of Complex Congenital Heart
operative monitoring.
Disease Using MRI: Segmental Approach The role of MRI has been greatly influenced by the per-
Visceroatrial Situs ceived success of echocardiography as the primary nonin-
Ventricular Loop vasive imaging technique in CHD. The major applications
Great Artery Relationships have been for lesions incompletely evaluated by echocar-
■■ Abnormalities of Atrioventricular diography. However, substantial technologic improvements,
especially fast cine MRI using steady-state free precession
Connections
(SSFP) and noncontrast- and contrast-enhanced magnetic
Discordant Atrioventricular Connections resonance angiography (MRA), make MRI at least equal
Double-inlet Ventricle (Single Ventricle; to echocardiography in the diagnosis of all types of CHD
Univentricular Heart) in children and adults. Echocardiography is still considered
Atresia of Atrioventricular Valve to be the more effective and easily applicable modality for
433
LWBK1209-ch29_p431-470.indd 433 5/18/13 12:54 PM

434 Part 4 ■ Congenital Heart Disease
infants. On the other hand, MRI is becoming recognized as blood images); breath-hold single-slice or multislice turbo
more effective in adolescent and adult CHD and for follow- spin-echo (black blood); balanced SSFP (white blood); and
up of corrected complex anomalies. noncontrast- and contrast-enhanced three-dimensional (3D)
MRA. The latter technique is usually applied for the evalua-
tion of anomalies of the great arteries, pulmonary veins, and
Major Clinical Indications surgical shunts.
At present, the major indications for MRI in CHD are as

Function
follows:
The techniques used in the evaluation of right and left ven-
1. Thoracic aortic anomalies, such as coarctation and arch
tricular function are generically called cine MRI. There are
anomalies.
a multitude of cine MRI sequences; the most frequently
2. Pulmonary arterial anomalies and pulmonary atresia.
applied one now is a balanced SSFP sequence and real-
3. Complex cyanotic disease, such as atresia of atrioventric-
time cine MRI. Because of optimized homogeneous con-
ular valves and double-inlet ventricles.
trast between cavitary blood and myocardium, balanced
4. Abnormalities of pulmonary venous connections.
SSFP sequences (balanced fast field-echo [FFE]; true fast
5. Postoperative evaluation of complex procedures.
imaging with steady precession [FISP]; and fast imaging
6. Coronary arterial anomalies.
employing steady-state acquisition [FIESTA]) are now pre-
7. Adolescent and adult CHD.
ferred. The so-called real-time sequence may be advanta-
8. Monitoring regurgitation and ventricular function before
geous in infants and young children who cannot hold their
and after surgery.
breath.
Cine MR images are used to quantify ventricular volumes
Techniques and function. Right and left ventricular volumes are indexed
to body surface areas (EDV/m2 and ESV/m2). A distinct ad-
MRI studies are directed toward precise display of the cardio- vantage of MRI compared to echocardiography is its pre-
vascular anatomy and quantification of ventricular function cision and reproducibility for quantifying right ventricular
and blood flow. An advantage of MRI in CHD is measure- (RV) volume and function. For ventricular volumetrics, im-
ment of ventricular volumes and function using cine MRI and ages are acquired in the cardiac short-axis plane.
velocity-encoded (phase-contrast) cine MRI. Cine MR images are also used to evaluate valvular
For children younger than 8 to 10 years and children un- function. Planes approximately parallel to the valve leaflet
willing or unable to remain immobile and cooperative during or cusps can demonstrate valve motion, such as the hori-
imaging, light anesthesia without intubation is used. This is zontal long-axis plane for assessing mitral valve motion
usually under the control of an anesthesiologist. The favored and RV outlet plane for visualizing infundibular and pul-
drug is intravenous propofol. Blood pressure, heart rate, and monary valvular stenosis. The high-velocity jet caused by
oxygen saturation are monitored during the procedure. valvular stenosis and regurgitation may be identified on
cine MR images as a flow void (Fig. 29.1). However, flow
voids may not be discernible if the echo time (TE) is less
Anatomy
than 6 milliseconds on GRE cine and on balanced SSFP
Assessment of anatomy is done with one or more of the fol- cine. On the latter cine, high flow velocity may cause in-
lowing techniques: ECG-gated multislice spin-echo (black creased signal.
Figure 29.1. Postoperative tetra

logy of Fallot. Sagittal cine MR images
in systole (left) and diastole (right)
demonstrate a signal void (arrowhead)
projected into the pulmonary artery
(pulmonary stenosis) and a signal void
(arrow) into the right ventricle (pulmo-
nary regurgitation).
LWBK1209-ch29_p431-470.indd 434 16/05/13 9:27 PM

Chapter 29 ■ Congenital Heart Disease: Magnetic Resonance Evaluation of Morphology and Function 435
Pulmonary regurgitation Figure 29.2. A: Postoperative

600
tetralogy of Fallot. Sagittal spin-echo
image (top left), axial magnitude image
(top right), and phase images in systole
400
(bottom left) and diastole (bottom
right). Phase-contrast images show
forward flow in systole (dark voxels)
Flow (mL/s)
200
and retrograde flow in diastole (bright
voxels; arrow). Region of interest for
0 flow measurements is shown on the
pulmonary artery. B: Flow versus time
curve displays the forward and retro-
–200 grade flow in the pulmonary artery.
Area under negative component of the
curve yields a direct quantification
–400 of the volume of regurgitation. PA,
B Phase of cardiac cycle pulmonary artery; RV, right ventricle.
that point in the cardiac cycle. The cross-sectional area of

Volume and Velocity of Blood Flow
the vessel can be multiplied by the spatial mean velocity to
Quantification of blood flow is done using velocity-encoded obtain the flow in the blood vessel.
(VEC; phase-contrast) MRI. In phase-contrast imaging, the VEC cine MRI is used to measure valvular regurgitant
signal intensity signifies the velocity of blood at each pixel. volume (Fig. 29.2); differential flow in central pulmonary ar-
ECG-gated phase-contrast images can be produced in which teries; systemic to pulmonary shunt flow; flow through con-
each image shows the velocity at a different time in the car- duits (Rastelli and Fontan conduits); and collateral flow in
diac cycle. Since these are cine images, they can be referred coarctation. By determining the voxel with the peak velocity
to as VEC cine MRI. A region of interest drawn around a just beyond a stenosis, this technique can estimate the peak
blood vessel will give the mean velocity within the vessel at gradient across a valvular or conduit stenosis.
LWBK1209-ch29_p431-470.indd 435 16/05/13 9:28 PM

Acyanotic Lesions
Left-to-Right Shunts
Echocardiography provides essential diagnostic informa-
tion in most left-to-right shunts. MRI is used for a few spe-
cific purposes. The major indications are for the following
suspected lesions: Sinus venosus atrial septal defect (ASD),
partial anomalous pulmonary venous connection (PAPVC),
and supracristal ventricular septal defect (VSD). VEC cine
MRI may also be employed to measure the systemic-to-
pulmonary flow ratio (1–4). A new indication in patients
with ASD is to provide precise measurement of the defect
and its location for planning of transcatheter ASD closure
device placement (5,6).
Atrial Septal Defect

Spin-echo and cine MR images in the transaxial or four-
chamber planes demonstrate the site of the ASD (Fig. 29.3).
MRI clearly depicts the defect in the portion of the septum
separating the superior vena cava from the left atrium,
which is diagnostic of a superior type of sinus venosus ASD
Figure 29.3. Secundum atrial septal defect. Cine MR image in the (Fig. 29.4A). It also shows the anomalous connection of the
axial plane demonstrates the defect in the atrial septum (arrowhead). right upper lobe pulmonary vein to the superior vena cava
RA, right atrium; LA, left atrium.
Figure 29.4. A: Superior

type of sinus venosus atrial
septal defect. Cine MR images
in three axial planes (top row
and bottom left) and the coronal
plane (bottom right) show the
defect (arrow) in the portion of
the atrial septum between the
superior vena cava and the left
atrium. The right upper lobe
pulmonary vein (arrowheads)
connecting at the junction of the
superior vena cava (SVC) with
the right atrium is associated
with the sinus venosus defect.
Coronal image shows the dilata-
tion of the SVC at the site of
connection of the anomalous
A pulmonary vein.
LWBK1209-ch29_p431-470.indd 436 16/05/13 9:28 PM

Figure 29.4. (continued)

B: Inferior type of sinus veno-
sus defect. Cine MR image in
axial planes at level of defect
(arrowhead) (left) and at the
level of inferior vena cava (IVC)
(right). Arrows represent inferior
pulmonary veins connecting to
IVC. LA, left atrium; RA, right
atrium. B
a djacent to the septal defect. In the inferior type of sinus a flow across the defect on cine MRI or phase-contrast cine
venous ASD, the defect in the septum separating the inferior MRI (5,6).
vena cava (IVC) from the bottom of the left atrium can be Secundum ASD is situated in the middle of the atrial septum
discerned on MR images in the axial or horizontal long-axis (Fig. 29.3). If very large, it may approach the aortic or mitral
planes (Fig. 29.4B). The anomalous connections of inferior valve annulus (Fig. 29.5). Primum ASD is situated just adjacent
pulmonary vein(s) to the top of the IVC can be displayed on to the atrioventricular valves (Fig. 29.6). A diagnostic character-
images in these planes (Fig. 29.4B) or using contrast MRA istic of primum ASD is that the mitral and tricuspid valves are at
in the coronal plane. These features are characteristic for the same level (distance) from the cardiac apex (Fig. 29.6). The
the much rarer inferior type of sinus venosus ASD. demonstration of the location and size of an ASD is facilitated
On spin-echo MR images, the thin fossa ovalis can be by cine MRI in the horizontal long-axis (four-chamber) plane.
mistaken for an ASD. To avoid this error, the defect should VEC (phase-contrast) cine in a plane parallel to the
be evident at two adjacent anatomic levels or confirmed by ASD is used to determine the dimensions of the defect and
Figure 29.5. Large secundum atrial septal defect. Axial cine MR Figure 29.6. Primum atrial septal defect. Axial cine MR image
image shows a large secundum ASD (arrow) approaching the aortic shows a defect in the primum atrial septum (arrow). The atrioventricular
valve annulus. RA, right atrium; LA, left atrium; Ao, aorta. valve is attached to the foreshortened ventricular septum (arrowhead).
LWBK1209-ch29_p431-470.indd 437 16/05/13 9:28 PM

Figure 29.7. ASD size. Magnitude (left) and phase (right) images from velocity-encoded cine MR sequence
acquired parallel to the atrial septum (en face) delineates the flow through the defect (arrows) allowing precise
measurement of the defect dimensions.
if it is amenable for closing with an ASD closure device MRI and MRA are the procedures of choice for identi-
(Fig. 29.7). For this purpose, a low velocity encoding value fying the presence and connections of this anomaly (7,8).
(40 cm/s) is used in order to include flow at the margins of MRA is crucial in this regard (Figs. 29.4, 29.9, and 29.10).
the defect. For placement of a closure device, the defect The connection of the right-sided pulmonary veins to one or
should be less than 3.5 to 4 cm in diameter. Cine MRI is more veins that connect to the IVC can be displayed in the
also used to measure the distance of the margins of the scimitar syndrome (Fig. 29.11).
defect from the mitral and aortic valves and pulmonary Cine MRI is done in the horizontal long plane to exclude
veins (5,6). a concomitant ASD. Cine MRI is also performed in the
VEC cine MRI measurement in the main pulmonary ar- short-axis plane in order to document the severity of RV
tery and proximal ascending aorta can be used to calculate enlargement. VEC cine MRI is performed perpendicular to
pulmonary-to-systemic flow ratio (Qp/Qs) (Fig. 29.8). The the direction of flow in the proximal pulmonary artery and
imaging planes are placed perpendicular to the direction of aorta, in order to calculate Qp/Qs (Fig. 29.8) (1–4).
blood flow in each artery using sagittal images to site their
planes. Good correlations have been found for Qp/Qs, mea-
sured by VEC cine MRI, oximetric and indicator dilution Ventricular Septal Defect
techniques at catheterization, and with radionuclide angio VSDs are classified according to the portion of the septum
graphy (1–4). with the defect: Inlet, perimembranous, outlet-infracristal,
outlet-supracristal, and trabecular (muscular) (9). MRI
in the transaxial, horizontal long-axis, or four-chamber
Partial Anomalous Pulmonary Venous Connection
plane can precisely demonstrate the site of single or mul-
PAPVC produces volume overload of right-sided cardiac tiple VSDs (Fig. 29.12). Inlet VSD is situated just beneath
chambers. The hemodynamic effects are identical to ASD. The the atrioventricular valves. The tricuspid and mitral valves
most common type is connection of right upper lobe pulmo- are equidistant from the cardiac apex. This relationship and
nary vein to superior vena cava. This type of connection can the defect are displayed optimally on cine MRI in the hori-
occur alone (Fig. 29.9) or accompany superior type of sinus zontal long-axis plane. The supracristal VSD is characterized
venosus ASD (Fig. 29.4). Nearly equally frequent is left upper by a defect that lies between the RV outlet and the base of
lobe vein drainage via a vertical vein to the left innominate the aorta (10) (Fig. 29.13). On cine MRI, a flow void appar-
vein (Fig. 29.10). There is an increased incidence of secundum ently passes from the base of the aorta (it actually passes
ASD as well as sinus venosus ASD in patients with PAPVC. In from just below the aortic valve) to the RV outlet region and
the rare inferior type of sinus venous defect, one or more lower into the proximal pulmonary artery. VEC cine MRI can be
pulmonary veins connect to the IVC (Fig. 29.4B). used to calculate Qp/Qs and quantify the left-to-right shunt.
LWBK1209-ch29_p431-470.indd 438 16/05/13 9:28 PM

Qp/Qs
400
350
300
250
Flow (mL/s)
PA
200
150
100
Aorta
50
0
B Phase of cardiac cycle
Figure 29.8. A: Atrial septal defect. Magnitude (left) and phase (right) images from VEC cine MRI in planes
perpendicular to the long axis of the pulmonary artery (top) and the proximal ascending aorta (below). Regions
of interest surround the aorta and pulmonary artery. B: Flow versus time curves for the aorta and pulmonary
artery. Because of the left-to-right shunt, the area under the pulmonary artery curve is greater. The area under
each curve provides a direct measurement of the pulmonary to systemic flow ratio (2.1:1).
LWBK1209-ch29_p431-470.indd 439 16/05/13 9:28 PM

Figure 29.9. Partial anomalous pulmonary venous connection.

Contrast-enhanced 3D MRA shows anomalous connection of the right
upper pulmonary vein (arrow) to the superior vena cava.
Aorta to Right-Sided Shunts Figure 29.11. Partial anomalous pulmonary vena connection in
MRI is infrequently used to demonstrate patent ductus arte- scimitar syndrome. Contrast-enhanced 3D MRA shows the scimitar
vein (arrow) connecting with the inferior vena cava, hypoplastic right
riosus or aorticopulmonary window (11,12) (Fig. 29.14).
pulmonary artery, and dextroposition of the heart. IVC, inferior vena
With recent improvements in coronary MRA, it is effec- cava; RA, right atrium. (Courtesy of Gus Bis, Detroit, Michigan.)
tive for demonstrating the presence and sites of emptying
of coronary arteriovenous fistulas. MRI and MRA are very
effective for demonstrating sinus of Valsalva aneurysms and
fistulas.
Atrioventricular Septal Defect/
Atrioventricular Canal
(Endocardial Cushion Defect)
These terms denote a spectrum of abnormalities that have in
common an abnormal septation between the atria and ven-
tricles (13). They are also called endocardial cushion defects
because the defects are considered to be abnormalities of the
embryologic endocardial cushions, which grow together in
the center of the heart and divide the atria from the ven-
tricles.
In the normal heart, the atrioventricular septum separates
the right atrium from the left ventricle. The atrioventricular
septum lies between the more apically located normal tricus-
pid valve and mitral valve (Fig. 29.15). In all cases of atrio-
ventricular septal defect, the tricuspid and mitral valves lie at
the same level, and the atrioventricular septum is defective.
This abnormal relationship can be shown on MRI in the
transaxial or horizontal long-axis plane (14). In the mild-
est form of atrioventricular septal defect, the shunt is from
the left ventricle to the right atrium, which can be evident
on cine MRI. In other cases, the atrial septum adjacent to
the atrioventricular valve orifice may also be absent, result-
ing in an ostium primum ASD (Fig. 29.16). Some patients
Figure 29.10. Partial anomalous pulmonary venous connection. have an inlet VSD, usually located in the same axial image
Contrast-enhanced 3D MRA shows anomalous connection of the left as the atrioventricular valve or valves. In the most severe
upper vein (arrowhead) to a vertical vein and the right upper vein cases, both the atrial and ventricular portions of the septum
(arrow) to the superior vena cava. AO, ascending aorta; PA, pulmo- around the valve origins are absent, a condition referred to
nary artery; SVC, superior vena cava; V, vertical vein. as complete atrioventricular canal. This creates a common
LWBK1209-ch29_p431-470.indd 440 16/05/13 9:28 PM

Figure 29.12. Ventricular septal

defect. Spin-echo (left) and gradient-
echo (right) axial images show a defect
in the perimembranous ventricular
septum. A signal void at the defect is
projected into the right ventricle.
Figure 29.13. Supracristal ven-

tricular septal defect. Spin-echo (left)
and cine MR (right) images in the
axial plane demonstrate a defect (white
arrow) in the outlet portion of the sep-
tum. A flow void (black arrow) projects
into the upper region of the right ven-
tricular outflow region.
Figure 29.14. Aorticopulmonary window. Coronal (left) and axial (center and right) spin-echo images
demonstrate a defect (arrows) between the proximal ascending aorta and the pulmonary artery. Ao, aorta; RV,
right ventricular.
LWBK1209-ch29_p431-470.indd 441 16/05/13 9:28 PM

Figure 29.15. Atrioventricular septum. Axial spin-echo image

shows the small portion (arrow) of ventricular septum separating the
left ventricle and the right atrium. Note the more ventral position of
the tricuspid valve relative to the mitral valve.
Figure 29.17. Atrioventricular canal (endocardial cushion) defect.
Axial spin-echo image shows a large defect in the inlet portion of the
atrioventricular valve orifice with continuous, common ventricular septum. A single atrioventricular valve (arrows) spans the
atrioventricular valve leaflets (Fig. 29.17). inlet of both ventricles.
Coarctation of the Aorta

for postoperative evaluation of recurrent or persistent hyper-
MRI is the procedure of choice for definitive diagnosis and tension (15–19). To evaluate coarctation, spin-echo or cine
assessment of the severity of coarctation. MRI has been shown MR sequences are obtained in the axial and oblique sagittal
to be effective for preoperative assessment of coarctation and planes. The diameter of the narrowing can be precisely mea-
sured with MRI, especially with the use of thin (3-mm) sec-
tions through the center of the stenosis. Thin oblique sagittal
images through the long-axis plane of the aorta show the
diameter of the stenosis and provide an accurate measure-
ment of its length (Fig. 29.18).
The oblique sagittal images also display the dimension
of aortic isthmus (region between the left subclavian artery
and the ligamentum arteriosum) and the aortic arch. In
some cases, a single 3-mm slice in this plane may not display
the coarctation and arch because of tortuosity of the arch
and proximal descending aorta, but the assessment can be
done from adjacent images. Gadolinium-enhanced 3D MRA
can display the entire thoracic aorta on a single image with
the use of maximum intensity projection or volume-ren-
dering reconstruction techniques (Fig. 29.19). Gadolinium-
enhanced MRA is also effective for demonstrating collateral
vessels (Fig. 29.20).
VEC cine MRI can be applied to demonstrate the pres-
ence of and estimate the volume of collateral circulation to
the descending aorta below the coarctation site (15,16) (Fig.
29.21). This is accomplished by using two imaging planes
perpendicular to the aorta; one is about 2 cm beyond the
coarctation and the other at the level of the diaphragm.
In the normal aorta, volume flow is greater (about 5% to
7% higher) at the proximal site. On the other hand, in a
Figure 29.16. Primum atrial septal defect. Cine MR image in the hemodynamically significant coarctation, volume flow is
horizontal long-axis plane shows a defect at the level of the septum greater at the diaphragm because of retrograde flow through
primum (arrow). RA, right atrium; LA, left atrium. the intercostal and mammary arteries and other collaterals
LWBK1209-ch29_p431-470.indd 442 16/05/13 9:28 PM

Figure 29.18. Coarctation of aorta. Axial (left) and oblique sagittal (right) spin-echo images. A line bisect-
ing the ascending and descending aorta at the coarctation prescribes the oblique sagittal plane (long-axis plane
of thoracic aorta). This latter image displays the coarctation site along with the arch.
Figure 29.19. Coarctation of aorta. Contrast-enhanced 3D MRA in the sagittal plane shows a discrete
juxtaductal coarctation (arrow) on the maximum intensity projection (left) and the volume-rendered 3D
reconstruction (right).
LWBK1209-ch29_p431-470.indd 443 16/05/13 9:28 PM

ones encountered with any frequency are complete (patent)

double arch, double arch with atretic posterior component
of the left arch (Fig. 29.23), right arch with aberrant (retro-
esophageal) left subclavian artery (Fig. 29.24), and left arch
with aberrant right subclavian artery. The first three produce
complete vascular rings that narrow the trachea and esopha-
gus. Compression of the trachea is by the aortic component
situated between the vertebral body and the esophagus.
The vascular ring with right arch and aberrant left subcla-
vian artery is completed by the posterior component that
is tethered anteriorly by a left-sided ligamentum arterio-
sum. With mirror-image right aortic arch, the ligamentum
courses between the left innominate artery and the proximal
left pulmonary artery. On the other hand, with nonmirror-
image right aortic arch, the ligamentum courses between the
descending aorta and the proximal left pulmonary artery. At
the site of attachment of the ligamentum, there is an enlarge-
ment or diverticulum of the descending aorta. This local-
ized dilatation tethered anteriorly by the posterior left-sided
ligamentum is the structure that compresses the trachea and
esophagus rather than the left subclavian artery itself. These
anatomic arrangements can be readily discerned on MR
images in the transaxial plane.
For the evaluation of arch anomalies, spin-echo images
are acquired in sagittal (5-mm slice thickness) and transaxial
(3-mm slice thickness) planes (see Fig. 29.23). These images
display the aortic anomaly and verify that the anomalous
component produces airway compression (21,22). Cine MR
images may be used to display the pulsatile nature of the
airway obstruction. In some cases, the site of maximal air-
way compression is at the carina or even involves a proximal
bronchus.
Figure 29.20. Coarctation of aorta. Contrast-enhanced 3D Contrast-enhanced 3D MRA may be used to demonstrate
MRA in the sagittal plane shows long-segment coarctation caused by the arch anomaly. The MRA acquisition is done in either the
Takayasu’s arteritis. Note the abundant collateral arterial connections
sagittal or coronal planes.
(arrows).
Pulmonary Arterial Anomalies

into the distal aorta. The presence of greater volume flow Pulmonary arterial anomalies are evaluated using spin-echo
at the diaphragmatic level is considered a functional indica- and cine MRI (Fig. 29.25) and contrast-enhanced 3D MRA
tor of hemodynamic significance of the coarctation. There (Fig. 29.26) for depicting morphology and VEC (phase-
is a rough linear relationship between the percentage of ste- contrast) MRI for measuring blood flow. Spin-echo and cine
nosis and the volume of collateral circulation (15,16). After MR images are acquired in the transaxial plane followed by
stenting of the coarctation, VEC cine MRI has demonstrated images along the long axis of the right (oblique coronal plane)
reversal of the collateral flow since volume flow at the proxi- and left (oblique sagittal plane) pulmonary arteries. Contrast-
mal site becomes greater than at the distal one. enhanced 3D MRA is usually acquired in the coronal plane;
Theoretically, VEC cine MRI can also be used to estimate a short acquisition period is optimal to depict the pulmonary
the gradient across the coarctation (17,20). Using planes arteries before enhancement of the left heart and aorta. VEC
perpendicular and parallel to the coarctation, peak velocity cine sequences are obtained perpendicular to the long axis of
of flow can be estimated. Applying the modified Bernoulli the right, left, and main pulmonary arterial segments.
formula (peak pressure gradient = 4 × [peak velocity]2), the Compared to echocardiography, MRI is more useful for
peak gradient is estimated. The voxel(s) with highest velocity examining the right and left pulmonary arteries, as well as
in systole can be identified on phase images in both planes. the lobar and segmental arteries. MRI is especially useful for
The highest in either plan is considered the peak velocity demonstrating central and peripheral pulmonary artery ste-
(Fig. 29.22). noses (see Fig. 29.27), absent pulmonary artery (Fig. 29.28),
and pulmonary artery sling (Fig. 29.29). MRI and MRA are
the preferred techniques for the evaluation of residual pul-
Arch Anomalies
monary arterial stenoses after repair of tetralogy of Fallot
There are numerous types of arch anomalies resulting from (TOF) (Fig. 29.30).
abnormal resorption of the anterior or posterior segments of The hemodynamic significance of pulmonary arterial ste-
the embryonic double-arch configuration. However, the only noses is assessed by VEC cine MRI measurement of blood
LWBK1209-ch29_p431-470.indd 444 16/05/13 9:28 PM

Severe coarctation of the aorta

70
60
Distal aorta
50
Flow (mL/s)
40
30
20
10 Proximal aorta
0
Figure 29.21. A: Coarctation of the aorta. Sites of planes for magnitude and phase images in the proximal
(top) and distal aorta (bottom) used to estimate the volume of collateral flow. B: Flow versus time curves for the
proximal and distal descending aorta show a larger flow volume in the distal compared with the proximal aorta
caused by retrograde flow in aortic branches below the coarctation. The volume of collateral flow is estimated as
the difference in flow volumes (areas under the curves) at the two sites.
LWBK1209-ch29_p431-470.indd 445 16/05/13 9:28 PM

Figure 29.22. Pressure gradient in aortic coarctation. Magnitude (top left) and phase (top right) images
from VEC-MR acquisition perpendicular to the descending aorta in a level distal to the coarctation dem-
onstrate an estimated pressure gradient of 21 mm Hg. Magnitude (bottom left) and phase (bottom right)
images from an in-plane VEC MRI aligned to the descending aorta show an estimated pressure gradient of
33 mm Hg.
LWBK1209-ch29_p431-470.indd 446 16/05/13 9:28 PM

Figure 29.23. Double aortic arch. Spin-echo images in sagittal (left) and axial (right) planes display pos-
terior compression of the trachea by a retroesophageal component (arrow) of a double arch. The axial plane
shows that the right arch is the larger of the two and also shows a short atretic posterior segment (arrowhead)
of the left arch.
Figure 29.24. Right aortic arch with aberrant left subclavian artery. Spin-echo axial images (left) show the
right arch (A) and aberrant left subclavian artery (arrowhead). Another axial image shows the diverticulum
(arrow) of the descending arch, which is the site of origin of the left subclavian artery.
Figure 29.25. Branch pulmonary

artery stenoses. Transaxial spin-echo
images display stenosis (arrow) of the
left pulmonary artery. Note the size of
the right pulmonary artery (arrowhead)
for comparison.
LWBK1209-ch29_p431-470.indd 447 16/05/13 9:28 PM

Figure 29.28. Absent pulmonary artery. Spin-echo axial image

shows absence of the right pulmonary artery. There is no pulmonary
artery coursing between the ascending aorta and the right bronchus.
Figure 29.26. Pulmonary valvular stenosis. Three-dimensional

reconstructions from an MRA demonstrate enlargement of the MPA benefit, although such measurements may not be possible in
and LPA pulmonary arteries, with normal dimension of the right the presence of stainless steel stents.
pulmonary artery (RPA).
Pulmonary Valvular Stenosis

flow separately for each pulmonary artery (2,23) (Fig.
29.30). Since flow is also measured for the main pulmonary Pulmonary valvular stenosis is characterized morphologically
artery, the values can be expressed as the percentage of total on MRI as enlargement of the main and left pulmonary artery
pulmonary blood flow to each lung. The measurement can be (Figs. 29.26 and 29.31). There is a systolic high-velocity
done before and after angioplasty to document therapeutic flow jet across the valve on cine MRI. Cine MRI also dis-
plays hypertrabeculation and hypertrophy of the infundibu-
lum. Cine MRI in the right ventricular outflow tract (RVOT)
plane can be used to exclude or determine the severity of any
concomitant infundibular stenosis. This plane is also used to
measure the diameter of the annulus. Velocity-encoded cine
MR sequences in planes both parallel (RVOT plane) and per-
pendicular to flow across the valve are used to determine peak
velocity of flow in order to estimate transvalvular gradient
(24) (Fig. 29.32).
After pulmonary valvuloplasty, cine MRI and VEC MRI
are applied for quantifying any resultant pulmonary regur-
gitation. For this purpose, VEC MRI is acquired in a plane
perpendicular to flow in the proximal pulmonary artery.
Cine MRI in the short-axis plane is used to measure RV vol-
umes and ejection fraction.
Congenital Aortic Stenosis

MRI is not used typically to evaluate aortic stenosis; this
is done with echocardiography. The major role of MRI is
to evaluate the severity of poststenotic dilatation of the
ascending aorta. Severe aortoannular ectasia of the proxi-
mal ascending aorta in association with aortic stenoses
or even nonobstructive bicuspid valve can occur in child-
hood (Fig. 29.33). MRI and MRA in the sagittal plane are
Figure 29.27. Branch pulmonary artery stenosis. Contrast-enhanced optimal for demonstrating the dimensions of the aorta.
3D MRA in the coronal plane shows stenosis (arrow) of the left pul- VEC cine MRI may be used to quantify concomitant aortic
monary artery. regurgitation.
LWBK1209-ch29_p431-470.indd 448 16/05/13 9:28 PM

Figure 29.29. Pulmonary artery sling. Gradient-echo images in the axial plane arranged from cranial
(top left) to caudal (bottom right) demonstrate the origin of the left pulmonary artery (arrows) from the proxi-
mal right pulmonary artery. The left pulmonary artery curves around and posterior to the right bronchus.
Coronary Artery Anomalies

course of the anomalous artery passes ventral to the RV outlet
Coronary MRA has now become recognized as the most region or behind the aorta (retroaortic course). While selec-
reliable technique for demonstrating anomalies of the ori- tive (x-ray) coronary arteriography can demonstrate ectopic
gin and course of the coronary arteries (25,26). Coronary origin, it usually cannot confidently and precisely define the
anomalies may occur as isolated lesions or in association proximal course of the artery.
with other congenital cardiac anomalies, especially TOF. The current approach to coronary MRA employs a respi-
Coronary artery anomalies can be classified as major (ori- ratory navigator-compensated 3D free-breathing technique.
gin of a coronary artery from the pulmonary artery) (Fig. This technique clearly displays the origin of the coronary
29.34) or minor (ectopic origin from the aorta). Ectopic artery from the sinus of Valsalva and its proximal course.
aortic origin may be innocuous (Fig. 29.35) or potentially A single coronary artery may arise from either the right or
lethal (Fig. 29.36). The potentially lethal anomalies have a left sinus of Valsalva, or both coronary arteries may arise
proximal course between the base of the aorta and the pul- individually from one sinus of Valsalva (Fig. 29.36). In a
monary artery or the RV outlet region (interarterial course). child with chest pain or syncope during exercise or aborted
The course between the aorta and proximal pulmonary is the sudden death, coronary MRA must exclude a proximal in-
most dangerous. With innocuous ectopic origin, the proximal terarterial course.
LWBK1209-ch29_p431-470.indd 449 16/05/13 9:28 PM

Pulmonary arteries
500
400
LPA
Flow (mL/s)
300
200
RPA
100
A 0
Figure 29.30. A: Sites of acquisition of velocity-encoded (phase-contrast) cine MR sequence for the right
and left pulmonary arteries. B: Flow versus time curves for the right and left pulmonary arteries demonstrate
drastically impaired flow in the right pulmonary artery.
Figure 29.31. Pulmonary valve

stenosis. Cine MR image during early
systole in the right ventricular outflow
tract plane (left) and during end-systole
in the short-axis plane (right). Note the
poststenotic flow jet (arrow) emanating
from the pulmonary valve (arrowhead).
Right ventricular hypertrophy is seen
as almost complete cavity obliteration
during end-systole. RV, right ventricle;
LV, left ventricle.
Figure 29.32. Pulmonary valve

stenosis. Magnitude (left) and phase
(right) images from VEC-MR sequence
perpendicular to the blood flow in the
main pulmonary artery. Note the flow
jet (arrows) representing high-velocity
flow distal to the stenotic pulmonary
valve.
LWBK1209-ch29_p431-470.indd 450 16/05/13 9:28 PM

Figure 29.33. Poststenotic dilata-

tion of the ascending aorta. Axial (left)
and sagittal (right) spin-echo images
show aneurysmal dilatation of the
sinus of Valsalva and the ascending
aorta. Ao, aorta; PA, pulmonary artery.
Figure 29.34. Anomalous origin of the left coronary artery. Axial cine MR images from the base of the
heart (bottom right) to the level of the great vessels (top left) demonstrate anomalous origin of the left coronary
artery (arrow) from the main pulmonary artery (MPA). Note the dilated right coronary artery (arrowhead) con-
necting normally to the aorta (Ao).
LWBK1209-ch29_p431-470.indd 451 16/05/13 9:28 PM

Figure 29.35. Anomalous origin

of the left coronary artery from the
right coronary artery with a retroaortic
course. Two adjacent sections from
coronary MRA display the ectopic ori-
gin of the left coronary artery from the
right coronary (arrowhead) and pas-
sage of the left artery behind the aorta
(arrow).
Cyanotic Lesions (oblique sagittal plane) pulmonary arteries is used to assess

the severity of central pulmonary arterial stenoses, which
Tetralogy of Fallot are common in this anomaly. Contrast-enhanced 3D MRA
is also employed to demonstrate pulmonary arterial steno-
TOF consists of obstruction to the RV outlet region (usually ses and the systemic source of collateral flow to the lungs
multilevel), malalignment outlet VSD, enlarged aorta over- (28,31,32) (Fig. 29.38). VEC cine MRI can be used to define
riding the VSD, and a hypertrabeculated, hypertrophied right differential blood flow to the two lungs.
ventricle (27). These features are clearly defined on spin-echo TOF is usually repaired by relieving the pulmonic ste-
and cine MR images (28–31) (Fig. 29.37). Transaxial and nosis and closing the VSD. Stenosis of the main or branch
oblique coronal images demonstrate the VSD and the posi- arteries, the size of the RVOT, and the size of the right ven-
tion of the aorta overriding the defect. Cine MRI in the tricle are important to assess in preoperative planning and
RVOT plane is usually optional for defining the narrowed are well delineated by MRI. Coronary anomalies may occur
infundibulum and in some cases stenotic annulus and/or in patients with TOF. The most important is the origin of
valve. Sagittal and transaxial images are used to assess the the left anterior descending coronary artery from the right
size of the main and central pulmonary arteries and to dis- coronary artery with anomalous artery passing anterior to
play focal stenoses. Imaging oriented in a plane parallel to the RVOT. This important complicating anatomy can be as-
the long axis of the right (oblique coronal plane) and left sessed preoperatively by coronary CTA (25,26). Adequate
sizes of the central pulmonary arteries and the presence of
a central confluence may signify that the patient with severe
multilevel stenoses is a candidate for a Rastelli procedure
connecting the right ventricle to the pulmonary artery.
The major role of MRI in this anomaly is the postopera-
tive monitoring of RV volumes and function, and quantita-
tion of pulmonary regurgitation. This is discussed later in
this chapter and in Chapter 30.
Pulmonary Atresia
Pulmonary atresia with VSD is an extreme form of TOF in
which a direct connection from the right ventricle to the pul-
monary arteries is lacking (27). On axial MRI, a solid layer
of muscle in the region of the RVOT indicates an infun-
dibulum with a blind end. No connection between the right
ventricle and the pulmonary artery confluence (if present) is
evident on sequential images (Fig. 29.39). The atresia can
be focal, limited to the valve level, or more extensive. The
length of the atresia can be determined by inspecting sequen-
tial axial tomograms. Focal membranous pulmonary atre-
Figure 29.36. Ectopic origin of the right coronary artery with sia may be indistinguishable from severe stenosis on axial
an interarterial course. Coronary MRA shows the origin of the right MR images because of partial volume averaging. Cine MRI
coronary artery (arrow) from the right end of the left sinus of Valsalva in the axial or sagittal planes can establish the presence or
in a position where the proximal portion of the artery is compressed absence of flow across the valve. An enlarged aorta is seen
between the aorta and the right ventricular outlet region. overriding a perimembranous VSD (see Figs. 29.37–29.39).
LWBK1209-ch29_p431-470.indd 452 16/05/13 9:28 PM

Figure 29.37. Tetralogy of Fallot. Oblique coronal (left), sagittal (center), and axial (right) cine MR
images show ventricular septal defect (arrows), aorta overriding the defect, and pulmonic stenosis (arrowhead
on flow void).
Blood is usually delivered to the lungs via systemic to pul- pulmonary arteries. The collateral channels are especially
monary collateral channels, which can be seen as abnormal well seen on contrast-enhanced MRA (Fig. 29.38).
vessels originating from the descending aorta and traveling It is important to assess the sizes of the central pulmo-
toward the lungs or connecting with the pulmonary arteries nary arteries in patients who have TOF with severe stenosis
(Fig. 29.38). These are shown best on contrast-enhanced 3D or pulmonary atresia. Thin (3-mm) transaxial MR images
MRA (31,32) (Fig. 29.38). can readily depict the sizes of the main, right, and left pul-
The surgical correction of pulmonary atresia with VSD monary arteries. The right pulmonary artery is observed on
usually consists of placing a conduit from the right ventricle the image that contains the right main bronchus, coursing
to the central pulmonary arteries (if present) or to a surgi- in front of the right bronchus, and the left pulmonary artery
cally created confluence of pulmonary arteries and larger passes over the left main bronchus and is seen on the image
systemic to pulmonary collateral vessels (unifocalization containing the left bronchus or on the one just above. The
procedure). Therefore, it is important for the surgeon to identification of central pulmonary arteries and of a cen-
know whether a native confluence of the pulmonary arter- tral confluence of the right and left pulmonary arteries is a
ies is present, its size, and the number and size of collateral unique capability of MRI since opacification of the vessels
vessels. Axial plane MRI is excellent for defining the main with contrast is not required. The pulmonary arteries are
frequently hypoplastic, or the central or peripheral arteries
may contain one or more stenoses.
Cine MRI and contrast-enhanced MRA can be used
to identify the blood supply to the lungs. Pulmonary and
bronchial arteries have bright signal on cine MRI. On
transaxial images at the level of the carina, pulmonary ar-
teries can often be differentiated from bronchial arteries.
Bronchial arteries or systemic to pulmonary artery collat-
erals arise from the aorta or its branches and are usually
located dorsal to the bronchi, whereas pulmonary arteries
are ventral to the bronchi. Occasionally, a bronchial artery
originating from a subclavian artery is seen ventral to the
bronchi.
Pulmonary Atresia with

Intact Ventricular Septum
A VSD is not identified in this form of pulmonary atresia
with intact ventricular septum. In this variety of pulmonary
atresia, MRI is effective for demonstrating the size of the
RV, which varies from markedly hypoplastic to dilated (Figs.
29.40 and 29.41). The pulmonary arteries are usually nor-
mal or nearly normal in size and do not contain stenoses.
Figure 29.38. Pulmonary atresia with ventricular septal defect. Cine MR images can be acquired in the short-axis plane to
Contrast-enhanced 3D MRA in the coronal plane displays several quantify the volumes of the right and left ventricles. The size
large systemic to pulmonary artery collaterals (arrows) arising from of the right ventricle is critical for determination of the surgi-
the descending aorta. cal approach. A right ventricle of adequate size is necessary to
LWBK1209-ch29_p431-470.indd 453 16/05/13 9:28 PM

Figure 29.39. Pulmonary atresia with ventricular septal defect. Spin-echo axial images arranged from cau-
dal (top left) to cranial (bottom right) demonstrate a single large artery (aorta) at the base of the heart, which
overlies the ventricular septal defect (arrow). Right ventricular wall is hypertrophied. Note the central conflu-
ence (arrowhead) of the right and left pulmonary arteries distal to the atretic segment.
consider treatment with a conduit from the right ventricle to ummary of Anatomy of
S
the pulmonary arteries.
Table 29.1 Ventriculoarterial
Exit of blood from the right ventricle in this anomaly is
Connection Abnormalities
by tricuspid regurgitation or retrograde flow through myo-
cardial sinusoid and coronary arteries into the aorta. Cine COMPLETE TRANSPOSITION
MRI can demonstrate and give some insight into the severity Situs solitus (right atrium on right side of chest)
of tricuspid regurgitation. D-ventricular loop (right ventricle to the right of left ventricle)
d-TGA (aorta anterior and to right of pulmonary artery)
Abnormalities of Ventriculoarterial CORRECTED TRANSPOSITION

Connections Situs solitus
L-ventricular loop (right ventricle to the left of left ventricle)
These abnormalities consist of complete transposition of the l-TGA (aorta anterior and to the left of pulmonary artery)
great arteries (TGA), corrected transposition of the great DOUBLE-OUTLET RIGHT VENTRICLE
arteries (CTGA), double-outlet right ventricle (DORV) (33), Aorta orifice >50% overlies the right ventricle
and double-outlet left ventricle (Table 29.1). The last anom- Pulmonary orifice >50% overlies the right ventricle
aly is exceedingly rare. Truncus arteriosus will also be con- Subaortic, subpulmonary, doubly committed or noncommitted VSD
sidered under this heading.
DOUBLE-OUTLET LEFT VENTRICLE
A critical step in evaluating these anomalies is deter- Aorta orifice >50% overlies the left ventricle
mining the morphology of the ventricles. This is readily Pulmonary orifice >50% overlies the left ventricle
accomplished using transaxial images that show the char-
acteristics of a right ventricle: Infundibulum (tunnel of TRUNCUS ARTERIOSUS
One enlarged great artery (truncus) overlies both ventricles
myocardium) separating the atrioventricular and semilunar
VSDs immediately beneath orifice of truncus
valves; moderator band; and corrugated surface of the RV
LWBK1209-ch29_p431-470.indd 454 16/05/13 9:28 PM

Figure 29.40. Pulmonary atresia with intact ventricular septum Figure 29.41. Pulmonary atresia with intact ventricular septum,
after pulmonary valvotomy. Axial cine MRI shows enlarged right- status postvalvotomy. Axial cine MR images show a distorted right
sided chambers as a consequence of the associated tricuspid regurgita- ventricle (RV) with lack of normal trabeculation (bald RV). RA, right
tion. RA, right atrium; RV, right ventricle. atrium; LV, left ventricle.
aspect of the septum (Fig. 29.42). The left ventricle shows and connected to the left atrium (L-ventricular loop). The
direct fibrous continuity between the two valves, the papil- most frequent types of TGA are complete TGA (situs soli-
lary muscles, and the smooth surface of the left ventricular tus, D-ventricular loop, d-TGA; see Figs. 29.43 and 29.44)
aspect of the septum. and corrected TGA (situs solitus, L-ventricular loop, l-TGA;
Ventriculoarterial connections can be concordant or dis- see Fig. 29.45). The latter anomaly is considered to be cor-
cordant (34,35). Concordant connections are right ventricle rected (corrected transposition) in terms of blood flow in the
to pulmonary artery and left ventricle to aorta. Discordant central circulation since systemic venous blood flows from
connections occur in a diverse group of anomalies in which the right atrium into the left ventricle and out to the pulmo-
the great arteries are inverted (transposition). Both great nary artery for normal oxygenation, precluding obligatory
arteries can arise predominantly from one of the ventricles cyanosis. Transaxial MR images also demonstrate anomalies
(double-outlet ventricle). The ventriculoarterial connections frequently associated with complete and corrected transposi-
and the arterial relationships are depicted on coronal and tion, such as ASD, VSD, subvalvular and valvular pulmonic
transverse MR images. stenosis, tricuspid atresia, and Ebstein anomaly.
A great artery is considered connected to a ventricle if Transaxial MR images demonstrate connection of both
more than half of its orifice arises from that ventricle. A great arteries to the right ventricle in DORV (36–39) (Fig.
series of transaxial images extending from the aortic arch 29.46). Rarely, both great arteries connect to the ana-
to the diaphragm demonstrates these connections. The ini- tomic left ventricle, indicating double-outlet left ventricle.
tial determination is to identify the aorta unequivocally by Transaxial images are very effective for demonstrating the
following one of the great arteries to the arch. If this great relationship of the obligatory VSD to the great arteries in
artery (aorta) connects to the right ventricle and the pul- DORV (36,37). The defect may be just below the aortic ori-
monary artery to the left ventricle, then the diagnosis of fice (subaortic VSD) or pulmonary artery orifice (subpulmo-
complete TGA is established (Figs. 29.43 and 29.44). At the nary VSD); beneath both orifices (doubly committed VSD);
base of the heart, the aorta is anterior and to the right of or removed some distance from both (noncommitted VSD).
the pulmonary artery in most cases and is called d-TGA.
The transaxial images also demonstrate the position of the Truncus Arteriosus
ventricles in relation to each other and their connections
to the atria (atrioventricular connections). Typically, in the Truncus arteriosus was classified by Collet and Edwards (40)
presence of situs solitus, the images show the right ventricle based on the origin of the pulmonary artery from the com-
situated to the right of the left ventricle and connected to mon arterial trunk. In type I, a septum divides the origin of
the right atrium (D-ventricular loop). If the ventricles are the aorta and pulmonary trunk. In type II, the right and left
inverted, the right ventricle is to the left of the left ventricle pulmonary arteries are close to each other but arise separately
LWBK1209-ch29_p431-470.indd 455 16/05/13 9:28 PM

Figure 29.42. Right ventricular characteristics. Axial cine MR images from the level of the great vessels
(top left) to the midventricular level (bottom right) show a hypertrophied systemic right ventricle (RV) in a patient
with transposition of the great arteries. Note the muscular infundibulum (I), moderator band (arrow), and irregular
surface of the ventricular septum (arrowhead) at the apical level. Ao, aorta; PA, pulmonary artery; LV, left ventricle.
from the pulmonary trunk. In type III, right and left pulmo- images can show the relative sizes of the ventricles. Since
nary arteries arise further laterally. The lesion labeled type it can demonstrate a small infundibular chamber in pulmo-
IV truncus is in fact pulmonary atresia with VSD rather than nary atresia, MRI can be used to distinguish truncus arterio-
truncus arteriosus. sus from pulmonary atresia.
Axial, sagittal, and coronal MR images at the base of The relative sizes and confluence of the pulmonary arter-
the heart can show the truncus arteriosus aligned over the ies are useful pieces of information because surgical treatment
VSD (Fig. 29.47). The origins of the main pulmonary artery involves excision of the pulmonary arteries from the common
from the truncus in type I can be delineated with MRI. Axial trunk and the creation of a conduit from the right ventricle to
LWBK1209-ch29_p431-470.indd 456 16/05/13 9:29 PM

Figure 29.43. d-transposition of great arteries. Spin-echo images arranged from cranial (left) to caudal
(right) show the aorta (Ao) ventral and slightly to the right of the pulmonary artery (PA). The aorta is shown
connecting to the right ventricle (RV) and the pulmonary artery to the left ventricle (LV). There is a baffle
(Mustard procedure) in the atrium, separating the mitral valve from the pulmonary venous atrium (PVA).
RVOT, right ventricular outflow tract.
the pulmonary arteries (Rastelli procedure). Also important Analysis of Complex Congenital
in the evaluation is the demonstration of a right aortic arch Heart Disease Using MRI:
(35%) and other arch anomalies. Cine MRI and VEC cine Segmental Approach
MRI can be employed to identify and quantify the severity
of truncal regurgitation. Cine MRI can also be used to quan- The analysis of complex CHD requires a systematic approach.
tify the volume and mass of the two ventricles. The approach most widely used today is the segmental one, in
which the heart is partitioned into three main segments (atria,
ventricles, and great vessels) and the connections between
them (atrioventricular and ventriculoarterial connections).
Transaxial images from the aortic arch to the upper abdomen
clearly demonstrate the segmental cardiovascular anatomy
and connections of one segment to the other (atrioventricu-
lar connections and ventriculoarterial connections) and the
types of situs.
Visceroatrial Situs
The right atria and left atria are described by their morpho-
logic structure and not necessarily their position (Table 29.2).
An atrium with the morphologic features of a left atrium,
I dentification and
Table 29.2
Position of Atria
LEFT ATRIUM
“Finger-shaped” appendage
Narrow orifice of appendage
RIGHT ATRIUM
Triangular-shaped appendage
Wide orifice of appendage
Figure 29.44. MRA status postatrial switch procedure. Connects to IVC
Gadolinium-enhanced MR angiography in the coronal plane is Right atrium on right side—situs solitus
shown in a patient status postatrial switch procedure for correction Right atrium on left side—situs inversus
of transposition of great arteries. The inferior vena cava (IVC) is Bilateral morphologic right atria—right sidedness (frequently
dilated and connects to the inferior limb of the baffle (B), which is asplenia)
then attached to the mitral valve. The superior limb of the baffle is Bilateral morphologic left atria—left sidedness (frequently
occluded (arrow). Note an enlarged azygos vein (arrowhead). LV, polysplenia)
left ventricle; PA, pulmonary artery.
LWBK1209-ch29_p431-470.indd 457 16/05/13 9:29 PM

Figure 29.45. l-transposition of great arteries. Spin-echo axial images arranged from cranial (left) to
caudal (right) demonstrate the aorta (Ao) ventral and to the left of the pulmonary artery. It is connected to
the morphologic right ventricle (RV), which lies to the left of the left ventricle (LV). The right ventricle is con-
nected to the left atrium. The pulmonary artery (PA) is connected to the left ventricle, which is connected to
the right atrium. RVOT, right ventricular outflow tract.
Figure 29.46. Double-outlet right ventricle. Spin-echo axial images arranged from cranial (top left) to
caudal (bottom right). At the base of the heart, the aorta (Ao) and pulmonary artery (PA) lie side by side. Both
great arteries are connected to the right ventricle (RV). There are coni beneath both great arteries, and trabecu-
lation of the right ventricular side of the ventricular septum. LV, left ventricle.
LWBK1209-ch29_p431-470.indd 458 16/05/13 9:29 PM

Figure 29.47. Truncus arteriosus,

type I. Spin-echo images in the coronal
(left) and axial (right) planes dem-
onstrate a large single artery arising
from the heart. The pulmonary artery
(arrow) arises from the left side of the
truncus, and the aortic arch is right
sided.
which may rarely be located to the right of midline, is called ventricle on the left side has an L-ventricular loop. In situs sol-
a morphologic left atrium. Transaxial MR images can clearly itus, a D-loop is normal. In situs inversus, L-loop is normal.
depict the characteristic morphologic features of the atria. Atrioventricular connections can be concordant or discor-
The most definite features of the atria are the appendages. dant. In normal, concordant atrioventricular connections, the
On an axial MR image, the right atrial appendage appears right atrium is connected to the right ventricle and the left
as a triangular structure with a broad-based opening into atrium to the left ventricle. The atrioventricular valves remain
the right atrium. This is in distinction to the left atrial ap- with their respective ventricles, regardless of the type of ven-
pendage, a long, narrow, finger-like projection with a nar- tricular loop. The mitral valve resides with the left ventricle
row orifice. The atrial appendages are the most constant and the tricuspid valve is part of the right ventricle, except in
part of the atria, even in complex abnormalities. If the atrial patients with double-inlet ventricle. Identification of ventricu-
appendages are difficult to identify, the next most reliable lar morphology indicates the type of atrioventricular valve
structure is the drainage of the IVC. The atrium connect- within the ventricle.
ing to the IVC is considered the morphologic right atrium. Discordant atrioventricular connections are right atrium
Superior vena caval and pulmonary venous drainage is vari- to left ventricle and left atrium to right ventricle. Congenitally
able and is not used to identify atrial morphology. corrected TGA is an example of an anomaly with atrioven-
tricular discordance.
Ventricular Loop
Several morphologic features identify right versus left ventri- Great Artery Relationships
cle (Table 29.3). The normal rightward bending of the primi- Transaxial images at the base of the heart clearly define the
tive cardiac tube places the morphologic right ventricle on normal relationship of the pulmonary artery anterior and to
the right side of the heart. This rightward bending is called the left of the aorta. Such images also identify concordant
D-looping (D for dextro, right). If the primitive heart tube (aorta connected to left ventricle and pulmonary artery con-
bends to the left, the result is called L-looping (L for levo, nected to right ventricle) and discordant (aorta connected to
left), in which the morphologic right ventricle is placed on right ventricle and pulmonary artery connected to left ven-
the left side of the heart. A heart with the morphologic right tricle) ventriculoarterial connections (39).
Complete TGA is an example of atrioventricular concor-
dance and ventriculoarterial discordance. Corrected TGA is an
orphologic Features
M example of atrioventricular and ventriculoarterial discordance.
Table 29.3 Double discordance results in blood flowing in a normal pat-
of Ventricles
tern serially through the pulmonary and systemic circulation.
RIGHT VENTRICLE
Infundibulum
No fibrous continuity between atrioventricular and semilunar
valves Abnormalities of
Moderator band Atrioventricular Connections
Corrugated surface of ventricular septum near apex
LEFT VENTRICLE Discordant Atrioventricular
No infundibulum Connections
Fibrous continuity between atrioventricular and semilunar valves
The right atrium connects with the left ventricle and the left
Papillary muscles
Smooth septal surface near apex
atrium connects with the right ventricle, constituting discor-
dant connections, in corrected transposition (see Fig. 29.45).
LWBK1209-ch29_p431-470.indd 459 16/05/13 9:29 PM

Figure 29.48. Situs solitus with corrected transposition and isolated dextrocardia. Spin-echo axial images
arranged from cranial (top left) to caudal (bottom right). The right atrium (RA) is right sided and the left
atrium (LA) is left sided, indicating situs solitus. The right ventricle (RV) is positioned to the left of the left
ventricle (LV). The aorta (Ao) is anterior and leftward to the pulmonary artery (PA). Note the conus (arrow)
and moderator band on the left-sided ventricle, indicating it is a morphologic right ventricle in an L-ventricular
loop. Thus, this arrangement of connections is situs solitus, L-ventricular loop, l-transposition, which consti-
tutes corrected TGA. The cardiac apex is right sided, which indicates isolated dextrocardia.
One of the most frequent anomalies in patients with situs is the optimal method for quantifying the volumes, mass,
inversus is corrected transposition. In this anomaly, the ana- and function of the dominant and rudimentary ventricle.
tomic right atrium (situated left of midline) connects dis- Volumetric data may be essential for surgical planning
cordantly to the left ventricle and the anatomic left atrium (Fontan procedure vs. two-ventricle repair) (41). After first-
(situated right of midline) connects to the right ventricle. stage repair (bidirectional Glenn), and Fontan procedure,
Mirror-image dextrocardia occurs with situs inversus. In cine MRI is used to monitor the volumes and ejection frac-
contrast, isolated dextrocardia consists of situs solitus with a tion of the dominant ventricle (23,42–44). Blood flow to the
right-sided cardiac apex (see Fig. 29.48). two lungs is assessed anatomically with contrast-enhanced
MRA and quantified with velocity-encoded (phase-contrast)
cine MRI (23,45).
Double-Inlet Ventricle (Single Ventricle;
Univentricular Heart)
Atresia of Atrioventricular Valve
This complex anomaly consists of one adequate-sized ven-
tricle and one rudimentary ventricle (41,42). The most fre- Either the tricuspid or mitral valves can be atretic; tricus-
quent type is double-inlet left ventricle: Both atrioventri pid atresia is more frequent. Tricuspid atresia is associated
cular valves connect to the morphologic left ventricle. The with d-TGA in nearly 50% of patients. It is also frequently
enlarged left ventricle is attached to a rudimentary RV outlet associated with valvular or subvalvular pulmonary steno-
via a bulboventricular foramina. In many instances, l-TGA sis or atresia. The right ventricle is invariably hypoplastic.
is an associated anomaly. Transaxial MR images demonstrate a bar of muscle and fat
Cine MRI can provide a set of end-diastolic and end-sys- across the atrioventricular inlet of the right ventricle (46)
tolic images encompassing the entire heart; consequently, it (Fig. 29.49–29.51). Both spin-echo and cine MR images
LWBK1209-ch29_p431-470.indd 460 16/05/13 9:29 PM

Figure 29.49. Classic Fontan procedure. Axial black blood images from the level of the great vessels (top left)
to the midventricular level (bottom right) in a patient with tricuspid atresia status postclassic Fontan opera-
tion. The right pulmonary artery (R) connects to the right atrial appendage (Ap). A dominant left ventricle
(LV) is seen. Note the fatty replacement of the right atrioventricular grove (arrow). L, left pulmonary artery;
Ao, aorta; RA, right atrium.
display the degree of RV hypoplasia, interatrial communica- (superior vena cava), Fontan conduit (IVC to pulmonary ar-
tion, VSD, and ventriculoarterial connections. In a few cases, tery conduit), and pulmonary arteries (23,44,45).
a thick membrane separates the right atrium and right ven-
tricle (imperforate tricuspid valve). The atrial and ventricular
Additional Cyanotic Lesions
septal openings permit systemic venous blood to flow into
the ventricles.
Ebstein Malformation
MRI is frequently used to evaluate tricuspid atresia after
various stages of surgical correction. Spin-echo MR images Ebstein malformation is a primary abnormality of the tricus-
and 3D contrast-enhanced MRA are effective for display- pid valve in which the septal and anterior leaflets of the valve
ing the anatomy of the Glenn anastomosis (Figs. 29.51 and adhere to the right ventricle wall (47). The leaflets become free
29.52) and the Fontan procedure and the status of the cen- at a variable distance, at a location more apical than usual, so
tral pulmonary arteries (Figs. 29.51–29.53). VEC cine MRI that the tricuspid valve orifice is displaced toward the apex.
may be employed to quantify blood flow in the Glenn shunt The right atrioventricular ring still defines the border of the
LWBK1209-ch29_p431-470.indd 461 16/05/13 9:29 PM

anatomic right ventricle; however, because the valve orifice is

more apical than usual, the functional part of the right ven-
tricle is truncated. The portion of the right ventricle that is
basal to the valve orifice becomes “atrialized,” meaning that
it functions as part of the atrium rather than as part of the
right ventricle. The atrialized portion of the right ventricle is
thin walled and nontrabeculated, and it may become dilated.
Ebstein anomaly is frequently associated with an ASD.
MRI can demonstrate the position of tricuspid valve
(48) (Fig. 29.54), and cine MRI can be used to quantify
chamber size and the ejection fraction of the functional
right ventricle. Cine MRI also demonstrates the systolic
signal void projected into the right atrium, indicating tri-
cuspid regurgitation, which is invariably present with this
anomaly.
Hypoplastic Left Heart Syndrome

The term hypoplastic left heart syndrome refers to several
different anomalies, all of which lead to underdevelopment
of the left ventricle. It is usually caused by aortic stenosis
or atresia, mitral stenosis or atresia, or both. As in tricuspid
atresia, the degree of ventricular hypoplasia varies, depend-
Figure 29.50. Tricuspid atresia. Spin-echo axial image shows a bar ing on the location and severity of the obstruction. For
of fat and muscle (arrow) separating the right atrium (RA) from the example, in mitral atresia without a large ASD and VSD,
hypoplastic right ventricle (RV). The left ventricle (LV) is enlarged. the left ventricle may be a small mass of muscle with no
Figure 29.51. Tricuspid atresia

with Fontan. Axial images, from the
ventricular level (bottom right) to the
level of the great vessels (top left),
show a Fontan conduit (F) connect-
ing the inferior vena cava (IVC) to
the right pulmonary artery (RPA). A
Glenn shunt (G) connects the superior
vena cava to the RPA. A dominant left
ventricle (LV) is seen. LPA, left pulmo-
nary artery; RA, right atrium; LA, left
atrium.
LWBK1209-ch29_p431-470.indd 462 16/05/13 9:29 PM

Figure 29.52. Glenn shunt. Coronal (upper left) and axial spin-echo images from cranial (top right) to
caudal (bottom right). The superior vena cava (S) is attached to the right pulmonary artery, with flow to both
pulmonary arteries (bidirectional Glenn shunt). PA, pulmonary artery.
visible lumen (Fig. 29.55). On the other hand, if the mitral diameters of the great arteries are clearly defined on axial
valve is patent and the problem is primarily one of aortic MR images.
valvular hypoplasia, the left ventricle may be normal in size Hypoplastic left heart syndrome is treated by the
and hypertrophied. In all cases, the right atrium tends to be Norwood procedure, in which the large right ventricle is
enlarged and the right ventricle to be dilated and hypertro- made to pump blood to the systemic circulation. This is ac-
phied. Axial MRI can readily depict the chamber enlarge- complished by severing the main pulmonary artery and anas-
ment and ventricular hypertrophy. tomosing the proximal pulmonary stump to the ascending
In most cases of hypoplastic left heart syndrome, little aorta. The pulmonary circulation is re-established by routing
blood flows through the ascending aorta. Blood tends to flow systemic blood into the pulmonary circulation, initially with
from the pulmonary artery through the ductus arteriosus a systemic to pulmonary shunt and later with a Fontan pro-
into the aorta, then in retrograde fashion to the aortic root cedure. An important role for MRI in hypoplastic left heart
to supply the coronary arteries. As a result, the ascending syndrome has been the evaluation of the morphology (43)
aorta is usually very small and the main pulmonary artery, and function of the various stages of the Norwood procedure
which receives most of the cardiac output, is very large. The (23,44).
LWBK1209-ch29_p431-470.indd 463 16/05/13 9:29 PM

Figure 29.53. Fontan procedure, extra-atrial type. Spin-echo

coronal image shows the conduit (arrows) between the inferior vena
cava and the right pulmonary artery (RP). The superior vena cava (S)
Figure 29.54. Ebstein anomaly. Gradient-echo axial image shows
that the attachments of the septal leaflets (black arrow) of the tricuspid
is connected to the top of the right pulmonary artery (Glenn anasto-
valve are displaced into the right ventricle. The level of the tricuspid
mosis). The conduit is located lateral to the right atrium (R), thus the
annulus (white arrow) is the normal site of attachment. Right-sided
connotation extra-atrial Fontan.
chambers are enlarged due to tricuspid regurgitation. RA, right atrium;
aRV, atrialized right ventricle; fRV, functional right ventricle.
Total Anomalous Pulmonary Venous

venous connection (TAPVC). TAPVC is usually classified
Connection
according to the location of the venous insertion. It may
In the embryo, the pulmonary veins grow from the lung buds be supracardiac (type I), in which the pulmonary venous
and come together to form a confluence, which normally confluence usually drains into the superior vena cava or an
is incorporated into the posterior wall of the left atrium. anomalous left-sided “vertical vein,” which in turn drains
However, if this confluence joins the circulatory system else- into the left brachiocephalic vein. TAPVC at the cardiac
where, a situation is created in which the entire pulmonary level (type II) usually drains into the right atrium. In type
venous return is anomalous: Total anomalous pulmonary III TAPVC, the pulmonary venous confluence drains below
Figure 29.55. Hypoplastic left heart syndrome. Spin-echo axial images demonstrate the severely hypoplastic
ascending aorta (arrow) and lack of cavitation of the left ventricle (L) at the level displayed.
LWBK1209-ch29_p431-470.indd 464 16/05/13 9:29 PM

Figure 29.56. Right-sided isomerism. Spin-echo axial image at the Figure 29.57. Left-sided isomerism. Spin-echo axial image at the
level of the main bronchi (arrows) shows that both pulmonary arteries level of the carina (arrow). Both pulmonary arteries course above the
are situated ventral to the bronchi. main bronchi.
the diaphragm into the truncus venosus or the portal venous syndrome. The liver is large and crosses the midline, and the
system. As the pulmonary veins pass below the diaphragm, spleen is often absent.
obstruction to flow frequently results in pulmonary venous In left-sided isomerism, both lungs have two lobes and
hypertension. both the right and left pulmonary arteries pass over their
MRI has been shown to be highly accurate for the diag- respective bronchi. This type of bronchial pattern is known
nosis of anomalous pulmonary venous connection. TAPVC as hyparterial because the bronchus passes below the pul-
can be diagnosed when MRI demonstrates that no pulmo- monary artery (Fig. 29.57). Left-sided isomerism is associ-
nary veins drain into the left atrium. Transaxial MRI and ated with polysplenia syndrome, characterized by multiple
gadolinium-enhanced MRA can demonstrate the location spleens. The IVC is frequently interrupted with azygos con-
of the anomalous connection along with enlargement of tinuation (Fig. 29.58).
the superior vena cava or coronary sinus, or the presence Coronal and transaxial MR images are very effective for
of a common pulmonary vein posterior to or above the left showing the anatomy of the pulmonary arteries in order to
atrium. establish the diagnosis of isomerism (49). MRI also clearly
defines the systemic, pulmonary venous anatomy and the
morphology of the atrial appendages.
Situs Abnormalities
The right-sided abdominal and thoracic structures are usu- Postoperative Evaluation
ally on the same side as the morphologic right atrium; like-
wise, the left-sided organs are usually on the same side as MRI is excellent for the postoperative evaluation of CHD
the morphologic left atrium. In a developmental anomaly (see Chapter 30). The wide field of view permits effec-
in which both atria have morphologic features of a right tive depiction of complex repairs used to correct complex
atrium (RA isomerism), the visceral and thoracic structures lesions. A description of many of the surgical procedures
tend to be right sided. Likewise, bilateral left atria (LA isom- used for palliation and correction of cyanotic heart disease
erism) are associated with bilateral left-sided visceral and are described in Table 29.4.
thoracic structures. Contrast-enhanced 3D MRA and tomographic images
The bronchi and pulmonary arteries have a characteristic demonstrate surgical anastomoses, conduits, and shunts.
relationship in isomerism syndromes. In RA isomerism, both Moreover, the 3D data set available from contiguous cine MR
lungs usually have three lobes and both bronchi have the images at multiple phases in the cardiac cycle permits precise
pattern of a right bronchus, that is, the pulmonary artery quantification of left and right ventricular volumes, mass, and
runs in front of and beneath the main bronchus (a type of function. Since cine MRI has high interstudy reproducibility
bronchus known as an eparterial bronchus) (Fig. 29.56). (50–52), it is the best technique available for serial monitoring
Right-sided isomerism is often associated with asplenia of ventricular volumes, mass, and function after surgery.
LWBK1209-ch29_p431-470.indd 465 16/05/13 9:29 PM

Figure 29.58. Polysplenia with interrupted inferior vena cava (IVC). Multiple axial cine MR images from
the great vessels (top left) to the liver (bottom left) demonstrate a single ventricle (SV) and single atrium (SA).
Note multiple hepatic veins (arrows) without an intrahepatic inferior vena cava (IVC). Azygos vein (Az) is
dilated. Bilateral morphologic left pulmonary arteries (top left). Multiple spleens (asterisks) are seen in the
axial black blood image (bottom right). Ao, aorta.
Spin-echo MRI and 3D contrast-enhanced MRA have the Rastelli procedure for pulmonary atresia (Fig. 29.61);
demonstrated the anatomy of complex repairs such as the and various repairs of TOF (see Figs. 29.1, 29.2, 29.61, and
Norwood procedure for hypoplastic left heart (Fig. 29.59); 29.62).
the Jatene procedure (arterial switch) for TGA (Fig. 29.60); VEC cine MRI is employed for quantification of the vol-
Senning and Mustard procedures (atrial baffle redirection of ume and fraction of pulmonary regurgitation (53) (see Fig.
venous flow) for TGA (Fig. 29.44); the Fontan procedure 29.2), volume flow through various components of the Fontan
for functional single ventricle (see Fig. 29.49 and 29.51); repair (23), and differential flow in the central pulmonary
LWBK1209-ch29_p431-470.indd 466 16/05/13 9:29 PM

urgical Procedures for

S arteries (see Fig. 29.30). After various surgical procedures
in cyanotic CHD, residual stenosis of the central pulmonary
Palliation and Correction of
Table 29.4 arteries is frequent. VEC cine MRI can be used to demon-
Cyanotic Congenital Heart strate disparity in flow between the right and left pulmo-
Disease nary arteries and changes after angioplasty.
TETRALOGY OF FALLOT
Palliation: Blalock–Taussig shunt (graft or anastomosis of Monitoring of Ventricular Function
subclavian to pulmonary artery) and Pulmonary Regurgitation in
Waterston shunt (anastomosis of ascending aorta to right Postoperative Tetralogy of Fallot
pulmonary artery)
Potts shunt (anastomosis of descending aorta to left pulmonary Several surgical procedures are employed in total correc-
artery) tion of TOF (Table 29.4). Correction may be preceded
Central shunt (synthetic tube graft from the ascending aorta to by surgically creating a connection between the aorta and
main pulmonary artery) one of the central pulmonary arteries. This is usually done
Correction: RV infundibulectomy with or without RV outflow or with a graft between the aorta or subclavian artery to the
pulmonary transannular patch main or a central pulmonary artery. If the graft originates
Closure of VSD in the ascending aorta, it is called a central shunt. If the
Rastelli procedure (tube graft from the RV to distal MPA or central graft is between a subclavian artery and central pulmo-
pulmonary arteries)
nary artery, it is called a modified Blalock–Taussig shunt.
PULMONARY ATRESIA WITH VSD (EXTREME TETRALOGY Alternatively, the subclavian artery contralateral to the
OF FALLOT) aortic arch is directly anastomosed to a central pulmonary
Palliation: Same as tetralogy of Fallot artery. At total correction, the shunt is closed and the VSD
Correction: Rastelli procedure with or without previous
is closed. The RVOT obstruction is removed using one or
unifocalization of multiple pulmonary central and segmental
arteries and systemic to pulmonary collateral arteries.
more of several techniques: Infundibulectomy, transan-
nular patch, or conduit from RV to the main or bifurca-
TRANSPOSITION OF GREAT ARTERIES tion of the pulmonary artery (Figs. 29.1, 29.2, 29.61, and
Palliation: Blalock–Hanlon procedure (surgical excision of portion
29.62).
of atrial septum)
Balloon septostomy (transcatheter procedure)
MRI is now used to identify, assess the severity, monitor
Correction: Jatene procedure (arterial switch) any residual abnormality(ies), and to quantify RV function
Mustard and Senning procedures (intra-atrial baffle redirecting (53–55). A typical study involves the following: Axial spin-
superior and inferior vena flow across mitral valve). echo or cine MR images to evaluate anatomy; cine MRI in
DOUBLE-OUTLET RIGHT VENTRICLE
the short-axis plane to quantify RV volumes and ejection
Correction: Intraventricular baffle connecting LV to aorta and fraction; cine MRI in the RVOT plane to identify sites of
Rastelli procedure connecting RV to pulmonary artery. RV infundibular, valvular, or supravalvular stenosis(es);
contrast-enhanced MRA in coronal plane to evaluate ste-
TRICUSPID ATRESIA
Glenn procedure (SVC to RA anastomosis) followed later with
nosis of main, central, or branch pulmonary arteries or in
Fontan procedure (conduit from IVC to RPA). conduits; VEC (phase-contrast) cine MRI perpendicular to
direction of flow in main, right, and left pulmonary arteries
FUNCTIONAL SINGLE VENTRICLE
to quantify pulmonary regurgitation and differential pulmo-
Hypoplastic left heart; hypoplastic right heart; double-inlet ventricle
Norwood and Damus procedures (anastomosis of proximal
nary flow to the lungs. Delayed contrast-enhanced inversion
pulmonary artery to distal ascending aorta)—early recovery gradient-echo sequence may be done to identify
Fontan procedure—later fibrosis (scar) at site of RV outlet resection (Fig. 29.63) or
elsewhere in the ventricles (56).
Figure 29.59. Norwood proce-

dure. Spin-echo axial (left) and sagittal
(right) images show the anastomosis
of the proximal pulmonary artery (P)
to the ascending aorta (A). Sagittal
view shows the pulmonary artery (P)
connected to a reconstructed distal
ascending aorta (A) and arch. Blood
flow to the distal pulmonary artery is
reconstituted with a central shunt (not
shown).
LWBK1209-ch29_p431-470.indd 467 16/05/13 9:29 PM

Figure 29.60. Jatene procedure.

Spin-echo images in the sagittal (left)
and axial (right) planes show the arte-
rial correction. The aorta (Ao) above
the anastomosis is situated between the
right and left pulmonary arteries. This
position can cause narrowing of the
branch pulmonary arteries. PA, pulmo-
nary artery.
The most frequent residual abnormality is moderate-to- Residual or recurrent stenosis of the RV infundibu-
severe pulmonary regurgitation. This nearly always presents lum, pulmonary valve, or valved conduit is also a frequent
after any repair but is particularly problematic after repair residual lesion. There may also be both residual stenosis and
by transannular patch. Severity of regurgitation is stated regurgitation. If stenosis is the dominant lesion, then the RV
as follows using regurgitation fraction: Less than 20% is shows hypertrabeculation and hypertrophy. Moreover, the
mild; 20% to 40% is moderate; greater than 40% is severe. combination of stenosis and regurgitation produces less se-
Moderate and severe degrees are always associated with vere RV enlargement and better preservation of RV ejection
enlarged RV end-diastolic (>90 mL/m2) and end-systolic fraction.
volumes (>60 mL/m2) (Fig. 29.64). End-diastolic volumes Regional contraction abnormalities of the RV are frequent
greater than 150 to 165 mL/m2 and/or end-systolic volume after repair of tetralogy. Akinesis or dyskinesis of the RV out-
greater than 90 to 100 mL/m2 are associated with reduced let region can be displayed on cine MR images in the sagittal
ejection fraction (<40%) and may prompt consideration of or RVOT planes. After placement of a generous outflow tract
pulmonary valve replacement. The preoperative RV volumes patch there may be aneurysm of this region (Fig. 29.62). The
as quantified by cine MRI are the major determinant of RV inlet and body portions of the RV show no or less severe
functional recovery after pulmonary valve replacement (57,58). contractile dysfunction.
Figure 29.61. Rastelli procedure.

Contrast-enhanced 3D MRA in oblique
axial (left) and sagittal (right) planes
shows a conduit (C) between the right
ventricle (R) and the pulmonary artery
(P). The oblique axial plane depicts a
stenosis (arrow) of the right pulmonary
artery.
LWBK1209-ch29_p431-470.indd 468 16/05/13 9:29 PM

Figure 29.62. Tetralogy of Fallot with right ventricular outflow Figure 29.64. Tetralogy of Fallot with pulmonary regurgitation.
tract aneurysm. Short-axis cine MR image at end-diastole demon- Short-axis cine MR image demonstrates an enlarged right ventricle
strates aneurysm (A) of the right ventricular outflow tract and an over- (RV) due to the volume overload. The left ventricle (LV) is normal
all enlarged right ventricle (RV). in size.
Adult Congenital Heart Disease in the adult. The most frequent is PAPVC, which is read-
ily shown on contrast-enhanced MRA in the coronal plane.
MRI is frequently used in the evaluation of CHD in grown-up Occasionally, an ASD is not recognized until adulthood.
children and adults (see Chapter 31). The major advantage The most frequent use of MRI in adult CHD is to moni-
for this purpose is precision and interstudy reproducibility of tor the severity of a residual lesion after corrective surgery
cine MRI for monitoring right and left ventricular volumes and to quantifying ventricular function. As discussed above,
and ejection fraction. In addition, VEC (phase-contrast) MRI MRI is now universally employed to monitor pulmonary
is important for quantifying valvular regurgitation and flow regurgitation and RV function in postoperative TOF.
in the pulmonary circulation. There are several other complex operations for which
Infrequently, MRI is used to initially identify lesions, monitoring of ventricular function is important. These are
which have gone undetected in childhood. These are usually operations or lesions in which the right ventricle connects
lesions, which have caused right-sided chamber enlargement to the aorta and supports the systemic circulation. These
Figure 29.63. Tetralogy of Fallot

after transannular patch. Cine MRI
(left) and delayed enhancement (right)
images in the short-axis plane show
thinning of the right ventricular free
wall (arrowheads) with associated
delayed enhancement (arrows). These
findings are consistent with myocardial
scar related to the surgical procedure.
RV, right ventricle; LV, left ventricle.
LWBK1209-ch29_p431-470.indd 469 16/05/13 9:29 PM

include the following: Atrial switch (Senning and Mustard 28. Beekman RP, Beek FJ, Meijboom EJ. Usefulness of MRI for the pre-operative evaluation of
the pulmonary arteries in tetralogy of Fallot. Magn Reson Imaging. 1997;15:1005–1015.
procedures) for TGA; corrected transposition; and single 29. Holmqvist C, Hochbergs P, Bjorkhem G, et al. Pre-operative evaluation with MR in tetralogy
ventricle with RV configuration. The end stage of some of Fallot and pulmonary atresia with ventricular septal defect. Acta Radiol. 2001;42:63–69.
30. Kersting-Sommerhoff BA, Sechtem UP, Higgins CB. Evaluation of pulmonary blood supply by
operated CHD is right or biventricular failure. MRI is nuclear magnetic resonance imaging in patients with pulmonary atresia. J Am Coll Cardiol.
now used to monitor ventricular function in order to rec- 1988;11:166–171.
ognize the need for reoperation in order to avert such an 31. Powell AJ, Chung T, Landzberg MJ, et al. Accuracy of MRI evaluation of pulmonary blood
supply in patients with complex pulmonary stenosis or atresia. Int J Card Imaging. 2000;
outcome. 16:169–174.
32. Geva T, Greil GF, Marshall AC, et al. Gadolinium enhanced 3D MRA of pulmonary blood
supply in patients with complex pulmonary stenosis or atresia: Comparison with x-ray angi-
ography. Circulation. 2002;106:473–478.
References 33. Freedom RM, Bawson JB, Yoo SJ, et al. Double-outlet right ventricle. In: Freedom
RM, Bawson JB, Yoo SJ, Benson LN, eds. Congenital Heart Disease: A Textbook of
1. Brenner LD, Caputo GR, Mostbeck G, et al. Quantification of left-to-right shunts with Angiocardiography. Vol. 2. Armonk, NY: Futura Publishing Co., Inc., 1997:1119–1169.
velocity-encoded cine nuclear magnetic resonance imaging. J Am Coll Cardiol. 1992;20: 34. Freedom RM, Bawson JB, Yoo SJ, et al. The segmental and sequential approach to con-
1246–1250. genital heart disease. In: Freedom RM, Bawson JB, Yoo SJ, Benson LN, eds. Congenital
2. Reddy GP, Higgins CB. Congenital heart disease: Measuring physiology with MRI. Semin Heart Disease: A Textbook of Angiocardiography. Vol. 1. Armonk, NY: Futura Publishing
Roentgenol. 1998;33:228–238. Co., Inc.; 1997:95–120.
3. Hundley WG, Li HF, Lange RA, et al. Assessment of left to right shunting by velocity 35. Freedom RM, Bawson JB, Yoo SJ, et al. Corrected transposition of the great arteries
encoded, phase difference MR imaging: A comparison with oximetric and indicator dilu- (Atrioventricular and ventriculoarterial discordance). In: Freedom RM, Bawson JB, Yoo
tion techniques. Circulation. 1995;91:2955–2960. SJ, Benson LN, eds. Congenital Heart Disease: A Textbook of Angiocardiography. Vol. 2.
4. Arheden H, Holmquist C, Thilen U, et al. Left to right cardiac shunts: Comparison of Armonk, NY: Futura Publishing Co., Inc.; 1997:1071–1117.
measurement with MR velocity mapping and radionuclide angiography. Radiology. 1999; 36. Yoo SJ, Lim TH, Park IS, et al. MR anatomy of ventricular septal defect in double-
211:453–458. outlet right ventricle with situs solitus and atrioventricular concordance. Radiology. 1991;
5. Beerbaum P, Korperich H, Esdorn H, et al. Atrial septal defects in pediatric patients: 181:501–505.
Noninvasive sizing with CMR. Radiology. 2003;228;361–369. 37. Mayo JR, Roberson D, Higgins CB. MR imaging of double outlet right ventricle. J Comput
6. Durongpisitkul K, Tang NC, Soongswang J, et al. Predictors of successful transcatheter Assist Tomogr. 1990;14:336–342.
closure of atrial septal defect size and morphology by CMR imaging. Pediatr Cardiol. 38. Niezen RA, Beekman RP, Helbring WA, et al. Double outlet right ventricle assessed by
2004;25;124–130. MRI. Int J Card Imaging. 1999;15:323–329.
7. Ferrari VA, Scott CH, Holland GA, et al. Ultrafast three-dimensional contrast-enhanced 39. Donnelly LF, Higgins CB. MR imaging of conotruncal abnormalities. AJR Am J Roentgenol.
magnetic resonance angiography and imaging in the diagnosis of partial anomalous pul- 1996;166:925–928.
monary venous drainage. J Am Coll Cardiol. 2001;37:1120–1128. 40. Collett RW, Edwards JE. Persistent truncus arteriosus: A classification according to ana-
8. Masui T, Seelos KC, Kersting-Sommerhoff BA, et al. Abnormalities of the pulmonary tomic types. Surg Clin North Am. 1949;29:1245–1270.
veins: Evaluation with MR imaging and angiography and echocardiography. Radiology. 41. Anderson RH, Ho SY. Which hearts are unsuitable for biventricular correction? Ann
1991;181:645–649. Thorac Surg. 1998;66:621–626.
9. Didier D, Higgins CB. Identification and localization of VSDs by gated magnetic resonance 42. Fellows KE, Fogel MA. MR imaging and heart function in patients pre- and post-Fontan
imaging. Am J. Cardiol. 1986;57:1363–1368. surgery. Acta Paediatr Suppl. 1995;410:57–59.
10. Bremerich J, Reddy GP, Higgins CB. MRI of supracristal ventricular septal defects. J 43. Kondo C, Hardy C, Higgins SS, et al. Nuclear magnetic resonance imaging of the palliative
Comput Assist Tomogr. 1999;23:13–15. operation for hypoplastic left heart syndrome. J Am Coll of Cardiol. 1991;18:817–823.
11. Schmidt M, Theissen P, Deutsch HJ, et al. Magnetic resonance imaging of ductus arteriosus 44. Brown DW, Gauvreau K, Powell A, et al. Cardiac MR versus routine cardiac catheteriza-
Botalli apertus in adulthood. Int J Cardiol. 1999;68:225–229. tion before bidirectional Glenn anastomosis in infants with single ventricle. Circulation.
12. Wang ZJ, Reddy GP, Gotway MB, et al. Cardiovascular Shunts: MR imaging evaluation. 2007;116:2718–2725.
Radiographics. 2003;23:S181–S194. 45. Fratz S, Hess J, Schwaiger M, et al. More accurate quantification of pulmonary blood
13. Freedom RM, Bawson JB, Yoo SJ, et al. Atrioventricular septal defect. In: Freedom flow by MRI than by lung perfusion scintigraphy in patients with Fontan. Circulation.
RM, Bawson JB, Yoo SJ, Benson LN, eds. Congenital Heart Disease: A Textbook of 2002;106:1510–1513.
Angiocardiography. Vol. 1. Armonk, NY: Futura Publishing Co., Inc; 1997:133–188. 46. Fletcher BD, Jacobstein MD, Abramowsky CR, et al. Right atrioventricular valve atresia:
14. Parsons JM, Baker EJ, Anderson RH, et al. Morphological evaluation of atrioventricular Anatomic evaluation with MR imaging. AJR Am J Roentgenol. 1987;148:671–674.
septal defects by magnetic resonance imaging. Br Heart J. 1990;64:138–145. 47. Freedom RM, Bawson JB, Yoo SJ, et al. Ebstein’s malformation. In: Freedom RM, Bawson
15. Steffens JC, Bourne MW, Sakuma H, et al. Quantification of collateral blood flow in JB, Yoo SJ, Benson LN, eds. Congenital Heart Disease: A Textbook of Angiocardiography.
coarctation of the aorta by velocity encoded cine magnetic resonance imaging. Circulation. Vol. 1. Armonk, NY: Futura Publishing Co., Inc.; 1997:349–366.
1994;90:937–943. 48. Eustace S, Kruskal JB, Hartnell GG. Ebstein’s anomaly presenting in adulthood: The role
16. Araoz PA, Reddy GP, Tarnoff H, et al. MR findings of collateral circulation are more of cine magnetic resonance imaging in diagnosis. Clin Radiol. 1994;49:690–692.
accurate measures of hemodynamic significance than arm-leg blood pressure gradient after 49. Applegate KE, Goske MJ, Pierce G, et al. Situs revisited: Imaging of the heterotaxy syn-
repair of coarctation of the aorta. J Magn Reson Imaging. 2003;17:177–183. drome. Radiographics. 1999;19:837–852; discussion 853–854.
17. Mohiaddin RH, Kilner PJ, Roes S, et al. MR volume flow and jet velocity mapping in aortic 50. Semelka RC, Tomei E, Wagner S, et al. Normal left ventricular dimensions and function:
coarctation. J Am Coll Cardiol. 1993;22:1515–1521. Interstudy reproducibility of measurements with cine MR imaging. Radiology. 1990;174:
18. Konen E, Marchant N, Provost Y, et al. Coarctation of the aorta before and after correc- 763–769.
tion: The role of cardiovascular MRI. Am J. Roentgenol. 2004;182:1333–1339. 51. Pattymana PM, Lamb H, Van del Velde EA, et al. Left ventricular measurement with
19. Nielsen JC, Powell AJ, Gauvreau K, et al. Magnetic resonance imaging predictors of coarc- spin echo MR imaging: A study of reproducibility with variance component analysis.
tation severity. Circulation. 2005;111:622–628. Radiology. 1993;187:261–268.
20. Oshinski JN, Parks WJ, Markow CP, et al. Improved measurement of pressure gradient by 52. Grothues F, Moon JC, Bellenger WG. Interstudy reproducibility of right ventricular vol-
MRI. J Am Coll Cardiol. 1996;28:1818–1826. umes, function and mass with cardiovascular MR. Am Heart J. 2004;147:218–223.
21. Bisset GS 3rd, Strife JL, Kirks DR, et al. Vascular rings: MR imaging. AJR Am J Roentgenol. 53. Rebergen SA, Chin JGJ, Ottenkamp J, et al. Pulmonary regurgitation in the late postopera-
1987;149:251–256. tive follow-up of tetralogy of Fallot: Volumetric quantification by MR velocity mapping.
22. Kertsing-Sommerhoff BA, Sechtem UP, Fisher MR, et al. MR imaging of congenital anoma- Circulation. 1993;88:2257–2266.
lies of the aortic arch. AJR Am J Roentgenol. 1987;149:9–13. 54. Geva T, Sandweiss BM, Gauvreau, et al. Factors associated with impaired clinical status
23. Fogel MA, Hubbard AM, Fellows KE, et al. MRI for physiology and function in congenital in long term survivors of tetralogy of Fallot repair evaluated by MR. J Am Coll. Cardiol.
heart disease: Functional assessment of the heart preoperatively and postoperatively. Semin 2004;14:1068–1074.
Roentgenol. 1998;23:239–251. 55. Niezen RA, Helbing WA, Van der Wall EE, et al. Biventricular systolic function and mass
24. Reid SA, Walker PG, Fisher J, et al. The quantification of pulmonary valve hemodynamics studied with MRI in children with pulmonary regurgitation after repair of tetralogy of
using MRI. Int J Cardiovasc. Imaging. 2002;18:217–225. Fallot. Radiology. 1996;201:135–140.
25. McConnell MV, Ganz P, Selwyn A, et al. Identification of anomalous coronary arteries and 56. Babu-Narayan SV, Kilner PJ, Li W, et al. Ventricular fibrosis suggested by cardiovascular
their anatomic course by MR coronary angiography. Circulation. 1995;92:3158–3162. MR in adults repaired tetralogy of Fallot and its relationship to adverse markers of clinical
26. Post JC, Van Rossum AC, Bronzwaer A, et al. MR angiography of anomalous coronary outcome. Circulation. 2006;113:405–413.
arteries. Circulation. 1995;92:3163–3171. 57. Vliegen HW, Van Straten A, de Roos A, et al. MRI to assess the hemodynamic effects of
27. Freedom RM, Bawson JB, Yoo SJ, et al. Tetralogy of Fallot and pulmonary atresia and pulmonary valve replacement in adults late after repair of tetralogy of Fallot. Circulation.
ventricular septal defect. In: Freedom RM, Bawson JB, Yoo SJ, et al., eds. Congenital Heart 2002;106:1703–1707.
Disease: A Textbook of Angiocardiography. Vol. 1. Armonk, NY: Futura Publishing Co., 58. Van Straten A, Vliegen HW, Hazelcamp MG, et al. Right ventricular function after pulmo-
Inc.; 1997:493–533. nary valve replacement in patients with tetralogy of Fallot. Radiology. 2004;233:824–829.
This Chapter is modified from Chapter 31 of W.R. Webb and C.B. Higgins Thoracic Imaging,
Second Edition. Lippincott Williams Wilkins, Philadelphia, 2011.
LWBK1209-ch29_p431-470.indd 470 16/05/13 9:29 PM

Chapter
30
Arno A.W. Roest
Lucia J.M. Kroft
Albert de Roos
Magnetic Resonance Imaging of

Function and Flow in Postoperative
■■ Technical Options The focus of this chapter is to discuss the current role
Black Blood Imaging of MRI in the evaluation of patients with CHD after cor-
Balanced Gradient-echo Imaging and SENSE rection or palliation. This is of paramount importance as
Technology many patients have residual defects after intervention or
will develop cardiovascular dysfunction during follow-up.
Flow Mapping
Technical options that are available with MRI are described,
Contrast-enhanced Magnetic Resonance Imaging and the practical applications of these options will be dis-
Magnetic Resonance Imaging Stress Testing cussed for the four most common referrals: Tetralogy of
■■ Contribution of Magnetic Resonance Fallot (TOF), transposition of the great arteries (TGA), aor-
Imaging in Specific Entities tic coarctation, and the Fontan circulation.
Tetralogy of Fallot
Transposition of the Great Arteries Technical Options
Aortic Coarctation
Fontan Circulation A wide array of MRI techniques is available for detailed
determination of cardiac morphology and quantification of
■■ Future Perspective cardiac function. The technical aspects have been discussed
in more detail in other chapters of this book.
Congenital heart disease (CHD) survival rates have increased
dramatically because of improvements in diagnostic tools, Black Blood Imaging
surgical techniques, and postoperative care for patients with
Black blood imaging provides clear depiction of the cardiac
CHD during the last 40 years. Traditionally, these patients
anatomy. The electrocardiographically triggered multislice spin-
are monitored over time by clinical assessment, electro-
echo acquisition is performed during breath-holds. The black
cardiography, echocardiography, and x-ray angiography.
blood appearance is caused by a double-inversion pulse with
Transthoracic echocardiography is the most commonly used
an appropriate delay, creating a signal void for flowing blood
technique. However, after surgical intervention the imag-
(black) with high tissue-to-blood contrast. It provides two-
ing quality is often restricted because scar, bone, or lung
dimensional (2D) images in three different directions (sagittal,
tissue may interfere with the acoustic window, while chest
transverse, and coronal) and in any other direction tailored
deformations may be present as well. X-ray angiography
to the morphology (1). Thin-slice (3 to 5 mm) imaging is pre-
has advantages for vascular imaging quality, but overpro-
ferred to avoid partial volume effects. For example, the assess-
jection of structures, radiation burden, and invasiveness
ment of focal stenosis as in coarctation of the aorta is best
are important drawbacks. These limitations do not occur
accomplished with 3-mm slice thickness in double-oblique
with magnetic resonance imaging (MRI). With this imag-
planes along the course of the vessel of interest.
ing technique, both superior depiction of cardiac anatomy
and reproducible measurement of cardiac function can be
Balanced Gradient-Echo Imaging
achieved in one session. Over the last 15 years MRI has
and Sense Technology
been increasingly accepted as an important imaging tool for
assessment of postoperative cardiovascular anatomy and Gradient-echo balanced fast-field echo techniques can be
function in patients with corrected or palliated CHD. used to assess the function of both ventricles in a highly
471
LWBK1209-ch30_p471-488.indd 471 17/05/13 5:15 PM

LV LV
* *
RV RV
A B
Figure 30.1. Gradient-echo (A) and phase-contrast (B) magnetic resonance imaging (MRI) of a ventricu-
lar septal defect. The black void (asterisk) on the gradient-echo MRI image (A) is caused by the dephasing of
spins in the turbulent left-to-right shunt. The bright white pixels within the black shunt flow (B) indicate high
velocities (>250 cm/s) through a restrictive ventricular septal defect. The amount of left-to-right shunt can be
quantified by subtracting the aortic volume measurement from the pulmonary artery volume measurement,
obtained by phase-contrast MRI. LV, left ventricle; RV, right ventricle. (Reprinted from Roest AAW, Helbing
WA, van der Wall EE, et al. J Magn Reson Imaging. 1999;10:656–666, with permission.)
reproducible way (2–4). A stack of consecutive slices can be measurement of regurgitation and/or stenosis at any level in
applied in the transverse plane or in short-axis orientation, the heart. With phase-contrast MRI the flow velocity of blood
covering both ventricles. Endocardial and epicardial borders can be measured, based on velocity-induced phase shifts of
can then be traced by using a dedicated software package. moving protons in the presence of a magnetic field gradient.
Parameters like end-diastolic volume (EDV) and end-systolic Across the area of interest, flow volumes of both the forward
volume (ESV), stroke volume (SV), cardiac output, ejection flow and any regurgitation can be assessed, whereas peak
fraction (EF), and wall mass (the latter to assess the degree velocities can be used to estimate pressure gradients over ste-
of myocardial hypertrophy) can be obtained. Wall motion notic segments, using the modified Bernoulli equation. The
abnormalities can be easily depicted when these images are advent of three-dimensional (3D) flow mapping with retro-
displayed in movie-loop. Gradient-echo multislice multi- spective valve tracking allows more detailed evaluation of
phase MRI also allows identification of recurrent or resid- valvular function. Valve tracking is especially important to
ual septal defects by displaying a signal void at the site of study flow over the tricuspid en mitral valve, as they move
turbulent and/or high-velocity flow through a defect. The up to 24 mm during ventricular contraction (6). Based on
signal void is caused by dephasing of the protons within a this technique flow mapping can also provide information
jet, resulting in a local loss of signal intensity observed (Fig. on the diastolic filling pattern of the ventricle and allows the
30.1). Such flow voids can also indicate valvular regurgita- construction of a ventricular time–volume curve (Fig. 30.2)
tion or stenosis. (7). In patients with corrected TOF and pulmonary regurgi-
So-called SENSE (SENSitivity Encoding) technology allows tation (PR), right ventricular (RV) inflow curves have been
considerable scan-time reduction in most MRI techniques, in- used to demonstrate impaired relaxation and restriction to
cluding black blood spin-echo and balanced fast-field gradient- filling as markers of abnormal RV diastolic function (7).
echo imaging. Simultaneously operated coils are used to Shunt quantification in the case of a left-to-right shunt is pos-
improve the signal-to-noise ratio. New techniques like the sible by comparing the aortic and pulmonary net forward
recently introduced k-t broad-use linear acquisition speed-up flow (8–10). It can also be assessed by comparison of the RV
technique (k-t BLAST) will also provide increased spatial and and left ventricular (LV) SVs calculated from gradient-echo
temporal resolution, in which a complete stack of short-axis imaging, with the exception of the presence of a ventricular
images can be obtained in a single breath-hold (5). These septal defect (VSD), because RV and LV SVs appear to be the
faster imaging sequences will especially be of benefit for the same then. Furthermore, shunt quantification is complicated
pediatric population, because small children often have dif- when multiple shunts are present or with coexisting valvular
ficulties with holding their breath. regurgitation. More recently phase-contrast MRI was used to
assess myocardial velocity, providing information on regional
systolic and diastolic peak velocities and differences in tim-
Flow Mapping
ing of these events within the left and right ventricle, which
Assessment of valvular and vascular function with phase- is of importance if resynchronization therapy is considered
contrast MRI is another essential tool, allowing accurate (Fig. 30.3) (11,12).
LWBK1209-ch30_p471-488.indd 472 17/05/13 5:15 PM

Chapter 30 ■ Magnetic Resonance Imaging of Function and Flow in Postoperative Congenital Heart Disease 473
A B
C D E
Flow (mL/s)
Time (ms)
F G
Figure 30.2. The use of a 3D-velocity-encoded MR imaging sequence with retrospective valve track-
ing allows assessment of all four cardiac valves during one acquisition. Figure shows post-processing of flow
through the tricuspid valve (TV) of a healthy subject. A,B: Cine images show retrospective tracking of the TV
performed by positioning the plane of interest perpendicular to the flow through the TV during every cardiac
phase in four-chamber (A) and two-chamber (B) views. C–E: Velocity encoded MR images from three orthogo-
nal directions, reconstructed from the position of the valve determined with retrospective valve tracking based
on views in (A) and (B). AP, anterior–posterior velocity encoded; FH, feet-to-head velocity-encoded; RL, right–
left velocity encoded. F: Through-plane velocity-encoded MR images are obtained by reformatting of center
valvular plane in each cardiac phase. Inner border of tricuspid annulus is traced for flow analysis. Region within
free wall of right ventricle is traced for background correction. G: Tricuspid flow velocity (in time) is presented
in a curve, from which flow volume can be calculated. (Reprinted from van der Hulst AE, Westenberg JJ, Kroft
LJ, et al. Tetralogy of fallot: 3D velocity-encoded MR imaging for evaluation of right ventricular valve flow and
diastolic function in patients after correction. Radiology. 2010;256(3):724—734, with permission.)
Contrast-Enhanced Magnetic between bolus injection and bolus arrival in the vessel of
Resonance Imaging interest, a 2D gradient-echo MRA sequence is performed.
Three-dimensional reconstructions can subsequently be
made of the vessel, clearly depicting any region of interest.
Gadolinium-enhanced magnetic resonance angiography Gadolinium-enhanced MRA correlates excellently with con-
(MRA) is well suited to detect morphologic abnormalities ventional angiography (13).
of the great vessels, like pulmonary branch stenoses in TOF A rapidly advancing technique is coronary MRA, for
and coarctation of the aorta (Fig. 30.4). The gadolinium the depiction of the origin, course, and diameters of the
shortens the T1-relaxation time of blood, resulting in a high coronary arteries. Congenital or acquired coronary anoma-
signal intensity of the blood pool. This allows rapid 3D lies, such as after the arterial switch operation for TGA
acquisition of the entire thoracic aorta within one breath- and Kawasaki disease, can be detected without the need
hold. After a bolus-timing acquisition to determine the delay for the conventional catheter-based coronary angiography
LWBK1209-ch30_p471-488.indd 473 17/05/13 5:15 PM

Figure 30.3. Tissue velocity MRI of

the myocardium of the right ventricle.
A: Example of the anatomic location
of the myocardium of the RV free wall,
where measurements can be performed.
Region of interest (ROI; white oval also
on B,D) indicates myocardial segments
where tissue velocity measurements can
be performed. B: Example of the ana-
A B tomic location of the myocardium of the
right ventricular outflow tract (RVOT)
segment where measurements can be
performed. C: Velocity encoded image
of the four-chamber view. Velocity is
encoded in the longitudinal direction.
ROI is positioned at the basal RV free
wall in all cardiac phases. The average
velocity at the ROI is calculated in every
phase and can be displayed in a time–
velocity curve from which peak velocity
at the RV free wall and time to peak sys-
tolic velocity at the RV free wall can be
derived (E). D: Velocity encoded image
of the RVOT. Velocity is encoded in the
longitudinal direction. ROI is positioned
at the lateral wall of the RVOT in all
cardiac phases. The average velocity
at the ROI is calculated in every phase
C D and can be displayed in a time–velocity
curve from which peak velocity at the
RVOT and time to peak systolic velocity
at the RVOT can be derived (F).Time
in graphs shown as percentage of
RR-interval. Ao, aorta; LA, left atrium;
PA, pulmonary artery; RA, right atrium;
LV, left ventricle; RV, right ventricle.
(Reprinted from van der Hulst AE,
Roest AA, Delgado V, et al. Corrected
tetralogy of Fallot: Comparison of tissue
Doppler imaging and velocity-encoded
MR for assessment of performance and
temporal activation of right ventricle.
Radiology. 2011;260(1):88–97, with
E F permission.)
(14,15). A recent MRI study showed the feasibility in Delayed Enhancement Imaging
patients with TGA, with clear depiction of the coronary
ostia and proximal coronary arteries (Fig. 30.5). Coronary Myocardial delayed enhancement can be used to determine
MRA combined with delayed enhancement (described next) viable and nonviable myocardium within regions of inter-
can provide accurate insight into the myocardial perfusion est. Fifteen minutes after the administration of gadolinium,
status (15). the nonviable cardiac regions show hyperenhancement on
LWBK1209-ch30_p471-488.indd 474 17/05/13 5:15 PM

Figure 30.4. A 15-year-old male with

recent onset headache. Physical examination
revealed hypertension. Aortic coarctation was
suspected. Thin (A) and thick (B) maximum
intensity projection images of gadolinium
contrast-agent enhanced MR-angiography in
left anterior oblique coronal orientation. Focal A B
tight aortic coarctation on the classic location
distal from the left subclavian artery (arrow, A).
Note extensive collateral circulation bypassing
the coarctation as to supply the distal part of
the body with blood; hypertrophic mammar-
ian artery (MA, B) connecting to and supply-
ing the epigastric artery (EA) as collateral to
the distal part of the body, and hypertrophic
intercostal arteries (IA) serving as collaterals
to the aorta. Double-oblique transverse images
of the aortic valve (C,D) showing a bicuspid
slit-like opening aortic valve in anatomic
(C) and flow-sensitive acquisition (D). Patient
had severe aortic coarctation with associated
bicuspid aortic valve, and mild (poststenotic)
dilatation of the ascending aorta (AA, in
A,B). Patient was treated by coarctectomy C D
and graft placement.
T1-weighted images compared with the normal myocar- suited to assess global fibrosis and only macroscopic scar-
dium. This technique has already been proven useful in ring can be visualized. With the use of contrast-enhanced
patients with coronary artery disease, nonischemic cardio- T1-mapping diffuse interstitial fibrosis can be detected. In a
myopathies, and myocarditis. Recent reports indicate that heterogeneous group of patients with corrected CHD it was
it can be used in CHD as well, specifically in anomalies demonstrated that significantly more interstitial fibrosis was
associated with perfusion defects and congenital coronary present in the myocardium of the left ventricle (LV) as com-
artery anomalies (15). Delayed enhancement may be used pared to controls (17).
to identify postoperative scar tissue in the right ventricu-
lar outflow tract (RVOT) in patients with TOF. Fibrosis
in the RVOT appeared to be associated with RVOT
Stress Testing
dilatation, which adversely affected RV hemodynamics
(16). Delayed enhancement in the RV myocardium of a Detection of cardiac dysfunction at an early stage is important
patient with TGA after Mustard repair is illustrated in to allow timely (re-)intervention when necessary. Stress testing,
Figure 30.6. either pharmacologic or physical, may reveal subtle abnormal
ventricular function that may not be apparent at rest (18–20).
Pharmacologic stress testing using MRI has successfully been
T1-mapping
applied in the evaluation of ischemic heart disease (21), and in
Since with delayed enhancement imaging the inversion time CHD patients (22). Physical stress testing with MRI does not
is deliberately chosen to null “normal” myocardium it is not require infusion of inotropes and is feasible in children and
LWBK1209-ch30_p471-488.indd 475 17/05/13 5:15 PM

Figure 30.5. Coronary MRA in

patients with transposition of the great
arteries (TGA) after the arterial switch
procedure, clearly indicating coronary
ostia and proximal coronary arteries.
A: Oblique sagittal view demonstrates
an anomalous origin of the left cir-
cumflex artery from the right coronary
artery. B: Coronal view demonstrates
anomalous accessory left anterior
descending artery arising from the same
ostia as right coronary artery. No ostial
stenoses or proximal coronary artery
kinking is seen. (Both pictures pub-
lished with permission from AM Taylor,
Institute of Child Health and Great
Ormond Street Hospital for Children,
A B
London, UK.)
adults with CHD (18–20). Despite certain practical problems, CHD (23). Anterior displacement of the outflow septum
especially related to space limitations in the scanner when per- is the primary defect, resulting in a VSD, overriding of the
forming bicycle exercise and the induction of motion-related aorta, RVOT obstruction, and RV hypertrophy (24). In
artifacts that might degrade image quality, the use of an 1955 Lillehei et al. (25) introduced the intracardiac repair of
MRI-compatible bicycle ergometer and a dedicated exercise TOF, aiming at closure of the VSD and relief of the RVOT
protocol allows measurement of large vessel flow and biven- obstruction by resection or patch placement. This surgical
tricular function in response to exercise. For example, in a approach is now performed early in childhood mostly with-
study of TOF patients after correction exercise MRI revealed out prior placement of palliative shunts (26).
that although resting EF of the right ventricle (RV) was normal Despite good survival rates (27), residua and complica-
in 80% of patients, the majority of 93% showed an abnormal tions after TOF repair frequently occur. PR is a common
RV response to exercise. This is most likely to be caused by finding after surgical repair (28). Direct quantification is
longstanding volume overload because of PR (19). possible using phase-contrast MRI in the pulmonary artery
(Fig. 30.7). The resulting volume overload of the RV may
be tolerated for years without signs of RV failure. However,
longstanding volume overload of the RV may lead to se-
Contribution of Magnetic vere RV dilatation, which is associated with biventricular
Resonance Imaging in dysfunction (4,29,30) and ventricular arrhythmias (31).
Specific Entities Impaired exercise capacity can be observed with MRI during
physical stress testing, showing aggravation of RV dysfunc-
Tetralogy of Fallot tion compared with resting values (19,32). Pulmonary valve
replacement (PVR) resolves PR and subsequently improves
Clinical Considerations
cardiac hemodynamics (Figs. 30.8 and 30.9) (33,34). PVR
TOF is one of the most common types of cyanotic CHD, may lead to normalization of RV dimensions when it is per-
accounting for approximately 5.5% of all patients with formed before the RV reaches 150% of its normal size (35).
Figure 30.6. A 30-year-old male

patient with TGA after (intra-atrial)
Mustard repair. Axial fast gradient-echo
image (A) and 2D delayed enhance-
ment image (B). The right ventricle (RV)
sustains the systemic circulation and is
RV RV hypertrophied. Small focus with high
signal intensity on fast gradient-echo
imaging, probably representing focal
fatty degeneration of the right ventricular
(RV) myocardium (A). Delayed enhance-
ment showing a larger focus of enhance-
ment (B). Delayed enhancement repre-
sents increased interstitial volume and is
A B
associated with dead myocardial cells.
LWBK1209-ch30_p471-488.indd 476 17/05/13 5:15 PM

Figure 30.7. Short-axis images at end RV LV RV LV

diastole (A) and end systole (B) obtained with
a balanced fast-field echo MR sequence in a
patient after correction of tetralogy of Fallot
(TOF). Note the marked RV dilatation. LV,
A B
left ventricle.
Timing of PVR remains difficult, because the expected flow (derived from a pulmonary artery phase-contrast study)
favorable hemodynamic effects have to be weighed against by the RV EDV, representing the actual efficiency of the RV
the risk of repeated reoperations for homograft failure. A re- (34). The background of this is that traditional RV EF does
cent report showed a dramatic improvement of RV function not reflect the RV systolic function in patients with PR,
as expressed by RV EF corrected for PR and the RV ESV. RV because loading conditions are different and may change
EF was calculated in this study by dividing the net pulmonary over time (e.g., after PVR). The RV EF is not capable of dem-
onstrating these changes, because both RV SV and RV EDV
incorporate the regurgitant volume. Since RV SV is smaller
than RV EDV, the effect of PR on RV SV is relatively larger
Preoperative pulmonary flow
than the effect on RV EDV, leading to an overestimation
700 of the actual systolic function. Similarly, after relief of PR,
the percentual decrease of RV SV is larger than the per-
500 Systolic forward flow centual decrease of RV EDV. The RV EF will subsequently
decrease, whereas the actual RV systolic function improves,
300
expressed by a decrease of RV ESV (34).
Flow (mL/s)
100 Infundibular, valvular, or supravalvular pulmonary ste-

nosis (Fig. 30.10) can also occur, which can be depicted by
–100 200 400 600 800 contrast-enhanced MRA of the pulmonary vascular tree
Time (ms) as well as by flow mapping across the pulmonary valve.
–300
Pulmonary stenosis may result in RV hypertrophy as an
–500 adaptive mechanism. Since there is an inverse relationship
Diastolic regurgitation between RV mass and EF, this hypertrophy will eventually
A –700 impair RV contractility (36). Furthermore, the presence of
PR seems to aggravate RV dysfunction (37,38). RVOT an-
Postoperative pulmonary flow eurysms and fibrosis of the RVOT, which may follow after
500 RVOT reconstruction during initial repair, also have a nega-
tive impact on RV function (16,36,38). The importance of
400 Systolic forward flow fibrosis imaging using delayed enhancement in TOF patients
was furthermore stressed by the independent relation be-
300 tween RV LGE score and clinical arrhythmia (39).
Flow (mL/s)
With MRI, LV function has been identified as an indepen-

200 dent predictor for long-term prognosis after TOF repair as
well (40,41). LV function may be hampered because of ad-
100 verse RV-to-LV interaction, in which a pressure- or volume-
Time (ms) overloaded RV causes dysfunction of the interventricular
0 septum (36,42).
200 400 600 800 A recently recognized feature associated with TOF is
B –100 dilatation of the aortic root and related aortic regurgitation
(43). This phenomenon is probably caused by previous long-
Figure 30.8. Flow mapping in the pulmonary artery in a 33-year-
standing aortic volume overload, with increased aortic flow
old male patient with TOF with repair in the past. A: Severe diastolic
regurgitation of 73% before re-repair. Also, increased flow velocity was
attributable to the right-to-left shunting before repair (44).
measured with an estimated pressure gradient of 25 mm Hg (Bernoulli In addition to hemodynamic factors, intrinsic wall abnor-
equation), indicating pulmonary stenosis as well. B: Situation after malities of the aortic wall (similar to patients with Marfan
pulmonary valve replacement (PVR) with homograft: Normal pulmo- syndrome) have been observed in these patients as well (45).
nary artery flow curve without regurgitation. Both an aneurysmatic aortic root and aortic regurgitation
LWBK1209-ch30_p471-488.indd 477 17/05/13 5:15 PM

ANEU RV
Ao PA
PA RA LV
LA
RV
Ao PA
RA LV
LA
Figure 30.9. Same patient as in Figure 30.8. TOF with repair in the past. Black blood spin-echo imaging
before re-repair (upper row) and after re-repair with pulmonary homograft and reduction of the right ventricular
outflow tract (RVOT) aneurysm (lower row). The RV end-diastolic volume (EDV) reduced from 500 mL preoper-
atively to 260 mL postoperatively. After the procedure, the effective RV ejection fraction (EF) remained moderate,
although RV EF was significantly increased from 11% preoperatively to 36% postoperatively. RV, right ventricle;
LV, left ventricle; RA, right atrium; LA, left atrium; ANEU, RVOT aneurysm; Ao, aorta; PA, pulmonary artery.
A B
Figure 30.10. An 18-year-old female patient

with a history of pulmonary atresia (extreme form
of TOF) with pulmonary conduit repair at the age of
6 years. At follow-up, conduit stenosis was observed
on flow mapping, represented by severe aliasing
caused by increased flow velocity through the pul-
monary conduit “valvular stenosis” (arrows, A,B).
Also, a moderate pulmonary regurgitation (PR) of
30% was observed (not shown). In addition, there
was stenosis of the proximal left pulmonary artery
on gadolinium-enhanced MRA with a vessel diam-
C eter of 7 mm at the stenosis (C).
LWBK1209-ch30_p471-488.indd 478 17/05/13 5:15 PM

A B C
Figure 30.11. The same patient as in

Figure 30.10, showing dilatation of the
ascending aorta as a feature associated
with TOF. There were no signs of aortic
valve stenosis or regurgitation. Note the
Ao normal aortic root diameter on trans-

verse black blood spin-echo imaging
Ao (A) with a normal aspect of the closed

and opened aortic valve on oblique sag-
ittal fast gradient-echo imaging (B,C).
Dilatation of the ascending aorta (Ao)
shown on transverse black blood spin-
echo imaging (D) and oblique sagittal
D E
fast gradient-echo imaging (E).
may necessitate aortic root replacement (46) and can be as- Transposition of the Great Arteries
sessed with black blood images of the aortic root and flow
mapping across the aortic valve. Figure 30.11 shows an exam-
ple of dilatation of the ascending aorta associated with TOF. TGA is defined as atrial situs solitus, normal (concordant)
Timely detection and monitoring of these morphologic connection between atria and ventricles, and abnormal (dis-
and functional abnormalities requires an integrated imaging cordant) ventriculoarterial connections. The aorta arises
protocol, focused on assessment of biventricular function, from the RV, and the pulmonary artery arises from the LV. It
depiction of the pulmonary vascular tree, and assessment of accounts for 4.5% of all congenital cardiac malformations
function of the pulmonary, aortic, and tricuspid valves. (47). Since the pulmonary circulation is isolated from the
systemic circulation, TGA is a life-threatening CHD requir-
ing palliative treatment after birth, followed by corrective
Imaging Protocol
surgery during the first weeks of life.
After scout images are obtained, an axial stack of black Before the introduction of the arterial switch operation
blood turbo spin-echo images is acquired to outline cardiac in 1975 (48), TGA was corrected at the atrial level, redi-
and noncardiac anatomy. A double-oblique transverse stack recting systemic venous blood from the superior vena cava
of images along the aortic root can be used for measure- (SVC) and inferior vena cava (IVC) to the LV and pulmo-
ment of aortic root diameters to exclude possible dilatation nary venous blood from the pulmonary veins to the RV
of the aortic root. Next, axial gradient-echo multislice multi- (Fig. 30.12), using artificial or pericardial tissue (Mustard
phase imaging is performed to quantify LV and RV function, technique) or atrial tissue (Senning technique, Fig. 30.13).
with special focus on RV function and the magnitude of RV Although a “physiologic” correct circulation is created,
hypertrophy and dilatation. Fifteen minutes after a contrast- normal anatomic relations are not restored, and the RV
enhanced MRA has been obtained to visualize the pulmo- remains subjected to systemic loading conditions, followed
nary vascular tree, delayed enhancement imaging can be by compensatory hypertrophy (Fig. 30.14). This results
performed to identify scarring of both ventricles, especially in late RV failure in up to 10% of patients with an atrial
at the site of the RVOT. While one waits for the delayed redirection operation (49). Other adverse effects like
enhancement sequence, flow mapping across the pulmonary obstruction and leakage of the venous pathways and
artery, aorta, and tricuspid valve is performed for calculation atrial arrhythmias frequently occur (50,51). Evaluation of
of pulmonary and aortic regurgitant volumes and/or magni- biventricular function and depiction of the venous pathways
tude of stenosis, and RV diastolic function, respectively. in patients with an atrial redirection or after the arterial
LWBK1209-ch30_p471-488.indd 479 17/05/13 5:15 PM

RV RV RV
* LV * LV * LV
A B C
Figure 30.12. Spin-echo (A), gradient-echo (B), and phase-contrast (C) MRI in the transverse plane in a
patient with a Mustard repair for TGA. A stenosis in the pulmonary venous conduit was observed (arrow-
heads, A,B) causing high flow velocities during systole (arrowhead, C). Asterisk indicates systemic venous
conduit. LV, left ventricle; RV, right ventricle.
switch procedure can be performed best by gradient-echo TGA compared with controls (18). A recent MRI study
MRI (Figs. 30.13–30.15). using the same bicycle exercise protocol in patients with
Exercise-MRI can be used to detect subtle ventricular TGA showed distinct regional wall motion disturbances of
dysfunction, which may not be apparent at rest. The sys- the anterior and free wall of the RV, with an increase of these
temic ventricular SV increase in response to bicycle exercise RV wall motion disturbances during exercise (52). This may
is significantly smaller in patients with atrially corrected be caused by regional perfusion abnormalities, as could be
Ao
RS
PA LS
SVC
LI
Ba Ba
RV
PVA LV
SVA
IVC
Ba
RI
Figure 30.13. A 36-year-old female patient with TGA and Senning repair in the past. Fast gradient-echo
images. The aorta arises from the RV. The RV is hypertrophied and sustains the systemic circulation. The pul-
monary artery arises from the LV and sustains the pulmonary circulation. Pulmonary veins: LS, left superior;
LI, left inferior; RS, right superior; RI, right inferior. Note how the pulmonary venous flow is directed to the
RV and how the flow from the superior vena cava (SVC) and inferior vena cava (IVC) is directed to the LV by
the intra-atrial baffle (Ba). PVA, pulmonary venous atrium; SVA, systemic venous atrium; Ao, aorta; PA,
pulmonary artery.
LWBK1209-ch30_p471-488.indd 480 17/05/13 5:15 PM

Figure 30.14. A 34-year-old female

patient with TGA after Mustard repair RV LV
in the past. Fast gradient-echo images
in short axis, at end-diastolic (A) and
end-systolic phases (B). The RV sus-
tains the systemic circulation and is
hypertrophied and enlarged: RV EDV
was 246 mL; LV EDV was 144 mL. RV
function was moderately impaired with
an EF of 45%. LV EF was 74%. LV, A B
left ventricle; RV, right ventricle.
explained by the hypertrophy of the RV itself (the need for arteries into the neo-aorta. The originally posterior branch
oxygen is not met by the supply) (53) or by preoperatively pulmonary arteries are brought anterior to the aorta with
preexisting perfusion defects (54). The relevance of stress the so-called Lecompte maneuver (Fig. 30.16).
testing was recently shown by the observation that in pa- Direct postoperative complications include myocardial
tients after atrial correction of TGA an inadequate reaction perfusion defects, caused by problems in the transfer of cor-
to either physical or pharmacologic stress was predictive for onary arteries (62). During the first year after surgery the in-
adverse outcome (55). Abnormal diastolic ventricular inflow cidence of coronary artery events ranges between 7.2% and
patterns may suggest early manifestation of ventricular dys- 7.8% (59,62,63). A recent study using contrast-enhanced
function, even in the absence of systolic dysfunction (56). At MRA of the coronaries and delayed enhancement suggested
an early stage, diastolic dysfunction may precede the onset that coronary artery problems are not a long-term problem,
of systolic dysfunction (57). showing no significant coronary stenoses and the absence
Currently, the arterial switch procedure is the operation of unexpected myocardial infarction areas (Fig. 30.5) (15).
of choice with which normal anatomic relations are restored. Evaluation of perfusion status was suggested to be restricted
Good mid-term (58,59) and long-term (60,61) results have to patients with known complications after coronary reim-
been reported. With this procedure the great vessels are tran- plantation (15). Long-term complications after the arterial
sected and switched, with reimplantation of the coronary switch operation also include dilatation of the aortic root
Figure 30.15. An 18-year-old male

LV
patient with TGA after arterial switch
repair. Fast gradient-echo images in RV
short axis, at end-diastolic (A) and
end-systolic phases (B). LV sustains
the systemic circulation. RV appears
slightly hypertrophied but not enlarged.
LV EDV was 185 mL, and RV EDV
was 163 mL. Good biventricular sys-
tolic function: LV EF was 60%, and
RV EF was 65%. Note the differences
compared with Figure 30.14. LV, left
A B
ventricle; RV, right ventricle.
LWBK1209-ch30_p471-488.indd 481 17/05/13 5:15 PM

400 Aortic flow
300
Flow (mL/s)
200
100
0
Ao A –100
100 200 300 400
Time (ms)
500 600 700
400 Pulmonary flow
300
Flow (mL/s)
200
100
0
Figure 30.16. Same patient as in Figure 30.15 with TGA after 100 200 300 400 500 600 700
arterial switch repair with Lecompte maneuver at early infancy. –100 Time (ms)
Gadolinium-enhanced MRA showing the typical arrangement of B
pulmonary arteries in relationship to the ascending aorta (Ao). The
Figure 30.18. Same patient as in Figures 30.17 and 30.19. TGA
pulmonary arteries are well developed and show no stenoses.
after arterial switch repair. Flow mapping across the aorta and the
main pulmonary artery. Note the broad and collapsed flow curve
for pulmonary flow compared with the aortic flow curve, which is
with subsequent aortic regurgitation (64–66). An echocardio explained by the increased resistance to flow through the main pulmo-
graphy study showed an incidence of 35% of neo-aortic regur- nary artery caused by the peripheral pulmonary stenosis at the proxi-
gitation, which was progressive in 23% of the patients (64). mal left and right pulmonary arteries.
After the arterial switch operation, stenoses at the level of
the pulmonary trunk and the peripheral pulmonary branches
occur at a rate of 1% per year (67). Black blood imaging and shown in Figure 30.20, with slight hypertrophy of the RV,
gadolinium-enhanced MRA proved to be superior in the de- but without signs of pulmonary stenosis.
tection of stenoses in the great vessels when compared with
transthoracic echocardiography, especially for stenoses in Imaging Protocol
the peripheral pulmonary branches (Figs. 30.17 and 30.18)
(13,67–69). An example of compensatory hypertrophy of After initial scout images and an axial stack of black blood
the RV is shown in Figure 30.19. Information on specific turbo spin-echo images, a double-oblique transverse stack
hemodynamic changes in the great vessels after the arterial of images along the aortic root is used for measurement of
switch procedure can be obtained with gradient-echo mul- aortic root diameters, identical to the Fallot protocol. Axial
tislice multiphase MRI (70). With this technique, narrowing or short-axis gradient-echo multislice multiphase imaging
of the (mainly right) pulmonary artery has been observed is performed to quantify biventricular function, with spe-
during systole, caused by expansion of the ascending aorta cial attention to possible wall motion abnormalities and the
during systole (70). Velocity mapping can additionally be patency of the surgically created connections: Both conduits
used to evaluate possible hemodynamic significance of these after atrial repair and the site of anastomosis of the great ves-
pulmonary artery stenoses (Fig. 30.18). sels after the arterial switch. In patients undergoing arterial
An abnormal diastolic RV inflow pattern may also be switch, a cine gradient-echo sequence is applied along the
observed after the arterial switch procedure. An example is pulmonary trunk to identify the spatial relationship between
Figure 30.17. A 7-year-old

PA female child with TGA after arte-
rial switch repair. Axial black blood
Ao spin-echo image (A) and correspond-
Ao ing gadolinium-enhanced maximum
intensity projection MRA (B). The
arterial switch repair is complicated
with severe pulmonary stenosis at the
proximal left and right pulmonary
arteries (left > right). See also Figures
A B
30.18 and 30.19.
LWBK1209-ch30_p471-488.indd 482 17/05/13 5:15 PM

Figure 30.19. Same patient as in

Figures 30.17 and 30.18. TGA after
arterial switch repair. Fast gradient-
echo images in short axis, at end- RV LV
diastolic (A) and end-systolic phases
(B). The LV sustains the systemic
circulation. Severe RV hypertrophy to
compensate for the severe pulmonary
A B
stenosis as shown in Figure 30.17.
the pulmonary branches and aorta. Temporary stenoses of associated or not associated with intracardiac defects like a
pulmonary branches during systole can be detected when the VSD or aortic valve pathology. Classic symptoms are heart
aorta “swells” during the expulsion phase and compresses failure and an increased blood pressure proximal to the nar-
the adjacent pulmonary branches. Flow mapping across the rowing, as well as a low perfusion status of the body distally
pulmonary artery and the aorta are acquired for assessment to the coarctation (71,72).
of pulmonary or aortic regurgitation or stenosis. In addition, Treatment consists of surgical intervention with resec-
flow mapping across the mitral valve and tricuspid valve is tion and end-to-end anastomosis or usage of a subclavian
obtained to evaluate ventricular inflow patterns and possible flap, or balloon angioplasty. Serial follow-up of these pa-
regurgitation. If indicated, an MRA of the pulmonary vascu- tients is mandatory because of the considerable incidence of
lar tree and delayed enhancement imaging of both ventricles aneurysm formation and re-coarctation, which varies from
can finally be performed. 8% to 67%, depending on the procedure of choice, time of
follow-up, and the age at which the intervention was per-
Aortic Coarctation formed (71,72). One-third of all patients become or remain
hypertensive, despite their lesion being corrected, with an
Clinical Considerations increased risk of accelerated atherosclerosis and end-organ
Coarctation of the aorta accounts for 5% of all CHD and is damage (73). Compensatory LV hypertrophy can be evalu-
defined as a congenital narrowing of the aorta, most com- ated by gradient-echo multislice multiphase imaging. MRI
monly located in a juxtaductal position just distally to the proved to be the most cost-effective approach to diagnose
origin of the left subclavian artery (Fig. 30.4) (71,72). A wide complications at the site of repair, by comparison of sensi-
spectrum of narrowing of the aorta can be observed, from a tivity and specificity of various imaging modalities and their
discrete narrowing to a hypoplastic aortic arch, whether it is costs (72).
High numbers of bicuspid aortic valves have been re-
ported in patients with coarctation, ranging between 25%
Ventricular flow and 85% (Fig. 30.4 C,D) (73–75). In addition to the clear
Time (ms) association between bicuspid aortic valves and aortic valve
0 pathology, the incidence of aortic valve pathology is higher
100 200 300 400 500 600 in studies of patients with coarctation and a bicuspid aortic
–100
valve compared with those without coarctation, emphasiz-
Flow (mL/s)
–200 ing the additional role of the coarctation in the develop-

–300 ment of aortic valve problems (73,75,76). Dilatation of the
ascending aorta during follow-up can be found in 28% to
–400
MV
44% of patients after repaired coarctation, without previous
–500 TV aortic valve intervention (73,76). Patients with coarctation
E A and a bicuspid aortic valve are at risk for ascending aortic
–600
dissection, probably caused by dilatation of the ascending
Figure 30.20. Same patient as in Figures 30.15 and 30.16. TGA aorta and persistent hypertension (45). Close follow-up is
after arterial switch repair. Flow mapping across the mitral valve (MV) necessary to time surgical intervention for aortic aneurysm.
and tricuspid valve (TV) for diastolic ventricular inflow patterns. Recently four-dimensional (4D) flow mapping has been
Normally, a biphasic inflow pattern is recognized with two peaks introduced to assess flow patterns in patients with coarcta-
representing early inflow (E) and late atrial kick (A). At this age, the
tion and bicuspid aortic valves (77–79). Although the clini-
E/A ratio should be >1, that is, the first peak (E) larger than the last
cal significance of this technique needs to be established it
peak (A). This is the case for the left ventricular inflow across the MV,
indicating normal relaxation of the LV. However, for the RV inflow, gains insight into the complex flow patterns within the aorta
the E/A ratio was <1, indicating abnormal relaxation for the RV after in these patients and may lead to improved understanding
arterial switch repair and slight RV hypertrophy. The pulmonary valve of pathophysiologic mechanisms leading to aortic dilatation
was normal, as were the pulmonary arteries. and aneurysm formation (77–79).
LWBK1209-ch30_p471-488.indd 483 17/05/13 5:15 PM

Imaging Protocol shunt quantification may be performed by measuring just

before the level of the coarctation or more distally. It should
After initial scout images and an axial stack of black blood be noted, however, that collateral flow might be missed in
turbo spin-echo images, a double-oblique transverse stack the latter. Finally, 4D flow mapping provides insight in the
of images along the aortic root is used for depiction of complex flow pattern over a bicuspid aortic valve and at the
aortic valve morphology and measurement of aortic root site of coarctation within the descending aorta.
diameters. Next, axial gradient-echo multislice multiphase
imaging is performed to assess biventricular function. A
stack of thin consecutive black blood slices in the oblique Fontan Circulation
sagittal direction through the plane of the aortic arch allows
detailed visualization of the ascending aorta, aortic arch,
and descending aorta, which is sensitive for the detection In 1971 Fontan (80) described a surgical correction for the
and follow-up of re-coarctation and aneurysms. Contrast- treatment of tricuspid atresia. This type of operation has
enhanced MRA of the thoracic aorta with 3D reconstruc- become the treatment of choice in a heterogeneous group of
tion can clearly demonstrate the spatial relationship between congenital cardiac malformations characterized by the pres-
the stenosis, the other arch vessels, and possible collateral ence of a single (systemic) ventricle. The procedure establishes
vessels. Flow mapping is performed across the aortic valve a circulation in which the systemic venous return directly
for detection of aortic valve stenosis or regurgitation. In the enters the pulmonary arteries, bypassing the heart except for
case of re-coarctation, velocity mapping across the aorta at the atrioventricular variant (Figs. 30.21 and 30.22). Different
the level of the ascending aorta, just distal to the level of types of connections can be implemented, including total
the re-coarctation and at the level of the diaphragm, can be cavopulmonary connection, such as an extracardiac conduit
obtained for quantification of the pressure gradient across (Fig. 30.23), atriopulmonary connection, and atrioventri
the re-coarctation and for calculation of the magnitude of cular connection. The introduction of a fenestration in the
possible collateral flow, by subtraction of the proximal aor- systemic venous tunnel wall improved the outcome of the pro-
tic flow from the distal one (71). Flow mapping may be dif- cedure in high-risk cases by decreasing the incidence of early
ficult to perform just distal to the coarctation site because pericardial effusions after the procedure and maintaining LV
of turbulent flow and increased flow velocity. In that case, output in times of hemodynamic stress (81).
F
PA SVC
hRV
RA LV
LA CS
Figure 30.21. A 27-year-old male patient with tricuspid atresia and hypoplastic RV after correction by
Fontan circulation. Axial fast-field gradient-echo images showing the Fontan circulation consisting of a patch
connecting the dilated right atrium with the hypoplastic RV and main pulmonary artery. Note the dilated
coronary sinus and superior vena cava, despite the open Fontan connection. Normal pulmonary arteries. LV,
left ventricle; hRV, hypoplastic right ventricle; F, Fontan circulation; RA, right atrium; LA, left atrium; PA, pul-
monary artery; CS, coronary sinus; SVC, superior vena cava.
LWBK1209-ch30_p471-488.indd 484 17/05/13 5:15 PM

tree. Unobstructed flow passage is of eminent importance

considering the low propelling pressure of the vena cava sys-
tem (87).
Flow mapping can be used to provide information on
flow patterns within the different types of Fontan connec-
tions (88,89). In patients with an atriopulmonary connec-
tion, flow was biphasic and pulsatile during the course of the
cardiac cycle, but with a relatively small difference between
the maximal and minimal velocities. After atrioventricular
RA PA connection, a systolic arterial flow pattern can be found in
the pulmonary arteries of half of the patients, indicating that
the RV significantly sustains pulmonary blood flow (88).
LV Recently, the evaluation of myocardial fibrosis as as-
sessed with late gadolinium enhancement in Fontan patients
revealed that fibrosis is frequently observed and is associated
with lower ejection and an increased incidence of nonsus-
tained ventricular tachycardia (90).
hRV
Imaging Protocol
F
After initial scout images, axial and sagittal stacks of black
blood turbo spin-echo images are acquired for depiction of
the often complex cardiac anatomy. Axial gradient-echo mul-
Figure 30.22. The same patient as in Figure 30.21. Gadolinium- tislice multiphase imaging is performed to quantify systemic
enhanced MRA of the Fontan circulation, consisting of a patch (F)
connecting the dilated right atrium with the hypoplastic RV and main
ventricular function and to assess the morphologic dynamics
pulmonary artery (after correction of infundibular pulmonary stenosis of the Fontan circulation (Fig. 30.21). Flow mapping across
and correction of atrium septum defect and ventricular septum defect). the aorta, the left and right pulmonary arteries, and the atrio-
Note the open Fontan connection between the right atrium and the ventricular valve is performed to evaluate flow patterns and
hypoplastic RV and pulmonary artery. Normal pulmonary arteries. LV, to determine the magnitude of possible stenoses within the
left ventricle; RA, right atrium; hRV, hypoplastic right ventricle; PA, different types of Fontan connections, which are character-
pulmonary artery. ized by turbulent flow and increased flow velocity. Finally,
contrast-enhanced MRA of the pulmonary vascular tree
is performed for optimal depiction of the spatial relation-
ship of the Fontan connections (Fig. 30.22). Diameters of
The state of the systemic ventricle is important in the the right and left pulmonary arteries for the McGoon ratio
presurgical planning because ventricular hypertrophy and can be obtained from axial imaging or the 3D reconstructed
abnormal wall motion impair the surgical outcome (82). In MRA, and the diameter of the descending aorta can be
the postoperative period ventricular function may deterio- obtained from axial spin-echo or gradient-echo imaging.
rate over time and should be followed carefully (82,83). EF Following the administration of gadolinium late gadolinium
may deteriorate 1 to 2 years after a Fontan operation com- enhancement of the myocardium can be evaluated to assess
pared with preoperative data (83). Gradient-echo multislice myocardial fibrosis.
multiphase MRI is ideally suited to assess the complex mor-
phology and function of the systemic ventricle before and
after construction of the Fontan circulation (84,85). Both Future Perspective
3D reconstructed MRA and spin-echo MRI can be used for
detection of conduit obstruction (86). In adult patients, a Future applications of MRI in the setting of CHD will
diameter of the conduit measured with spin-echo MRI of include new MRI sequences, subsequent improvement
less than 15 mm suggests that a significant pressure gradi- of clinical decision making in patients with corrected or
ent within the conduit can be expected, whereas a diameter palliated CHD, and the emerging field of interventional
of more than 20-mm diameter is regarded as the optimal MRI.
diameter (86). New techniques will provide faster MRI protocols, mak-
The dimensions of the right and left pulmonary arteries, ing it more attractive for widespread clinical usage, and will
expressed as the McGoon ratio, are one of the most power- (partially) overcome the limitation of long breath-holds that
ful risk factors for death or a dysfunctional Fontan circuit. apply to the younger patient population (5). Subsequently,
This includes summation of the diameter of the distal but real-time MRI allows the assessment of dynamic processes
immediately pre-branching portion of the right and the left such as the cardiac response to exercise. Furthermore, the
pulmonary artery, divided by the diameter of the descending evaluation of pulse-wave velocity within the aorta in various
aorta measured just above the diaphragm (87). An increased CHD patient groups may further improve our understand-
risk for death or a dysfunctional Fontan circuit can be ex- ing of the ventricular–vascular interaction (91), and the
pected when the ratio is less than 1.8, because this is associ- relation between altered pulse-wave velocity and cerebral
ated with flow obstruction through the pulmonary vascular lesions (92).
LWBK1209-ch30_p471-488.indd 485 17/05/13 5:15 PM

A B
C D E
F G H
Figure 30.23. A 16-year-old male with mitral atresia and transposition of the great arteries, after Fontan
circulation with an extracardiac conduit. White blood MR images in right anterior oblique coronal orientation
(A), left anterior oblique coronal orientation (B), and axial orientation (C–H). Pulmonary arteries have normal
sizes (D,E). The inferior vena cava (IVC) is connected to the right pulmonary artery (RPA) and the superior
vena cava (SVC) to the left pulmonary artery (LPA). No obstruction of the Fontan circulation; the IVC, SVC,
and liver veins were not dilated. Functionally single ventricle connected to the aorta (Ao) that supplies the sys-
temic circulation (note the large ventricular septum defect [asterisk, G] between the right ventricle [RV]
and hypoplastic left ventricle [LV]). Pulmonary veins (F,G) enter in the pulmonary venous atrium (PVA) that
is connected to the ventricles. Patient had a total end-diastolic volume of 230 mL with a moderate ejection
fraction of 34%. Note transposition of the great arteries where the aorta is located at the left and in front of
the pulmonary artery.
MRI may play an increasing role in clinical decision mak- may lead to normalization of RV volume when performed
ing in various patient groups with CHD, by serial assessment before the RV reaches 150% of its normal size (13,33,93).
of cardiac parameters at fixed time points during follow-up. This kind of information derived from MRI studies may be
For example, the timing of PVR after total correction of TOF helpful for improvement of patient-management strategies.
is a controversial issue, but it has been observed that PVR With decreasing CHD-mortality rates, focus is shifting to
LWBK1209-ch30_p471-488.indd 486 17/05/13 5:16 PM

evaluation of morbidity, especially neurologic outcome. It is 21. de RA, Niezen RA, Lamb HJ, et al. MR of the heart under pharmacologic stress. Cardiol
Clin. 1998;16(2):247–265.
increasingly recognized that CHD and subsequent correction 22. Romeih S, Groenink M, Roest AA, et al. Exercise capacity and cardiac reserve in chil-
or palliation affects the brain and may influence future cogni- dren and adolescents with corrected pulmonary atresia with intact ventricular septum
after univentricular palliation and biventricular repair. J Thorac Cardiovasc Surg. 2012;
tive and motor outcome (91). 143(3):569–575.
Cardiac catheterization procedures guided with MRI 23. Hoffman JI, Kaplan S. The incidence of congenital heart disease. J Am Coll Cardiol. 2002;
are exciting new option, combining the unique advantages 39(12):1890–1900.
24. Anderson RH, Tynan M. Tetralogy of Fallot—a centennial review. Int J Cardiol. 1988;21
of MRI with catheter-driven interventions. Razavi et al. (3):219–232.
(94) showed that cardiac catheterization guided by MRI 25. Lillehei CW, Cohen M, Warden HE, et al. Direct vision intracardiac surgical correction of
with radiographic support is feasible, moving a stage fur- the tetralogy of Fallot, pentalogy of Fallot, and pulmonary atresia defects; report of first ten
cases. Ann Surg. 1955;142(3):418–442.
ther toward complete MRI-guided cardiac catheterization. 26. Van Arsdell GS, Maharaj GS, Tom J, et al. What is the optimal age for repair of tetralogy
This will include advantages especially for the pediatric of Fallot? Circulation. 2000;102(19 suppl 3):III123–III129.
27. Murphy JG, Gersh BJ, Mair DD, et al. Long-term outcome in patients undergoing surgical
population, considering the essential role of cardiac cath- repair of tetralogy of Fallot. N Engl J Med. 1993;329(9):593–599.
eterization in the diagnosis and treatment of CHD, and the 28. Gatzoulis MA, Balaji S, Webber SA, et al. Risk factors for arrhythmia and sudden car-
concern for long-term complications associated with radia- diac death late after repair of tetralogy of Fallot: A multicentre study. Lancet. 2000;356
(9234):975–981.
tion burden. 29. Bove EL, Byrum CJ, Thomas FD, et al. The influence of pulmonary insufficiency on ven-
tricular function following repair of tetralogy of Fallot. Evaluation using radionuclide ven-
triculography. J Thorac Cardiovasc Surg. 1983;85(5):691–696.
30. Waien SA, Liu PP, Ross BL, et al. Serial follow-up of adults with repaired tetralogy of
References Fallot. J Am Coll Cardiol. 1992;20(2):295–300.
31. Gatzoulis MA, Till JA, Somerville J, et al. Mechanoelectrical interaction in tetralogy of
1. Jara H, Barish MA. Black-blood MR angiography. Techniques, and clinical applications. Fallot. QRS prolongation relates to right ventricular size and predicts malignant ventricular
Magn Reson Imaging Clin N Am. 1999;7(2):303–317. arrhythmias and sudden death. Circulation. 1995;92(2):231–237.
2. Helbing WA, Bosch HG, Maliepaard C, et al. Comparison of echocardiographic methods 32. Kondo C, Nakazawa M, Kusakabe K, et al. Left ventricular dysfunction on exercise long-
with magnetic resonance imaging for assessment of right ventricular function in children. term after total repair of tetralogy of Fallot. Circulation. 1995;92(suppl 9):II250–II255.
Am J Cardiol. 1995;76(8):589–594. 33. Therrien J, Siu SC, McLaughlin PR, et al. Pulmonary valve replacement in adults late after
3. Helbing WA, Rebergen SA, Maliepaard C, et al. Quantification of right ventricular func- repair of tetralogy of fallot: Are we operating too late? J Am Coll Cardiol. 2000;36(5):
tion with magnetic resonance imaging in children with normal hearts and with congenital 1670–1675.
heart disease. Am Heart J. 1995;130(4):828–837. 34. Vliegen HW, van SA, de RA, et al. Magnetic resonance imaging to assess the hemodynamic
4. Niezen RA, Helbing WA, van der Wall EE, et al. Biventricular systolic function and mass effects of pulmonary valve replacement in adults late after repair of tetralogy of fallot.
studied with MR imaging in children with pulmonary regurgitation after repair for tetra Circulation. 2002;106(13):1703–1707.
logy of Fallot. Radiology. 1996;201(1):135–140. 35. Therrien J, Provost Y, Merchant N, et al. Optimal timing for pulmonary valve replacement
5. Tsao J, Boesiger P, Pruessmann KP. k-t BLAST and k-t SENSE: Dynamic MRI with in adults after tetralogy of Fallot repair. Am J Cardiol. 2005;95(6):779–782.
high frame rate exploiting spatiotemporal correlations. Magn Reson Med. 2003;50(5): 36. Davlouros PA, Kilner PJ, Hornung TS, et al. Right ventricular function in adults with
1031–1042. repaired tetralogy of Fallot assessed with cardiovascular magnetic resonance imaging:
6. Westenberg JJ, Roes SD, Ajmone MN, et al. Mitral valve and tricuspid valve blood flow: Detrimental role of right ventricular outflow aneurysms or akinesia and adverse right-to-
Accurate quantification with 3D velocity-encoded MR imaging with retrospective valve left ventricular interaction. J Am Coll Cardiol. 2002;40(11):2044–2052.
tracking. Radiology. 2008;249(3):792–800. 37. Chaturvedi RR, Kilner PJ, White PA, et al. Increased airway pressure and simulated branch
7. van der Hulst AE, Westenberg JJ, Kroft LJ, et al. Tetralogy of fallot: 3D velocity-encoded pulmonary artery stenosis increase pulmonary regurgitation after repair of tetralogy of
MR imaging for evaluation of right ventricular valve flow and diastolic function in patients Fallot. Real-time analysis with a conductance catheter technique. Circulation. 1997;95(3):
after correction. Radiology. 2010;256(3):724–734. 643–649.
8. Beerbaum P, Korperich H, Gieseke J, et al. Rapid left-to-right shunt quantification in chil- 38. Tulevski II, Hirsch A, Dodge-Khatami A, et al. Effect of pulmonary valve regurgitation on
dren by phase-contrast magnetic resonance imaging combined with sensitivity encoding right ventricular function in patients with chronic right ventricular pressure overload. Am
(SENSE). Circulation. 2003;108(11):1355–1361. J Cardiol. 2003;92(1):113–116.
9. Brenner LD, Caputo GR, Mostbeck G, et al. Quantification of left to right atrial shunts 39. Babu-Narayan SV, Kilner PJ, Li W, et al. Ventricular fibrosis suggested by cardiovascu-
with velocity-encoded cine nuclear magnetic resonance imaging. J Am Coll Cardiol. lar magnetic resonance in adults with repaired tetralogy of fallot and its relationship to
1992;20(5):1246–1250. adverse markers of clinical outcome. Circulation. 2006;113(3):405–413.
10. Rebergen SA, van der Wall EE, Helbing WA, et al. Quantification of pulmonary and sys- 40. Geva T, Sandweiss BM, Gauvreau K, et al. Factors associated with impaired clinical sta-
temic blood flow by magnetic resonance velocity mapping in the assessment of atrial-level tus in long-term survivors of tetralogy of Fallot repair evaluated by magnetic resonance
shunts. Int J Card Imaging. 1996;12(3):143–152. imaging. J Am Coll Cardiol. 2004;43(6):1068–1074.
11. van der Hulst AE, Westenberg JJ, Delgado V, et al. Tissue-velocity magnetic resonance 41. Ghai A, Silversides C, Harris L, et al. Left ventricular dysfunction is a risk factor for sud-
imaging and tissue Doppler imaging to assess regional myocardial diastolic velocities at den cardiac death in adults late after repair of tetralogy of Fallot. J Am Coll Cardiol. 2002;
the right ventricle in corrected pediatric tetralogy of fallot patients. Invest Radiol. 2012;47 40(9):1675–1680.
(3):189–196. 42. Muzzarelli S, Ordovas KG, Cannavale G, et al. Tetralogy of Fallot: Impact of the excursion
12. van der Hulst AE, Roest AA, Delgado V, et al. Corrected tetralogy of Fallot: Comparison of the interventricular septum on left ventricular systolic function and fibrosis after surgical
of tissue Doppler imaging and velocity-encoded MR for assessment of performance and repair. Radiology. 2011;259(2):375–383.
temporal activation of right ventricle. Radiology. 2011;260(1):88–97. 43. Ishizaka T, Ichikawa H, Sawa Y, et al. Prevalence and optimal management strategy
13. Kondo C, Takada K, Yokoyama U, et al. Comparison of three-dimensional contrast- for aortic regurgitation in tetralogy of Fallot. Eur J Cardiothorac Surg. 2004;26(6):
enhanced magnetic resonance angiography and axial radiographic angiography for diagnos- 1080–1086.
ing congenital stenoses in small pulmonary arteries. Am J Cardiol. 2001;87(4):420–424. 44. Niwa K, Siu SC, Webb GD, et al. Progressive aortic root dilatation in adults late after
14. Flamm SD, Muthupillai R. Coronary artery magnetic resonance angiography. J Magn repair of tetralogy of Fallot. Circulation. 2002;106(11):1374–1378.
Reson Imaging. 2004;19(6):686–709. 45. Bonderman D, Gharehbaghi-Schnell E, Wollenek G, et al. Mechanisms underlying aortic
15. Taylor AM, Dymarkowski S, Hamaekers P, et al. MR coronary angiography and late- dilatation in congenital aortic valve malformation. Circulation. 1999;99(16):2138–2143.
enhancement myocardial MR in children who underwent arterial switch surgery for trans- 46. Dodds GA III, Warnes CA, Danielson GK. Aortic valve replacement after repair of pulmo-
position of great arteries. Radiology. 2005;234(2):542–547. nary atresia and ventricular septal defect or tetralogy of Fallot. J Thorac Cardiovasc Surg.
16. Oosterhof T, Mulder BJ, Vliegen HW, et al. Cardiovascular magnetic resonance in the follow- 1997;113(4):736–741.
up of patients with corrected tetralogy of Fallot: A review. Am Heart J. 2006;151(2):265–272. 47. Hoffman JI. Incidence of congenital heart disease: I. Postnatal incidence. Pediatr Cardiol.
17. Broberg CS, Chugh SS, Conklin C, et al. Quantification of diffuse myocardial fibrosis and 1995;16(3):103–113.
its association with myocardial dysfunction in congenital heart disease. Circ Cardiovasc 48. Jatene AD, Fontes VF, Paulista PP, et al. Successful anatomic correction of transposition of
Imaging. 2010;3(6):727–734. the great vessels. A preliminary report. Arq Bras Cardiol. 1975;28(4):461–464.
18. Roest AA, Kunz P, Helbing WA, et al. Prolonged cardiac recovery from exercise in asymp- 49. Williams WG, Trusler GA, Kirklin JW, et al. Early and late results of a protocol for sim-
tomatic adults late after atrial correction of transposition of the great arteries: Evaluation ple transposition leading to an atrial switch (Mustard) repair. J Thorac Cardiovasc Surg.
with magnetic resonance flow mapping. Am J Cardiol. 2001;88(9):1011–1017. 1988;95(4):717–726.
19. Roest AA, Helbing WA, Kunz P, et al. Exercise MR imaging in the assessment of pulmo- 50. Roest AA, Lamb HJ, van der Wall EE, et al. Cardiovascular response to physical exercise
nary regurgitation and biventricular function in patients after tetralogy of fallot repair. in adult patients after atrial correction for transposition of the great arteries assessed with
Radiology. 2002;223(1):204–211. magnetic resonance imaging. Heart. 2004;90(6):678–684.
20. Roest AA, de RA, Lamb HJ, et al. Tetralogy of Fallot: Postoperative delayed recovery of 51. Sampson C, Kilner PJ, Hirsch R, et al. Venoatrial pathways after the Mustard operation
left ventricular stroke volume after physical exercise assessment with fast MR imaging. for transposition of the great arteries: Anatomic and functional MR imaging. Radiology.
Radiology. 2003;226(1):278–284. 1994;193(1):211–217.
LWBK1209-ch30_p471-488.indd 487 17/05/13 5:16 PM

52. Tops LF, Roest AA, Lamb HJ, et al. Intraatrial repair of transposition of the great arteries: 73. Roos-Hesselink JW, Scholzel BE, Heijdra RJ, et al. Aortic valve and aortic arch pathology
Use of MR imaging after exercise to evaluate regional systemic right ventricular function. after coarctation repair. Heart. 2003;89(9):1074–1077.
Radiology. 2005;237(3):861–867. 74. Presbitero P, Demarie D, Villani M, et al. Long term results (15–30 years) of surgical repair
53. Hornung TS, Kilner PJ, Davlouros PA, et al. Excessive right ventricular hypertrophic of aortic coarctation. Br Heart J. 1987;57(5):462–467.
response in adults with the mustard procedure for transposition of the great arteries. Am J 75. Ward C. Clinical significance of the bicuspid aortic valve. Heart. 2000;83(1):81–85.
Cardiol. 2002;90(7):800–803. 76. Nistri S, Sorbo MD, Marin M, et al. Aortic root dilatation in young men with normally
54. Redington AN, Rigby ML, Oldershaw P, et al. Right ventricular function 10 years after the functioning bicuspid aortic valves. Heart. 1999;82(1):19–22.
Mustard operation for transposition of the great arteries: Analysis of size, shape, and wall 77. Hope MD, Hope TA, Meadows AK, et al. Bicuspid aortic valve: Four-dimensional MR
motion. Br Heart J. 1989;62(6):455–461. evaluation of ascending aortic systolic flow patterns. Radiology. 2010;255(1):53–61.
55. Winter MM, Scherptong RW, Kumar S, et al. Ventricular response to stress predicts out- 78. Hope MD, Meadows AK, Hope TA, et al. Clinical evaluation of aortic coarctation with 4D
come in adult patients with a systemic right ventricle. Am Heart J. 2010;160(5):870–876. flow MR imaging. J Magn Reson Imaging. 2010;31(3):711–718.
56. Rebergen SA, Helbing WA, van der Wall EE, et al. MR velocity mapping of tricuspid 79. Hope MD, Hope TA, Crook SE, et al. 4D flow CMR in assessment of valve-related ascend-
flow in healthy children and in patients who have undergone Mustard or Senning repair. ing aortic disease. JACC Cardiovasc Imaging. 2011;4(7):781–787.
Radiology. 1995;194(2):505–512. 80. Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax. 1971;26(3):240–248.
57. Nishimura RA, Abel MD, Hatle LK, et al. Assessment of diastolic function of the heart: 81. Bridges ND, Lock JE, Mayer JE Jr, et al. Cardiac catheterization and test occlusion of the
Background and current applications of Doppler echocardiography. Part II. Clinical stud- interatrial communication after the fenestrated Fontan operation. J Am Coll Cardiol. 1995;
ies. Mayo Clin Proc. 1989;64(2):181–204. 25(7):1712–1717.
58. Kirklin JW, Blackstone EH, Tchervenkov CI, et al. Clinical outcomes after the arterial 82. Chin AJ, Franklin WH, Andrews BA, et al. Changes in ventricular geometry early after
switch operation for transposition. Patient, support, procedural, and institutional risk fac- Fontan operation. Ann Thorac Surg. 1993;56(6):1359–1365.
tors. Congenital Heart Surgeons Society. Circulation. 1992;86(5):1501–1515. 83. Fogel MA, Weinberg PM, Chin AJ, et al. Late ventricular geometry and performance
59. Pretre R, Tamisier D, Bonhoeffer P, et al. Results of the arterial switch operation in neo- changes of functional single ventricle throughout staged Fontan reconstruction assessed by
nates with transposed great arteries. Lancet. 2001;357(9271):1826–1830. magnetic resonance imaging. J Am Coll Cardiol. 1996;28(1):212–221.
60. Haas F, Wottke M, Poppert H, et al. Long-term survival and functional follow-up in 84. Altmann K, Shen Z, Boxt LM, et al. Comparison of three-dimensional echocardiographic
patients after the arterial switch operation. Ann Thorac Surg. 1999;68(5):1692–1697. assessment of volume, mass, and function in children with functionally single left ventricles
61. Losay J, Touchot A, Serraf A, et al. Late outcome after arterial switch operation for trans- with two-dimensional echocardiography and magnetic resonance imaging. Am J Cardiol.
position of the great arteries. Circulation. 2001;104(12 suppl 1):I121–I126. 1997;80(8):1060–1065.
62. Tzifa A, Tulloh RM. Coronary arterial complications before and after the arterial switch 85. Fogel MA, Weinberg PM, Fellows KE, et al. Magnetic resonance imaging of constant total
operation: Is the future clear? Cardiol Young. 2002;12(2):164–171. heart volume and center of mass in patients with functional single ventricle before and after
63. Legendre A, Losay J, Touchot-Kone A, et al. Coronary events after arterial switch opera- staged Fontan procedure. Am J Cardiol. 1993;72(18):1435–1443.
tion for transposition of the great arteries. Circulation. 2003;108 (suppl 1):II186–II190. 86. Sampson C, Martinez J, Rees S, et al. Evaluation of Fontan’s operation by magnetic reso-
64. Formigari R, Toscano A, Giardini A, et al. Prevalence and predictors of neoaortic regur- nance imaging. Am J Cardiol. 1990;65(11):819–821.
gitation after arterial switch operation for transposition of the great arteries. J Thorac 87. Fontan F, Fernandez G, Costa F, et al. The size of the pulmonary arteries and the results of
Cardiovasc Surg. 2003;126(6):1753–1759. the Fontan operation. J Thorac Cardiovasc Surg. 1989;98(5 Pt 1):711–719.
65. Hourihan M, Colan SD, Wernovsky G, et al. Growth of the aortic anastomosis, annu- 88. Be’eri E, Maier SE, Landzberg MJ, et al. In vivo evaluation of Fontan pathway flow dynam-
lus, and root after the arterial switch procedure performed in infancy. Circulation. 1993; ics by multidimensional phase-velocity magnetic resonance imaging. Circulation. 1998;
88(2):615–620. 98(25):2873–2882.
66. Hutter PA, Thomeer BJ, Jansen P, et al. Fate of the aortic root after arterial switch opera- 89. Morgan VL, Graham TP Jr, Roselli RJ, et al. Alterations in pulmonary artery flow patterns
tion. Eur J Cardiothorac Surg. 2001;20(1):82–88. and shear stress determined with three-dimensional phase-contrast magnetic resonance
67. Williams WG, Quaegebeur JM, Kirklin JW, et al. Outflow obstruction after the arterial imaging in Fontan patients. J Thorac Cardiovasc Surg. 1998;116(2):294–304.
switch operation: A multiinstitutional study. Congenital Heart Surgeons Society. J Thorac 90. Rathod RH, Prakash A, Powell AJ, et al. Myocardial fibrosis identified by cardiac mag-
Cardiovasc Surg. 1997;114(6):975–987. netic resonance late gadolinium enhancement is associated with adverse ventricular
68. Beek FJ, Beekman RP, Dillon EH, et al. MRI of the pulmonary artery after arterial switch mechanics and ventricular tachycardia late after Fontan operation. J Am Coll Cardiol.
operation for transposition of the great arteries. Pediatr Radiol. 1993;23(5):335–340. 2010;55(16):1721–1728.
69. Hardy CE, Helton GJ, Kondo C, et al. Usefulness of magnetic resonance imaging for evalu- 91. Roest AA, de RA. Imaging of patients with congenital heart disease. Nat Rev Cardiol.
ating great-vessel anatomy after arterial switch operation for D-transposition of the great 2012;9(2):101–115.
arteries. Am Heart J. 1994;128(2):326–332. 92. van Elderen SG, Brandts A, van der Grond J, et al. Cerebral perfusion and aortic
70. Gutberlet M, Boeckel T, Hosten N, et al. Arterial switch procedure for D-transposition of the stiffness are independent predictors of white matter brain atrophy in type 1 dia-
great arteries: Quantitative midterm evaluation of hemodynamic changes with cine MR imag- betic patients assessed with magnetic resonance imaging. Diabetes Care. 2011;34(2):
ing and phase-shift velocity mapping-initial experience. Radiology. 2000;214(2):467–475. 459–463.
71. Konen E, Merchant N, Provost Y, et al. Coarctation of the aorta before and after correc- 93. Hazekamp MG, Kurvers MM, Schoof PH, et al. Pulmonary valve insertion late after repair
tion: The role of cardiovascular MRI. AJR Am J Roentgenol. 2004;182(5):1333–1339. of Fallot’s tetralogy. Eur J Cardiothorac Surg. 2001;19(5):667–670.
72. Therrien J, Thorne SA, Wright A, et al. Repaired coarctation: A “cost-effective” approach 94. Razavi R, Hill DL, Keevil SF, et al. Cardiac catheterisation guided by MRI in children and
to identify complications in adults. J Am Coll Cardiol. 2000;35(4):997–1002. adults with congenital heart disease. Lancet. 2003;362(9399):1877–1882.
LWBK1209-ch30_p471-488.indd 488 17/05/13 5:16 PM

Chapter
Philip Kilner
Sonya V. Babu-Narayan 31
Adult Congenital Heart Disease
■■ Techniques Cardiovascular MRI can make outstanding contributions to

Multislice Scout Imaging the assessment of CHD in adults, in whom body size and the
Cine and Multislice Cine Imaging results of surgery tend to limit echocardiographic access. The
Flow Velocity Acquisitions costs of imaging should be weighed against the potential costs
of inappropriate management, which can entail complicated
Three-dimensional Acquisitions with and without
repeated surgery and extended hospitalization.
Contrast Imaging specialists need not be deterred by the anatomic
Measurement of Right Ventricular Function and variability found in these patients as discussion with the
Mass CHD specialist can clarify questions to be answered, and the
Late Gadolinium Enhancement comprehensive anatomic coverage offered by MRI almost
■■ Applications of Magnetic Resonance Imaging/ always provides useful diagnostic information. Variations of
Magnetic Resonance Angiography in Specific anatomy and function do, however, mean that decisions on
cine imaging planes and sequences generally must be made
Diseases
during acquisition. It is not only anatomic structures that
Left Ventricular Outflow Tract Obstruction and vary; surgical techniques have also evolved and changed.
Aortic Stenosis Operations for conditions such as transposition of the great
The Outflow Tracts after the Ross Operation arteries (TGA) and heart with only one effective ventricle
Aortic Coarctation, Recoarctation, and Aneurysm have changed radically during the last few decades, and
Patent Ductus Arteriosus those investigating older patients need to be aware of both
Atrial and Ventricular Septal Defects previous and current surgical practices. A number of reviews
Pulmonary Hypertension of MRI of CHD have been published (4–6), and information
Right Ventricular Outflow Tract and Conduit on the investigation and management of malformations has
Obstruction been put together by the Canadian Consensus Conference
Repaired Tetralogy of Fallot on Adult Congenital Heart Disease (7), conveniently acces-
sible through the Web site of the Canadian Adult Congenital
Ebstein Anomaly
Heart Network (see www.cachnet.org/res_managing.shtml
Transposition of the Great Arteries Repaired by for information on managing congenital heart defects).
Arterial Switch, Mustard, Senning, and Rastelli The diagnostic contributions of MRI tend to complement
Operations those of transesophageal echocardiography. The latter can
Fontan Operation for Functionally Single Ventricle be more effective than MRI for imaging thin, mobile intra-
■■ Summary cardiac structures such as the atrial septum, valve structures,
and the vegetations of endocarditis. Diagnostic cardiac cath-
eterization is rarely needed now, and interventional catheter-
This chapter considers the uses of magnetic resonance imag- ization may be expedited after MRI and transesophageal
ing (MRI), magnetic resonance angiography (MRA) and, echocardiography. Further, information from MRI can direct
where specifically important, computed tomography (CT) cardiac catheterization and thus shorten radiation exposure.
in the assessment of adolescents and adults with congeni- However, adequate assessments of coronary artery stenoses
tal heart disease (CHD). The number of such patients is or pulmonary vascular resistance still rely on invasive inves-
increasing because of the success of surgical, catheter, and tigations. Uses of MRI-guided catheterization (8) include at-
other interventions. The effects of congenital disease, sur- tempted measurement of pulmonary vascular resistance by
gery, and acquired disease may combine to create varied and MRI flow combined with catheter pressure measurement (9).
challenging pathology (1–3). Residual structural and func- Since many adults with CHD have a complex clinical his-
tional abnormalities generally must be followed on a lifelong tory with multiple complementary investigations, collaborative
basis, preferably at a specialist referral center. Appropriate presentation and discussion can be important, allowing a com-
imaging facilities represent a key aspect of such a centre. prehensive picture to be built up that can clarify questions
489
LWBK1209-ch31_p489-502.indd 489 5/18/13 1:01 PM

Figure 31.1. Systolic jet of moder-

ate aortic stenosis shown by steady
state free precession (SSFP) cine imaging
(echo time = 1.6 milliseconds) in a
coronal plane (left) and transecting the
jet and aortic root (right). The arrows
point to lines of signal loss caused by
shear, outlining the relatively bright jet
core.
of clinical management and further the process of learning. characterization, a full transaxial stack of half-Fourier acqui-
At such meetings, MRI studies should be shown with inter- sition single-shot turbo spin-echo (HASTE) images should
active interrogation of multislice, cine, velocity, and three- be acquired. We recommend following this with multislice
dimensional (3D) acquisitions. steady-state free precession (SSFP) image stacks in trans-
Cardiovascular MRI can provide functional as well as axial, coronal, and sagittal orientations. These bright blood
anatomic information, including the location and severity of acquisitions provide an overview of cardiovascular anatomi-
stenoses (e.g., aortic coarctation or pulmonary artery steno- cal structures and connections, including the pulmonary
sis), severity of regurgitation (e.g., pulmonary regurgitation), veins. These stacks should, ideally, each be acquired during a
size and function of the heart chambers (the right as well as breath-hold to assist accurate location of subsequent breath-
the left ventricle [LV]), and measurements of shunt flow. In hold cine acquisitions.
adults with CHD, the underlying cardiac anatomy is likely
to have been assessed previously. Although it is important Cine and Multislice Cine Imaging
to understand the descriptions, they should not necessar-
ily be assumed to be correct and should be reviewed in the SSFP imaging, which gives good contrast between blood and
light of new information gained from the MRI study. The myocardium, is well suited for imaging and measuring ven-
terminology used for malformed cardiovascular anatomy is tricular function and mass and for visualizing valve leaflets.
described in Chapter 29. Where there is jet flow through a stenotic or regurgitant ori-
CT offers unrivalled 3D spatial resolution in shorter fice, SSFP cine imaging can outline a coherent jet core, if
acquisition times than MRI. In patients with CHD, it has present, because of the localized loss of signal from shear
particular roles in the visualization of anomalous coronary layers (Fig. 31.1). Rapid acquisition makes it possible to
arteries and coronary arteries in relation to RV to PA con- interrogate a jet area precisely and repeatedly. The approach
duits where percutaneous dilatation and valve placement are of “cross-cutting”—locating an orthogonal slice through a
considered, and in the visualization of aortopulmonary col- partially visualized feature such as a valve orifice or jet—
lateral arteries, calcification, and stents in various locations. is an effective way of “homing in” on a particular feature
ECG-gated cine CT allows measurements of biventricular of interest by MRI. In addition, or alternatively, contigu-
size and function, albeit at a lower temporal resolution than ous stacks of cine images are easy to acquire and extremely
MRI, and subject to adequate opacification of each intraven- valuable for relatively comprehensive coverage of anatomi-
tricular blood volume. In patients with a pacemaker or an cal regions or structures. In the case of thin structures such
ICD, CT provides an alternative to MRI. However, exposure as the leaflets of the mitral valve or the atrial septum and
to ionizing radiation can be an issue, especially in relatively any defect in it, the orientation of cine slices, which typically
young patients who may need repeated follow-up (10). have 5- to 7-mm slice thickness, is a key consideration. Due
to possible partial volume effects, a thin structure is only
likely to be visualized clearly by cine slices oriented perpen-
Techniques dicular or near perpendicular to it.
Multislice Scout Imaging Flow Velocity Acquisitions

In CHD, it is advisable to acquire coronal and sagittal as Phase-shift velocity mapping, if correctly implemented, can
well as transaxial multislice images. To help with tissue be an accurate and versatile way of measuring flow (11,12).
LWBK1209-ch31_p489-502.indd 490 16/05/13 9:24 PM

Chapter 31 ■ Adult Congenital Heart Disease 491
However, the development of MRI systems with rapidly to map velocities in a plane aligned with the direction of
switching gradients for cardiovascular acquisition has led flow, with velocity encoded in the read-gradient direction,
to potential inaccuracies in velocity mapping, particularly which must also be aligned with the direction of jet flow.
when used for measurements of shunt flow and regurgi- This allows depiction of the jet in relation to upstream and
tant fraction. In these cases, a slight shift of the background downstream regions. It is necessary, however, to locate the
phase, due to eddy currents or any uncorrected Maxwell image plane correctly by cross-cutting.
gradients, can introduce significant errors in flow calcula- Jet velocity mapping can be particularly valuable for assess-
tion. Manufacturers have gradually become aware of these ing stenoses in which ultrasonic access is limited, such as in
problems and have implemented modifications of hardware aortic coarctation (native or repaired), ventriculopulmonary
and software to avoid or counteract them in more recent conduits, pulmonary artery branch stenoses, and obstructions
systems. Unfortunately, however, it cannot be assumed that at the atrial or atriopulmonary level following a Mustard,
all systems in current clinical use are accurate in this respect Senning, or Fontan operation.
(13). The acquisition of all three components of velocity
Phase-shift velocity mapping is based on the frequency relative to voxels distributed throughout a 3D volume for
changes experienced by nuclei moving relative to applied each of 20 or more phases through the cardiac cycle is also
magnetic gradients (11,12). The direction, steepness, and feasible. Such four-dimensional (4D) flow datasets require
timing of the applied gradients can be chosen, so that relatively long acquisition times of about 10 to 25 minutes
flows with a range of velocities, from low-velocity venous (19). This of course means that any instabilities or other
to high-velocity poststenotic flows, can be measured accu- beat-to-beat variations of flow are effectively averaged out,
rately. Clinical uses include measurements of cardiac out- which makes attempted calculations of streaming, mix-
put, valvar heart disease (14), shunt flow (15), collateral ing, and energy dissipation of fluid-wall shear problematic.
flow (16), regurgitant flow (17), and jet velocities through However, an important potential advantage, particularly in
stenoses (18). However, for successful clinical application, complex CHDs, is that multiple cavities, valves, and ves-
the operator should have an understanding not only of the sels can be covered in a single acquisition. The relative vol-
anatomy and pathophysiology of the operated CHD, but umes of flow passing through multiple valve orifices and
also of the available technical choices for velocity mapping. vessels can theoretically be calculated retrospectively (20).
It is necessary to select a plane, echo time, velocity-encoding This requires relatively advanced postprocessing, which is
direction, and velocity sensitivity appropriate for a particu- not yet widely available.
lar investigation.
Velocity can be encoded in directions that lie either
Three-Dimensional Acquisitions
in or through an image plane. The mapping of velocities
with and Without Contrast
through a plane transecting a vessel (velocity encoded in
the direction of the slice-selection gradient) allows the Gadolinium-enhanced 3D MRA combines fast acquisition
measurement of flow volume. The cross-sectional area of with good spatial resolution. Important applications of MRA
the lumen and the mean axially directed velocity within in adults with CHD are depiction of anomalous pulmonary
that area are measured for each phase through the heart or systemic venous return, depiction of the geometry and
cycle. From these data, a flow curve is plotted, and systolic dimensions of the aorta in the presence of coarctation with
forward flow and any diastolic reversed flow are computed a view to balloon dilatation and stenting, and depiction of
by integration. the size and distribution of aortopulmonary collaterals and
To calculate shunt flow, both aortic and pulmonary surgical shunts in pulmonary atresia (Fig. 31.2). Although
artery flows are measured. Two separate acquisitions are there may be some trade-off between spatial resolution and
generally needed, except in cases of TGA, in which the rapidity of acquisitions, shorter acquisitions allow several
aorta and pulmonary trunk usually run almost parallel to phases of contrast passage to image sequentially in a single
each other. For aortic flow measurement, a suitable plane breath-hold, which is valuable in the presence of complex
transects the aortic root at the level of the sinotubular cardiovascular connections or shunts.
junction, and for pulmonary flow measurement, a suit- Contrast-enhanced MRA is generally acquired in a
able plane transects the pulmonary trunk proximal to its breath-hold without cardiac gating, but this approach is
bifurcation. Pulmonary regurgitation is common after re- not suitable for coronary angiography due to the relative
pair of tetralogy of Fallot, and in such cases, the repaired importance of cardiac motion. For angiographic visualiza-
outflow tract and pulmonary trunk are tethered by scar- tion of the proximal course of coronary arteries, ECG-gated
ring. However, if the mobility of the aortic and pulmonary 3D SSFP acquisitions can be used, either in a single breath-
valves is normal, especially in cases in which the root is hold or with diaphragm monitoring that accepts only data
dilated, the diastolic regurgitant flow can be significantly acquired during a limited part of respiratory cycles over a
underestimated because of compliance and movement of period of minutes. The suppression of signal from fat im-
the arterial root relative to the imaging plane. A potential proves the delineation of coronaries. In the diagnosis and
solution to this problem will be movement of the acqui- assessment of Kawasaki disease, CVMR angiography can
sition plane to follow valve movement through the heart be used to visualize coronary artery aneurysms or stenosis.
cycle—“motion tracking.” ECG-gated 3D SSFP acquisitions are also widely used for
Through-plane velocity mapping can be used to delin- relatively high resolution anatomic evaluation in CHD and,
eate an atrial septal defect and measure the velocities of jets given the relatively bright signal from blood in all cavities,
through stenotic orifices (18). Alternatively, it can be helpful as the basis of 3D “road maps” to guide electrophysiologic
LWBK1209-ch31_p489-502.indd 491 16/05/13 9:24 PM

more reliable than those obtained by any other imaging

modality.
Late Gadolinium Enhancement

Late gadolinium enhancement (LGE) imaging, which makes
use of the property of intravenously injected gadolinium
chelate to linger in extracellular spaces in scarred myocar-
dial tissue after its concentration has begun to fall in the
bloodstream, is described more fully elsewhere. Ventricular
fibrosis imaging by LGE has been used for investigation of
patients with CHD, particularly after surgery. Scarring has
been found to be common in distinct locations. These are
often directly attributable to the interventions performed,
for example, scarring of the wall of the partially resected
infundibulum in patients after repair of Tetralogy of Fallot
(23). Scarring of the systemic RV in patients who had previ-
ously undergone Mustard or Senning procedures for TGA
was found to relate to adverse ventricular function and inci-
dence of arrhythmias (24). Evidence is accumulating that
abnormally extensive LGE correlates with, and may poten-
tially predict, arrhythmia.
Applications of Magnetic
Resonance Imaging/Magnetic
Resonance Angiography in
Specific Diseases
Figure 31.2. Contrast-enhanced MRA showing left and right Left Ventricular Outflow Tract
Blalock–Taussig shunts (asterisks) and an aortopulmonary collateral Obstruction and Aortic Stenosis
(arrow) supplying abnormal pulmonary artery branches in a patient
with pulmonary atresia. In congenital valvar and subvalvar aortic stenoses, it is
important to clarify the level(s) and the severity of stenosis
by appropriate acquisitions. In Shone’s syndrome, there may
be both subaortic and aortic stenoses, sometimes combined
interventions (21). Acquisition can be in more than one with mitral valve stenosis, and all should be looked for and
phase of the cardiac cycle (22). assessed. Our approach is to acquire cine images of the left
ventricular outflow tract (LVOT) in several planes: A stack
of cines in three-chamber orientation covering the LVOT, a
Measurement of Right Ventricular
coronal LVOT view orthogonal to these, and a stack of aor-
Function and Mass
tic valve cines transecting the sinus of valsalva, one of which
Ventricular measurements have been described in greater may be suitable for the direct planimetry of valve opening.
detail elsewhere. In CHD, measurements of the right ven- Velocity mapping can be performed with in-plane veloc-
tricle (RV) are at least as important as those of the LV, but ity encoding in the three-chamber LVOT view followed by
they pose particular challenges. The RV cavity is crossed through-plane velocity mapping at one or more levels, ap-
by coarse trabeculations, particularly near the apex. propriately labeled for subsequent analysis, through the sub-
These become thick and significant in summed volume aortic and aortic valves and immediate supra-aortic regions.
when the RV is hypertrophied, but even if clearly visual- When there is significant subaortic stenosis, it can be dif-
ized, trabeculations are difficult to outline individually ficult or impossible to be sure whether there is aortic valve
because of their complex structure. Furthermore, the base stenosis also. Even a mobile valve will not open fully where
of the RV tends to be more difficult to delineate than that a narrow jet passes through it, but comparison of jet ve-
of the LV. After repair of tetralogy of Fallot, the RV out- locities and different levels beneath, at, and above the valve
flow tract can be dilated and dyskinetic and may have no should allow determination of the level of the main stenosis.
effective pulmonary valve. All these factors can make it
difficult to decide on the limits of the RV. It is usual to
The Outflow Tracts After
include a dilated or aneurysmal outflow tract as part of the
The Ross Operation
RV, where it lies beneath the level of the pulmonary valve
annulus (4). A particular unit or researcher must establish The Ross operation is performed for replacement of a diseased
a reproducible protocol for the acquisition and analysis of aortic valve with the patient’s own pulmonary valve (a pul-
RV function, which, despite difficulties, are likely to prove monary autograft in aortic position). This in turn necessitates
LWBK1209-ch31_p489-502.indd 492 16/05/13 9:24 PM

replacement of the pulmonary valve, usually with a pulmo- (diastolic “tail”), which is a useful indicator of the obstruc-
nary homograft. It is the homograft in pulmonary position tive significance of coarctation.
that most often gives a cause for concern after Ross proce- When aortic coarctation is investigated in an adult (unop-
dure. The homograft may have a tendency to shrink and erated, balloon-dilated, or operated), a number of questions
become stenotic, with up to 10% of patients developing a must be considered:
peak velocity approaching 3 m/s in the first year or two after
•• Is blood pressure high in one or both arms?
Ross operation. This shrinkage is usually self-limiting. It can
•• To what extent is it caused by obstruction resulting from
be at one or both suture lines, or along the length of the tube
coarctation?
of the conduit (25). MRI with in- and through-plane veloc-
•• If operated, what type of surgery?
ity mapping is useful to determine the level and severity of
•• Is diastolic prolongation of forward flow present?
stenosis. Cine imaging of the LVOT and neoaortic sinus is
•• Do blood pressure and “gradient” rise markedly with
also required to look for any dysfunction of the autograft
exercise?
valve and possible dilatation of the autograft root in aortic
•• Is associated aortic valve disease present?
position.
•• Is the LV hypertrophied?
•• How extensive is the collateral flow?
Aortic Coarctation, Recoarctation, •• What is the location and severity of the coarctation?
and Aneurysm •• Of what type is it (e.g., membrane or narrow segment)?
•• What is the geometry of the aorta and its branches in the
Aortic coarctation consists of narrowing or occlusion of the
vicinity of the coarctation?
proximal descending thoracic aorta, in most cases just distal
•• Is dissection, aneurysm, or false aneurysm present?
to the left subclavian artery branch, but occasionally proxi-
mal to it. The geometry of the aorta is highly variable in No single modality answers all questions, but MRI with
adults with aortic coarctation after different types of repair. cine imaging and velocity mapping can generally deter-
The risks associated with repaired aortic coarctation are mine the nature and severity of coarctation (26) and iden-
related to systemic hypertension in the upper body, which tify any dissecting or false aneurysms (27). In cases likely
may be exacerbated by any residual coarctation; rupture of to require intervention or reintervention for either stenosis
an aneurysm or false aneurysm is also a risk. In this setting, or aneurysm formation, gadolinium-enhanced MRA of the
a resting peak velocity of 3 m/s is significant, particularly region should be performed to provide 3D visualization
if associated with a diastolic prolongation of forward flow (Fig. 31.3).
A B C
Figure 31.3. Unoperated aortic coarctation imaged by three different magnetic resonance techniques.
The turbo spin-echo image (A) gives no blood signal and differentiated signal from other types of tissue. The
SSFP cine image (B) gives bright blood with visualization of flow effects when viewed in cine mode. The 3D
gadolinium contrast angiogram (C) shows the lumen of the aorta and collateral arteries. The 3D dataset can
be rotated and viewed from any angle. Although there was no visible orifice through the coarctation on 3D
angiography, this patient was successfully treated by transcatheter balloon dilatation and stenting, the size and
potential suitability of the stent having been determined from these images.
LWBK1209-ch31_p489-502.indd 493 16/05/13 9:24 PM

A B D
Figure 31.4. Cardiac computed tomography can provide images of the stent structure as seen in this case
(A,B). Where there is concern to investigate for stent fracture cardiac CT has an important role. In this patient,
there was associated bicuspid valve (C) and ascending aortic root dilatation (D). (Images by kind permission of
Dr Michael Rubens, Royal Brompton Hospital.)
Poststenotic dilatation is common. It appears as fusi- Patent Ductus Arteriosus

form dilatation beyond a stenosed or previously stenosed
Patent ductus arteriosus is identifiable by MRI if sought. Flow
region and is usually distinguishable by its location and
through a patent ductus arteriosus, usually directed anteriorly
smooth contours from more sinister false aneurysmal dila-
in the top of the left pulmonary artery close to the pulmonary
tation, which may require reoperation or protection with
artery bifurcation, is detectable on appropriately located cine
a covered stent (Fig. 31.4). True or false aneurysms may
images. The amount of shunting can be assessed by measuring
complicate repairs, particularly those incorporating patches
flow in the pulmonary trunk and aorta. Importantly, although
of noncompliant fabric, such as Dacron. We recommend
excessive flow arrives via the duct to the branch pulmonary
that patients with such patches be followed regularly with
arteries, “pulmonary flow” measured at the level of the main
MRI examinations. Leakage of blood through a false an-
pulmonary artery (MPA) is less than that in the ascending
eurysm can lead to hemoptysis. In such cases, para-aortic
aorta if duct flow is from the aorta to the pulmonary artery
hematoma is generally well visualized by MRI, appearing
bifurcation.
bright, usually with diffuse edges, on spin-echo or SSFP im-
ages. Postoperative hematoma is common, however, and
Atrial and Ventricular Septal Defects
sometimes leaves a region of signal adjacent to the aorta,
which may be distinguished from a developing false aneu- Although atrial septal defects (ASDs) and ventricular septal
rysm only if a comparison of images over time is possible. defects (VSDs) can often be assessed satisfactorily by echo-
For this reason, it is worth acquiring baseline postoperative cardiography, MRI offers unrestricted access and makes it
images in adults who have undergone repeated surgery for possible to determine shunt flow from the difference between
coarctation. Dissecting aneurysm can result from attempted measurements of flow in the pulmonary artery and aorta. In
balloon dilatation (27). both VSDs and ASDs, a good way of searching for the defect
LWBK1209-ch31_p489-502.indd 494 16/05/13 9:24 PM

Figure 31.5. Double-chambered

right ventricle in diastole (A) and sys-
tole (B) seen in a routine short-axis
cine acquisition. Stenosis in the cav-
ity of the right ventricle is caused by
hypertrophied myocardial ridges of
bands, often associated with a peri-
membranous VSD. The arrow points
to the jet through the narrowing in A B
systole.
is with a stack of multiple contiguous cine images. These can which the RV obstruction is subinfundibular, resulting from
be acquired as a ventricular short-axis stack, working down hypertrophied muscular ridges or bands beneath an infun-
the ventricles from base to apex, followed by an atrial short- dibulum that is not hypertrophied and a pulmonary valve
axis stack, working back into the atria as far as the superior that is not stenosed. This can be corrected by surgical resec-
vena cava (SVC) from the basal slice. Each cine should be tion without the need for incision or replacement of the pul-
studied carefully for evidence of jet flow through a septal monary valve (28).
defect, with a further long-axis cine located to visualize any
jet seen, followed by through-plane velocity mapping to
Repaired Tetralogy of Fallot
measure the width and velocity of any trans-septal jet.
Pulmonary regurgitation is common 20 to 30 years after
repair of tetralogy of Fallot and is considered a risk factor
Pulmonary Hypertension
for progressive RV dysfunction, arrhythmia, and late sud-
MRI cannot measure pulmonary vascular pressure or resis- den cardiac death. MRI has important roles in the follow-
tance directly, but it can give important indirect evidence of up of tetralogy of Fallot (24). If an RV to PA conduit has
pulmonary arterial hypertension. For example, abnormal been placed and has become stenosed or regurgitant, the
hypertrophy of the walls and trabeculations of the RV and percutaneous placement of a stented valve within the con-
dilatation of the pulmonary trunk and proximal PA branches, duit may be considered. Prior to this, contrast CT imaging is
with reduction of pulsatility, are typical appearances of indicated to assess the degree of calcification and narrowing
pulmonary hypertension. With chronic pressure load, the of the conduit, and the proximity of a potentially vulner-
hypertrophied and perhaps increasingly dilated RV shows able coronary artery branch. Implanted bioprosthetic valves
laborious and prolonged systolic contraction relative to the have a finite lifespan and the optimal timing of pulmonary
LV, which leads to a paradoxical shift of the ventricular sep- valve replacement remains controversial. Information from
tum to the left at the onset of LV diastole. In assessing and MRI can contribute to the decision-making process (24).
monitoring pulmonary hypertension, CVMR allows measure- Late gadolinium imaging can contribute information on RV
ment of RV mass and function, and of cardiac output, which fibrosis (23) (Fig. 31.6). MRI studies have shown evidence
may be indexed to body surface area. of functional improvements after the elimination of pulmo-
nary regurgitation by surgical replacement of the valve (29).
Interestingly, in free pulmonary regurgitation, without a
Right Ventricular Outflow Tract
functional pulmonary valve at all, the regurgitant fraction is
and Conduit Obstruction
usually only about 40% to 45%, where the regurgitant flow
The RV outflow tract and conduits from the RV to the pul- volume is expressed as a percentage of the forward flow vol-
monary artery are often inadequately visualized by ultraso- ume. The pulmonary regurgitant volume originates mainly
nography, so MRI can make important contributions in this from the recoil of the elastic pulmonary arteries in diastole,
region (24). For cine imaging and in-plane velocity mapping, and the amount of regurgitation depends on the size and
an oblique sagittal slice located with respect to transaxial compliance of these vessels and the location of any stenosis
scouts is usually suitable. In some cases, it is also helpful to relative to the compliance, not only on the incompetence of
orient the plane with respect to coronal scouts. It is important the valve itself (30) (Fig. 31.7).
to identify the level of stenosis. The level of the pulmonary The long-term follow-up of patients with a repaired te-
valve can usually be identified on cine imaging by visualiz- tralogy of Fallot should include measurement of the pulmo-
ing the valve cusps as they close at the onset of diastole—for nary regurgitant fraction by through-plane velocity mapping
example, by the location of a regurgitant jet at the moment of and measurement of the RV function and volume, particu-
valve closure. Obstruction may be subvalvar or supravalvar— larly the end-systolic volume increase which is probably the
for example, at the distal suture line of a homograft conduit. best marker of RV failure in response to volume overload.
An important variant is double-chambered RV (Fig. 31.5), in However, measurement of RV volume and function is not
LWBK1209-ch31_p489-502.indd 495 16/05/13 9:24 PM

Figure 31.6. Typical late gado-

A B linium enhancement patterns late
after repair of tetralogy of Fallot.
Fibrosed regions appear white in con-
trast to the black appearance of normal
myocardium. Typical sites of enhance-
ment in adults after repaired tetralogy
of Fallot include the anterior wall of
the RVOT (A; upward pointing arrow,
B), the region of surgical patching of
the ventricular septal defect (C), the
site of apical vent insertion in the LV
where this was part of previous surgery
(D), and the inferior (lower arrow) and
superior (upper arrow) RV–LV insertion
points. (From Babu-Narayan SV, Kilner
PJ, Li W, et al. Ventricular fibrosis
suggested by cardiovascular magnetic
resonance in adults with repaired tetral-
ogy of fallot and its relationship to
adverse markers of clinical outcome.
C D
Circulation. 2006;113:405–413.)
straightforward. Reproducible measurements require time right atrium (RA) may become grossly enlarged. An ASD,
and care. The complex blood–myocardial borders of a vol- VSD, tricuspid stenosis, or pulmonary stenosis may be pres-
ume-loaded and hypertrophied RV, which, for consistency, ent. In adults, MRI can be useful to assess the degree of RA
should be taken to include any thin-walled aneurysmal region dilatation, the size and function of the RV, the severity of
of outflow tract up to the expected level of the pulmonary tricuspid regurgitation, and the mobility of the dysplastic
valve annulus, have to be painstakingly traced. A second valve leaflets.
observer is only likely to reproduce the measurements of a A simple initial approach to imaging these patients is to
first observer if an agreed method of boundary recognition, acquire a transaxial stack of cines, starting at the inferior bor-
including the RA–RV and the RV–PA boundaries, is adhered der of the ventricles and working up to the level of the pul-
to. Assessment should also include study of the pulmonary monary arteries. This allows relatively comprehensive cover-
arteries and measurement of any aortic dilatation. age to search for the jet or stream of tricuspid regurgitation,
Assessment of a stenotic or regurgitant RV to PA conduit which may be surprisingly displaced through malposition of
with a view to percutaneous dilatation and valve placement parts of the tricuspid valve. Once the regurgitant stream has
should include CT as well as MRI (Fig. 31.8). been located and imaged in several slices, through-plane ve-
locity mapping is valuable to determine the cross-sectional
area of the regurgitant jet, which can be as broad as 1 cm
Ebstein Anomaly
or more in diameter. A transaxial stack of cines may also
Ebstein anomaly is a malformation of the tricuspid valve, be more suitable than the conventional short-axis stack for
with displacement of the septal and posterolateral leaflets attempting to measure volumes of the functional part of the
toward the apex from their usual position at the atrioven- RV in some Ebstein patients, although the location of the
tricular junction (31). This is associated with atrialization tricuspid valve may not be easy to determine in either.
of part of the RV and varying degrees of tricuspid regurgi- Cine imaging vertically (oblique coronal) in alignment
tation. Depending on the severity of the malformation, the with the right atrioventricular axis is valuable for assessment
LWBK1209-ch31_p489-502.indd 496 16/05/13 9:24 PM

Figure 31.7. Different amounts of pulmonary regurgitation in two patients, both lacking competent pulmonary valves. Patient A (left panels)
is a 20-year-old man with tetralogy of Fallot repaired with a homograft right ventricle to pulmonary artery conduit and graft augmentation of the
LPA. Patient B (right panels) is a 25-year-old man of similar body surface area, born with pulmonary and subpulmonary stenosis and a VSD, who
had patch reconstruction of the RVOT. Cine imaging (upper row) showed differences of size and expansion of the proximal pulmonary arteries,
replaced by a conduit and graft in patient A, but dilated and expansile in patient B, as indicated by the diastolic > systolic diameter measurements
shown. In-plane, vertically encoded velocity maps aligned with the right ventricular outflow tract (second row) show no effective valve action in
diastole. Arrows represent the direction of flow. Flow curves (third row) were plotted from acquisitions of velocities through planes transecting the
proximal MPA or conduit, giving the regurgitant volumes and fractions shown. The peak systolic velocity was 3 m/s in the conduit of patient A
and 1 m/s in the MPA of patient B. Four-chamber cine images are shown (bottom row). M , male; BSA, body surface area; MPA, main pulmonary
artery; RPA, right pulmonary artery; LPA, left pulmonary artery; RV, right ventricle; PR, pulmonary regurgitation; EDV, end diastolic volume; ESV,
end systolic volume; SV, stroke volume; EF, ejection fraction. (From Kilner PJ, Balossino R, Dubini G, et al. Pulmonary regurgitation: The effects of
varying pulmonary artery compliance, and of increased resistance proximal or distal to the compliance. Int J Cardiol. 2009;133:157–166.)
LWBK1209-ch31_p489-502.indd 497 16/05/13 9:24 PM

Conduit
A B C
D E F
Figure 31.8. Conduit assessment and suitability for transcatheter approaches. Three-dimensional SSFP
MRI (A–C) and corresponding cardiac CT postcontrast (D–F) in a patient status post redo surgery with RV–
PA homograft repair of fallot-type pulmonary atresia and VSD being considered for percutaneous pulmonary
valve implantation (PPVI) for pulmonary regurgitation and stenosis. From the interventionalist’s viewpoint, the
proximity of the coronary arteries to the calcified and stenosed region of the conduit is important information,
indicating a potential risk of coronary compression by the deployment and expansion of a stent. (Non-CVMR
data with kind permission from Dr. Michael Rubens and Dr. Anselm Uebing, Royal Brompton Hospital.) LCA,
left coronary artery; RCA, right coronary artery.
of the posterolateral leaflet and its displacement. A three- discordance. About two-thirds of such patients have no major
chamber LVOT cine can show compression of the LV by the additional malformations, and are said to have “simple” trans-
distended right side of the heart, which may compromise LV position. The other third has associated abnormalities, such as
filling in diastole. A four-chamber view should be acquired, a VSD and pulmonary or subpulmonary stenosis. Since unop-
aligned with the long axis of the LV, and tilted down ante- erated simple transposition is not compatible with long-term
riorly to pass through the abnormal tricuspid orifice. Cine survival, patients seen as adults are likely to have undergone
imaging in additional planes through the tricuspid valve surgery. The current surgical approach to simple transposi-
and RV outflow tract may be helpful toward decision mak- tion is usually to switch the great arteries with respect to the
ing regarding possible surgical repair of the tricuspid valve. ventricles in the first years of life, a technically difficult opera-
Although echocardiography may better define the valve leaf- tion that requires reinsertion of the coronary arteries in the
lets, MRI has the advantage of providing wide fields of view neoaortic root (formerly the pulmonary root) (32–34). Such
across the enlarged volume of the heart. patients may be referred for MRI if there is concern regarding
right ventricular outflow tract (RVOT) or branch PA obstruc-
tion, which is a relatively common complication. Imaging of
Transposition of the Great Arteries
both the RVOT to PA, the PA branches, and the neoaortic
Repaired by Arterial Switch, Mustard,
root all give clinically important information (Fig. 31.9).
Senning, and Rastelli Operations
Older surviving adults are likely to have had surgery by
In complete TGA, the atria and ventricles are normally either a Mustard or a Senning procedure (Fig. 31.10), in
related but the aorta and pulmonary artery are transposed which the flow paths are switched at the level of the atria;
with respect to the ventricles resulting in ventriculoarterial in this case, the ventriculoarterial connections having been
LWBK1209-ch31_p489-502.indd 498 16/05/13 9:24 PM

A B
C D
Figure 31.9. Transposition of the great arteries after arterial switch operation. Panels A and B are from 3D
bSSFP acquisitions. Panels C and D are from SSFP cine imaging in a 17-year-old postneonatal arterial switch
operation and ventricular septal defect closure. The aortic and main pulmonary artery positions have been
switched so that the aorta comes off the left ventricle (A) and the pulmonary artery from the right ventricle
(B). The outflow tracts are seen in image C. There is mild neoaortic root dilatation (41 × 36 mm at sinus level
in systole) with mild supravalvar aortic stenosis. After a Lecompte maneuver, the branch pulmonary arteries
straddle the aorta (D). In this case there is no evidence of associated branch pulmonary artery obstruction.
Ao, aorta; LV, left ventricle; PA, pulmonary artery; RV, right ventricle; RPA, right pulmonary artery; LPA, left
pulmonary artery.
left discordant (35). Blood is redirected at the atrial level takes experience to assess the patency of all three atrial flow
with a baffle (Mustard operation) or with part of the in- paths. Oblique planes for cine imaging and velocity mapping
turned atrial wall (Senning operation). A physiologic cor- may be located with respect to sagittal scouts. Since it can
rection is achieved, although the RV continues to support be difficult to align a single plane with both the superior and
the systemic circulation. The imaging of patients after a inferior caval pathways, cross cuts may be needed to decide
Mustard or Senning operation requires an assessment of the whether or not the pathways are stenosed. Obstruction can
pulmonary venous flow paths which pass from the pulmo- occur in one or more of the atrial flow paths, typically be-
nary veins to the right of the curved baffle, to the tricuspid tween the baffle and a ridge that may represent tissue where
valve. Both channels of the systemic venous flow paths must the native atrial septum was resected. Such obstruction at
also be assessed, which can be challenging. In order that no the atrial level does not generate jets with a very high veloc-
parts of these channels are missed, a stack of slightly oblique ity; a peak diastolic velocity between 1 and 2.5 m/s may be
coronal cine images, aligned with the entries of both caval found distal to a narrowing that is significantly obstructive.
veins to the atrial regions, is recommended. Channels from Gradual obstruction of one of the two caval flow paths is
the superior and inferior venae cavae converge on the mitral generally well tolerated as the azygous vein(s) dilates to di-
valve, which lies to the left of the saddle-shaped baffle. It vert flow to the other caval pathway. Since the hypertrophied
LWBK1209-ch31_p489-502.indd 499 16/05/13 9:24 PM

A B C
D E F
G H I
Figure 31.10. Transposition of the great arteries after atrial redirection surgery. Images A–C from SSFP
cine MRI. The parallel arrangement of the anterior aorta and posterior pulmonary artery and the subaortic RV
are shown in A. This patient underwent the Mustard operation whereby oxygenated blood from the pulmo-
nary veins drains the region of the original left atrium (asterisk, B,C). The atrial septum has been removed and
pulmonary venous blood is now directed anteriorly, to the right of the baffle (solid arrows, B,C) to the right
ventricle and the aorta. Deoxygenated blood from the SVC and IVC passes to the left of the baffle to the left
ventricle and the pulmonary artery. The atrial pathways are unobstructed in this case. Images D,G are selected
from 3D bSSFP imaging and show that the coronary origins can be imaged. Images E and F are 2D inver-
sion recovery images late after gadolinium administration. A small transmural area of systemic RV fibrosis is
shown in two cross-cutting views (dotted arrows) also seen with 3D inversion recovery (H). A cine following
gadolinium in the same plane as the fibrosis confirms thinning and regional wall motion abnormality (H; dot-
ted arrow). Ao, aorta; LCA, left coronary artery; LV, left ventricle; PA, pulmonary artery; RCA, right coronary
artery; RV, right ventricle.
RV delivers systemic pressure in these patients, it may be im- located aortic root. Either of the reconstructed outflow paths
portant to assess its function by volume measurements and may become obstructed—for example, if the VSD becomes
assess any tricuspid regurgitation. restrictive or if the external conduit is compressed behind
In the small proportion of patients with TGA (<10%) the sternum.
who also have a VSD and pulmonary or subpulmonary ste-
nosis, a Rastelli operation (Fig. 31.11) may be performed
Fontan Operation for Functionally
(36). Blood is redirected at the ventriculoarterial level so
Single Ventricle
that the LV supports the systemic circulation. A valved con-
duit is placed from the anterior wall of the RV to the pulmo- Infants born with only one adequately sized ventricle may
nary artery, the native pulmonary valve being closed, and LV survive, initially, through the mixing of systemic and pul-
outflow is guided via the VSD and a patch to the anteriorly monary venous blood, but sooner or later their condition is
LWBK1209-ch31_p489-502.indd 500 16/05/13 9:24 PM

Figure 31.11. Reconstructed ven-

triculoarterial anatomy after Rastelli
operation for transposition of the great
arteries with ventricular septal defect
and pulmonary stenosis. The patch
(black arrow) is positioned to direct
flow from the left ventricle to the aorta
via the ventricular septal defect. The
right ventricle-to-pulmonary artery
conduit (white arrow) is located
retrosternally. In this case, it is mildly
narrowed, with a peak velocity of
2.8 m/s. The native pulmonary valve
was closed at surgery. Ao, aorta; LV,
left ventricle; RV, right ventricle; PA,
pulmonary artery.
likely to deteriorate because of ventricular volume overload, to assess contractile function of the ventricle, competence of
outflow obstruction, or inadequate or excessive pulmonary its inflow valve, and patency of its outflow tract.
flow. The Fontan operation aims to eliminate shunting, rout-
ing systemic venous return to the pulmonary arteries with-
out passage through an intervening ventricle, so that the Summary
systemic and pulmonary flows are in series, propelled by
the one ventricle (37). In this abnormal circulation, pressure Cardiovascular MRI, MRA, and CT offer unrestricted access
is elevated in the systemic veins and this residual systemic to structures throughout the chest, including the RV and
pressure maintains flow through the lungs back to the left great arteries. They therefore have very important contribu-
atrium. Any obstruction of the systemic vein to pulmonary tions to make in the diagnosis and follow-up of adults with
artery flow path easily raises the systemic venous pressure to CHD. Examples of the principal applications have been pro-
an unsustainable level (38). vided. However, investigations can be costly and time consum-
The Fontan operation was originally performed by con- ing with regard to both acquisition and analysis. These costs
necting the RA to the pulmonary arteries, either via a con- should be weighed against the potential expense of incomplete
duit or by direct anastomosis and patching of the region of assessment and inappropriate management. Variations of
the right atrial appendage to the pulmonary arteries. In the underlying anatomy and surgical procedures between patients
last two decades, total cavopulmonary connection (TCPC), mean that planes and sequences must be selected during acqui-
either by intra-atrial tunnel or extracardiac conduit, has sition. It is therefore important for imaging specialists to gain
been used (39). The superior vena cava is connected to the experience in the growing subspecialty of adult CHD, and in
upper wall of the right pulmonary artery, and from below, this field, collaboration between specialists is essential.
flow from the inferior vena cava is channeled by a patch,
flap, or conduit up the side of the RA to the pulmonary ar-
tery. The right and left pulmonary arteries communicate, References
and the pulmonary trunk is disconnected from the heart.
1. Baumgartner H, Bonhoeffer P, De Groot NM, et al. ESC Guidelines for the management of
A small residual atrial shunt may be created deliberately grown-up congenital heart disease (new version 2010). Eur Heart J. 2010;31:2915–2957.
at operation to moderate systemic venous pressure at the 2. Gatzoulis MA, Webb GD, Daubeney PEF, eds. Diagnosis and Management of Adult
Congenital Heart Disease. London, UK: Churchill Livingstone; 2010.
cost of a small systemic venous–pulmonary venous shunt 3. Moons P, Meijboom FJ, Baumgartner H, et al. Structure and activities of adult congenital
in the postoperative period. This fenestration either closes heart disease programmes in Europe. Eur Heart J. 2010;31:1305–1310.
spontaneously in due course or is occluded by a device. 4. Kilner PJ, Geva T, Kaemmerer H, et al. Recommendations for cardiovascular magnetic
resonance in adults with congenital heart disease from the respective working groups of the
Whichever variant of the Fontan operation is used, it is European Society of Cardiology. Eur Heart J. 2010;31:794–805.
crucial that cavopulmonary flow paths are visualized and 5. Roest AA, de Roos A. Imaging of patients with congenital heart disease. Nat Rev Cardiol.
shown to be unobstructed by cine imaging or angiography. 2011;9:101–115.
6. Gupta-Malhotra M. An approach to imaging adult congenital heart disease: Pitfalls and
Phase velocity mapping to measure velocities in regions of pearls. Methodist Debakey Cardiovasc J. 2011;7:18–25.
any suspected narrowing is important. It must be remem- 7. Silversides CK, Marelli A, Beauchesne L, et al. Canadian Cardiovascular Society 2009
Consensus Conference on the management of adults with congenital heart disease: Executive
bered that peak velocity at an atriopulmonary connection co- summary. Can J Cardiol. 2010;26:143–150.
incides with atrial systole, making retrospective ECG gating 8. Muller L, Saeed M, Wilson MW, et al. Remote control catheter navigation: Options for
mandatory, and that apparently minor elevations of velocity guidance under MRI. J Cardiovasc Magn Reson. 2012;14:33.
9. Muthurangu V, Taylor A, Andriantsimiavona R, et al. Novel method of quantifying pul-
between 1 and 2 m/s are significant in this setting. Evidence monary vascular resistance by use of simultaneous invasive pressure monitoring and phase-
of thrombus should be sought in the RA. It is also important contrast magnetic resonance flow. Circulation. 2004;110:826–834.
LWBK1209-ch31_p489-502.indd 501 16/05/13 9:24 PM

10. Siripornpitak S, Pornkul R, Khowsathit P et al. Cardiac CT angiography in children with 24. Geva T. Repaired tetralogy of Fallot: The roles of cardiovascular magnetic resonance
congenital heart disease. Eur J Radiol. 2011. [Epub ahead of print] in evaluating pathophysiology and for pulmonary valve replacement decision support.
11. Firmin DN, Nayler GL, Kilner PJ, et al. The application of phase shifts in NMR for flow J Cardiovasc Magn Reson. 2011;13:9.
measurement. Magn Reson Med. 1990;14:230–241. 25. Carr-White GS, Kilner PJ, Hon JK, et al. Incidence, location, pathology, and significance of
12. Srichai MB, Lim RP, Wong S, et al. Cardiovascular applications of phase-contrast MRI. pulmonary homograft stenosis after the Ross operation. Circulation. 2001;104:16–20.
Am J Roentgenol. 2009;192:662–675. 26. Muzzarelli S, Meadows AK, Ordovas KG, et al. Prediction of hemodynamic severity of
13. Gatehouse PD, Rolf MP, Graves MJ, et al. Flow measurement by cardiovascular magnetic coarctation by magnetic resonance imaging. Am J Cardiol. 2011;108:1335–1340.
resonance: A multi-centre multi-vendor study of background phase offset errors that can 27. Fawzy ME, Fathala A, Osman A, et al. Twenty-two years of follow-up results of balloon
compromise the accuracy of derived regurgitant or shunt flow measurements. J Cardiovasc angioplasty for discreet native coarctation of the aorta in adolescents and adults. Am Heart
Magn Reson. 2010;12:5. J. 2008;156:910–917.
14. Dall’Armellina E, Hamilton CA, Hundley WG. Assessment of blood flow and valvular 28. Singh NK, Karn JP, Gupt A, et al. Double chambered right ventricle with ventricular septal
heart disease using phase-contrast cardiovascular magnetic resonance. Echocardiography. defect presenting in adulthood. J Assoc Physicians India. 2011;59:451–453.
2007;24:207–216. 29. Therrien J, Provost Y, Merchant N, et al. Optimal timing for pulmonary valve replacement
15. Prompona M, Muehling O, Naebauer M, et al. MRI for detection of anomalous pulmo- in adults after tetralogy of Fallot repair. Am J Cardiol. 2005;95:779–782.
nary venous drainage in patients with sinus venosus atrial septal defects. Int J Cardiovasc 30. Kilner PJ, Balossino R, Dubini G, et al. Pulmonary regurgitation: The effects of varying
Imaging. 2011;27:403–412. pulmonary artery compliance, and of increased resistance proximal or distal to the compli-
16. Muzzarelli S, Ordovas KG, Higgins CB, et al. Collateral flow measurement by phase-contrast ance. Int J Cardiol. 2009;133:157–166.
magnetic resonance imaging for the assessment of systemic venous baffle patency after atrial 31. Tobler D, Yalonetsky S, Crean AM, et al. Right heart characteristics and exercise param-
switch repair for transposition of the great arteries. J Thorac Imaging. 2012;27:175–178. eters in adults with Ebstein anomaly: New perspectives from cardiac magnetic resonance
17. Gabriel RS, Renapurkar R, Bolen MA, et al. Comparison of severity of aortic regurgi- imaging studies. Int J Cardiol. 2011. [Epub ahead of print]
tation by cardiovascular magnetic resonance versus transthoracic echocardiography. Am 32. Warnes CA. Transposition of the great arteries. Circulation. 2006;114:2699–2709.
J Cardiol. 2011;108:1014–1020. 33. Giardini A, Khambadkone S, Taylor A, et al. Effect of abnormal pulmonary flow distri-
18. Garcia J, Marrufo OR, Rodriguez AO, et al. Cardiovascular magnetic resonance evalu- bution on ventilatory efficiency and exercise capacity after arterial switch operation for
ation of aortic stenosis severity using single plane measurement of effective orifice area. transposition of great arteries. Am J Cardiol. 2010;106:1023–1028.
J Cardiovasc Magn Reson. 2012;14:23. 34. Ho JG, Cohen MD, Ebenroth ES, et al. Comparison between transthoracic echocardio
19. Markl M, Kilner PJ, Ebbers T. Comprehensive 4D velocity mapping of the heart and great graphy and cardiac magnetic resonance imaging in patients status post atrial switch proce-
vessels by cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2011;13:7. dure. Congen. Heart Dis. 2012;7:122–130.
20. Valverde I, Nordmeyer S, Uribe S, et al. Systemic-to-pulmonary collateral flow in patients 35. Lu JC, Dorfman AL, Attili AK, et al. Evaluation with cardiovascular MR imaging of baf-
with palliated univentricular heart physiology: Measurement using cardiovascular mag- fles and conduits used in palliation or repair of congenital heart disease. Radiographics.
netic resonance 4D velocity acquisition. J Cardiovasc Magn Reson. 2012;14:25. 2012;32:E107–E127.
21. Ernst S, Babu-Narayan SV, Keegan J, et al. Remote-controlled magnetic navigation and 36. Vouhe PR, Tamisier D, Leca F, et al. Transposition of the great arteries, ventricular septal
ablation with 3D image integration as an alternative approach in patients with intra-atrial defect, and pulmonary outflow tract obstruction. Rastelli or Lecompte procedure? J Thorac
baffle anatomy. Circ Arrhythm Electrophysiol. 2012;5(1):131–139. Cardiovasc Surg. 1992;103:428–436.
22. Hussain T, Lossnitzer D, Bellsham-Revell H, et al. Three-dimensional dual-phase whole-heart 37. Gewillig M, Brown SC, Eyskens B, et al. The Fontan circulation: Who controls cardiac
MR imaging: Clinical implications for congenital heart disease. Radiology. 2012;263(2): output? Interact Cardiovasc Thorac Surg. 2010;10:428–433.
547–554. 38. Angeli E, Pace Napoleone C, Balducci A, et al. Natural and modified history of single-
23. Babu-Narayan SV, Kilner PJ, Li W, et al. Ventricular fibrosis suggested by cardiovascu- ventricle physiology in adult patients. Eur J Cardiothorac Surg. 2012;42:996–1002.
lar magnetic resonance in adults with repaired tetralogy of fallot and its relationship to 39. de Leval MR, Deanfield JE. Four decades of Fontan palliation. Nat Rev Cardiol. 2010;7:
adverse markers of clinical outcome. Circulation. 2006;113:405–413. 520–527.
LWBK1209-ch31_p489-502.indd 502 16/05/13 9:24 PM

PART
5
Vascular Diseases
LWBK1209-ch32_p503-544.indd 503 16/05/13 9:21 PM

LWBK1209-ch32_p503-544.indd 504 16/05/13 9:21 PM
C h apt e r
Rossella Fattori
Vincenzo Russo 32
MRI and CT of Thoracic Aorta
■■ Introduction With its ability to delineate the intrinsic contrast

between blood flow and vessel wall, and acquire images
■■ Imaging Techniques in multiple planes with a wide field of view, MRI provides
MRI Techniques a high degree of reliability in the diagnosis of aortic dis-
CT Techniques eases, acute and chronic. MRI is totally noninvasive and
■■ Acquired Diseases of the Thoracic Aorta can be repeated, so that the progression of the disease over
Aortic dissection the time can be evaluated. Functional information can
Intramural hematoma be obtained by gradient-echo sequences and phase map-
ping, which quantify blood flow volume and velocity, so
Aortic ulcers
that our knowledge of aortic function can be expanded.
Postoperative evaluation of aortic dissection The new MRA techniques have enhanced the noninvasive
Aortic aneurysms evaluation of vascular pathology, providing a high degree
Aortic trauma of spatial and contrast resolution and in many instances
Indication and follow-up of endovascular aortic making invasive x-ray angiography an obsolete procedure
repair with CT technique for the detection of aortic diseases. In addition, the ability
Aortitis to differentiate tissue structures at a power of resolution in
■■ Congenital Aortic Diseases the order of micrometers provides an incomparable accu-
racy for the analysis of atherosclerotic plaque. At present,
Aortic arch anomalies MRI microscopy, intravascular MRI, and spectroscopy are
Aortic Coarctation emerging techniques that will be used to understand ath-
Aortic pseudocoarctation erosclerotic disease and its contributing pathogenic mecha-
Aneurysms of the Valsalva sinus nisms more clearly.
CT study of congenital aortic disease Noninvasive vascular system imaging using computed
MRI and CT of human atherosclerotic tomography angiography (CTA) has become an impor-
plaque tant technique in the evaluation of vessels and has already
been proven to yield high accuracy in the assessment of the
thoracic–abdominal aortae and its major branches. Helical
Introduction single detector CT (SDCT), introduced to clinical routine
imaging in the early 1990s, revolutionized body imaging
In the past few years, considerable interest in aortic diseases through the use of slip-ring technology, demonstrating
has been shown in the medical literature. The prevalence its superiority over conventional angiography in different
of aortic disease appears to be increasing in the Western applications (1–3). It created volume acquisition of image
population, likely corresponding to aging of the population data using continuous patient translation during gantry
in addition to heightened clinical awareness. The continu- rotation. The advantage of three-dimensional (3D) data-
ous advances in our understanding of aortic pathology have set allows for visual vessel assessment from any angle and
been based on molecular and cellular studies elucidating the rapidly became a standard diagnostic tool for large vessel
mechanisms of many pathologic conditions of the aorta and investigation. However, the visualization of small aortic
the complex interaction of this vessel with the cardiovascu- branches and extended vascular territories has been limited
lar system. However, increased clinical observation of aortic due to the restricted volume coverage (20 to 30 cm in one
diseases, demonstrated in epidemiologic studies, and a more breath-hold with moderate slice collimation). Another limi-
appropriate definition of its pathologic substrate, reported tation of this technique was breathing and pulsation arti-
in recent literature, may also reflect concurrent outstanding facts, due to the long scan time (often more than 30 seconds)
progress in imaging techniques. Among the imaging modali- and to the slow gantry rotation time (1 second), without
ties, MRI and CT offer the greatest versatility. any cardiac gating.
505
LWBK1209-ch32_p503-544.indd 505 16/05/13 9:21 PM

506 Part 5 ■ Vascular Diseases
The development of multidetector-row CT (MDCT) in the

late 1990s represented the most significant recent advance-
ment in helical CT. Thinner collimations, faster gantry rota-
tion times, large detector arrays, powered x-ray tubes, and
increased table speed dramatically improved image quality
and expanded the applications and indications of CT non-
invasive vascular imaging. Multidetector CT angiography
(MDCTA) provides a low-risk, efficient and cost-effective
evaluation of the arterial vascular system, allowing greater
vessel length and smaller diameter to be visualized. In a single
study, information of the vessel lumen and wall may be ob-
tained together with extravascular information (4).
Latest development of MDCT scanners, in the early
2000s, with the use of 8, 16, 32, or even 64 detector-
rows or more, now give a submillimeter, isotropic 3D data
acquisition of extended anatomical ranges, enabling high-
quality vascular imaging with 2D–3D artifact-free recon-
structions from virtually any angle and in any desirable
plane (5–10).
Imaging Techniques Figure 32.1. Fast spin-echo axial image of a large atheroma-
tous plaque protruding in the aortic lumen of the descending aorta
(arrows).
MRI Techniques
Spin-echo MRI
Conventional spin-echo T1-weighted imaging provides the substantial saving of time and improvements of image qual-
best anatomic detail of the tissue of the aortic wall and is ity. Conventional T2-weighted spin-echo sequences are of
still the basis of any aortic study (11,12). With the spin-echo little utility in the study of the thoracic aorta. Because of the
technique, a presaturation pulse is used to null the signal low signal-to-noise ratio and long acquisition time, image
of the blood pool. The “signal void” produced by flowing quality is affected, mainly by motion artifacts. On the con-
blood throughout most of the cardiac cycles provides a nat- trary, with fast spin-echo techniques it is possible to obtain
ural contrast between the lumen and the layers of the ves- high-resolution T2-weighted images that provide a further
sel wall. Electrocardiographic (ECG) triggering is essential option for evaluating the aortic wall. With a short acqui-
to minimize motion and pulsatility artifacts. An echo time sition time, respiratory motion can be suppressed using a
(TE) of 20 to 30 milliseconds is standard, and the repetition breath-hold technique.
time (TR) is determined from the R–R interval of the ECG. Usually, a conventional study of the thoracic aorta is first
Slice thickness of 3 to 8 mm and high-resolution parameters acquired in the axial plane (Table 32.1) for a display of the
(matrix size and signal average) ensure a detailed definition orientation of the great arteries and optimal visualization
of the morphology of the aorta and surrounding structures. of mural lesions perpendicular to their long axes. Images in
T1-weighted ECG-gated spin-echo images reveal the ana- additional planes, depending on the anatomy and diagnos-
tomic details of the aortic wall and pathologic conditions, tic problems, are then acquired; for example, the oblique
such as atheromatous plaques, intimal flaps, or intramural sagittal view shows the entire extent of the thoracic aorta
hemorrhage (Fig. 32.1). In spin-echo imaging, each section and supra-aortic vessels in a single image. Fast spin-echo
corresponds to a different cardiac phase. Diastolic slow flow sequences allow the acquisition of high-quality T2 images
and entry or exit slice phenomena may produce high sig- useful in tissue characterization of the aortic wall and blood
nal intensity in the aortic lumen that can mask underlying components.
luminal pathology or erroneously simulate mural plaques
or thrombosis. A shorter acquisition time can be achieved
Gradient-echo MRI
with fast spin-echo pulse sequences. In a fast spin-echo
sequence, a long train of echoes is acquired by using a series Gradient-echo techniques provide dynamic and functional
of 180-degree radiofrequency (RF) pulses; as a result, wash- information, although with fewer details of the vessel wall.
out effects are even more substantial than those obtained In gradient-echo imaging, data are acquired with monitor-
with conventional spin-echo techniques. A superior black ing of the ECG such that multiple images at the same slice
blood effect is achieved by using preparatory pulses (13). are reconstructed to represent multiple phases of the cardiac
Preparatory pulses such as presaturation, dephasing gra- cycle. The bright signal of the blood pool on gradient-echo
dients, and preinversion pulses involve the application of images results from flow-related enhancement obtained by
one or more additional RF pulses outside the plane to sup- applying RF pulses to saturate a volume of tissue. With a
press the signal intensity of in-flowing blood and nullify the short TR (20 to 40 milliseconds) and low flip angle (30 to
blood signal. The replacement of conventional spin-echo 40 degrees), maximal signal is emitted by blood flowing in
sequences with fast spin-echo sequences has resulted in the voxel. An ECG signal is acquired with the imaging data
LWBK1209-ch32_p503-544.indd 506 16/05/13 9:21 PM

Chapter 32 ■ MRI and CT of Thoracic Aorta 507
Flow Mapping
eneral Strategy of MR
G
Table 32.1
Study of Thoracic Aorta Gradient-echo imaging permits the semiquantitative mea-
surement of flow turbulence by sizing the signal void in the
CONGENITAL ANOMALIES (COARCTATION, AORTIC heart chambers and great vessels. However, various imag-
ARCH ANOMALIES) ing parameters can influence this measure and produce erro-
Spin-echo MRI Axial (slice thickness 5–7 mm)
neous results. Accurate quantitative information on blood
— Sagittal (coarctation)
— Coronal (aortic arch anomalies)
flow is obtained from modified gradient-echo sequences
GRE/flow mapping — with parameter reconstruction from the phase rather than
(coarctation) the amplitude of the MR signal, known as flow mapping or
MRA — velocity-encoded cine MRI (15–17). Most of the MR meth-
MARFAN SYNDROME —
ods of measuring flow velocity are based on the flow depen-
Spin-echo Axial (slice thickness 5–7 mm) dence of MR signal. In each pixel of velocity images, the
— Sagittal (slice thickness 3–5 mm) phase of the signal is related to the velocity component in
GRE/flow mapping Axial, ascending aorta (aortic the direction of a bipolar velocity phase-encoding gradient.
distensibility) In the phase image, the velocity of blood flow can be deter-
CHRONIC AORTIC DISEASE (ANEURYSM, POSTOPERATIVE) mined for any site of the vascular system. Flow velocity is
Spin-echo MRI Sagittal (thoracic; slice thickness calculated using a formula in which velocity is proportional
3–7 mm) to change in phase angle of protons in motion. MR maps of
— Axial (region of interest, slice flow velocity are obtained two dimensionally which is par-
thickness 5–7 mm, black blood ticularly important in profiles of nonuniform flow, such as
high-resolution parameters) that in the great vessels. On phase images, the gray value
MRA (thoracic and — of a pixel depends on velocity and direction with respect to
abdominal) the imaging plane. Signals below a defined range are con-
ACUTE AORTIC DISEASE — sidered noise and are eliminated by a subtraction process.
Trauma spin-echo MRI Sagittal (slice thickness 3–5 mm) Quantitative data on flow velocity and flow volume are
MR angiography (thoracic) — obtained from the velocity maps through a region of inter-
Dissection spin-echo MRI Axial (slice thickness 7–10 mm) est (ROI). The mean blood flow is estimated by multiplying
— Sagittal (slice thickness 3–5 mm)
the spatial mean velocity and the cross-sectional area of the
GRE/flow mapping Coronal (aortic valve, entry site)
MR angiography (thoracic —
vessel.
and abdominal) MR velocity mapping has been validated both in vivo
and in vitro as an accurate technique for detection of flow
pattern. The accurate analysis of aortic elastic properties
so that the images are reconstructed in the different phases provides specific information on alterations in aortic wall
of the cardiac cycle. Gradient-echo images are acquired with structure. Vector mapping has been used to describe flow
a high degree of temporal resolution throughout the cardiac patterns in physiologic conditions (e.g., patterns in the nor-
cycle (8 to 16 frames per second) and can be displayed in cine mal aorta as a function of age), and aortic diseases (e.g.,
format. Flow-related enhancement is produced by inflow of Marfan syndrome, coarctation, hypertension, aneurysms,
unsaturated blood exposed to only one RF pulse. As a result, dissection) (18–20).
the laminar-moving blood displays bright signal in contrast
to stationary tissues. Signal can be reduced if the flow is low, MRA
as in aortic aneurysms. Mural thrombi can be identified by
persistently low signal intensity in different phases of the The “gold standard” for many manifestations of vascular
cardiac cycle. Turbulent flow produces rapid spin dephas- disease, especially arterial occlusive disease, has long been
ing and results in a signal void. This phenomenon makes considered catheter angiography, which is invasive, costly,
it possible to detect anomalous turbulence, such as aortic and potentially hazardous procedure. MRA may represent an
or mitral insufficiency or jetlike communication between the alternative, noninvasive approach. A variety of MRA tech-
true and the false lumen in aortic dissection. Despite the use niques including various pulse sequences, methods of data
of flow compensation, turbulent flow can be also observed acquisition, and postprocessing have been developed (21).
in the normal aorta especially on the inner wall of the aor- Bright blood techniques, which constitute the basis of
tic arch. Recently, high performance gradient systems with MRA, can be subcategorized into time-of-flight (TOF) and
even faster acquisition (TE, 2 to 3 milliseconds; TR, 4 to phase-contrast (PC) techniques. All bright blood MRA tech-
8 milliseconds; flip angle, 20 degrees), have provided high- niques rely on the visualization of flowing blood for lumi-
quality images of the entire aorta at 15 to 25 different levels nal imaging of the vessel. The essence of MRA is the abil-
in less than 10 minutes. Gradient-echo images can provide ity to portray blood vessels in a projective format similar
additional information in many pathologic conditions such to conventional angiography, with the use of postprocessing
as coarctation, aortic valve insufficiency, and aortic aneu- methods, such as the maximum intensity projection (MIP)
rysm and dissection (14). Particularly, in aortic dissection algorithm. In a MIP algorithm, the brightest pixels along
the detection of entry and re-entry sites is a special capabil- a user-defined direction are extracted to create a projec-
ity of functional MRI that can be helpful in planning both tion image. In the TOF method, flow signal is enhanced by
surgical and endovascular therapy. inflow effects, whereas the background (stationary tissue)
LWBK1209-ch32_p503-544.indd 507 16/05/13 9:21 PM

is saturated by the rapid repeated application of RF pulses. MRA (22–28) in which bright blood images are not strictly
To provide a flow signal before phase dispersion, the TE is dependent on blood flow. Gadolinium diethylenetriamine
kept very short (<10 milliseconds) and the 180-degree re- pentaacetic acid (Gd-DTPA) is infused during data ac-
focusing pulse is eliminated. Fresh blood flowing into the quisition. The technique relies on the contrast-induced
plane of imaging produces a bright signal in a dark back- T1-shortening effects of the contrast medium, and satura-
ground. Venous signal, which can confuse the arterial image, tion problems with slow flow or turbulence-induced signal
can be eliminated with presaturation pulses. The use of a voids are avoided. During the short intravascular phase, the
segmented gradient-echo sequence with cardiac triggering paramagnetic contrast agent provides signal in the arterial
is helpful to eliminate arterial pulsation artifacts. TOF im- or venous system, enhancing the vessel-to-background
ages can be acquired in two-dimensional (2D) or 3D fash- contrast-to-noise ratio irrespective of flow patterns and veloc-
ion. Two-dimensional TOF imaging is very fast but poor in ity. Flow-induced artifacts are mostly eliminated. Pulsatility
resolution, whereas high-resolution 3D images require 10 to artifacts are minimized, even in ascending aorta and ECG
20 minutes of acquisition time. The advantages of both 2D gating is not required. Contrast-enhanced MR aortography
and 3D techniques are gained with a series of thin-slab 3D is performed following acquisition of axial localizer images
acquisitions. The sequential 3D (multiple overlapping thin- to determine the maximum anterior and posterior extension
slab acquisition [MOTSA]) technique provides better flow of the thoracic and abdominal aortae as well as the left and
enhancement than single-slab 3D techniques and causes less right border. If the main area of interest is limited to the tho-
dephasing than 2D techniques. However, with sequential 2D racic aorta, a sagittal acquisition is advisable. The paramag-
or 3D acquisitions, even slight movement by the patient can netic contrast agent (0.2 mmol/kg of bodyweight) is gener-
generate discontinuities in the vessel contour. ally administered via an intravenous catheter placed into the
The basis for PC-MRA is that the flow of blood along a antecubital vein. It is necessary to time the gadolinium bolus
magnetic field gradient causes a shift in the phase of the MR so that the peak enhancement occurs during the middle of
signal. In this technique, information about flow direction MR acquisition. The flow rate should be adjusted to ensure
and velocity are encoded in the image data set. With PC, that the contrast volume is injected in a period not exceed-
pairs of images are acquired that have different sensitivities ing the acquisition time. Imperfectly timed acquisition can
to flow. These are then subtracted to cancel background sig- result in variable degrees of venous enhancement in the MR
nal, so that only signal from flowing blood is left. PC-MRA aortogram. Improved gradient systems allow a considerable
requires a determination of the range of velocities in the ves- reduction of the minimum TRs and TEs and the acquisi-
sel of interest and this information strongly affects the image tion of complex 3D data sets within a breath-hold interval
quality. In patients with aneurysm or dissection with non- under 30 seconds. With the support of MIP images and of
laminar turbulent flow, this technique is often not effective. the 3D multiplanar reformation (MPR), this technique de-
The recent implementation of faster gradient systems has lineates all the morphologic details of the aorta and its side
made possible the development of 3D contrast-enhanced branches in any plane in a 3D format (Figs. 32.2 and 32.3).
Figure 32.2. Gadolinium-enhanced MRA in a patient with type A

dissection (maximum intensity projection algorithm). The intimal flap Figure 32.3. Gadolinium-enhanced 3D MRA of the thoracic aorta
is visible in both, the ascending and descending aorta (asterisk). (surface-shaded display algorithm).
LWBK1209-ch32_p503-544.indd 508 16/05/13 9:21 PM

Table 32.2 Example of SDCT and MDCT Scan Protocols in Aortic Evaluation
Features SDCT 4-DCT 8-DCT 8-DCT 16-DCT 16-DCT 64-DCT
Number of detectors 1 4 8 8 16 16 64
Collimation 1 × 10 mm 4 × 2.5 mm 8 × 1.25 mm 8 × 2.5 mm 16 × 0.75 mm 16 × 1.5 mm 64 × 0.6 mm
Gantry rotation 1,000 ms 500 ms 500 ms 500 ms 375 ms 375 ms 330 ms
Tube voltage 120 kV 100–120 kV 100–120 kV 100–120 kV 120 kV 120 kV 120 KV
Table feed 10 mm/s 20 mm/s 20 mm/s 40 mm/s 32 mm/s 64 mm/s 116 mm/s
Slice thickness 10 mm 3 mm 1.25 mm 2.5 mm 1 mm 2 mm 0.75 mm
Increment 5 mm 1.5 mm 0.6 mm 1.25 mm 0.5 mm 1 mm 0.4 mm
Acquisition time 30–70 s 15–35 s 15–35 s 8–18 s 10–22 s 5–11 s 2.5–6 s
Calculated for a scan range of 300 to 700 mm, with pitch factor 1.
SDCT, single-detector CT; MDCT, multidetector CT (number of detector CT).
Ultrafast 3D acquisition in conjunction with fast table feeds acquisition of multiple parallel slices with each gantry rota-
also makes it possible to chase the contrast bolus through tion, instead of a single slice per rotation as generated by
several vascular districts. Since aortic aneurysm and dissec- SDCT (32,36,37). The increased number of slices per rota-
tion are mainly related to atherosclerotic vascular disease, a tion, combined with a faster gantry rotation time allows
noninvasive concomitant assessment of the entire vascular shorter scan time, wider volume coverage, and thinner col-
system is advisable before surgical planning. limation. Compared with 1-second rotation SDCT scanner,
MDCT technique allows for up to 40 times faster data acqui-
sition. Today, a very high-quality submillimeter scanning of
Noncontrast MRA
the whole thoracic–abdominal vasculature has become pos-
Noncontrast MRA techniques are increasingly applied, sible within comfortable breath-hold duration. An example
stimulated by concerns over the safety of gadolinium-based of the differences between SDCT and MDCT scan protocols
contrast in patients at risk for renal insufficiency or nephro- is showed in Table 32.2.
genic systemic fibrosis. One of the most employed is the bal- Another very important feature of MDCT is the capabil-
anced steady-state free precession (SSFP) MRA, also known ity to have an ECG gating of the scan sequence (35,38,39).
as FIESTA (fast imaging employing steady-state acquisition), This mark permits to synchronize cardiac and thoracic vas-
balanced FFE (fast field-echo), TrueFISP (fast imaging with cular data acquisitions to the phase of least cardiac and
steady-state precession), and TrueSSFP, which is a gradient-echo– arterial systolic motion achieving a reasonable temporal reso-
based sequence that maintains steady-state longitudinal and lution which minimizes any motion artifact, especially in the
transverse magnetization by applying a series of equidistant ascending aorta.
RF pulses (29,30). With this technique, image contrast is Unlike MRI techniques, CTA scan sequences are mostly just
determined by T2/T1 ratios, which leads to inherently bright two: Precontrast sequence and contrast-enhanced sequence;
blood images with little dependence upon blood inflow sometimes in CTA it is mandatory to perform a delayed acqui-
(29–31). Three-dimensional acquisition is used to produce sition after the arterial phase, useful for slow blood flow like in
angiographic images with a high signal-to-noise ratio. dissection, inflammation, or after surgical–endovascular aortic
The major advantages of SSFP angiography are the short repair (leakage).
acquisition times and the high signal-to-noise ratio (30). However, constant high image quality in CTA requires an
In addition, SSFP angiography with 3D acquisition is rela- adaptation of scan parameters as well as the contrast injec-
tively independent of flow, including flow direction, because tion regimen, to simultaneously ensure high spatial resolu-
it is fully flow compensated in all three directions (29–31). tion and high signal within the vessel of interest.
However, background signals, which interfere with exclusive Radiation dose can be adjusted to the body contour and
detection of vessels, are abundant, and arteries and veins give to the phase of the cardiac cycle and is up to four times lower
high signal intensity (31). Therefore, additional preparatory than conventional digital subtraction angiography (DSA) (40).
pulses, including T2 preparation pulses to suppress myocar-
dial and venous signals or arterial spin labeling (31), are
needed to selectively enhance arteries. Otherwise, postpro- Contrast Injection
cessing is needed to remove additional background signals.
While in nonvascular MDCT imaging, contrast application
has not changed dramatically, in MDCTA with faster data
CT Techniques acquisition, contrast agent delivery becomes more critical.
The principal factors influencing contrast bolus geometry
Technique
(Table 32.3) are contrast bolus timing, iodine concentra-
Unlike its single-detector predecessor, MDCT permits increased tion of contrast agent (mgI/mL), injection rate (mL/s), and
longitudinal volume coverage and spatial resolution (5,32–35). saline chaser (41). The timing of the contrast delivery is
The basic advantage of MDCT is the simultaneous the major issue in CTA. Although diagnostically adequate
LWBK1209-ch32_p503-544.indd 509 16/05/13 9:21 PM

Table 32.3 SDCT and MDCT Contrast Injection Parameters in Study of the Aorta
Features SDCT 4-DCT 8-DCT 16-DCT 64-DCT
Collimation 1 × 10 mm 4 × 2.5 mm 8 × 1.25 mm 16 × 0.75 mm 64 × 0.6 mm

Table feed 10 mm/s 20 mm/s 20 mm/s 32 mm/s 116 mm/s
Contrast volume 120–180 mL 100–140 mL 100–140 mL 80–100 mL 60–80 mL
Injection rate 2 mL/s 3 mL/s 3 mL/s 4 mL/s 5 mL/s
Injection time 60–90 s 30–45 s 30–45 s 20–25 s 12–16 s
Iodine concentration 200–250 mgI/mL 250–300 mgI/mL 250–300 mgI/mL 300–350 mgI/mL 350–400 mgI/mL
Bolus chaser — 40 mL; 3 mL/s 40 mL; 3 mL/s 50 mL; 4 mL/s 50 mL; 5 mL/s
Total injection time 60–90 s 45–60 s 45–60 s 32–38 s 22–26 s
Acquisition time 30–70 s 15–35 s 15–35 s 10–22 s 2.5–6 s
Acquisition delay 30 s 25–30 s 25–30 s Bolus triggered Bolus triggered
Calculated for a scan range of 300 to 700 mm, with pitch factor 1.
enhancement has been reported using a standard (empirical) contour, thus is possible to have the entire vessel midline
start delay (42), individual tailoring of the bolus injection is on a single 2D image. MIP is a technique closer to DSA; it
required in order to maximize arterial enhancement in dif- shows aortic opacification and caliber, but may provide only
ferences of circulation time (43,44). This can be performed endoluminal information. Like a threshold technique, den-
either by test bolus acquisitions or by bolus tracking algo- sity values below that of contrast such as plaques or throm-
rithm sequences. The first technique requires an additional bus and above that of contrast such calcifications are dif-
injection of a small contrast volume (10 to 20 mL) before ficult to discriminate; MIP images will not allow depth
the acquisition of CTA data, followed by the use of the perception or understanding of interstructural relationships.
same injection parameters. Bolus tracking modality permits SSD provides a good overall view of aortic anatomy but
the initial injection of the whole contrast agent volume, while does not provide internal detail, can overestimate vessel ste-
the start of data acquisition is triggered by an automated noses and is also a threshold technique. VR is the latest and
detection of the contrast bolus arrival, using preconfigured most powerful method for 3D reconstruction. With high
trigger thresholds. An ROI is placed into the target vessel fidelity to the original dataset all tissues are represented
and the attenuation level (expressed in Hounsfield unit) is based on their Hounsfield values so that, unlike a thresh-
monitored within the ROI in real time during a single-level old technique, simultaneous depiction of different kind of
dynamic scan. tissues is possible in 3D projection. With this feature, com-
Because of the fast scan time of MDCT, the delay between bined with adequate filters, metallic stents or surgical clips
injection and scan start generally needs to be extended, so do not present a problem and both calcifications and throm-
that the scan does not outrun the contrast material bolus bus can be discriminated from the vessel lumen.
(45). Shorter acquisition time also allows a reduction of the
contrast agent volume and higher flow rates, even if a more
concentrated solution often must be used, together with a Acquired Diseases of the
saline chaser to compact the bolus and to push it toward Thoracic Aorta
the right atrium by the use of dual syringe CT injectors
(6,40,46,47). Aortic Dissection
Aortic dissection is characterized by a laceration of the aor-
Postprocessing
tic intima and inner layer of the aortic media that allows
Acquisitions are exclusively in axial plane; other planes blood to course through a false lumen in the outer third of
are reconstructed from the scanned volume by multiplanar the media. Dissection can occur throughout the length of the
reconstruction (MPR) and with MDCT are as clear as a aorta and the two most common classifications are based
multiplanar acquisition due to the isotropic 3D data scan. on the anatomic location and extension of intimal flap.
In the evaluation of aortic diseases other frequently used According to the De Bakey classification, in type I dissec-
postprocessing techniques are curved MPR (CMPR), MIP, tion the intimal tear originates in the ascending aorta and
shaded surface display (SSD), and volume rendering (VR) the intimal flap extends below the origin of the left subcla-
(7,48,49). MPR re-format image datasets to alternate 2D vian artery; type II dissection is confined to the ascending
planes, in general, sagittal, coronal, or oblique. However, aorta; and in type III dissection the entry tear develops after
with thinner slice the aorta is not completely visualized in a the origin of the subclavian artery and extends distally. The
single image, due to the aortic curvature that cannot be re- Stanford classification simply classifies an aortic dissection
duced to any single plane. Moreover, with modern software irrespective of the site of the entry tear as type A, if the
it is possible to thicken any selected plane (MPR thick) and ascending aorta is involved, and as type B if the ascending
look at the entire vessel. Curved MPR is an extension of aorta is spared. The Stanford classification is fundamen-
this feature that displays a curved plane traced along aortic tally based on prognostic factors: Type A dissection requires
LWBK1209-ch32_p503-544.indd 510 16/05/13 9:21 PM

Figure 32.4. A: Spin-echo sagit-

tal image of type B dissection: The
double aortic lumen is 4 cm below the
left subclavian artery. The false lumen
(high signal intensity) is severely
dilated. B: MRA of the same patient
after stent graft treatment (arrow):
Aortic remodeling with complete
thrombosis of the false lumen. The
metallic structure of the stent graft
produces minimal artifacts in the
A B
upper portion of the descending aorta.
MRI Technique and Findings

urgent surgical repair whereas most of type B dissections can
be successfully managed with medical therapy. In a suspected case of aortic dissection, the standard exami-
Acute aortic dissection is a life-threatening condition re- nation should begin with spin-echo sequences acquired with
quiring prompt diagnosis and treatment (50). The 14-day high-resolution parameters and preparatory pulses to nullify
period after onset has been designated as an acute phase the blood signal and obtain a better definition of the aortic
because the rates of morbidity and mortality are highest dur- wall structures (Table 32.4). In the axial plane the intimal flap
ing this period. The estimated mortality rate of untreated aor- is detected as a straight linear image inside the aortic lumen.
tic dissection is 1% to 2% per hour in the first 24 hours after The true lumen can be differentiated from the false by the
onset and 80% within 2 weeks. Early and accurate detection anatomic features and flow pattern. The true lumen shows a
of the dissection and a delineation of its anatomic details are signal void, whereas the false lumen has a higher signal inten-
critical for successful management. However, because physi- sity. In addition, the visualization of remnants of the dissected
cal findings may be absent or misleading and symptoms may media as cobwebs adjacent to the outer wall of the lumen
mimic those of other disorders, such as myocardial ischemia may help to identify the false lumen. The leakage of blood
and stroke, the diagnosis of aortic dissection is often missed at from the descending aorta into the periaortic space, which
initial evaluation (51,52). The anatomic characteristics of the can appear with high signal intensity and result in a left-sided
dissection indicate the type of surgical technique, and affect pleural effusion, is usually better visualized on axial images.
both the surgical success rate and long-term results. Thus, in A high signal intensity of a pericardial effusion indicates a
dissection, the diagnostic goal, regardless the imaging modal- bloody component and is considered a sign of impending
ity, used a clear delineation not only of the intimal flap and rupture of the ascending aorta into the pericardial space. A
its extension but also detection of the entry and re-entry sites, detailed anatomic map of aortic dissection must indicate the
presence and degree of aortic insufficiency, and flow in the aor- type and extension of dissection but also distinguish the ori-
tic branches (53). Transcatheter endovascular reconstruction gin and perfusion of branch vessels (arch branches, celiac,
of type B aortic dissection is a new option for the treatment superior mesenteric, renal arteries and coronary arteries) from
of both acute and chronic dissections (54,55). In endovascular the true or false channels. Therefore, a further spin-echo
techniques, the success of the procedure is strictly related to sequence on the sagittal plane should be performed to define
a detailed anatomic definition of the features of the dissected the extension of the dissection in the thoracic and abdominal
aorta. The identification of the entry and re-entry sites, the aortae and in the aortic arch branches (Fig. 32.5).
relationship between true and false lumina and the visceral Adjunctive gradient-echo sequences or phase-contrast im-
vessels, and any involvement of the iliac arteries are crucial in ages can be instrumental in identifying aortic insufficiency and
patient selection and stent-graft design (Fig. 32.4). entry or re-entry sites (Fig. 32.6), as well as in differentiating
LWBK1209-ch32_p503-544.indd 511 16/05/13 9:21 PM

Table 32.4 MR Study Modality in Acute Aortic Syndromes
Sequence/Plane Diagnostic Findings Anatomic Details
AORTIC DISSECTION
Spin-echo axial/sagittal Intimal flap/true–false lumen Periaortic hematoma
Pericardial effusion
Gradient-echo axial/sagittal Intimal flap/true–false lumen Thrombosis false lumen/entry and re-entry sites
Aortic insufficiency
MR angiography Intimal flap/true–false lumen Origin/perfusion of supra-aortic, coronary,
abdominal vessels
INTRAMURAL HEMATOMA
Spin-echo axial/sagittal, T1-weighted Abnormal wall thickness, crescent shape. Periaortic hematoma
High signal intensity (T1) Pericardial effusion
Spin-echo axial, T2-weighted High signal intensity (recent) Pericardial, pleural, mediastinal effusion:
Low signal intensity (old) Increased signal intensity
MR angiography No utility —
PENETRATING AORTIC ULCER
Spin-echo axial/sagittal Craterlike outpouching/circumscribed Periaortic hematoma/pleural effusion
dissection/intramural hemorrhage Diffuse aortic wall atherosclerosis
Gradient-echo sagittal No utility —
MR angiography Craterlike outpouching/saccular Relationship with aortic arch or abdominal
pseudoaneurysm vessels
Figure 32.5. Spin-echo sagittal image of type A dissection: The Figure 32.6. Gradient-echo sagittal image of type B dissection:
intimal flap is visible as subtle linear image in the ascending and Flow turbulence (signal void, arrow) in the descending aorta indicates
descending aorta. the entry site.
LWBK1209-ch32_p503-544.indd 512 16/05/13 9:21 PM

Figure 32.7. MRA 3D (A) and MIP

(B) images of aortic dissection (abdomi-
nal aorta): Celiac, mesenteric, and left
renal arteries originate from the ante-
A B
rior true lumen (arrows).
slow flow from thrombus in the false lumen (56,57). diagnosis and anatomical definition (58). However, two
However, because the diagnosis of aortic dissection is not cases of intramural hematoma (IMH) missed by MRA (28)
dependent on functional gradient-echo images, these se- raise concern about using MRA as the sole modality for sus-
quences should be reserved to clinically stable patients. pected aortic dissection.
The third step in the diagnosis of aortic dissection and the At present, MRI is one of the most accurate tools in
definition of its anatomic detail relies on the use of gadolinium- the detection of aortic dissection. A high degree of spatial
enhanced 3D MRA. Since 3D MRA is rapidly acquired resolution and contrast and the capability for multiplanar
without any need of ECG triggering, this technique may be acquisition provide excellent sensitivity and specificity that
used for even severely ill patients. Since it is not nephrotoxic approximate 100% in the published series (58–61). With
and causes no other adverse effects, gadolinium can be used modern scanner, a comprehensive study of the entire aorta is
in patients with renal failure or low cardiac output. With completed in less than 10 minutes, and patient’s ECG, blood
spin-echo sequences, artifacts caused by imperfect ECG gat- pressure, and oxygen saturation can be monitored, even dur-
ing, respiratory motion, or slow blood pool can result in ing assisted ventilation. The implementation of open systems
intraluminal signal, simulating, or obscuring an intimal flap. may soon allow a wider use of MRI even in acute pathology.
In gadolinium-enhanced 3D MRA, the intimal flap is eas- Aortography has long been considered the method of
ily detected and the relationship with aortic vessels clearly choice in suspected aortic dissection, despite the risk of
depicted (Figs. 32.7 and 32.8). Entry and re-entry sites catheter manipulation and injection of high flow contrast in
appear as a segmental interruption of the linear intimal a dissected aorta. With the advent of noninvasive imaging
flap on axial or sagittal images (Fig. 32.9). The analysis modalities, its low accuracy has been demonstrated; the re-
of MRA images should not be limited to viewing MIP imported sensitivity is 77% to 90% and the specificity is 90%
ages or SSD; it should also include a complete evaluation of to 100%. The superiority of transesophageal echocardio
reformatted images in all three planes, to confirm or improve graphy (TEE), CT, and MRI in comparison to angiography
spin-echo information and exclude artifacts. In MRA post- has been widely reported in the literature (51–53,59).
processing displays, the appearance of the dissected aorta is In general, TEE is a reliable method with excellent sensi-
similar to that on conventional catheter angiograms, but di- tivity, and a great advantage is that it can be performed at the
agnostic information such as the intimal flap can be masked. bedside in patients too unstable for transportation. However,
Combining the spin-echo with MRA images completes the artifacts and “blind areas,” such as the distal portion of the
LWBK1209-ch32_p503-544.indd 513 16/05/13 9:21 PM

A B C D
Figure 32.8. MRA 3D (A,C) and MIP (B,D) images of aortic dissections; the false lumen is patent in (A,B)
and partially thrombosed in (C,D) (asterisk).
ascending aorta, can influence specificity in an operator-

dependent manner. Since TEE information is limited to the
thoracic aorta, sometimes with suboptimal display also of
the aortic arch, a second imaging modality encompassing
the entire aorta is advisable in stable patients.
CT Technique and Findings

CT has a difficult and crucial role in the diagnostic workup for
aortic dissection. According to the results of the International
Registry of Acute Aortic Dissection (62), CT resulted to be
the initial modality to confirm the clinical suspicion of aortic
dissection in the majority of cases (63%), followed by TEE
(32%), aortography (4%), and MRI (1%). Sensitivity rates
of these four imaging modalities, in the set of suspected acute
aortic dissection, are 100% for MRI, 93% for CT, 88% for
angiography, and 87% for TEE.
The critical clinical issue required of any imaging test ap-
plied to a patient suspected of having an aortic dissection is
the identification of an intimal flap and its localization to
the ascending (type A) or descending (type B) aorta. This
fundamental diagnostic feature that determines the need for
emergent repair can be addressed by at least four imaging
modalities: Angiography, CT, MRI, and TEE. The relative
accuracy of these modalities has been debated in the medical
literature and is confounded by the fact that technical im-
Figure 32.9. MRA of aortic dissection (thoracic aorta): The entry provements in CT, MRI, and TEE have outpaced our ability
site is visible as segmental interruption of the linear intimal flap. to compare them appropriately. Recent opinion has shifted
LWBK1209-ch32_p503-544.indd 514 16/05/13 9:21 PM

omparison between CT—Standard (S) or Helical (H)—

C
Table 32.5
and TEE, MR, and Angiography in Aortic Dissection
Author Year Technique Sensitivity (%) Specificity (%) PPV (%)
Erbel 1989 S CT 77 — —
TEE 98 — —
Aortography 89 — —
Barbant 1992 S CT — — 40–95
TEE — — 40–95
Aortography — — 40–65
MR — — 90–00
Nienaber 1993 H CT 94 87 —
TEE 98 — —
MR 98 98 —
Sommer 1996 H CT 100 94 —
TEE 100 94 —
MR 100 100 —
toward MRI or TEE as the most sensitive tests for aortic dis- dissections are diagnosed based on identifiable flow in the
section. Unfortunately, much of this opinion is based upon false lumen. When slower flow is present, false lumen filling
comparative studies where MRI or TEE is compared with defects that represent strand of thrombus are observed, and
relatively obsolete conventional CT or SDCT techniques complete thrombosis can be reliably diagnosed on follow-up
(53,59,60,63–65) (Table 32.5). To date there have been no examinations (Fig. 32.12) (71,72). One unusual type of aor-
comparisons of multidetector CT to either MRI or TEE and tic dissection is the intimo–intimal intussusception produced
is useless to consider these prior studies while CT, more than by circumferential dissection of the intimal layer, which sub-
MRI or TEE, got substantial advances and features. sequently invaginates like a wind sock; CT scan may show
Helical CT allows diagnosis of acute aortic dissection one lumen wrapped around the other lumen in aortic arch,
with sensitivity and specificity rates respectively of 83% with the inner lumen invariably being the true one (73,74).
to 94% and 87% to 100% (66–68). Imaging of dissection Sometimes an aortic aneurysm with intraluminal throm-
requires a volume of coverage from supra-aortic branches bus may be difficult to distinguish from a dissection with
superiorly to the femoral arteries inferiorly, and MDCT in a thrombosed false lumen; may help us the fact that dissec-
few seconds can acquire this extent. Imaging sensitivity is tion generally has a spiroidal shape, whereas a thrombus
enhanced by greater temporal resolution and ECG-gating se- tends to maintain a constant circumferential relationship
quences which minimize pulsation artifacts at the aortic root with the aortic wall and, furthermore, a mural thrombus
(Fig. 32.10). On unenhanced CT, it is possible to see internal usually has a smooth internal border. Calcification in aneu-
displacement of intimal calcifications and this finding could rysm is typically located at the periphery of the aorta (75).
be confused with an aneurysm with calcified mural thrombus. Visceral and supra-aortic vessel’s involvement can account
High attenuation of the false lumen at unenhanced CT may for high mortality and MDCT has the spatial and contrast
help differentiate between the two conditions (69). The main resolution to reliably diagnose branch vessel involvement
and characteristic finding of aortic dissection on contrast- and document true or false lumen supply (Figs. 32.11
enhanced CT scan is an intimal flap that separates true from and 32.12). General strategy of CT study of the thoracic–
the false lumen. This usually appears as a thin linear luminal abdominal aortae and main CT findings of aortic diseases
filling defect and its appearance is determined by the circum- are showed in Table 32.6.
ference and length of dissection, the relative lumen flow, and Apart from axial images and MPR, which provides an
aortic pulsation. Accurate differentiation between true and overall view of the aortic dissection and demonstrate the
false lumen became fundamental with the advent of endo- anatomic relationships between the flap and adjacent great
vascular procedure for treatment planning (70). The slender vessels, VR is preferred to MIP and SSD for dissection 3D
linear areas of low attenuation that occasionally appear in postprocessing as it preserve the variable enhancement pat-
the false lumen on CT images, known as the cobweb sign, terns of the lumina and is more sensitive for visualization
are specific to the false lumen and may aid in its recogni- of the flap. The accurate localization of entry and re-entry
tion. These findings correspond to residual ribbons of the sites remains a difficulty for all imaging techniques, but
media, incompletely sheared away during the dissection pro- the high resolution of submillimeter MDCT acquisitions
cess (71). Two other useful indicators of the false lumen are and cardiac gating may be enough reliable to depict them
a larger cross-sectional area and the beak sign. This one is (Figs. 32.11 and 32.12). The last generation of scanners
the manifestation of the wedge of hematoma that cleaves a may also depict coronary artery involvement but, to date,
space for the propagation of the false lumen (72). However, determinate or quantify aortic insufficiency with CT is still
on most contrast-enhanced CT scans the true lumen may be impossible.
identified by its continuity with an undissected portion of the In CT study for suspected aortic dissection, especially with
aorta (Fig. 32.11). Communicating and noncommunicating SDCT scanners, there are a variety of pitfalls mimicking aortic
LWBK1209-ch32_p503-544.indd 515 16/05/13 9:21 PM

A B
C D
Figure 32.10. ECG-gated MDCT 3D (A), axial (B), and MIP (C,D) images of a type A aortic dissection
showing the entry tear (asterisk) and the intimal flap (arrows).
dissection. These pitfalls are attributable to technical fac-

Intramural Hematoma
tors (improper timing or rate of contrast administration),
streak artifacts (high-attenuation materials, high-contrast IMH was first described in 1920 as “dissection without inti-
interfaces, and cardiac motion), periaortic structures (aor- mal tear” (77), but it was rarely recognized in the clinical set-
tic arch branches, mediastinal veins, pericardial recess, ting before the advent of high-resolution imaging modalities.
thymus, atelectasis, and pleural thickening or effusion), Spontaneous rupture of aortic vasa vasorum of the media
aortic wall motion, aortic anomalies, and the aforemen- layer is considered the initiating process, which is confined
tioned aortic aneurysm with thrombus (68,76). At present in the aortic wall without intimal tear. This results in a
with an optimal scan protocol which requires a precontrast circumferentially oriented blood containing space seen on
study, an ECG-gated arterial phase sequence and a delayed tomographic imaging studies. IMH may occur spontane-
scan for slower blood flows, most of these artifacts/pitfalls ously or as a consequence of penetrating aortic ulcer (PAU)
are minimized. in intrinsically diseased media; it has also been described
LWBK1209-ch32_p503-544.indd 516 16/05/13 9:21 PM

Figure 32.11. MDCT 3D

(A,B) and MIP (C) images of a
type B aortic dissection: True
lumen is clearly visible (asterisk)
and recognizable by its continu-
ity with undissected aorta at the
level of the aortic arch and the A B C
iliac bifurcation.
A B
Figure 32.12. MPR thick (A) and VR (B–D)

MDCT images of type B aortic dissection showing a
partially thrombosed false lumen in descending thoracic
aorta (arrow). Entry and re-entry tears are also illus-
trated (arrowhead). C D
LWBK1209-ch32_p503-544.indd 517 16/05/13 9:21 PM

eneral Strategy and CT Study Findings of the

G
Table 32.6
Thoracic–abdominal Aorta Diseases
Disease Type of Scan Findings
Aortic dissection Precontrast Intimal flap

Contrast-enhanced True–false lumen flow
Delayed False lumen filling defects (thromboses)
Size of true and false lumen
Visceral–epiaortic vessel’s involvement
Entry and re-entry sites
Intramural Precontrast Wall thickening (crescent/circumferential)
hematoma Contrast-enhanced Displaced calcified intima (subintimal location)
(IMH) Delayed Concentric area of high density around lumen
+/− Aortic dilatation
Aortic ulcer Precontrast “Collar button” contrast-filled ulceration
Contrast-enhanced Dislodgment of the intimal calcifications
+/− Wall enhancement, pseudoaneurysms, dissection, rupture
Aortic aneurysms Precontrast Size (caliber, length, tortuosity)
Contrast-enhanced Extension
Neck
Mural thromboses
Calcifications
Aortic trauma Precontrast Mediastinal hemorrhage
Contrast-enhanced Extravasation
Aortic contour abnormality (intima–media disruption)
Focal caliber change or dissection flap
Aortitis Precontrast Irregular calcified aortic wall
Contrast-enhanced Periaortic density (inflammation) and adjacent gas
collection
Dilated aortic lumen
following blunt chest trauma (78). As in aortic dissection differentiation of IMH from aortic atherosclerotic plaque
arterial hypertension is the most frequent predisposing factor. may be difficult and false positive and false negative results
Clinical signs and symptoms and the prognosis do not differ have been reported (82,83). The displacement of intimal
to classic aortic dissection, and IMH should be regarded as a calcification represents a useful marker in distinguishing
variant of dissection with similar or more serious prognostic IMH from mural thrombus or plaque with both TEE and
and therapeutic implications (79). Typical complications of CT. In a comparison of different imaging modalities, MRI
dissection such as fluid extravasation with pericardial, pleu- demonstrated the best sensitivity in the detection of IMH
ral, and periaortic hematoma may occur in IMH as well. In a (83). T1-weighted images reveal a crescent-shaped area
retrospective analysis of the Yale experience on 214 patients of abnormal signal intensity within the aortic wall (Fig.
with acute aortic syndromes, Coady et al. (80) found 47.1% 32.13). Moreover, MRI is the only imaging method which
rate of aortic rupture for IMH, higher than that for classic may allow the assessment of the age of hematoma on the
type A (7.5%) or type B dissection (4.1%). However, after basis of the different degradation products of hemoglobin.
the acute phase, the evolution of IMH may be also favorable; In the acute phase (0 to 7 days after the onset of symp-
in the series described by Yamada et al. (81), nine of ten sur- toms) on T1-weighted spin-echo images, oxyhemoglobin
vivors to initial presentation showed complete resolution of shows intermediate signal intensity whereas in the subacute
the aortic hematoma within 1 year. phase (>8 days), methemoglobin shows high signal inten-
sity. However, when the signal intensity is medium to low,
it can be difficult to distinguish IMH from mural thrombus.
MRI Findings
T2-weighted spin-echo sequences may help in differentiat-
The diagnosis of IMH relies on the visualization of intra- ing the two entities: Signal intensity is high in recent hem-
mural blood manifested as localized thickened aortic wall. orrhage but low in chronic thrombosis. In a retrospective
The abnormal wall thickness, symmetric or asymmetric, study of 22 cases, Murray et al. (84) described three cases
can vary in dimension from 3 mm to more than 1 cm. with recurrence of symptoms and unfavorable evolution in
The mural involvement can encompass the entire aortic which MR signal intensity changes consistently with recur-
circumference. However, with all the imaging modalities rent bleeding. The progression of IMH to overt dissection
IMH can be confused with clot or plaque especially if local- and rupture has been reported in 32% of cases, in particu-
ized in the descending aorta. In particular with TEE, the lar, with the involvement of ascending aorta (Fig. 32.14).
LWBK1209-ch32_p503-544.indd 518 16/05/13 9:22 PM

A B C
Figure 32.13. A,B: T1-weighted spin-echo axial image of intramural hematoma of the ascending and
descending aorta; the abnormal wall thickening (arrows) present intermediate signal intensity in (A) (oxyhe-
moglobin, acute phase) and high signal intensity in (B) (methemoglobin, subacute phase). C: T2-weighted
spin-echo image signal intensity is high in acute phase (recent hemorrhage).
Instability of the hematoma and recurrent bleeding which circumferential relationship with aortic wall. Another finding
can be detected by MRI are important parameters in assess- that can be seen in the setting of IMH is intense enhancement
ing the need for surgical repair. and thickening of the aortic wall external to the hematoma,
which may represent adventitial inflammation.
Several investigators have attempted to assess the useful-
CT Findings
ness of CT findings for predicting the progression of aor-
Unenhanced CT imaging finding of IMH is a crescent-shaped tic IMH to overt dissection, but relationship between them
area of hyperdensity in the aortic wall or circumferential remains unclear. However, findings such as IMH type A,
wall thickening, corresponding to a hematoma in the medial aortic diameter of more than 50 mm, thick hematoma with
layer which extends cephalic and caudal beneath a displaced compression of the aortic lumen, and pericardial or pleural
calcified intima, with a constant relationship to the wall effusion are useful for predicting progression to aortic dis-
(subintimal location) (80,85–88) (Fig. 32.15). The hema- section (87,90–93).
toma may, or may not, compress the aortic lumen (89). It Axial images alone will suffice for detection of IMH or
is important to perform unenhanced CT as the first imag- penetrating ulcers in most cases, but 3D review is useful to
ing evaluation, because contrast agent within the vessel may discriminate both from irregular mural thrombus and to
obscure IMH. Unlike the false lumen in typical aortic dissection, plan endovascular or surgical therapy and to map the full
the crescent-shaped area of IMH remains unenhanced after extent of IMH.
contrast material administration, and no intimal tear is seen
on contrast-enhanced CT scan. Aortic dilatation could be
Aortic Ulcers
present. One useful observation that may help differentiate
IMH from thrombosed false lumen of a classical intimal dis- In 1934, Shennan (94) was the first to describe penetrating
section is that the latter tends to longitudinally spiral around atheromatous ulcers of the thoracic aorta. In elderly, hyper-
the aorta, whereas the former tends to maintain a constant tensive patients with severe atherosclerotic involvement of
Figure 32.14. Spin-echo axial

images of intramural hematoma
evolved in classic dissection in the acute
(A) and chronic (B) phase. A: The high
signal intensity of hematoma is visible
in ascending and descending aorta.
B: Three months later the hematoma
presents low signal intensity and partial
area of reabsorption: An intimal flap A B
is visible.
LWBK1209-ch32_p503-544.indd 519 16/05/13 9:22 PM

A B C
Figure 32.15. IMH. A,B: CT image showing a crescent-shaped area of hyperdensity within the aortic wall with-
out any contrast enhancement (arrows). C: T1-weighted MR spin-echo image presenting a crescent-shaped area with
intermediate signal intensity within the aortic wall (arrow). Both findings are evocative for IMH in acute phase.
the aortic wall, usually in the descending aorta, a plaque IMH. MRA is particularly suitable for depicting aortic
may ulcerate into the media. Aortic ulcer is character- ulcers (Fig. 32.17) along with the irregular aortic wall pro-
ized by rupture of the atheromatous plaque disrupting file seen in diffuse atherosclerotic involvement (96). The aor-
the internal elastic lamina. Extension of the ulcerated ath- tic ulcer is easily recognized as a contrast-filled outpouching
eroma into the media may result in an IMH or localized
intramedial dissection, or the plaque may break through
to the adventitia and form a saccular pseudoaneurysm.
The adventitia may also rupture, in which case only the
surrounding mediastinal tissues contain the hematoma.
Aortic ulcers occur almost only in the descending aorta
but location in the aortic arch or in the ascending aorta
has occasionally been reported. The clinical features of
penetrating atherosclerotic ulcers may be similar to those
of aortic dissection, but the ulcers should be considered a
distinct entity with a different management and prognosis.
Hypertension, advanced age, and systemic atherosclerosis
are common predisposing factors. Persistent pain, hemo-
dynamic instability, and sign of expansion should trigger
surgical treatment whereas asymptomatic patients can be
managed medically and monitored with imaging follow-
up. Movsowitz et al. (95) analyzed 45 cases of aortic ulcers
reported in the medical literature and found an incidence
of transmural rupture of 8%. Different prognostic profiles
were reported by Coady et al. (80): Among 19 patients with
a diagnosis of penetrating ulcer, the ulcers of 8 (42%) rup-
tured prior to surgical treatment. Since penetrating ulcer is
much less common than classic dissection and imaging find-
ings may be subtle, careful awareness of its insidious behav-
ior is particularly important for successful management.
MRI Findings
The MRI diagnosis of aortic ulcers is based on the visu-
alization of a craterlike ulcer located in the aortic wall
(Fig. 32.16). Mural thickening with high or intermediate Figure 32.16. Spin-echo sagittal image of severe aortic athero-
signal intensity on spin-echo sequences may indicate exten- matous plaques and a significant penetrating ulcer of the descending
sion of the ulcer into the media and the formation of an aorta (arrow).
LWBK1209-ch32_p503-544.indd 520 16/05/13 9:22 PM

intimal calcifications are also often seen. On CTA, penetrat-

ing ulcer appears as a discrete “collar button” contrast-
filled ulceration, similar to that of a peptic ulcer (Fig. 32.18)
(80,85–88,99). Lesions can be single or multiple. It seems
more eccentric than irregular mural thrombus and may be
associated with wall thickening and enhancement, pseu-
doaneurysms, dissection, or rupture (86). Atheromatous
ulcers that are confined to the intimal layer sometimes have
a radiologic appearance similar to that of penetrating ulcer.
Therefore, particular attention should be taken in making
diagnosis of PAU if the lesions are discovered incidentally
in an asymptomatic patient and if associated IMH is absent
(96). CT has the advantage, as well as in IMH, to visualize
dislodgment of the intimal calcifications that are very fre-
quently observed in aortic ulcers. CTA, as well as MRI,
has permitted accurate noninvasive follow-up of both PAU
and IMH with reliability for detecting change that allows
conservative management approaches to be explored and
early intervention to be indicated.
MDCT can depict small, penetrating atherosclerotic
ulcers and can demonstrate complex spatial relationships,
mural abnormalities, and extraluminal pathologic condi-
tions, which may offset this weakness (100).
Postoperative Evaluation
of Aortic Dissection
The objective of surgical treatment of aortic dissection is
to prevent aortic rupture in the proximal portion of the
ascending aorta. Early and accurate detection by new imag-
ing modalities and aggressive surgical approach have con-
tributed to a decrease in operative mortality from 40% to
50% in the 1970s to 5% to 7% in recent series (101,102).
Nevertheless, survivors after initial repair still remain at
considerable risk of future complications. A persistent distal
false lumen has been reported in 75% to 100% of cases.
Second entry tear in the descending aorta or in the aortic
arch, which is common, is responsible for patency of the dis-
tal dissection, which is associated with an unfavorable prog-
nosis. It is recognized that dilatation and subsequent rupture
of the distal aorta is the most common cause of death of
patients after surgery for aortic dissection. Prosthetic graft
degeneration or infection (Fig. 32.19), and malfunction of
the prosthetic aortic valves are additional causes of postop-
Figure 32.17. MRA of the same patient: The ulcer appears erative complications. Therefore, every patient after aortic
as contrast-filled outpouching protruding on the vessel profile dissection repair must be carefully monitored with imag-
(arrow). ing according to a strict schedule, and the timing is crucial.
Heinemann et al. (103) suggest a first imaging session at
discharge, so that a relatively normal baseline postoperative
of variable extent with jagged edges, which may result even image of the new anatomic situation is available. Afterward,
in large pseudoaneurysm (97). The disadvantage of MRI the subsequent examinations are scheduled on the basis
with respect to CT is failure to visualize dislodgment of the of absolute diameter (<5 cm once per year; >5 cm every
intimal calcifications, frequently observed in aortic ulcers. 6 months) and the expansion rate on two subsequent fol-
low-ups. Rupture is often preceded by a period of rapid
CT Findings aneurysm expansion and detection of this phenomenon can
identify patients at high risk of rupture, thereby decreas-
In penetrating atherosclerotic ulcers (PAU), the plaque rup- ing the risk-to-benefit ratio of prophylactic surgical repair.
ture and disruption of the internal elastic lamina is mani- The reported expansion rate for dissected segments ranges
fest on unenhanced CT as extensive atherosclerosis and between 1.2 and 4.3 mm/y. However, a yearly expan-
IMH of variable extent. Frequently the IMH is focal, due sion rate of 5.6 mm has been reported by Bonser et al.
to medial fibrosis caused by atherosclerosis (98). Displaced (104) in aneurysm with a diameter of more than 60 mm.
LWBK1209-ch32_p503-544.indd 521 16/05/13 9:22 PM

A B
C D
Figure 32.18. MDCT images illustrate penetrating atherosclerotic ulcer and intramural hematoma of the
aortic arch. A,B: Axial image of a double penetrating ulcer (arrow) with intramural hematoma (arrowhead).
C,D: Volume rendering (C) and virtual endoscopy (D) images showing ulcers (arrows).
The presence of partial thrombosis of the false lumen is an essential component of serial examinations, in which
seems to protect against dilatation; an aortic growth rate minimal change in dimension may represent a prognostic
of 3.4 mm/y in dissection with partial thrombosis of the finding or indicates a need of preventive surgical strategies.
false lumen versus an increase of 5.6 mm/y in dissection Residual dissection is easily detected on spin-echo images,
with no thrombosis of the false lumen has been reported and gradient-echo sequences or phase-display images can be
(105). Therefore, in patients who have undergone surgery used to distinguish thrombosis of the false lumen from slow
for aortic dissection, an appropriate follow-up should take flow.
into account an accurate measurement of the residual aorta Slight thickening around the graft is a common finding
to identify early high-risk patients. caused by perigraft fibrosis (109,110). However, large or
asymmetric thickening around the tube graft may repre-
sent localized hematoma caused by an anastomotic leakage.
MRI
Suture detachment with leakage has been reported in par-
MRI is recognized as one of the imaging modality of choice ticular after composite graft replacement of the ascending
in postoperative evaluation (106–109). MRI measurements aorta. Reoperation for bleeding at the site of repair has been
of parameters are highly reproducible, and reproducibility reported after composite graft operation in 8% of patients
LWBK1209-ch32_p503-544.indd 522 16/05/13 9:22 PM

after 30 days and in 4% of patients after 1 year. The higher

incidence of bleeding has been reported at the site of reim-
planted coronary arteries. Gadolinium-enhanced MRI with
standard spin-echo sequences can provide detailed informa-
tion on suture detachment (111); the site of bleeding appears
as high signal intensity within the hematoma. Moreover,
gadolinium-enhanced MRA is particularly effective in the
depiction of the complex postoperative anatomy and eluci-
dating the prosthetic tube, distal and proximal anastomoses,
residual distal dissection, and eventually dilated segments
(Fig. 32.20). Reimplanted coronary arteries or reimplanted
supra-aortic vessels can be also visualized by MRA, and par-
ticular attention should be paid to evaluating these sites for
possible postoperative weakness. Aneurysm of reimplanted
coronary ostium has been reported after composite graft re-
placement of ascending aorta with Bentall technique. In the
Cabrol technique, intraoperative thrombosis or distortion of
the prosthetic limb connecting the right coronary artery to
the main prosthetic tube may cause myocardial infarction.
Contrast-enhanced MRA can visualize the proximal seg-
ment of reimplanted coronary arteries and detect and moni-
tor proximal coronary aneurysms (Fig. 32.21).
Figure 32.19. Spin-echo axial image of a periprosthetic infective
collection (arrowheads) in a patient who underwent surgery for aortic
dissection. CT
The role of CTA as an appropriate imaging modality in post-
operative evaluation of surgical or endovascular aortic repair
is surely increased by the introduction of MDCT scanners.
Figure 32.20. Three-dimensional

(A) and MIP (B) MRA images of a
patient who underwent surgery for
type A dissection: The prosthetic tube
is visible in the ascending aorta and in
the aortic arch (arrows). Residual dis-
section is seen in the descending aorta,
with partial thrombosis the false lumen A B
(asterisk).
LWBK1209-ch32_p503-544.indd 523 16/05/13 9:22 PM

Because of the increasing age of population and environ-

mental factors, it is expected that aortic aneurysm will occur
more frequently in the next future.
Thoracic Aneurysm
In the period between 1951 and 1980, the incidence of
thoracic aortic aneurysm was estimated to be 3.5 cases per
100,000 persons per year. Recently, a population-based
study of thoracic and thoracoabdominal aortic aneurysms
reported an incidence of 10.9 cases per 100,000 persons per
year, revealing an increased rate of occurrence of the disor-
der (117). This phenomenon is not completely explained:
The complex multifactorial mechanism, which leads to an
alteration of the media and the formation of an aneurysm, is
still under investigation. Most aneurysms are atherosclerotic
in nature, usually fusiform and long. Saccular false aneu-
rysms may also develop in patients with atherosclerosis as
a result of a penetrating ulcer. The natural history is quite
diverse, reflecting a broad spectrum of underlying causes.
Although many studies have identified risk factors related to the
formation and progression of aortic aneurysms, none has fully
explained the etiology of the disorder. Atherosclerosis is less
commonly found in aneurysm of the ascending aorta than in
those of the descending aorta. However, atherosclerosis should
be considered a concomitant process and not a direct cause
of aneurysm formation and growth. Aortic medial degenera-
tion has been demonstrated in most aneurysms, regardless
of their cause and location. The structural integrity of the
adventitia is also lost as an aneurysm forms and expands.
In the ascending aorta, gradual degenerative changes of the
media can be related to congenital disorders of the extra-
cellular matrix associated with an alteration of the elastic
fibers, such as Marfan syndrome (Fig. 32.25). Moreover,
families with a genetic predisposition to the development
of thoracic and abdominal aneurysms, without evidence of
Figure 32.21. MRA of a patient who underwent multiple opera- collagen–vascular disease, have been documented.
tions for type A dissection: In the ascending aorta, the origin of reim- The overall survival of patients with thoracic aortic an-
planted coronary arteries is well depicted. The right coronary artery is eurysms has improved significantly in the past few years.
visible up to its distal portion. The estimated 5-year risk of rupture of a thoracic aneurysm
with a diameter between 4 and 5.9 cm is 16%, but it rises
to 31% for aneurysms 6 cm or more. Since the only well-
High anatomic details and image quality derived from thin- documented risk factor for aortic rupture is increasing size
ner collimations, volumetric acquisition, and superior aortic of aneurysm, the major goal in the imaging evaluation of an
enhancement are the hallmarks of this technique. Especially aortic aneurysm is accurate measurement of its size (104).
with MPR and 3D VR features, it is easy and clear to see
the postoperative anatomy and the success or complication MRI Findings
of any aortic repair modality, particularly during follow-up
(Figs. 32.22 and 32.23) (3,112–116). Artifacts from metal- MRI is effective in identifying and characterizing thoracic
lic stents and surgical clips do not represent a problem with and abdominal aortic aneurysms. Standard spin-echo
this new generation of scanners, because of thinner collima- sequences are helpful in evaluating alterations of the aortic
tions, fast gantry rotation time plus half-rotation reconstruc- wall and periaortic space (Fig. 32.26). Periaortic hematoma
tion algorithm (180 degrees), and kernel convolution filters and areas of high signal intensity within the thrombus may
(Fig. 32.24). indicate instability of the aneurysm and are well depicted on
spin-echo images. Atherosclerotic lesions are visualized as
areas of increased thickness with high signal intensity and
Aortic Aneurysms
irregular profiles. Use of the sagittal plane allows assess-
Aortic aneurysm is a localized or diffuse dilatation involv- ment of location and extent of the aneurysm and avoids
ing all layers of the aortic wall, exceeding the expected aor- partial volume effects. With fat suppression technique,
tic diameter by a factor of 1.5 or more. Aortic aneurysm is the outer wall of the aneurysm can be easily distinguished
the 13th most common cause of death in the United States. on MR images by the periadventitial fat tissue, so that the
LWBK1209-ch32_p503-544.indd 524 16/05/13 9:22 PM

A B
Figure 32.22. MDCT MPR (A–B)

and VR (C–D) images showing a post-
operative evaluation of a type A dissec-
tion treated with surgical prosthesis of
ascending aorta (Bentall technique) and
aortic arch (elephant trunk technique).
Surgical anastomosis (arrow), supra-
aortic vessels reimplant (arrowhead)
and prosthetic tube with coronary
reimplants (large arrow) are well recog- C D
nized by CT scan.
Figure 32.23. MDCT 3D (A) and

MIP (B) images of a patient who under-
went surgery for type A aortic dissec-
tion. CT easily detect the prosthetic
tube in ascending aorta and aortic
arch (arrows) together with proximal
and distal anastomoses (asterisk) and
supra-aortic vessel reimplant (island
A B
technique, hash)
LWBK1209-ch32_p503-544.indd 525 16/05/13 9:22 PM

A B C
Figure 32.24. MDCT VR (A) and MPR (B–C) showing postoperative evaluation after stent graft apposi-
tion (arrows) in type B dissection; metallic artifacts are minimized with such type of scanner.
aneurysm diameter can be accurately measured. The high aortic branches (Fig. 32.27) (21,24,26,28). The homoge-
level of reproducibility of MRI measurements ensures optimal neous enhancement of flowing blood within the lumen
reliability in monitoring expansion rate (109). Gadolinium- facilitates the delineation of thrombus. Stenosis of the aortic
enhanced T1-weighted MRI and MRA may play an impor- branch arteries can be detected with high rate of sensitivity,
tant role in preoperative evaluation. Contrast-enhanced and x-ray angiography is no needed (118,119).
3D MRA can provide precise topographic information Postoperative neurologic deficit secondary to spinal
about the extent of an aneurysm and its relationship to the cord ischemia is a serious and unpredictable complication
of surgery performed on descending aorta. Since selective
angiography of the spinal arteries is time consuming, dif-
ficult, and potentially hazardous, preoperative evaluation of
the Adamkievicz artery has been uncommon. The capabil-
ity of contrast MRA to visualize the artery of Adamkievicz
reported in the literature represents an important advance
in planning the surgical repair of a thoracic aneurysm (120).
CT Findings
Helical and, better, multidetector CT can easily detect aneu-
rysms and facilitate surgical planning by delineating the extent
of the aneurysm and the involvement of aortic branches
(Fig. 32.28). In 1996, Quint et al. (121) found that an analy-
sis of transverse sections and MPRs were 94% accurate, with
a positive predictive value of 95% and a negative one of 93%
for successfully predicting the need for hypothermic circula-
tory arrest. MDCT with enhanced resolution, image quality,
and features could be now close to 100% of accuracy.
Due to the tortuosity and curvature of the thoracic aorta,
Figure 32.25. Spin-echo axial and sagittal images of the thoracic aneurysm sizing is performed most accurately when double-
aorta of a Marfan patient: Severe dilatation of the Valsalva sinuses. oblique tomograms are generated perpendicular to the aortic
LWBK1209-ch32_p503-544.indd 526 16/05/13 9:22 PM

Figure 32.26. Axial (A) and

sagittal (B) MR spin-echo images
showing a dilated ascending aorta
A B
(arrows).
flow lumen. To date, information concerning the risk of an- tomograms, and compute accurate cross-sectional areas and
eurysm rupture and expansion rate is based upon measure- mean diameters, the creation of true vessel cross-sections is
ments made from transverse sections, where true diameters probably not practical for routine applications unless siz-
can be overestimated. ing endoluminal prostheses. MDCT and 3D VR applications
Until analysis tools are available to automatically iden- can help enhance customized stent design and assessment of
tify the center of the flow channel, create true perpendicular the approach vessels of the iliac–femoral system (patency,
Figure 32.27. Three-

dimensional (A) and MIP (B) MRA
images showing a dilated aortic
A B
root (arrows).
LWBK1209-ch32_p503-544.indd 527 16/05/13 9:22 PM

Figure 32.28. MDCT ECG-

gated MIP (A) and VR (B) images
of a severe dilatation of the ascend-
A B
ing aorta (arrows).
pathology, and tortuosity) (Fig. 32.29). The caliber, length, only technique available to perform this, differentiating
angle, calcification, and burden of mural thrombus of poten- thrombosed and atheromatous elements from the adjacent
tial “landing sites” are also documented. Length, shape, and structures, is a painstaking manual segmentation per-
angle of the aneurysm neck are also evaluated (1,113,122). formed by drawing ROI around the aorta on each cross-
The most sensitive measure of aneurysm size, however, section. Until this technique will not be standardized or
is not the determination of true aortic diameters or even automated, this approach is probably not practical for rou-
cross-sectional area, but aneurysm volume. Currently, the tine applications.
Figure 32.29. MDCT MIP (A)

and VR (B) images of a ruptured
thoracic aneurysm. CT shows the
A B rupture point (arrowheads) and the
mediastinal hemorrhage (arrows).
LWBK1209-ch32_p503-544.indd 528 16/05/13 9:22 PM

Abdominal Aneurysm administration of gadolinium, the periaortic cuff enhances

significantly, so that intraluminal thrombus and aortic wall
The incidence of abdominal aneurysm is 3% in persons are clearly defined, along with adjacent involved structures
aged 50 years or more. As in thoracic aneurysm, athero- embedded in the inflammatory cuff. Gadolinium-enhanced
sclerosis is considered the most common cause. A familial MRA will eventually replace catheter angiography in pre-
predisposition to atherosclerotic aneurysm development is operative assessment of AAAs (125). The extent and mor-
under investigation. The causes are thought to be multi- phology of the abdominal aorta are clearly depicted by 3D
factorial, including genetic, environmental, and physi- MRA. Postprocessing by surface rendering provides a 3D
ologic influences. The effects of normal aging, such as impression that can be helpful in planning surgical or endo-
elastic fiber fragment, decreases in the number of smooth vascular therapy. Targeted coronal MIP images display the
muscle cell, and alteration in collagen, are accelerated arterial anatomy and pathology of an aortic aneurysm and
by hypertension and smoking. Since the initial studies the abdominal vessels, so that vascular stenosis can be iden-
in the 1960s, several trials have investigated the natu- tified. Since the contrast agent is not nephrotoxic, contrast-
ral history of abdominal aortic aneurysms (AAA), and enhanced MRA is the method of choice for detection and
the results have been controversial. Large size is rec- monitoring AAAs in patients with renal failure.
ognized as the major risk factor for aortic rupture and
is the most relevant finding in the decision to intervene
on a nonemergent basis. In a study of 300 patients by CT Findings
Guirguis and Barber (123), the 6-year cumulative inci- CT has been advocated in the preoperative evaluation of
dence of rupture was 1% among patients with aneurysms AAAs (126). It is less invasive and more accurate than conven-
smaller than 4 cm and 2% in patients with aneurysm 4 to tional angiography for predicting abdominal aortic aneurysm
5 cm in size. For aneurysms larger than 5 cm, the inci- size (127,128) and is also superior in its ability to demonstrate
dence of rupture was considerably higher, approaching mural thrombus within an aneurysm, inflammatory aneu-
20%. rysms, perianeurysmal blood due to contained rupture, and
Abdominal aneurysm is usually monitored by ultraso- coexistent nonvascular abdominal disease (126).
nography or CT, both of which are accurate in depicting the Different studies demonstrated the reliability and accu-
dimensions of an aneurysm and its relationship to the renal racy of helical CT depicting size, location, and extent of
or iliac arteries. However, catheter angiography has long AAAs, close to 100% (1,129–131). The role of CT is essen-
been considered the standard for the preoperative evalua- tial in planning endovascular stent procedure. As well as in
tion; it provides detailed information about vessels of abdo- thoracic aneurysms, in abdominal aortic aneurysm MDCT
men or lower extremity, which is not always well defined can easily evaluate iliac–femoral access and, length, shape
with spiral CT. The use of endovascular procedures to treat and angle of the aneurysm neck, distance from renal arteries
abdominal aneurysms has further increased the need for and caliber, length, angle, calcification and burden of mural
precise definition of these anatomic details, which is strictly thrombus of potential “landing sites” (1,113,122).
related with procedural success (124). The distance of
the aneurysm from the renal arteries, the involvement of the
iliac arteries, and the angle of the aneurysm neck or of the Aortic Trauma
iliac–femoral axis are crucial in planning an endovascular A traumatic aortic rupture is a lesion caused by trauma
stent graft. that extends from the intima to the adventitia. Trauma is
the third leading cause of death in the United States and the
MRI Findings leading cause of death in individuals under the age of 40.
Among lethal traumatic lesions, aortic rupture is secondary
MRI was ineffectively used in the evaluation of the only to head trauma; one-fourth of the deaths resulting from
abdominal aneurysms before the advent of recent techni- motor vehicle accident are associated with aortic rupture,
cal improvements. The implementation of inversion recov- which occurs in 8,000 victims per year in the United States.
ery fast spin-echo sequences has made possible T1-weighted Air bags and seatbelts do not protect against this type of
breath-hold imaging with adequate suppression of the blood injury, which can be expected to increase in the statistic of
pool. The acquisition with this technique is rapid, so that road traffic injuries; the frequency of lethal injuries in head-
ECG gating is unnecessary. The aortic wall, mural throm- on collisions has been lowered by the mandatory use of
bus, and atheromatous plaques are well depicted in the restraints, which protect the victim from thoracic and head
axial plane, whereas the coronal plane allows definition of lesions but not from the mechanism producing aortic rup-
the diameter and longitudinal extension of the aneurysm. ture. The aortic segment subjected to the greatest strain by
T1-weighted and T2-weighted spin-echo sequences are also rapid deceleration forces is just beyond the isthmus, where
effective in detecting inflammatory changes in the mural the relatively mobile thoracic aorta is joined by the ligamen-
and periaortic space. Inflammatory AAAs are a variant of tum arteriosus. Aortic rupture occurs at this site 90% of
abdominal aneurysm characterized by thickened aortic wall, the time in clinical series. In the ascending aorta, the seg-
periaortic fibrosis, and adhesion to local structures. This ment close to the innominate artery or the proximal segment
form is known to be associated with a higher rate operative immediately superior to the aortic valve may be involved.
mortality, and a specific operative technique is required. On Other less common locations of trauma are the distal seg-
spin-echo images the periaortic cuff of inflammatory aneu- ments of the descending aorta or the abdominal infrarenal
rysms has intermediate signal intensity. After the intravenous segment. The lesion is transverse and involves all or part of
LWBK1209-ch32_p503-544.indd 529 16/05/13 9:22 PM

the aortic circumference, penetrating the aortic layers to vari- On the basis of positive chest x-ray findings, several imag-
ous degrees. Intimal hemorrhage without any laceration has ing modalities are currently available to confirm or exclude
been described in pathologic series but was not recognized the presence of the lesion (133). Particularly in a patient with
in the clinical setting before the advent of high-resolution multiple traumas, aortography should be considered an obso-
tomographic imaging modalities. When a laceration is pres- lete procedure because its sensitivity is low, the information
ent it may extend through the media into the adventitia with it provides is limited and the approach is potentially danger-
the formation of false aneurysm. Periaortic hemorrhage ous. The choice of imaging strategy has to take account of the
occurs irrespective of the type of lesion. Complete rupture patient’s clinical condition. In the case of severe hemodynamic
causes immediate death in 85% of cases. If the aorta is not instability, TEE has the advantage that it can be performed at
completely ruptured at the time of trauma, the aortic wall the bedside, without interrupting resuscitative and therapeutic
is stabilized by the formation of a hematoma in the adven- measures. In the more stable patients, the ideal modalities are
titia and surrounding structures. In such cases, effective those that provide high-definition images of the aortic wall
intravenous therapy with vasodilators and beta-blockers and information on the other organs and structures affected
reduces the aortic wall stress and the risk of rupture in the by the traumatic impact, like MR and CT.
acute phase. Delayed surgery can be planned after other
associated lesions have been treated (132). This type of
MRI Findings
management, reported in recent surgical series, has low-
ered the operative mortality to 0% to 10% in comparison A long examination time as well a difficult access to the
with the 20% to 50% mortality previously reported with patient has been considered the main limitation of MRI in
emergency surgery. acute aortic pathology. The development of fast MRI tech-
niques has shortened the examination time to a few minutes,
so that MRI can be used even in critically ill patients. The
Imaging Findings
value of MRI in detecting traumatic aortic rupture in com-
Although the new surgical strategies are associated with low parison with angiography and CT was reported in a series
rates of spontaneous mortality in patients with traumatic of 24 consecutive patients (134). The diagnostic accuracy of
aortic lesion observed in the clinical setting, traumatic aortic MRI was 100%, that of angiography was 84% (two false
rupture must be considered a potentially evolving lesion. For negatives, in two cases of limited partial lesion), and that of
these reasons, a prompt and accurate diagnosis is necessary to CT (single detector) was 69% (two false negatives and three
initiate a pharmacologic control of the arterial blood pressure false positives). The potential of MRI to detect the hemor-
and stratify the risk of delayed or emergency surgical repair. rhagic component of a lesion by its high signal intensity is
Chest x-ray is routinely performed in all victims of blunt beneficial in traumatized patients. On spin-echo image in
thoracic trauma victims and plays an essential role in deter- the sagittal plane (Fig. 32.30), longitudinal visualization of
mining when a traumatic aortic lesion should be suspected. the thoracic aorta makes it possible to distinguish a partial
A B
Figure 32.30. MR spin-echo sagittal images of acute traumatic aortic lesion: Partial laceration of the aortic
wall results in a diverticular aneurysm (asterisk, A and section sign, B).
LWBK1209-ch32_p503-544.indd 530 16/05/13 9:22 PM

Figure 32.31. Spin-echo sagittal

images of an unstable traumatic aortic
lesion. A: Circumferential laceration
with diverticular aneurysm and large
periaortic hematoma. B: One week
later the diverticular aneurysm is A B
enlarged.
lesion (a tear limited to the anterior or the posterior wall) CT Findings

from a lesion encompassing the entire aortic circumference.
This discrimination is of prognostic significance, because a CT scanning for the detection of aortic injury is widely
circumferential lesion may be more likely to rupture (135). employed, especially in emergency due to the availability and
The presence of periadventitial hematoma and of pleural and speed of this technique. CT is not operator dependent or sig-
mediastinal hemorrhagic effusion may also be considered a nificantly influenced by patient’s habitus, is not invasive and
sign of instability (Fig. 32.31). In the same sequence used to patient is easily monitored while in the scanner. In addition,
evaluate the aortic lesion, without the need of any additional CT allows quick assessment of not only the aorta, but also
time, the wide field of view of MRI provides a comprehen- the entire thorax and abdomen. Nonradiologists also find
sive evaluation of chest trauma, such as lung contusion and CT images the easiest to understand in the consultation and
edema, pleural effusion, and rib fractures. Furthermore, if operating room (89). CTA has been effective and efficient in
delayed surgery is considered, MRI may be used to moni- emergency posttraumatic aortic evaluation (136–138). With
tor thoracic and aortic lesions because it is noninvasive and conventional CT, the detection of subtle aortic injuries is still
repeatable. a problem because a small intimal tear extending in the axial
MRA provides an excellent display of the aortic lesion plane may be obscured by volume averaging with the normal
and its relationship with supra-aortic vessels (Fig. 32.32). aortic lumen. The principle application of conventional CT is
However, it does not add any diagnostic value to spin-echo the detection of mediastinal hemorrhage that has a speci-
MRI, and it cannot supply information on parietal lesions ficity of 87% and sensitivity of 99% to 100% for predicting
and hemorrhagic fluids outside the aortic vessel. aortic injury (137). Spiral MDCT with multiplanar and 3D
Figure 32.32. MRA images of

traumatic aortic lesion: The relation-
ship with arch vessels is well depicted
(arrows) as well as diverticular aneu- A B
rysms (asterisk, A and section sign, B).
LWBK1209-ch32_p503-544.indd 531 16/05/13 9:23 PM

A B C
D E F
Figure 32.33. MPR axial (A–C) and MIP oblique sagittal (D–F) CT images of three different traumatic
aortic lesions (A–D,B–E,C–F) showing the aortic injury (arrows) with pleural effusion (asterisk), periaortic
mediastinal hematoma (section sign) and acute thoracic-mediastinal hemorrhage (hash).
reconstruction and better contrast enhancement overcomes In 1999, Gavant (143) proposed a useful helical CT trau-
the limitations of conventional CT. Since it provides high- matic aortic injury grading system in which, for each grade
quality images with thinner collimation and drastic reduction of of aortic injury, CT findings and triage/clinical management
acquisition time and motion artifacts, spiral CT is a diagnostic were considered. There are four grades of injury: Normal
method of great value. Traumatic injuries manifest on CTA as aorta, minimal aortic injury, confined thoracic aortic injury,
frank contrast agent extravasation, aortic contour abnormal- and total aortic disruption (Table 32.7).
ity (representing intima and media disruption), focal caliber
change, and pseudoaneurysm or dissection flap (Fig. 32.33)
Indication and Follow-up of
(139,140). CTA is widely applied in the setting of acute chest
Endovascular Aortic Repair
trauma providing vessel information and a review of the chest
with CT Technique
wall, lungs, airways, and mediastinum (141). It is considered
to have high sensibility and specificity (close to 100%) and During the past few years, advances in medical treatment
good negative and positive predictive values for aortic injury and close imaging follow-up contributed to improved sur-
(142–144). These remain the only published comparative stud- vival of patients with aortic diseases. However, no medical
ies of helical CTA of blunt thoracic aortic trauma to date. therapy can arrest progression into expansion and rupture,
LWBK1209-ch32_p503-544.indd 532 16/05/13 9:23 PM

resley Trauma Center CT Grading System: Grades and

P
Table 32.7
CT Findings
Grade Subgrade CT Findings
GRADE I Ia Normal thoracic aorta

Normal aorta No mediastinal hematoma
Ib Normal thoracic aorta
Mediastinal hematoma (para-aortic)
GRADE II IIa Small (<1 cm) pseudoaneurysm
Minimal aortic injury Indeterminate <1 cm intimal flap or thrombus
No mediastinal hematoma
IIb Small (usually <1 cm) pseudoaneurysm
Indeterminate <1 cm intimal flap or thrombus
Mediastinal hematoma (para-aortic)
GRADE III IIIa >1 cm regular, well-defined pseudoaneurysm
Confined thoracic aortic injury Intimal flap or thrombus
No ascending aorta, arch, or great vessel involvement
Mediastinal hematoma
IIIb >1 cm regular, well-defined pseudoaneurysm
Intimal flap or thrombus
Ascending aorta, arch, or great vessel involvement
GRADE IV IV Irregular, poorly defined pseudoaneurysm
Total aortic disruption Intimal flap or thrombus
Modified from Gavant ML. Helical CT grading of traumatic aortic injuries. Impact on clinical guidelines for
medical and surgical management. Radiol Clin North Am. 1999;37(3):553–574.
which lead to 16% 5-year mortality for aneurysms smaller the aortic wall and the insufficient stent graft diameter may
than 6 cm in diameter and 31% 5-year mortality for aneu- result in a perigraft leakage. Therefore, a precise assessment
rysms larger than 6 cm (145,146). Therefore, preventive of aortic diameters is mandatory for planning the EVT.
surgical resection has been in the past the only chance to Evaluation of proximal aortic neck anatomy includes
improve survival, despite the considerable mortality and determination of length, diameter, shape, angulation, and
morbidity still reported in surgical series for descending presence of atheroma or thrombus within (wall assessment).
aorta repair (101,147–152). Since its introduction in the In thoracic aortic diseases, at least 1 cm of normal wall at
early 90s, endovascular aortic repair has rapidly established the aneurysm extremities is necessary for aneurysm sealing.
its place as an effective treatment for aortic diseases, and this Heavily calcified aorta, thrombus, or atherosclerosis may
approach offers a less invasive therapeutic option, even to preclude graft deployment.
patient classified in a high-risk surgical category (nonsurgi- The evaluation of iliac and femoral arteries is also very im-
cal candidates) (54,152–169). portant: The caliber should be 9 mm or larger and no severe
In aortic stent graft repair, CT scan has been accepted as tortuosity to allow access with the endovascular equipment.
the gold standard technique in preprocedure evaluation and, More specific information are necessary in aortic dissec-
especially, in postprocedure follow-up (170–172) (Figs. 32.34 tion like identification of entry and re-entry sites, relation-
and 32.35). ships among true and false lumina with visceral and femoral
vessels, true and false lumen dimension, and diameter of dis-
tal aortic arch as well as diameter and extension of the dissec-
Preoperative Evaluation
tion to the femoral and iliac arteries; entry and re-entry tears
Patients are selected for endovascular treatment (EVT) on can be well depicted with CT (173).
the basis of suitable anatomy for stent graft placement. With These anatomic criteria are used to determine whether
multidetector scanners and postprocessing techniques such a patient is a candidate for endovascular stent grafting:
as MPR, CMPR, MIP, SSD, and VR, we have the possibility Complete hemodynamic sealing and secure fixation at the
to assess the stent graft repair feasibility with good anatomic landing sites are the absolute requirements for permanent
details and measurements precision. aneurysm or dissection exclusion.
The efficacy of aneurysm sealing by stent graft apposition
is based on the radial forces of the stent graft. Ten to twenty
Postprocedure Follow-up
percent of graft oversizing with respect to the aortic diameter
at the aneurysm extremities is generally sufficient to ensure The CT stent graft surveillance after treatment is, as generally
a good stent graft anchoring and to exclude the aneurysm accepted (156,166,174), at 4 to 8 days and at intervals of 1,
sac from blood flow. The excessive oversizing may damage 3, 6, and 12 months after the procedure and yearly thereafter.
LWBK1209-ch32_p503-544.indd 533 16/05/13 9:23 PM

A B C
Figure 32.34. MDCT VR (A,C) and MIP (B) images of a pre- and postoperative evaluation (EVT of a
residual dissection, arrows) in a patient who underwent surgery for type A dissection (prosthetic ascending
aorta and aortic arch). In the postoperative images the false lumen is completely thrombosed and partially
shrinked (arrowhead).
Parameters that must be evaluated are the absence or treated as soon as possible with surgical or endovascular
presence (endoleak) of flow within the aneurismal sac or in procedure; type II leaks could be treated or not, depending
the false lumen, aneurysm/aortic dimension, and morphol- on sac expansion. Type III endoleaks associated with sub-
ogy of the stent graft (position and configuration of me- sequent aneurysm rupture must be treated quickly because
tallic frames, kinking or rotations, irregular aortic profile risk is the same as type I (aortic rupture) while in type IV
and diameter and wall morphology of proximal and distal leaks we behave as in type II. What must be done in type
necks). In patients treated for dissection particular atten- V endoleaks is still debated, even if the current opinion is
tion should be taken evaluating the perfusion of abdominal about a surgical or EVT. Other factors for rupture, not re-
vessels, especially those arising from the false lumen (Figs. lated to endoleaks, are device migration, stent frame break-
32.35–32.37). age, and fabric tears.
According to accepted nomenclature (175–180), endo In follow-up CT scan, due to delayed filling of the aneu-
leaks are classified into five groups. Type I endoleaks, proxi- rysm sac or false lumen from the side branch vessels, type
mal or distal, are in cases of involvement of the proximal or II leaks can be easily missed when using single phase fol-
distal attachment site of the endovascular graft to the vascu- lowing contrast agent administration; image acquisition
lar wall; type II in cases of the presence of liquid blood origi- should therefore be also delayed (biphasic scan) in order
nating from a tributary to the aortic lumen excluded by the to visualize feeding vessels supplying type II leaks. Biphasic
graft with or without outflow to a second tributary; type III helical CT scanning has been shown to improve detection
in cases of an endoleak that arose from a defect of the graft rate by 11% compared with single-phase scan alone (181).
itself as a laceration, a hole, or a disconnection of modular Disadvantages of CT imaging for lifelong surveillance of en-
graft; type IV is an endoleak caused by stent macroporos- dovascular aortic repair include repeated radiation exposure
ity; and type V is caused by endotension, in case of systemic and complications related to contrast injection. Aneurysm
pressurization of the aneurysm sac or false lumen despite ab- remodeling has been implicated in graft migration and sub-
sence of CT endoleak features (contrast material in the ex- sequent type I leaks or in the development of type III leaks.
cluded portion). Further distinction is between early and late Two-dimensional CT does not provide any information on
endoleaks, the former from the time of procedure through such morphologic changes (182,183). Volumetric analysis
6 months, the latter at the 12-month follow-up or later. of MDCTA may improve the surveillance of evolution and
At present, it is generally accepted that all type I endole- endoleak, but is still time consuming and not easily reliable
aks, even if sealed and regardless of the origin, need to be (184,185).
LWBK1209-ch32_p503-544.indd 534 16/05/13 9:23 PM

A B
Figure 32.35. MDCT pre- (A,B) and

postoperative (C,D) evaluation after EVT in
a posttraumatic pseudoaneurysm (arrows) of
the isthmus: The lesion is completely sealed
and retracted (asterisk), stent graft and the left
C D
subclavian artery are patent.
Aortitis medical treatment may be more effective. Conventional

angiography provides information mainly on late manifes-
Inflammatory diseases of the aorta can be classified into two
tation of the disease, such as vascular stenosis or localized
major subgroups: Aortitis of nonspecific or unknown etiol-
aneurysms.
ogy (Takayasu arteritis, Bechet disease, giant cell aortitis,
Kawasaki disease, ankylosing spondylitis), and specific aor-
MRI Findings
titis, in which the aortitis is the consequence of an inflam-
matory disease of known origin (e.g., syphilitic aortitis). Spin-echo images can detect mural thickening of the aorta.
Strong ethnic differences have been observed in the epide- High signal intensity of the aortic wall has been observed
miologic distribution of nonspecific aortitis. They are com- in giant cell aortitis whereas media with fibrotic replace-
mon in Asian countries and rare in Caucasian population, ment generally has lower signal intensity. Contrast-enhanced
so that the hypothesis of a genetic transmission is supported. T1- and T2-weighted spin-echo MRI have been shown to be
Histologically, marked irregular mural thickening of the highly effective in the evaluation of Takayasu arteritis (186),
aortic wall and a fibrous lesion are the result of the inflam- even in the early phases, providing important information on
matory process of the media. These alterations can lead to the activity of the disease. Active inflammatory disease appears
stenotic lesions (Takayasu disease), aneurysms of the aorta as variable thickening of the aortic wall that enhances after
and its major branches, or aortic insufficiency as a conse- gadolinium administration. Choe et al. (186) reported 88% of
quence of dilatation of the aortic root. concordance between contrast-enhanced MRI and clinical
The diagnosis of aortitis is often difficult because nonspe- and laboratory findings in 26 patients with active Takayasu
cific systemic symptoms are present in the early phase when arteritis; the chronic, quiescent stage is characterized by
LWBK1209-ch32_p503-544.indd 535 16/05/13 9:23 PM

A B
Figure 32.36. MDCT pre- (A,B) and

postoperative (C,D) evaluation after EVT of a
thoracic aneurysm (arrows): Stent graft is pat-
ent and the aneurysm sac is completely sealed
C D
(asterisk).
extensive perivascular fibrosis without contrast enhance- aortitis prior to luminal dilatation remains difficult because
ment. MRA can replace invasive angiography in the study CT features overlap with retroperitoneal fibrosis, hemor-
of aortic and branch vessel stenosis, allowing diagnostic rhage, and lymphadenopathy.
assessment and follow-up (187). Avoidance of angiography The CT appearance of abdominal aortic infection has
is highly desirable in patients affected by aortitis, because been described from different authors in the past years (188–
it carries a risk of pseudoaneurysm formation at the site of 193). It consists of an irregular calcified aortic wall, periaor-
arterial puncture. Moreover, MRI can be useful in evaluat- tic density representing inflammation and/or hematoma, and
ing the response to medical treatment by depicting decreases adjacent gas collection, mostly in presence of dilated aortic
in wall thickness of the involved arteries. In patients who lumen. Aneurysm results from weakening or disruption of
undergo surgery for aortitis, recurrence of the inflammatory the aortic wall by an infectious process, therefore, they rep-
process is frequently reported that caused dehiscence of the resent pseudoaneurysm, usually saccular (194). Associated
sutures. A strict, lifelong follow-up is therefore mandatory osteomyelitis, enlarged lymph nodes, and renal infarctions
in patients with aortitis, both before and after surgery. have also been reported (190–193). Early stages of aortic
wall infection are rarely observed (191–193) as periaortic
soft tissue densities with rim enhancement in predominant
CT Findings
anterior and left lateral distributions, sparing the retroaor-
CT is an effective technique for depicting aortic infections tic region (inflammation). In some cases, distributions could
and infected aneurysm, but an early diagnosis of infectious be circumferential and rim could be hypodense and without
LWBK1209-ch32_p503-544.indd 536 16/05/13 9:23 PM

A B C
Figure 32.37. MDCT VR (A), MIP (B) and MPR (C) images showing type III endoleaks (arrows) after a
thoracic aortic aneurysm EVT.
contrast enhancement, located between the periaortic mass vascular ring. The right subclavian artery may arise from an
and aortic mural calcification (edema or necrosis) (170). outpouching, known as diverticulum of Kommerell which
Subtle changes of the aortic wall during the early phase of represents persistence of the most distal portion of the em-
aortitis can be detected on spiral CT with contrast enhance- bryonic right arch. Usually, an aberrant right subclavian ar-
ment of the inflamed segments. However, the high density of tery does not cause any symptom and is an incidental finding
contrast medium inside the vessel lumen may cause artifacts on MRI or CT of the chest. Sometimes, with aging, the right
in the adjacent aortic wall, so that these images may be un- subclavian artery becomes tortuous and ectatic, and may
suitable for detecting early mural changes. In conclusion, the cause an esophagus indentation or obstruction (197).
CT findings of hazy periaortic density with gas adjacent to A right aortic arch passes to the right of the trachea and may
the aortic wall should suggest the presence of acute bacte- descend either to the right or to the left of the thoracic spine.
rial aortitis. This appearance may herald impending rupture It occurs approximately in 0.1% of adults. Two types of right
of the aorta, even in the absence of aneurismal dilatation aortic arch are described: Right aortic arch with mirror image
(170,195). brachiocefalic branching and right aortic arch with aberrant
left subclavian artery. The origin of the aberrant left subclavian
artery may have a focal enlargement. Both types are frequently
Congenital Aortic Diseases associated with other congenital cardiac anomalies. Usually, a
Aortic Arch Anomalies

During fetal life, six pairs of aortic arches form to join the nomalies of the Aortic
A
two dorsal aortae with the aortic sac that will become the Table 32.8
Arch Complex
ascending aorta. At the end of development, some of the
arches disappear, and the third, fourth, and sixth aortic arch 1. Anomalies of course or composition of the aorta
give rise to the adult vascular structures. Aortic arch anom- A. Double aortic arch
B. Aberrant right subclavian artery
alies result either from abnormal regression of an embry-
C. Aberrant innominate or left common carotid arteries with or
onic arch that normally remains patent, or from persistent without trachea compression
patency of a structure that normally regresses (Table 32.8) D. Subclavian steal
(196). E. Ductus arteriosum sling
Aberrant right subclavian artery, the most common type F. Circumflex retroesophageal aortic arch
of vascular anomaly, affects 0.5% of the population. The G. Right aortic arch with or without retroesophageal
reported prevalence ranges from 0.4% to 2%, with the con- component
dition occurring in approximately 1 in 200 births. The right 2. Anomalies of length, size, or continuity of the aorta
subclavian artery arises from the left embryonic aortic arch. A. Cervical aorta
At the end development, an aberrant right subclavian artery B. Pseudocarctation of the aorta
C. Hypoplasia of the aorta
originates in the proximal portion of the descending aorta
D. Complete interruption of the aortic arch
and passes posterior to the esophagus to form an incomplete
LWBK1209-ch32_p503-544.indd 537 16/05/13 9:23 PM

Figure 32.39. MRA of right aortic arch: The aberrant left subcla-
vian artery arise from the right arch.
systematic approach should be used to examine the anatomic

structures in each subsequent slice. Spin-echo images in the
Figure 32.38. Spin-echo axial image of right aortic arch with aber- axial plane can detect the abnormal vessel and its relationship
rant left subclavian artery joining the trachea and the esophagus. with the esophagus and trachea (Fig. 32.38). Additional infor-
mation can be derived from the coronal and sagittal planes
and thin slice thickness to demonstrate the origin of the aber-
rant vessels. Kersting-Sommerhoff et al. (198) demonstrated
right aortic arch with left subclavian artery does not in itself the ability of spin-echo MRI to detect aortic arch anomalies in
cause any symptoms. However, if the ligamentum arteriosum 16 patients. At present, contrast-enhanced MRA is the method
is on the left side, a vascular ring is formed by the right aortic of choice to assess aortic arch anomalies because it can provide
arch, anterior left common carotid artery, ligamentum arterio- an accurate overview of the aortic arch and associated vascu-
sum, and retroesophageal left subclavian artery. lar malformation noninvasively in a 3D format (Fig. 32.39)
Double aortic arch is characterized by the presence of (199,200). Postprocessing by MIP and surface rendering pro-
both a left and a right aortic arch; these arise from a branch- vides a 3D impression that can be useful in understanding the
ing of the ascending aorta, pass on both sides of the trachea abnormal mediastinal anatomy, which is particularly helpful
and esophagus and join posteriorly to form the descending in planning an optimal surgical approach. The combination
aorta, which may lie to the right or left of the vertebral col- of MRA and spin-echo MRI is more effective than catheter
umn. The luminal size of the two arches in relation to each angiography in preoperative evaluation because MRI can dis-
other varies considerably, and one of them, usually the left, play the abnormal aortic arch and arch vessels along with any
may be partially or completely atretic. The double aortic compression of the mediastinal structures.
arch is a vascular ring that can produce severe symptoms if
it compresses the trachea and esophagus.
Aortic Coarctation
Cervical arch is a rare anomaly in which the aortic arch
extends into the soft tissues of the neck before turning down Coarctation is a common congenital anomaly in which an
on itself forming the descending aorta. The anomalous high abnormal plication of the tunica media of the posterior
position of the aortic arch may sometimes produce symp- aortic wall proximal to the ligamentum arteriosum causes
toms related to tracheal compression, such as stridor or a fibrous ridge to form that protrudes into the aorta and
dyspnea. The presence of pulsatile neck mass is the most causes an obstructive lesion. The stenotic segment can be
characteristic finding. focal (aortic coarctation), diffuse (hypoplasic aortic isth-
mus), or complete (aortic arch interruption).
MRI Findings
MRI Findings
MRI is particularly valuable in detecting aortic arch anomalies
because of its ability to image multiple projections with a large Spin-echo sagittal images show the morphologic features
field of view. When an aortic arch anomaly is suspected, a of the coarctation. Obtaining an anatomical display of the
LWBK1209-ch32_p503-544.indd 538 16/05/13 9:23 PM

Figure 32.40. MRA of aortic

coarctation: Focal stenotic segment at the
isthmic zone below a large left subclavian
arterym (arrow, A,B). Where the stenosis
is more severe, it is possible to see a lot of
A B
collateral vessels (arrowhead in A).
extent and severity of the stenotic segment is the first impor- aortic arch and of the coarctation, and the age and clini-
tant step in the diagnosis and quantification of the disease. cal condition of the patient. Surgery for aortic coarctation is
Aortic coarctation is easier to identify on the sagittal plane; recommended at an early age because long-term results seem
on axial images, partial volume effects may lead to an under- to be better. Recently, interventional procedures and bal-
estimation of the severity of coarctation. The aortic arch and loon angioplasty have come into wide use and provide good
arch vessels are also well depicted on the sagittal plane. By results especially in mild or moderate cases. An accurate
the measurement of aortic diameter at the isthmus and above selection of favorable anatomy by high-resolution imaging
the diaphragm, morphologic indexes of the coarctation can modalities is particularly important in interventional proce-
be determined. The severity of the stenosis is expressed as the dure to ensure a low rate of complications and restenosis
ratio of diameter or cross-sectional area, at the isthmus to the (202).
same parameters measured in the abdominal aorta. However, Residual coarctation and aortic arch hypoplasia may be
although the detection of anatomic narrowing of the aorta responsible for postoperative hypertension, more frequently
establishes the diagnosis of coarctation, an assessment of associated with surgical techniques such as coarctation re-
its clinical significance depends on determining its hemody- section with end-to-end anastomosis. An increased risk for
namic effects. Cine MRI has been applied to evaluate flow aneurysm formation at the site of repair has been reported
turbulence across the coarctation; the severity of coarctation after both synthetic patch aortoplasty and subclavian flap
is quantified on the basis of the length of flow void. Further arterioplasty. Moreover, restenosis, aortic dissection, and
functional information can be provided by MR flow map- pseudoaneurysms have been reported after balloon angio-
ping, which can define the severity of the stenosis by measur- plasty. Therefore, careful follow-up is recommended for
ing velocity jets at the level of coarctation. Moreover, flow patients who have undergone repair of an aortic coarcta-
mapping is able to quantify the flow pattern and volume of tion, independently of surgical technique used and timing of
collateral flow in the descending aorta (201). The volume of the repair (203). Echocardiography is widely applied in the
collateral flow is another important parameter of the sever- postoperative evaluation of aortic coarctation repair. With
ity of coarctation, and this information may be crucial in the color Doppler, it can provide useful data on the gradient
choice of surgical strategy. 3D MRA can display the extent across the coarctation, which identifies restenosis. However,
and severity of the coarctation without partial volume errors evaluation of aortic arch anatomy can often be made dif-
and spin dephasing artifacts (Fig. 32.40). Collateral vessels, ficult by limited acoustic windows, especially in adults. MRI
which indicate the severity of hemodynamic effects of the has long been used in the follow-up of repaired aortic co-
anatomic coarctation, are also displayed, and this informa- arctation; it provides an optimal depiction of the thoracic
tion is also important in planning surgical repair. aorta with multiplanar standard spin-echo sequences.
Additional diagnostic information can be obtained with
contrast-enhanced MRA. MRA better visualizes the aortic
Postoperative Findings
arch and proximal portion of the descending aorta, which
Several therapeutic strategies are available for the treatment may have a tortuous or kinked course. Postoperative com-
of aortic coarctation, depending on the morphology of the plications, such as re-coarctation, Dacron patch aneurysm
LWBK1209-ch32_p503-544.indd 539 16/05/13 9:23 PM

A B C
Figure 32.41. Patients underwent surgery for aortic coarctation. Spin-echo sagittal image (A), MRA MIP
(B), and 3D MRA (C) showing a large Dacron patch aneurysm (arrows).
(Fig. 32.41), and anastomotic pseudoaneurysm, are well young patients in whom the disease was no suspected. In a
displayed by postprocessing methods (204). Since postoper- neonate with severely dilated Valsalva sinuses, the differen-
ative complications may not produce any symptoms, MRA tial diagnosis should include the neonatal variant of Marfan
of aortic coarctation after surgical repair should be recom- syndrome. The aortic abnormality is visible on spin-echo
mended routinely. MR images and MR angiograms (205). The abnormal dila-
tation of the aortic sinuses is typically asymmetric, whereas
in Marfan syndrome the dilatation involves the aortic root
Aortic Pseudocoarctation
uniformly. Rupture of sinus aneurysm is usually into the
Pseudocoarctation is a rare anomaly of the thoracic aorta right atrium; it creates a left-to-right shunt, which can be
that occurs when the third to seventh embryonic dorsal seg- visualized by gradient-echo sequences. Criteria for surgical
ments fail to fuse properly to form the aortic arch. It results timing in unruptured aneurysms have not been developed.
in elongation of the aortic arch and the first portion of the Regardless the absolute size of an aneurysm, progressive
descending aorta, fixed by the ligamentum arteriosum, so enlargement on serial studies may constitute indication for
that an abnormal kinking develops. Despite the abnormal surgical repair.
tortuosity and morphologic aspects, similar to those of aortic
coarctation, no significant gradient develops through the
CT Study of Congenital Aortic Disease
kinking. It is usually asymptomatic, but hypertension may
be occasionally present. With aging, turbulent flow can cause Before the advent of faster scanners, CT study of congeni-
progressive dilatation of the pseudocoarctation, and aortic tal aortic disease was unfeasible, due to low-image quality
dissection is frequently reported. and scanning planes (thick collimations and just transaxial
MRI is able to identify pseudocoarctation and is particu- plane), radiation dose, and iodinated contrast material.
larly useful in differentiating it from true coarctation. On With helical or, better, multidetector CT it is possible to
spin-echo images, the abnormal kinking may be visualized avoid image quality and scanning planes problems, is pos-
in the axial and sagittal plane. The morphology of the aortic sible to use a lower quantity of contrast agent and technical
tortuosity is similar to that of coarctation, but no fibrous improvements or tricks can reduce radiation dose.
ridge is present. The elongated and high position of the aor- CT scan is not as frequently used as in adults, due to the
tic arch and the absence of collateral vessels on MRA, and good reliability of echocardiography and MR imaging tech-
the absence of significant stenosis on reformatted images in niques in pediatric patients, but with the great anatomical
the axial and sagittal planes are characteristic features diag- detail of MDCT 2D and 3D reconstruction CT studies are
nostic of pseudocoarctation. growing up.
In literature, CT angiographic evaluations were per-
formed in cases of aortic arch anomalies and coarctation,
Aneurysms of the Valsalva Sinus
pulmonary artery anomalies, venous anomalies, and com-
Aneurysm of the Valsalva sinus is a rare congenital anomaly plex heart disease with success (206,207). CT has the ad-
of the structural layers of the aortic wall characterized by vantages of easy availability and very short scanning times.
the absence of the medial layer. This abnormality is usually It is also possible to set the lowest diagnostic tube current
limited to one of the Valsalva sinuses, most frequently the according to the patient weight and age and, in addition,
right coronary Valsalva sinus. Because of the absence of the doubling the pitch reduces radiation dose by half (207).
elastic components of the medial layer, the Valsalva sinus is Sedation times are significantly shorter than those required
asymmetrically dilated even in the neonatal age. Aneurysm for MR (2 to 10 minutes vs. 30 to 45 minutes) and is pos-
of the Valsalva sinuses does not produce any symptoms sible to use nonionic contrast material with medium iodine
and a high incidence of aortic rupture has been reported in concentration (200 to 300 mgI/mL) administered at a
LWBK1209-ch32_p503-544.indd 540 16/05/13 9:23 PM

dose of 2 mL/kg, better with bolus tracking technique for sensitivity and specificity rates respectively of 89% and 77%
contrast peak optimization (206). The decision to image for MDCT and 97% and 94% for MRI.
with CT versus MR imaging technique should be based on Both techniques allow reliable detection of noncalcified,
availability (equipment and scheduling), patients’ ability to atherosclerotic plaques. MRI is more accurate for tissue
co-operate (children) and clinical goal. characterization and MDCT offers the additive value of a
very short image acquisition time and the possibility to eval-
uate also calcified plaques.
MRI and CT of Human
However, further studies and technical progresses are
Atherosclerotic Plaque
needed for a better atherosclerotic plaque characterization
Autopsy studies have shown that the amount of atheroscle- with MDCT.
rotic plaque in the thoracic aorta directly correlates with
the degree of atherosclerotic disease in the coronary arter-
ies. Thoracic aortic atherosclerosis is a stronger predictor of References
coronary artery disease than conventional risk factors are,
1. Van Hoe L, Baert AL, Gryspeerdt S, et al. Supra- and juxtarenal aneurysms of the
and it is also a marker of an increased risk for mortality, abdominal aorta: Preoperative assessment with thin-section spiral CT. Radiology. 1996;
stroke, and visceral thromboembolic events. Examination of 198(2):443–448.
2. Kaatee R, Van Leeuwen MS, De Lange EE, et al. Spiral CT angiography of the renal arter-
the descending thoracic aorta by TEE and fast CT is used to ies: Should a scan delay based on a test bolus injection or a fixed scan delay be used
predict cardiovascular risk. MRI is a noninvasive imaging to obtain maximum enhancement of the vessels? J Comput Assist Tomogr. 1998;22(4):
modality that can visualize and characterize the composition 541–547.
3. Armerding MD, Rubin GD, Beaulieu CF, et al. Aortic aneurysmal disease: Assessment of
of atherosclerotic plaques and differentiates tissue structure stent-graft treatment-CT versus conventional angiography. Radiology. 2000;215(1):138–
on the basis of proton magnetic properties (208–210). High- 146.
resolution MRI has been increasingly used for assessing the 4. Katz DS, Jorgensen MJ, Rubin GD. Detection and follow-up of important extra-arterial
lesions with helical CT angiography. Clin Radiol. 1999;54(5):294–300.
vascular wall and characterizing atherosclerotic lesions. 5. Rubin GD, Shiau MC, Leung AN, et al. Aorta and iliac arteries: Single versus multiple
Continuous improvement in MR technology and software detector-row helical CT angiography. Radiology. 2000;215(3):670–676.
6. Rubin GD. MDCT imaging of the aorta and peripheral vessels. Eur J Radiol.
allow MR system to achieve a spatial resolution in the order 2003;45(suppl 1):S42–S49.
of micrometers and enable in vivo MR microscopy. The 7. Lawler LP, Fishman EK. Multi-detector row CT of thoracic disease with emphasis on 3D
capability of MRI to distinguish tissue components with an volume rendering and CT angiography. Radiographics. 2001;21(5):1257–1273.
8. Lawler LP, Fishman EK. Multidetector row computed tomography of the aorta and
excellent soft tissue contrast is unique for plaque character- peripheral arteries. Cardiol Clin. 2003;21(4):607–629.
ization. The different composition of stable and unstable 9. Wintersperger BJ, Nikolaou K, Becker CR. Multidetector-row CT angiography of the
lesion can be discriminated by measuring T1, T2, and pro- aorta and visceral arteries. Semin Ultrasound CT MR. 2004;25(1):25–40.
10. Wintersperger BJ, Herzog P, Jakobs T, et al. Initial experience with the clinical use of a 16
ton density changes (209). Lipid components present hyper- detector row CT system. Crit Rev Comput Tomogr. 2002;43(4):283–316.
intense regions within the plaque on both T1- and proton 11. Urban BA, Bluemke DA, Johnson KM, et al. Imaging of thoracic aortic disease. Cardiol
Clin. 1999;17(4):659–682.
density-weighted images and show hypointense signal on T2- 12. Reddy GP, Higgins CB. MR imaging of the thoracic aorta. Magn Reson Imaging Clin N
weighted images. Fibrocellular components are defined as Am. 2000;8(1):1–15.
hyperintense regions of the plaque on T1-, proton density-, 13. Stemerman DH, Krinsky GA, Lee VS, et al. Thoracic aorta: Rapid black-blood MR imag-
ing with half-Fourier rapid acquisition with relaxation enhancement with or without elec-
and T2-weighted images. Calcium deposits can be appreci- trocardiographic triggering. Radiology. 1999;213(1):185–191.
ated as hypointense regions within the plaque on T1-, proton 14. Sakuma H, Bourne MW, O’Sullivan M, et al. Evaluation of thoracic aortic dissection
density-, and T2-weighted images. The fibrous cap and lipid using breath-holding cine MRI. J Comput Assist Tomogr. 1996;20(1):45–50.
15. Niezen RA, Doornbos J, van der Wall EE, et al. Measurement of aortic and pulmonary flow
core, organized thrombus and fresh thrombosis, or calcifi- with MRI at rest and during physical exercise. J Comput Assist Tomogr. 1998;22(2):194–201.
cation and necrotic areas have been imaged in studies per- 16. Powell AJ, Maier SE, Chung T, et al. Phase-velocity cine magnetic resonance imaging
measurement of pulsatile blood flow in children and young adults: In vitro and in vivo
formed both in vitro and in vivo, and atherosclerotic lesions validation. Pediatr Cardiol. 2000;21(2):104–110.
have been assessed in carotid arteries, coronary arteries, and 17. Bogren HG, Buonocore MH. 4D magnetic resonance velocity mapping of blood flow patterns
in the thoracic aorta. MR methods for direct noninvasive in the aorta in young vs. elderly normal subjects. J Magn Reson Imaging. 1999;10(5):861–869.
18. Hopkins KD, Leheman ED, Gosling RG. Aortic compliance measurements: A noninvasive
assessment of aortic atherosclerotic plaque thickness, extent, indicator of atherosclerosis. Lancet. 1994;343:1447.
and composition allow serial evaluation of the progression 19. Groenink M, de Roos A, Mulder BJM, et al. Changes in aortic distensibility and pulse
and therapy-induced regression of atherosclerotic plaques. wave velocity assessed with magnetic resonance imaging following beta-blockers therapy
in the Marfan syndrome. Am J Cardiol. 1998;82:203–208.
Furthermore, MR spectroscopy with carbon 13 can pro- 20. Fattori R, Bacchi Reggiani L, Pepe G, et al. MRI evaluation of aortic elastic properties as
vide information on the chemical constituents of the lipid early expression of Marfan Syndrome. J Cardiovascul Magn Reson. 2000;4:43–48.
21. Debatin JF, Hany TF. MR-based assessment of vascular morphology and function. Eur
pool as a marker of the progression of the atherosclerotic Radiol. 1998;8:528–539.
plaque (211). The advent of MR intravascular imaging de- 22. Sodickson DK, McKenzie CA, Li W, et al. Contrast-enhanced 3D MR angiography
vices further enhance the spatial resolution and homogeneity with simultaneous acquisition of spatial harmonics: A pilot study. Radiology. 2000;217(1):
284–289.
by providing close proximity between imaging coil and the 23. Lee VS, Martin D J, Krinsky GA, et al. Gadolinium-enhanced MR angiography: Artifacts
vessel wall (212). The combination of tissue characterization and pitfalls. AJR. 2000;175:197–205.
and spectroscopy may have an extraordinary potential in as- 24. Neimatallah MA, Ho VB, Dong Q, et al. Gadolinium-enhanced 3D magnetic resonance
angiography of the thoracic vessels. J Magn Reson Imaging. 1999;10:758–770.
sessing the characteristics of plaque, so that new insight into 25. Goyen M, Ruehm SG, Debatin JF. MR-angiography: The role of contrast agents. Eur
plaque formation and evolution will be acquired. J Radiol. 2000;34(3):247–256.
26. Prince MR, Narasimham DL, Jacoby WT, et al. Three-dimensional gadolinium-enhanced
Recently, high-resolution MDCT imaging has been MR angiography of the thoracic aorta. AJR. 1996;166:1387–1397.
used in the attempt to characterize atherosclerotic human 27. Krinsky G, Rofsky N, Flyer M, et al. Gadolinium-enhanced three-dimensional MR angi-
plaques, even if focused on coronary artery plaques (213– ography of acquired arch vessels disease. AJR. 1996;167:981–987.
28. Krinsky G, Rofsky N, De Corato DR, et al. Thoracic aorta: Comparison of gadolinium-
218). A comparative study between MRI and CT in ath- enhanced three-dimensional MR angiography with conventional MR imaging. Radiology.
erosclerotic noncalcified plaque assessment (214) reported 1997;202:183–193.
LWBK1209-ch32_p503-544.indd 541 16/05/13 9:23 PM

29. Scheffler K, Lehnhardt S. Principles and applications of balanced SSFP techniques. Eur 67. Chung JW, Park JH, Im JG, et al. Spiral CT angiography of the thoracic aorta.
Radiol. 2003;13(11):2409–2418. Radiographics. 1996;16(4):811–824.
30. Fuchs F, Laub G, Othomo K. TrueFISP—technical considerations and cardiovascular 68. Sebastia C, Pallisa E, Quiroga S, et al. Aortic dissection: Diagnosis and follow-up with
applications. Eur J Radiol. 2003;46(1):28–32. helical CT. Radiographics. 1999;19(1):45–60.
31. Miyazaki M, Lee V. Nonenhanced MR angiography. Radiology. 2008;248(1):20–43. 69. Fisher RG, Chasen MH, Lamki N. Diagnosis of injuries of the aorta and brachioce-
32. Prokop M. Multislice CT angiography. Eur J Radiol. 2000;36(2):86–96. phalic arteries caused by blunt chest trauma: CT vs aortography. AJR Am J Roentgenol.
33. Heiken JP, Brink JA, Vannier MW. Spiral (helical) CT. Radiology. 1993;189(3):647–656. 1994;162(5):1047–1052.
34. Brink JA, Heiken JP, Balfe DM, et al. Spiral CT: Decreased spatial resolution in vivo due 70. Lee DY, Williams DM, Abrams GD. The dissected aorta: Part II. Differentiation of the
to broadening of section-sensitivity profile. Radiology. 1992;185(2):469–474. true from the false lumen with intravascular US. Radiology. 1997;203(1):32–36.
35. Fuchs T, Kachelriess M, Kalender WA. Technical advances in multi-slice spiral CT. Eur J 71. Williams MP, Farrow R. Atypical patterns in the CT diagnosis of aortic dissection. Clin
Radiol. 2000;36(2):69–73. Radiol. 1994;49(10):686–689.
36. Hu H. Multi-slice helical CT: Scan and reconstruction. Med Phys. 1999;26(1):5–18. 72. LePage MA, Quint LE, Sonnad SS, et al. Aortic dissection: CT features that distinguish
37. Berland LL, Smith JK. Multidetector-array CT: Once again, technology creates new true lumen from false lumen. AJR Am J Roentgenol. 2001;177(1):207–211.
opportunities. Radiology. 1998;209(2):327–329. 73. Nelsen KM, Spizarny DL, Kastan DJ. Intimointimal intussusception in aortic dissection:
38. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial and temporal resolution CT diagnosis. AJR Am J Roentgenol. 1994;162(4):813–814.
for electrocardiogram-gated multislice spiral CT of the heart. J Comput Assist Tomogr. 74. Karabulut N, Goodman LR, Olinger GN. CT diagnosis of an unusual aortic dissection
2001;25(6):907–923. with intimointimal intussusception: The wind sock sign. Comput Assist Tomogr. 1998;
39. Flohr T, Stierstorfer K, Bruder H, et al. Image reconstruction and image quality evaluation 22(5):692–693.
for a 16-slice CT scanner. Med Phys. 2003;30(5):832–845. 75. Rubin GD. Helical CT angiography of the thoracic aorta. J Thorac Imaging. 1997;12(2):
40. Rubin GD. Techniques for performing multidetector-row computed tomographic angiog- 128–149.
raphy. Tech Vasc Interv Radiol. 2001;4(1):2–14. 76. Batra P, Bigoni B, Manning J, et al. Pitfalls in the diagnosis of thoracic aortic dissection at
41. Cademartiri F, van der Lugt A, Luccichenti G, et al. Parameters affecting bolus geometry CT angiography. Radiographics. 2000;20(2):309–320.
in CTA: A review. J Comput Assist Tomogr. 2002;26(4):598–607. 77. Krukemberg E. Beiträge zur Frage des Aneurysma dissecans. Beitr Pathol Anat Allg
42. Macari M, Israel GM, Berman P, et al. Infrarenal abdominal aortic aneurysms at multi- Pathol. 1920;67:329–351.
detector row CT angiography: Intravascular enhancement without a timing acquisition. 78. Fattori R, Bertaccini P, Celletti F, et al. Intramural posttraumatic hematoma of the ascend-
Radiology. 2001;220(2):519–523. ing aorta in a patient with a double aortic arch. Eur Radiol. 1997;7:51–53.
43. Fleischmann D, Rubin GD, Bankier AA, et al. Improved uniformity of aortic enhance- 79. Nienaber CA, von Kodolitsch Y, Petersen B, et al. Intramural hemorrhage of the thoracic
ment with customized contrast medium injection protocols at CT angiography. Radiology. aorta. Diagnostic and therapeutic implication. Circulation. 1995;92:1465–1472.
2000;214(2):363–371. 80. Coady MA, Rizzo JA, Elefteriades JA. Pathologic variants of thoracic aortic dissec-
44. Bae KT, Tran HQ, Heiken JP. Multiphasic injection method for uniform prolonged vascu- tions. Penetrating atherosclerotic ulcers and intramural hematomas. Cardiol Clin. 1999;
lar enhancement at CT angiography: Pharmacokinetic analysis and experimental porcine 17(4):637–657.
model. Radiology. 2000;216(3):872–880. 81. Yamada T, Takamiya M, Naito H, et al. Diagnosis of aortic dissection without inti-
45. Fleischmann D. Use of high-concentration contrast media in multiple-detector-row CT: mal rupture by x-ray computed tomography. Nippon Igaku Hoshasen Gakkai Zasshi.
Principles and rationale. Eur Radiol. 2003;13(suppl 5):M14–M20. 1985;45(5):699–710.
46. Haage P, Schmitz-Rode T, Hubner D, et al. Reduction of contrast material dose and arti- 82. Keren A, Kim CB, Hu BS, et al. Accuracy of multiplane transesophageal echocardiogra-
facts by a saline flush using a double power injector in helical CT of the thorax. AJR Am phy in diagnosis of typical acute aortic dissection and intramural hematoma. J Am Coll
J Roentgenol. 2000;174(4):1049–1053. Cardiol. 1996;28:627–636.
47. Hopper KD, Mosher TJ, Kasales CJ, et al. Thoracic spiral CT: Delivery of contrast material 83. Moore A, Oh J, Bruckman D, et al. Transesophageal echocardiography in the diagnosis
pushed with injectable saline solution in a power injector. Radiology. 1997;205(1):269–271. and management of aortic dissection. An analysis of data from the International Registry
48. Fishman EK. High-resolution three-dimensional imaging from subsecond helical CT data of Aortic Dissection (IRAD). J Am Coll Cardiol. 1999;33-2(A):470 A.
sets: Applications in vascular imaging. AJR Am J Roentgenol. 1997;169(2):441–443. 84. Murray JG, Manisali M, Flamm SD, et al. Intramural hematoma of the thoracic aorta:
49. Kirchgeorg MA, Prokop M. Increasing spiral CT benefits with postprocessing applica- MR imaging findings and their prognostic implications. Radiology. 1997;204:349–355.
tions. Eur J Radiol. 1998;28(1):39–54. 85. Quint LE, Williams DM, Francis IR, et al. Ulcerlike lesions of the aorta: Imaging features
50. Coady MA, Rizzo JA, Goldstein LJ, et al. Natural history, pathogenesis and etiology of and natural history. Radiology. 2001;218(3):719–723.
thoracic aneurysms and dissection card. Clin North Am. 1999;17(4):615–633. 86. Kazerooni EA, Bree RL, Williams DM. Penetrating atherosclerotic ulcers of the descend-
51. Bansal RC, Krishnaswamy C, Ayala K, et al. Frequency and explanation of false nega- ing thoracic aorta: Evaluation with CT and distinction from aortic dissection. Radiology.
tive diagnosis of aortic dissection by aortography and transesophageal echocardiography. 1992;183(3):759–765.
J Am Coll Cardiol. 1995;25:1393–1401. 87. Sueyoshi E, Matsuoka Y, Imada T, et al. New development of an ulcerlike projection in
52. Spittel PC, Spittel JA, Joyce W, et al. Clinical features and differential diagnosis of aortic aortic intramural hematoma: CT evaluation. Radiology. 2002;224(2):536–541.
dissection: experience with 236 cases (1980 through 1990). Mayo Clin Proc. 1993;68: 88. Castaner E, Andreu M, Gallardo X, et al. CT in nontraumatic acute thoracic aortic
642–651. disease: Typical and atypical features and complications. Radiographics. 2003;23:S93–
53. Cigarroa JE, Isselbacher EM, De Sanctis, et al. Diagnostic imaging in the evaluation of sus- S110.
pected aortic dissection. Old standard and new direction. N Engl J Med. 1993;328:35–43. 89. Ledbetter S, Stuk JL, Kaufman JA. Helical (spiral) CT in the evaluation of emergent
54. Nienaber CA, Fattori R, Lund G, et al. Nonsurgical reconstruction of thoracic aortic dis- thoracic aortic syndromes. Traumatic aortic rupture, aortic aneurysm, aortic dissection,
section by stent-graft placement. New Eng J Med. 1999;140:1338–1345. intramural hematoma, and penetrating atherosclerotic ulcer. Radiol Clin North Am.
55. Dake MD, Kato N, Mitchell RS, et al. Endovascular stent-graft placement for the treat- 1999;37(3):575–589.
ment of acute aortic dissection. N Engl J Med. 1999;140:1546–1552. 90. Ide K, Uchida H, Otsuji H, et al. Acute aortic dissection with intramural hematoma: Possibility
56. Nitatori T, Yokoyama K, Hachiya J, et al. Fast dynamic MRI of aortic dissection: Flow of transition to classic dissection or aneurysm. J Thorac Imaging. 1996;11(1):46–52.
assessment by subsecondal imaging. Radiat Med. 1999;17(1):9–14. 91. Kaji S, Nishigami K, Akasaka, et al. Prediction of progression or regression of type A
57. Chang JM, Friese K, Caputo GR, et al. MR measurement of blood flow in the true and aortic intramural hematoma by computed tomography. Circulation. 1999;100(suppl
false channel in chronic aortic dissection. J Comput Assist Tomogr. 1991;15:418–423. 19):II281–II286.
58. Bogaert J, Meyns B, Rademakers FE, et al. Follow-up of aortic dissection: Contribution 92. Choi SH, Choi SJ, Kim JH, et al. Useful CT findings for predicting the progression of aortic
of MR angiography for evaluation of the abdominal aorta and its branches. Eur Radiol. intramural hematoma to overt aortic dissection. J Comput Assist Tomogr. 2001;25(2):295–299.
1997;7:695–702. 93. Sueyoshi E, Matsuoka Y, Sakamoto I, et al. Fate of intramural hematoma of the aorta: CT
59. Nienaber CA, von Kodolitsch Y, Nikolas V, et al. The diagnosis of thoracic aortic dissec- evaluation. J Comput Assist Tomogr. 1997;21(6):931–938.
tion by noninvasive imaging procedures. N Engl J Med. 1993;328:1–9. 94. Shennan T. Dissecting aneurisms. Great Britain Medical Research Council, Special Report
60. Sommer T, Fehske W, Holzknecht N, et al. Aortic dissection: A comparative study of Series no. 193:1934.
diagnosis with spiral CT, multiplanar transesophageal echocardiography, and MR imag- 95. Movsowitz HD, Lampert C, Jacobs LE, et al. Penetrating atherosclerotic aortic ulcers. Am
ing. Radiology. 1996;199:347–352. Heart J. 1994;128:1210–1217.
61. Fisher U, Vossherich R, Kopka L, et al. Dissection of the thoracic aorta: Pre- and post- 96. Hayashi H, Matsuoka Y, Sakamoto I, et al. Penetrating atherosclerotic ulcer of the aorta:
operative findings of turbo-FLASH MR images in the plane of the aortic arch. AJR. Imaging features and disease concept. Radiographics. 2000;20(4):995–1005.
1994;163:1069–1072. 97. Yucel EK, Steinberg FL, Egglin TK, et al. Penetrating atherosclerotic ulcers: Diagnosis
62. Moore AG, Eagle KA, Bruckman D, et al. Choice of computed tomography, transesoph- with MR imaging. Radiology. 1990;177:779–781.
ageal echocardiography, magnetic resonance imaging, and aortography in acute aortic 98. Welch TJ, Stanson AW, Sheedy PF 2nd, et al. Radiologic evaluation of penetrating aortic
dissection: International Registry of Acute Aortic Dissection (IRAD). Am J Cardiol. atherosclerotic ulcer. Radiographics. 1990;10(4):675–685.
2002;89(10):1235–1238. 99. Sawhney NS, DeMaria AN, Blanchard DG. Aortic intramural hematoma: An increasingly
63. Erbel R, Engberding R, Daniel W, et al. Echocardiography in diagnosis of aortic dissec- recognized and potentially fatal entity. Chest. 2001;120(4):1340–1346.
tion. Lancet. 1989;1(8636):457–461. 100. Chung JW, Park JH, Im JG, et al. Spiral CT angiography of the thoracic aorta.
64. Clague J, Magee P, Mills P. Diagnostic techniques in suspected thoracic aortic dissection. Radiographics. 1996;16(4):811–824.
Br Heart J. 1992;67(6):428–429. 101. Fann JI, Smith JA, Miller DC, et al. Surgical management of aortic dissection during a
65. Barbant SD, Eisenberg MJ, Schiller NB. The diagnostic value of imaging techniques for 30-year period. Circulation. 1995;92(suppl 9):II110–II121.
aortic dissection. Am Heart J. 1992;124(2):541–543. 102. Svensson LG, Crawford SE. Statistical analyses of operative results. In: Svensson LG,
66. Small JH, Dixon AK, Coulden RA, et al. Fast CT for aortic dissection. Br J Radiol. Crawford ES, eds. Cardiovascular and Vascular Disease of the Aorta. Philadelphia, PA:
1996;69(826):900–905. WB Saunders Company; 1997:432–455.
LWBK1209-ch32_p503-544.indd 542 16/05/13 9:23 PM

103. Heinemann M, Laas J, Karck M, et al. Thoracic aortic aneurysms after acute type A aortic 138. Scaglione M, Pinto A, Pinto F, et al. Role of contrast-enhanced helical CT in the evalua-
dissection: Necessity for follow-up. Ann Thorac Surg. 1990;49:580–584. tion of acute thoracic aortic injuries after blunt chest trauma. Eur Radiol. 2001;11(12):
104. Bonser RS, Pagano D, Lewis ME, et al. Clinical and patho-anatomical factors affecting 2444–2448.
expansion of thoracic aortic aneurysms. Heart. 2000;84(3):277–283. 139. Creasy JD, Chiles C, Routh WD, et al. Overview of traumatic injury of the thoracic aorta.
105. Fattori R, Bacchi Reggiani ML, Bertaccini P, et al. Evolution of aortic dissection after Radiographics. 1997;17(1):27–45.
surgical repair. Am J Cardiol. 2000;86(8):868–872. 140. Marotta R, Franchetto AA. The CT appearance of aortic transection. AJR Am J Roentgenol.
106. Moore NR, Parry AJ, Trottman-Dickenson B, et al. Fate of the native aorta after repair of 1996;166(3):647–651.
acute type A dissection: A magnetic resonance imaging study. Heart. 1996;75:62–66. 141. Kuhlman JE, Pozniak MA, Collins J, et al. Radiographic and CT findings of blunt
107. Mesana TG, Caus T, Gaubert J, et al. Late complications after prosthetic replacement of chest trauma: Aortic injuries and looking beyond them. Radiographics. 1998;18(5):
the ascending aorta: What did we learn from routine magnetic resonance imaging follow- 1085–1106;discussion 1107–1108.
up? Eur J Cardiothorac Surg. 2000;18(3):313–320. 142. Gavant ML, Menke PG, Fabian T, et al. Blunt traumatic aortic rupture: Detection with
108. Kawamoto S, Bluemke DA, Traill TA, et al. Thoracoabdominal aorta in Marfan syn- helical CT of the chest. Radiology. 1995;197(1):125–133.
drome: MR Imaging findings of progression of vasculopathy after surgical repair. 143. Gavant ML. Helical CT grading of traumatic aortic injuries. Impact on clinical guidelines
Radiology. 1997;203:727–732. for medical and surgical management. Radiol Clin North Am. 1999;37(3):553–574.
109. Gaubert J, Moulin G, Mesana T, et al. Type A dissection of the thoracic aorta. Use of MR 144. Dyer DS, Moore EE, Mestek MF, et al. Can chest CT be used to exclude aortic injury?
imaging for long term follow-up. Radiology. 1995:363–369. Radiology. 1999;213(1):195–202.
110. Loubeyre P, Delignette A, Boneloy L, et al. MRI evaluation of the ascending aorta 145. Dapunt OE, de Asla RA, Griepp EB, et al. Computer-generated 3D representations of the
after graft-inclusion surgery: Comparison between an ultra-fast contrast-enhanced MR aorta: A new tool in the management of aortic aneurysm patients. Thorac Cardiovasc
sequence and conventional cine-MRI. J Magn Reson Imaging. 1996;6:478–483. Surg. 1994;42(1):25–28.
111. Fattori R, Descovich B, Bertaccini P, et al. Composite graft replacement of the ascend- 146. Griepp RB, Ergin MA, Galla JD, et al. Natural history of descending thoracic and
ing aorta: Leakage detection with gadolinium-enhanced MR imaging. Radiology. thoracoabdominal aneurysms. Ann Thorac Surg. 1999;67(6):1927–1930;discussion
1999;212:573–577. 1953–1958.
112. Quint LE, Francis IR, Williams DM, et al. Synthetic interposition grafts of the thoracic 147. Crawford ES, Crawford JL, Safi HJ, et al. Thoracoabdominal aortic aneurysms:
aorta: Postoperative appearance on serial CT studies. Radiology. 1999;211(2):317–324. Preoperative and intraoperative factors determining immediate and long-term results of
113. Rydberg J, Kopecky KK, Lalka SG, et al. Stent grafting of abdominal aortic aneurysms: operations in 605 patients. J Vasc Surg. 1986;3(3):389–404.
Pre- and postoperative evaluation with multislice helical CT. J Comput Assist Tomogr. 148. Schor JS, Yerlioglu ME, Galla JD, et al. Selective management of acute type B aortic dis-
2001;25(4):580–586. section: Long-term follow-up. Ann Thorac Surg. 1996;61(5):1339–1341.
114. Mita T, Arita T, Matsunaga N, et al. Complications of endovascular repair for thoracic and 149. Griepp RB, Ergin MA, Galla JD, et al. Looking for the artery of Adamkiewicz: A quest
abdominal aortic aneurysm: An imaging spectrum. Radiographics. 2000;20(5):1263–1278. to minimize paraplegia after operations for aneurysms of the descending thoracic and
115. Sakai T, Dake MD, Semba CP, et al. Descending thoracic aortic aneurysm: Thoracic CT thoracoabdominal aorta. J Thorac Cardiovasc Surg. 1996;112(5):1202–1213; discussion
findings after endovascular stent-graft placement. Radiology. 1999;212(1):169–174. 1213–1215.
116. Riley P, Rooney S, Bonser R, et al. Imaging the post-operative thoracic aorta: Normal 150. Kouchoukos NT, Dougenis D. Surgery of the thoracic aorta. N Engl J Med. 1997;336
anatomy and pitfalls. Br J Radiol. 2001;74(888):1150–1158. (26):1876–1888.
117. Clouse WD, Hallett JW, Shaff HV, et al. Improved prognosis of thoracic aortic aneurysm: 151. Svensson LG. Natural history of aneurysms of the descending and thoracoabdominal
A population based study. JAMA. 1998;9280:1926–1929. aorta. J Card Surg. 1997;12(2 suppl):279–284.
118. Weishaupt D, Ruhm SG, Binkert C, et al. Equilibrium-phase MR angiography of the 152. Ehrlich M, Grabenwoeger M, Cartes-Zumelzu F, et al. Endovascular stent graft repair for
aortoiliac and renal arteries using a blood pool contrast agent. Am J Roentgenol. aneurysms on the descending thoracic aorta. Ann Thorac Surg. 1998;66(1):19–24; discus-
2000;175(1):189–195. sion 24–25.
119. Holland AE, Barentsz JO, Skotnicki S, et al. Preoperative MRA assessment of the coro- 153. Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for
nary arteries in an ascending aortic aneurysm. J Magn Reson Imaging. 2000;11:324–326. abdominal aortic aneurysms. Ann Vasc Surg. 1991;5(6):491–499.
120. Yamada N, Okita Y, Minatoya K, et al. Preoperative demonstration of the Adamkiewicz 154. Dake MD, Miller DC, Semba CP, et al. Transluminal placement of endovascular stent-
artery by magnetic resonance angiography in patients with descending or thoracoabdomi- grafts for the treatment of descending thoracic aortic aneurysms. N Engl J Med. 1994;
nal aortic aneurysms. Eur J Cardiothorac Surg. 2000;18(1):104–111. 331(26):1729–1734.
121. Quint LE, Francis IR, Williams DM, et al. Evaluation of thoracic aortic disease with the 155. Dake MD, Kato N, Mitchell RS, et al. Endovascular stent-graft placement for the treat-
use of helical CT and multiplanar reconstructions: Comparison with surgical findings. ment of acute aortic dissection. N Engl J Med. 1999;340(20):1546–1552.
Radiology. 1996;201(1):37–41. 156. Kato N, Dake MD, Miller DC, et al. Traumatic thoracic aortic aneurysm: Treatment with
122. Bortone AS, De Cillis E, D’Agostino D, et al. Endovascular treatment of thoracic aortic endovascular stent-grafts. Radiology. 1997;205(3):657–662.
disease: Four years of experience. Circulation. 2004;110(11 suppl 1):II262–II267. 157. Mitchell RS, Miller DC, Dake MD, et al. Thoracic aortic aneurysm repair with an endo-
123. Guirguis EM, Barber GG. The natural history of abdominal aortic aneurysm. Am J Surg. vascular stent graft: The “first generation”. Ann Thorac Surg. 1999;67(6):1971–1974;
1991;162(5):481–483. discussion 1979–1980.
124. Hilfiker PR, Quick HH, Pfammatter T, et al. Three-dimensional MR angiography of a 158. Zarins CK, White RA, Schwarten D, et al. AneuRx stent graft versus open surgical
nitinol-based abdominal aortic stent graft: Assessment of heating and imaging character- repair of abdominal aortic aneurysms: Multicenter prospective clinical trial. J Vasc Surg.
istics. Eur Radiol. 1999;9(9):1775–1780. 1999;29(2):292–305; discussion 306–308.
125. Grist TM. MRA of the abdominal aorta and lower extremities. J Magn Reson Imaging. 159. Zarins CK, Heikkinen MA, Lee ES, et al. Short- and long-term outcome following endo-
2000;11:32–43. vascular aneurysm repair. How does it compare to open surgery? J Cardiovasc Surg
126. Papanicolaou N, Wittenberg J, Ferrucci JT Jr, et al. Preoperative evaluation of abdominal (Torino). 2004;45(4):321–333.
aortic aneurysms by computed tomography. AJR Am J Roentgenol. 1986;146(4):711–715. 160. Temudom T, D’Ayala M, Marin ML et al. Endovascular grafts in the treatment of thoracic
127. Bandyk DF. Preoperative imaging of aortic aneurysms. Conventional and digital subtrac- aortic aneurysms and pseudoaneurysms. Ann Vasc Surg. 2000;14(3):230–238.
tion angiography, computed tomography scanning, and magnetic resonance imaging. Surg 161. Grabenwoger M, Hutschala D, Ehrlich MP, et al. Thoracic aortic aneurysms: Treatment
Clin North Am. 1989;69(4):721–735. with endovascular self-expandable stent grafts. Ann Thorac Surg. 2000;69(2):441–445.
128. Pavone P, Di Cesare E, Di Renzi P, et al. Abdominal aortic aneurysm evaluation: 162. Beregi JP, Prat A, Gaxotte V, et al. Endovascular treatment for dissection of the descending
Comparison of US, CT, MRI, and angiography. Magn Reson Imaging. 1990;8(3):199–204. aorta. Lancet. 2000;356(9228):482–483.
129. Rubin GD, Walker PJ, Dake MD, et al. Three-dimensional spiral computed tomographic 163. Greenberg R, Resch T, Nyman U, et al. Endovascular repair of descending thoracic aortic
angiography: An alternative imaging modality for the abdominal aorta and its branches. aneurysms: An early experience with intermediate-term follow-up. J Vasc Surg. 2000;31
J Vasc Surg. 1993;18(4):656–664; discussion 665. (1 Pt 1):147–156.
130. Rubin GD, Dake MD, Napel SA, et al. Three-dimensional spiral CT angiography of the 164. Greenberg R, Khwaja J, Haulon S, et al. Aortic dissections: New perspectives and treat-
abdomen: Initial clinical experience. Radiology. 1993;186(1):147–152. ment paradigms. Eur J Vasc Endovasc Surg. 2003;26(6):579–586.
131. Zeman RK, Silverman PM, Berman PM, et al. Abdominal aortic aneurysms: Evaluation with 165. Fattori R, Caldarera I, Rapezzi C, et al. Primary endoleakage in endovascular treatment
variable-collimation helical CT and overlapping reconstruction. Radiology. 1994;193(2): of the thoracic aorta: Importance of intraoperative transesophageal echocardiography.
555–560. J Thorac Cardiovasc Surg. 2000;120(3):490–495.
132. Pate JW, Fabian TC, Walker W. Traumatic rupture of the aortic isthmus: An emergency? 166. Fattori R, Napoli G, Lovato L, et al. Descending thoracic aortic diseases: Stent-graft
World J Surg. 1995;19:119–126. repair. Radiology. 2003;229(1):176–183.
133. Mirvis SE, Shanmuganathan K. MR imaging of thoracic trauma. Magn Reson Imaging 167. Resch T, Koul B, Dias NV et al. Changes in aneurysm morphology and stent-graft config-
Clin N Am. 2000;8(1):91–104. uration after endovascular repair of aneurysms of the descending thoracic aorta. J Thorac
134. Fattori R, Celletti F, Bertaccini, et al. Delayed surgery of traumatic aortic rupture: Role of Cardiovasc Surg. 2001;122(1):47–52.
magnetic resonance imaging. Circulation. 1996;94:2865–2870. 168. White RA, Donayre CE, Walot I, et al. Endovascular exclusion of descending thoracic
135. Fattori R, Celletti F, Descovich B, et al. Evolution of post traumatic aneurysm in the subaortic aneurysms and chronic dissections: Initial clinical results with the AneuRx device.
acute phase: Magnetic resonance imaging follow-up as a support of the surgical timing. J Vasc Surg. 2001;33(5):927–934.
Eur J Cardiothorac Surg. 1998;13:582–587. 169. Demers P, Miller DC, Mitchell RS, et al. Stent-graft repair of penetrating atheroscle-
136. Mirvis SE, Kostrubiak I, Whitley NO, et al. Role of CT in excluding major arterial injury rotic ulcers in the descending thoracic aorta: Mid-term results. Ann Thorac Surg.
after blunt thoracic trauma. AJR Am J Roentgenol. 1987;149(3):601–605. 2004;77(1):81–86.
137. Mirvis SE, Shanmuganathan K, Miller BH, et al. Traumatic aortic injury: diagnosis with 170. Rozenblit A, Marin ML, Veith FJ, et al. Endovascular repair of abdominal aortic
contrast-enhanced thoracic CT—five-year experience at a major trauma center. Radiology. aneurysm: Value of postoperative follow-up with helical CT. AJR Am J Roentgenol.
1996;200(2):413–422. 1995;165(6):1473–1479.
LWBK1209-ch32_p503-544.indd 543 16/05/13 9:23 PM

171. Tummala S, Powell A. Imaging of endoleaks. Tech Vasc Interv Radiol. 2001;4(4): 196. Thiene G, Frescura C. Etiology and pathology of aortic arch malformations. In: Nienaber
208–212. CA, Fattori R, eds. Diagnosis and Treatment of Aortic Diseases. New York, NY: Kluwer
172. Veith FJ, Ohki T. Endovascular approaches to ruptured infrarenal aorto-iliac aneurysms. Academic Publishers; 1999:225–269.
J Cardiovasc Surg (Torino). 2002;43(3):369–378. 197. Bakker DA, Berger RM, Witsenburg M, et al. Vascular rings: A rare cause of common
173. Quint LE, Platt JF, Sonnad SS, et al. Aortic intimal tears: Detection with spiral computed respiratory symptoms. Acta Paediatr. 1999;88(9):947–952.
tomography. J Endovasc Ther. 2003;10(3):505–510. 198. Kersting-Sommerhoff BA, Sechtem UP, Fisher MR, et al. MR imaging of congenital anom-
174. Nienaber CA, Fattori R, Mehta RH, et al. International Registry of Acute Aortic Dissection. alies of the aortic arch. AJR Am J Roentgenol. 1987;149(1):9–13.
Gender-related differences in acute aortic dissection. Circulation. 2004;109(24): 199. Delabrousse E, Kastler B, Bernard Y, et al. MR diagnosis of a congenital abnormality
3014–3021. of the thoracic aorta with an aneurysm of the right subclavian artery presenting as a
175. White GH, Yu W, May J. Endoleak: A proposed new terminology to describe incomplete Horner’s syndrome in an adult. Eur Radiol. 2000;10(4):650–652.
aneurysm exclusion by an endoluminal graft. J Endovasc Surg. 1996;3(1):124–125. 200. Carpenter JP, Holland GA, Golden MA, et al. Magnetic resonance angiography of the
176. White GH, Yu W, May J, et al. Endoleak as a complication of endoluminal grafting of aortic arch. J Vasc Surg. 1997; 25: 145–151.
abdominal aortic aneurysms: Classification, incidence, diagnosis, and management. 201. Julsrud PR, Breen JF, Felmlee JP, et al. Coarctation of the aorta: Collateral flow assessment
J Endovasc Surg. 1997;4(2):152–168. with phase-contrast MR angiography. AJR. 1997;169:1735–1742.
177. White GH, May J, Petrasek P, et al. Endotension: An explanation for continued AAA 202. Paddon AJ, Nicholson AA, Ettles DF, et al. Long-term follow-up of percutaneous balloon
growth after successful endoluminal repair. J Endovasc Surg. 1999;6(4):308–315. angioplasty in adult aortic coarctation. Cardiovasc Intervent Radiol. 2000;23(5):364–367.
178. Wain RA, Marin ML, Ohki T, et al. Endoleaks after endovascular graft treatment of aor- 203. Therrien J, Thorne SA, Wright A, et al. Repaired coarctation: A “cost-effective” approach
tic aneurysms: Classification, risk factors, and outcome. J Vasc Surg. 1998;27(1):69–78; to identify complications in adults. J Am Coll Cardiol. 2000;35(4):997–1002.
discussion 78–80. 204. Bogaert J, Kuzo R, Dymorkovski S, et al. Follow-up of patients with previous treatment
179. Deaton DH, Makaroun MS, Fairman RM. Endoleak: Predictive value for aneurysm for coarctation of the thoracic aorta: Comparison between contrast-enhanced MR angi-
growth at 3 years. Ann Vasc Surg. 2002;16(1):37–42. ography and fast spin-echo MR imaging. Eur Radiol. 2000;10:1047–1054.
180. Veith FJ, Baum RA, Ohki T, et al. Nature and significance of endoleaks and endotension: 205. Baur LH, Vliegen HW, van der Wall EE, et al. Imaging of an aneurysm of the sinus of
Summary of opinions expressed at an international conference. J Vasc Surg. 2002;35(5): Valsalva with transesophageal echocardiography, contrast angiography and MRI. Int
1029–1035. J Card Imaging. 2000;16(1):35–41.
181. Golzarian J, Dussaussois L, Abada HT, et al. Helical CT of aorta after endoluminal stent- 206. Gilkeson RC, Ciancibello L, Zahka K. Pictorial essay. Multidetector CT evaluation of
graft therapy: Value of biphasic acquisition. AJR Am J Roentgenol. 1998;171(2):329–331. congenital heart disease in pediatric and adult patients. AJR Am J Roentgenol. 2003;
182. Matsumura JS, Pearce WH, McCarthy WJ, et al. Reduction in aortic aneurysm size: 180(4):973–980.
Early results after endovascular graft placement. EVT Investigators. J Vasc Surg. 1997; 207. Haramati LB, Glickstein JS, Issenberg HJ, et al. MR imaging and CT of vascular anom-
25(1):113–123. alies and connections in patients with congenital heart disease: Significance in surgical
183. Resch T, Ivancev K, Brunkwall J, et al. Midterm changes in aortic aneurysm morphology planning. Radiographics. 2002;122(2):337–347; discussion 348–349.
after endovascular repair. J Endovasc Ther. 2000;7(4):279–285. 208. Fayad ZA, Fallon JT, Shinnar M, et al. Noninvasive in vivo high resolution magnetic
184. Singh-Ranger R, McArthur T, Corte MD, et al. The abdominal aortic aneurysm sac after resonance imaging of atherosclerotic plaque in genetically engineered mice. Circulation.
endoluminal exclusion: A medium-term morphologic follow-up based on volumetric tech- 1998;98:1541–1547.
nology. J Vasc Surg. 2000;31(3):490–500. 209. Fayad ZA, Nahar T, Fallon JT, et al. In vivo magnetic resonance evaluation of atheroscle-
185. Pollock JG, Travis SJ, Whitaker SC, et al. Endovascular AAA repair: Classification of rotic plaques in the human thoracic aorta: A comparison with transesophageal echocar-
aneurysm sac volumetric change using spiral computed tomographic angiography. diography. Circulation. 2000;101(21):2503–2509.
J Endovasc Ther. 2002;9(2):185–193. 210. Worthley SG, Helft G, Fuster V, et al. Serial in vivo MRI documents arterial remodeling in
186. Choe YH, Kim DK, Koh EM, et al. Takayasu arteritis: Diagnosis with MR imaging experimental atherosclerosis. Circulation. 2000;101(6):586–589.
and MR angiography in acute and chronic active stages. J Magn Reson Imaging. 1999; 211. Toussaint J, Southern JF, Fuster V, et al. 13 C-NMR spectroscopy of human atheroscle-
10(5):751–757. rotic lesions. Relation between fatty acid saturation, cholesteryl ester content and luminal
187. Berkmen T. MR angiography of aneurysms in Beçhet disease: A report of four cases. obstruction. Circulation. 1994;14:1951–1957.
J Comput Assist Tomogr. 1998;202–206. 212. Zimmermann-Paul GG, Quick HH, Vogt P, et al. High resolution intravascular magnetic
188. Pripstein S, Cavoto FV, Gerritsen RW. Spontaneous mycotic aneurysm of the abdominal resonance imaging. Monitoring of plaque formation in heritable hyperlipidemic rabbits.
aorta. J Comput Assist Tomogr. 1979;3(5):681–683. Circulation. 1999;99:1054–1061.
189. Atlas SW, Vogelzang RL, Bressler EL, et al. CT diagnosis of a mycotic aneurysm of the 213. Fayad ZA, Sirol M, Nikolaou K, et al. Magnetic resonance imaging and computed tomog-
thoracoabdominal aorta. J Comput Assist Tomogr. 1984;8(6):1211–1212. raphy in assessment of atherosclerotic plaque. Curr Atheroscler Rep. 2004;6(3):232–242.
190. Vogelzang RL, Sohaey R. Infected aortic aneurysms: CT appearance. J Comput Assist 214. Viles-Gonzalez JF, Poon M, Sanz J, et al. In vivo 16-slice, multidetector-row computed
Tomogr. 1988;12(1):109–112. tomography for the assessment of experimental atherosclerosis: Comparison with mag-
191. Gonda RL Jr, Gutierrez OH, Azodo MV. Mycotic aneurysms of the aorta: Radiologic netic resonance imaging and histopathology. Circulation. 2004;110(11):1467–1472.
features. Radiology. 1988;168(2):343–346. 215. Schroeder S, Kuettner A, Wojak T, et al. Non-invasive evaluation of atherosclerosis with
192. Blair RH, Resnik MD, Polga JP. CT appearance of mycotic abdominal aortic aneurysms. contrast enhanced 16 slice spiral computed tomography: Results of ex vivo investigations.
J Comput Assist Tomogr. 1989;13(1):101–104. Heart. 2004;90(12):1471–1475.
193. Gomes MN, Choyke PL. Infected aortic aneurysms: CT diagnosis. J Cardiovasc Surg 216. de Feyter P, Mollet N, Nieman K, et al. Noninvasive visualisation of coronary atheroscle-
(Torino). 1992;33(6):684–689. rosis with multislice computed tomography. Cardiovasc Radiat Med. 2004;5(1):49–56.
194. Chan P, Lan CK, Wan YL. Salmonella cholerasuis bacteremia and mycotic aneurysm of 217. Achenbach S, Daniel WG. Imaging of coronary atherosclerosis using computed tomogra-
abdominal aorta—report of five cases. Changgeng Yi Xue Za Zhi. 1989;12(2):115–120. phy: Current status and future directions. Curr Atheroscler Rep. 2004;6(3):213–218.
195. Mantello MT, Panaccione JL, Moriarty PE, et al. Impending rupture of nonaneurysmal 218. Becker CR. Noninvasive assessment of coronary atherosclerosis by multidetector-row
bacterial aortitis: CT diagnosis. J Comput Assist Tomogr. 1990;14(6):950–953. computed tomography. Expert Rev Cardiovasc Ther. 2004;2(5):721–727.
LWBK1209-ch32_p503-544.indd 544 16/05/13 9:23 PM

Chapter
Michael D. Hope
Petter Dyverfeldt 33
Thoracic Aorta Disease:
Flow Evaluation by MR
■■ Background common than their thoracic counterparts. They are found in

Introduction to Aortic Imaging up to 5% of men over 65 years old (2). In addition to male
Current Clinical Flow Imaging Applications sex and age, other key risk factors are smoking, hyperten-
Emerging Flow Applications sion, and atherosclerosis. Unlike AAAs, thoracic aneurysms
are heterogeneous in etiology, patient demographics, and
■■ Multidimensional Flow Imaging regional progression of diseases (3). Like AAAs, however,
Scan Time and Acceleration basic vessel dimensions are the primary imaging measure-
Visualization ment used clinically to risk-stratify patients.
Data Quality Progression of aortic disease is partially understood.
General guidelines exist for managing patients in various
■■ Quantitative Hemodynamic Biomarkers
clinical contexts. The commonality is surveillance imaging
Pulse-wave velocity with elective aortic surgery at a threshold aortic diameter or
Turbulence interval growth rate (3,4). Intervention, in other words, is
Pressure mapping warranted when the risk of not intervening supersedes pro-
Wall Shear Stress cedural risk. For example, elective surgery should be consid-
Flow Eccentricity ered for ascending aortic diameters greater than 5.5 cm, or at
■■ Summary smaller dimensions if a growth rate over 0.5 cm/y is found.
However, for both the thoracic and abdominal aortae, a sig-
nificant percentage of morbidity and mortality is seen with
aortic diameters smaller than intervention thresholds (5–9).
Background The complexity of aortic disease is not fully revealed by basic
anatomic considerations alone.
Blood flow assessment with magnetic resonance imaging Current aortic surveillance imaging emphasizes anatomy
(MRI) is rapidly evolving. Advances in scanner hardware, at the expense of other important and unique considerations
pulse sequence design, and postprocessing of data have for the progression of aortic disease. Two such consider-
allowed the dynamic visualization of flowing blood through ations are (1) the structural status of the aortic wall and
large and complex cardiovascular territories (Fig. 33.1) and (2) the effect of hemodynamics. While not the focus of this
the quantification of many hemodynamic parameters that chapter, many advances have been made with molecular im-
influence vascular homeostasis. The aorta has been the focus aging that allow direct imaging of inflammation within the
of many of these developments. It is the largest artery in aortic wall, as well as specific extracellular matrix proteins
the body and subject to extreme hemodynamic forces as it and pathologic processes (10,11). At the same time, recent
receives and directs the systolic impulse of the left ventricle developments with MRI allow assessment of dynamic blood
with each heartbeat. flow and a range of associated hemodynamic parameters
that may contribute to understanding and predicting disease
progression. Hemodynamics has long been implicated in the
Introduction to Aortic Imaging
progression of aortic disease. For example, low wall-shear
Aortic aneurysmal disease is the 18th most common cause stress has been linked to atherosclerosis, aortic stenosis to
of death, accounting for at least 13,843 deaths annually in aneurysmal disease of the ascending aorta (i.e., postste-
the United States (1). An apparent increase in patients with notic dilatation), and focal aortic wall stress to common
aneurysms, in recent years, may largely reflect increased sites of aortic dissection (5,12,13); but until recent advances
detection with the more frequent use of cross-sectional in MRI blood flow imaging, routine evaluation of these
imaging. Abdominal aortic aneurysms (AAAs) are more hemodynamic parameters was not possible. The focus of
545
LWBK1209-ch33_p545-557.indd 545 16/05/13 9:14 PM

means that it “provides clinically relevant information and

is frequently useful, may be used as first-line imaging tech-
nique, and is usually supported by substantial literature”
(18). MRI has the additional advantage over echocardiogra-
phy of superior assessment of related left ventricular status,
with accurate and reproducible calculation of ventricular
size, function, and mass.
For assessment of aortic regurgitation, an imaging plane
perpendicular to the aorta is prescribed, typically 2 to 3 cm
above the aortic valve at the vertical, tubular portion of the
ascending aorta. The aortic lumen is segmented at each time
point for calculation of flow volumes. The regurgitant frac-
tion is the ratio of retrograde to antegrade flow. If there is
concomitant aortic stenosis, a higher velocity encoding range
(VENC) is needed to avoid aliasing. In some cases, a dual ve-
locity window may be used to obtain optimum flow sensitiv-
ity with a high systolic and relatively lower diastolic VENC.
The degree of aortic valve stenosis is estimated by using
the modified Bernoulli equation ΔP = 4ν 2, where ΔP is the
peak pressure gradient in millimeters of mercury and ν is the
peak blood flow velocity in meters per second. This tech-
Figure 33.1. Systolic blood flow in the great vessels of a normal nique is also used routinely for Doppler ultrasound. Good
volunteer as visualized by 4D flow. The 3D streamlines used are accuracy compared to Doppler echocardiography for aortic
aligned with the local velocity vector field at a given moment in time stenosis has been demonstrated (19). Unlike flow quantifica-
and provide a 3D perspective of instantaneous flow. Red indicates tion, images may be prescribed in a parallel or perpendicular
aorta, blue indicates pulmonary artery, and maroon indicates pulmo- orientation to the aorta to capture the peak systolic velocities
nary veins. The data for this image were collected in a single acquisi-
within the vena contracta downstream of the stenotic valve. A
tion of approximately 15 minutes.
theoretical advantage of the volumetric MRI flow sequences
described in the next section is a better three-dimensional
(3D) localization of the vena contracta. A disadvantage of
this chapter is advanced MRI evaluation of hemodynamics MRI, however, for pressure gradient estimations compared
and its potential clinical impact. to echocardiography is its lower temporal resolution.
Current Clinical Flow Imaging Aortic Coarctation

Applications
MRI has become the imaging modality of choice for evalu-
MRI blood flow imaging is currently used in select clinical ation of aortic coarctation (20–22). Coarctation refers to
setting. Two-dimensional (2D) phase-contrast (PC) MRI a narrowing of the distal aortic arch in the region of the
is the most common flow-sensitive cardiac sequence used ligamentum arteriosum that restricts forward flow. MRI is
(discussed in Chapter 5). A 2D plane is prescribed in the useful for assessment of both anatomy and hemodynamics.
appropriate orientation for evaluation of a given vessel or Aortic dimensions are measured with high-resolution, 3D
heart valve, and time-resolved PC data are acquired in a magnetic resonance angiography (MRA) sequences. Blood
single direction. The technique allows for quantification of flow imaging can be used in at least 3 ways for assessment of
cardiac output, valve regurgitation, severity of vascular and relevant hemodynamics: (1) Pressure gradient estimation can
valvular stenosis, pulmonary to systemic flow ratio (i.e., QP/ be performed using velocity data and the modified Bernoulli
QS ratio, which reflects shunting of blood), differential lung equation as discussed above; (2) quantification of collateral
perfusion, and coronary flow reserve. We will focus on aor- flow; and (3) evaluation of flow-versus-time profiles in the
tic valve disease and aortic coarctation. descending aorta.
Collateral flow occurs with the altered pressure dynamics
Aortic Valve Disease seen in aortic coarctation. Blood bypasses the region of nar-
rowing through lower pressure intercostal vessels to reach
PC MRI allows precise calculation of aortic regurgitation the descending thoracic aorta and beyond. The presence
and reasonable estimation of the degree of valve stenosis of collateral flow indicates a hemodynamically significant
(14–17). Echocardiography is the initial imaging modality lesion that may require intervention. MRI evaluation is per-
of choice for assessment of cardiac valves. It is both cheaper formed with perpendicular planes to the aorta just distal to
and faster than MRI, but MRI provides quantification of the coarctation and at the diaphragm. Normal blood flow
valve regurgitation whereas echocardiography does not. will drop by approximately 7% over this interval (23). With
Instead, echocardiography qualitatively estimates regur- a hemodynamically significant coarctation, however, flow
gitation based on apparent size of flow jets, which can be will increase rather than decrease over the interval through
affected by imaging parameters and orientation. The use- collateral pathways. The percentage increase in flow gives a
fulness of MRI for valve disease is deemed Class 1, which quantitative measure of the degree of collateralization (23–25).
LWBK1209-ch33_p545-557.indd 546 16/05/13 9:14 PM

Chapter 33 ■ Thoracic Aorta Disease: Flow Evaluation by MR 547
When performing this analysis, images must be carefully chapter, we will discuss how current research is approach-
reviewed for velocity aliasing. The presence of aliasing can ing this hurdle, and showcase the clinical potential of MRI
simulate a coarctation. Blood flow at the proximal plane hemodynamic imaging. Key features of the technique are the
will be underestimated and consequently, flow at the distal focus of the next section, and then attention will shift to the
plane may be incorrectly interpreted as increased. various hemodynamic parameters that can be assessed with
Flow profile analysis in the descending aorta is another these datasets. Evaluation of these hemodynamic parameters
straightforward imaging means of identifying the adverse may represent the true clinical value of the technique. The
hemodynamic consequences of an aortic coarctation (21). Normal ultimate goal is identification of risk with MRI well before
arterial flow exhibits a rapid systolic upstroke followed by a life-threatening aortic complications occur.
swift return to baseline (typically within 300 milliseconds).
With the obstruction of flow and collateralization seen in aor-
tic coarctation, both the upstroke and return to baseline are Multidimensional Flow Imaging
delayed. Significant flow persistence into diastole is found.
Evaluating these abnormal features of the aortic flow profile is Dynamic blood flow imaging with MRI most commonly
a simple, fast, and reliable means of identifying a hemodynam- employs a PC pulse sequence. Please refer to Chapter 5 for
ically significant coarctation. The analysis can be performed a more thorough description of the sequence and its use. The
at the diaphragm where flow turbulence, aliasing, and stent- terminology used to describe volumetric MRI blood flow
related artifacts seen more proximally are not present (26). imaging can be cumbersome. In this chapter, we will focus on
3D, time-resolved (cine), three-directional PC-MRI. For sim-
plicity, we will call it four-dimensional (4D) flow, where “4D”
Emerging Flow Applications
refers to 3D plus time, and “flow” to the comprehensive flow
Multidimensional MRI blood flow imaging has been increas- field information (mean velocity field and intensity of fluc-
ingly studied (Fig. 33.2) and proposed as a tool for the eval- tuating velocity field) obtained by three-directional PC-MRI
uation of many cardiovascular disease processes including (Fig. 33.3). Four-dimensional flow affords an appealing, intu-
atherosclerosis, aneurysm and, dissection. The technique itive visualization of blood flow and unique characterization
has become more widely available and easier to use. Time- of vessel hemodynamics not possible with less comprehensive,
saving measures like parallel imaging have been employed 2D sequences, or with any other imaging modality.
to reduce scan time to 15 minutes or less, making the tech- In conventional 2D cine PC-MRI, the through-plane
nique clinically feasible. Advantages over other modalities velocity component of blood flow is measured by apply-
and simpler 2D MRI sequences have been enumerated. The ing motion-sensitizing bipolar gradients in this direction.
technique substantially improves upon echocardiography by Three-directional VENC can be achieved by applying
capturing volumetric velocity data rather than single-point, motion-sensitizing gradients along multiple axes (27–31). As
single-direction velocity data. Compared to 2D imaging, in through-plane PC-MRI, the VENC determines the maxi-
benefits include complete temporal and spatial coverage of mum measurable velocity. However, different VENCs can
a cardiovascular territory, continuous breathing, no require- be used for the different axes, and, at the cost of increased
ment for prospective placement of 2D planes, and many scan time, datasets with multiple VENCs can be acquired.
visualization and quantification options for velocity data Work in the early 1990s with both nontime-resolved 3D PC-
that are not otherwise available. MRI (32) and time-resolved PC-MRI velocity field imaging
The next hurdle in the evolution of MRI aortic blood flow based on 2D planes stacked to achieve 3D datasets showed
imaging is the demonstration of clear clinical utility. In this the utility of this type of imaging for uncovering complex,
4D PC-MRI publications
VZ, σZ
VY, σY
VX, σX
V = average voxel velocity

σ = turbulence intensity
05 06 07 08 09 10 11 Figure 33.3. Three-dimensional cine phase-contrast MRI or 4D flow

20 20 20 20 20 20 20
permits measurements of volumetric three-directional velocity fields that
Figure 33.2. Four-dimensional flow-related publications have been are time-resolved over the cardiac cycle. In addition to three-directional
growing exponentially over the last few years. In 2011, there were over velocities, the technique also permits estimation of the standard devia-
tenfold the number of publications compared to 2005. tion of velocity within each image volume element (voxel).
LWBK1209-ch33_p545-557.indd 547 16/05/13 9:14 PM

secondary aortic blood flow characteristics such as helices that can severely degrade the appearance and usefulness of
and vortices (33). Subsequently, true 3D cine PC-MRI or the image when data are not fully sampled. Nevertheless,
4D flow acquisitions were developed (31,34,35) and applied with the use of multichannel coils, scan acceleration can be
to address previously unknown aspects of cardiovascular effectively employed by undersampling k-space in combi-
hemodynamics (36–38). nation with designated reconstruction algorithms. Parallel
imaging methods such as SENSE (SENSitivity Encoding)
(40) and GRAPPA (GeneRalized Autocalibrating Partially
Scan Time and Acceleration
Parallel Acquisitions) (41) reduce scan time by skipping
Four-dimensional flow has been limited in the past by long phase-encoding lines during the acquisition, and then restore
scan times. A large number of k-space lines are needed for the missing data in the image or k-space domain. Parallel
three-directional VENC with sufficient spatiotemporal resolu- imaging reduction factors of 2 are routinely used for 4D
tion and coverage. Full datasets can take well over 30 minutes flow. Higher reduction factors are being explored with new,
to acquire, which is too long for routine clinical medicine. We high channel count receiver coil arrays.
will discuss the various approaches that are used to reduce In recent years, several new methods for acceleration of
scan time so as to make 4D flow clinically feasible. Currently, dynamic MRI have been introduced (42–45). Some allow
typical aortic 4D flow protocols require about 10 to 20 min- high acceleration factors for applications other than PC, but
utes of scan time (Table 33.1). PC-MRI velocity mapping, unlike many other MRI meth-
One commonly used time-saving approach is k-space seg- ods, relies on accurate reconstruction of the phase of the
mentation, which enables a trade-off between scan time and MR signal. This can be a liability when using high accel-
temporal resolution via the acquisition of multiple k-space lines eration factors and advanced image reconstruction methods.
per cardiac cycle (39). In addition, substantial advancements Temporal smoothing introduced at higher k-t acceleration
have been made to speed up the 4D flow acquisition through factors, for example, can lead to inaccuracy in flow mea-
k-space undersampling. While undersampling normally causes surements and particle trace visualizations (46–49). Notably,
image artifacts, acquisition, and reconstruction strategies to al- imaging acceleration based on the theory of compressed
leviate, these artifacts exist. Both rectilinear Cartesian and non- sensing has rapidly gained popularity in the past few years
Cartesian approaches to k-space sampling can be used. (44). Compressed sensing goes a step beyond conventional
The Cartesian trajectory is more conventional and widely acceleration methods by exploiting the fact that images have
employed. It is relatively robust to imperfections such as off- inherent sparsity. Its application in 4D flow is underway and
resonance and eddy currents, but causes fold-over artifacts preliminary data show promising results (Fig. 33.4).
Table 33.1 Example of Parameter Settings for a Cartesian 4D Flow Aorta Protocol
Parameter Value Explanation
Velocity encoding range 150–300 VENC defines the maximum measurable velocity. Velocities that exceed the VENC are aliased.
(VENC) (cm/s) For some applications, the VENC can intentionally be set below the maximum velocity.
Repetition time (TR) (ms) 4.2–5 The time it takes to acquire one line of k-space. It is often set to the shortest possible value
to increase temporal resolution. Its value depends on the VENC. Lower VENCs lead to
longer TRs.
Echo time (TE) (ms) 2.5–3 The time from the beginning of the TR until the center of the echo. It is often set to the
shortest possible value to minimize various artifacts.
Receiver bandwidth 400–600 The range of frequencies accepted by the receiver to sample the MR signal. Shorter bandwidth
(Hz/pixel) not only increases the signal-to-noise ratio (SNR) but also increases chemical shift.
Orientation Sagittal– The orientation of the imaging volume.
oblique
3D field-of-view 320 × 320 × 70 Volumetric coverage in the readout-, phase-, and slice-encoding directions. FOV is adjusted to
(FOV) (mm3) encompass the aorta.
3D matrix size, 128 × 128 × 30 The number of frequency-, phase-, and slice-encoding steps, respectively.
Nx ¥ Ny ¥ Nz
K-space segmentation 2–3 The number of k-space lines acquired per cardiac cycle. Higher segmentation factors decrease
(Nk) scan time at the cost of temporal resolution.
Temporal resolution (ms) 35–45 The temporal resolution equals 4∙TR/Nk, where 4 is the number of encodings needed for three-
directional velocity data (phase reference + three-directional velocity encoding).
Spatial resolution (mm3) 2.5 × 2.5 × 2.5 The size of the image volume elements (voxels).
Respiratory efficiency, 60–80 The respiratory gating efficiency is highly dependent on the subject’s breathing pattern.
Reff (%)
Parallel imaging net 1.8–2 The parallel imaging undersampling rate.
acceleration factor, Rnet
Total scan time for a 11–24 (Ny × Nz) / (Nk × Rnet × Reff × HR) (min)
heart rate (HR) of
75 bpm
LWBK1209-ch33_p545-557.indd 548 16/05/13 9:14 PM

Fully sampled R=3 but in 50% to 70% less time (53). In the near future, further
scan time reductions can be expected from combinations of
acceleration techniques and non-Cartesian k-space trajecto-
ries, as well as from higher acceleration factors made fea-
sible by high channel count receiver coil arrays.
Another time-consuming aspect of aortic 4D flow is respira-
tory motion compensation. This is most commonly performed
using navigator gating. Other approaches include bellows and
self-gating. In navigator-gated scans, one-dimensional (1D)
images of the lung–liver interface are acquired once per car-
diac cycle. In this way, diaphragm motion can be monitored
during scanning and used to gate data acquisition with
respect to the respiratory cycle. Conventionally, data are
acquired during end-expiration. Depending on the subject’s
breathing pattern, navigator gating can significantly add to
an already long scan time. Several options for increasing
the gating efficiency are available. Most sequences use an
acceptance window that is dynamically updated based on
the subject’s breathing pattern. This can be combined with
techniques that take into account the fact that motion occur-
ring during the acquisition of the outer parts of k-space will
degrade image quality less than the same motion during ac-
quisition of central k-space. One such method is respiratory-
ordered phase encoding, which reorders the acquisition of
k-space based on the respiratory position (54). This method
Flow rate
has been suggested to reduce motion artifacts in 4D flow

and thereby allow the use of a larger gating window (55).
Another option is to use different gating windows for dif-
ferent parts of k-space, that is, narrower window for central
k-space (k-space inspired navigator gating [KING], Philips
Healthcare, Best, the Netherlands). An alternative to naviga-
tors is self-gating, which broadly refers to a sequence that
acquires data for motion compensation without the use of
Time navigators. A self-gated 4D flow sequence that repeatedly
Figure 33.4. Example of the application of compressed sensing acquires 1D projections of the entire imaging volume to de-
in 4D flow. Streamline visualization of 4D flow data obtained from tect motion has been implemented (56). To maximize sensi-
fully sampled k-space data (left) and threefold undersampled k-space tivity for respiratory motion, only the coil element closest to
data reconstructed using compressed sensing (right). Bottom: Flow the diaphragm is used and the readout direction is oriented
waveforms in the descending aorta measured in the fully sampled data in the feet-to-head direction, similar to conventional respira-
(solid line) and threefold undersampled data (dotted line). Note that tory navigators.
the image quality is well maintained in the threefold undersampled
data reconstructed with compressed sensing, in spite of three times
shorter scan time. (Courtesy of Jing Liu, UCSF.) Visualization
Multidimensional flow visualization is an integral part
of 4D flow (Fig. 33.5). Informative visualizations of the
Non-Cartesian k-space trajectories are an alternative ap- dynamic velocity fields are often generated using vector plots
proach that limits the severity of undersampling artifacts. or particle trace techniques such as streamlines and path-
For example, a radial acquisition can be employed where lines (32,36,57). Streamlines are the tangent to the velocity
k-space is traversed along radial lines through its center, field at a given moment in time. They allow visualization
rather than with rectilinear lines (50). Undersampling in ra- of instantaneous velocity field characteristics, and can nicely
dial acquisitions results in streak artifacts, but this artifact is depict secondary flow patterns such as helices and vortices.
typically more forgiving, at least visually, than fold-over arti- However, streamlines should be used with care. They do
facts. When streaks are acceptable, the radial pattern is faster not actually represent blood flow, and should be used as an
than a Cartesian approach without sacrificing spatial resolu- adjunct to other visualization techniques. Pathlines calcu-
tion. In-plane and 3D radial k-space sampling have been suc- late the trajectories of virtual blood particles over time by
cessfully applied to reduce the scan time of 4D flow (51,52). numerical integration of the 4D velocity field. They are the
Another approach is spiral k-space trajectories using oscillat- most all-inclusive of the blood flow visualization approaches
ing gradient waveforms. This results in reduced scan time by as they take into account all components of 4D flow data
using longer readouts and sampling at a higher bandwidth. (i.e., 3D velocity and time). Other approaches that have
Spiral 4D flow can generate data with similar quality and ac- been used to visualize MR velocity data include isosurface
curacy as that obtained with rectilinear Cartesian sampling, and volume rending of specific flow features (58–60).
LWBK1209-ch33_p545-557.indd 549 16/05/13 9:14 PM

A B C
Figure 33.5. A: Reference magnitude

image of the right atrium (RA), superior
vena cava (SVC), inferior vena cava (IVC),
and pulmonary vein (PV). B: Velocity vector
field overlaid in mid-diastole. C: Close-up
of vector plot; D: Streamlines of same data
nicely visualize the vortical pattern. E: A
single image from a dynamic pathline visu-
alization where flow is color-coded by vessel
D E
of origin (SVC in yellow, IVC in red).
Four-dimensional flow visualization has proved to be valu- their application is not always straightforward. For exam-
able in generating novel flow-based hypotheses. Examples of ple, the optimal method to correct for background phase
this include the detection of unexpected, potential transarte- offsets may depend on the vendor, sequence, and applica-
rial pathways for emboli (61); apparent inefficiencies in the tion (70). Moreover, phase unwrapping algorithms are only
transit of blood flow through the left ventricle (62); and the effective to a certain degree of VENC underestimation.
possible negative impact of flow eccentricity on the aortic Consequently, systematic validation and quality control are
wall (63). Standardized approaches for flow visualization important in 4D flow.
permit screening of large numbers of datasets and can help As different levels of accuracy are required for different
trained reviewers identify subtle differences between patients. indications, validation may need to be targeted for the par-
Regions of interest can be quickly appreciated and further ticular application at hand. One approach to validation is to
evaluation of quantitative flow parameters can be performed. perform systematic comparisons with data obtained using
In addition, standardized visualization strategies are valuable gold standard methods. As the availability of noninvasive
in identifying for data quality screening (64). methods for acquiring comprehensive hemodynamic data is
limited, these validations are typically performed in silico, in
vitro, or in animal models (71–75).
Data Quality
Even well-validated imaging protocols can result in variable
The accuracy of both the visualized flow patterns and the image quality. A range of factors may be responsible including
derived hemodynamic parameters depends on the fidelity of different orientations of the imaging volume, patient move-
the underlying data. Sources of error that need to be taken ment during scanning, and software updates. The acquired 4D
into account include concomitant gradient field effects (65), velocity fields, however, can be used as an “internal” quality
nonlinear gradients (66), phase wraps (67), and background control. For example, when both the aorta and the pulmo-
phase offsets primarily attributed to eddy currents (68,69). nary artery are covered by the imaging volume, the ratio of
While approaches to correct for these sources of error exist, flow to the pulmonary (QP) and systemic (QS) circulation can
LWBK1209-ch33_p545-557.indd 550 16/05/13 9:14 PM

be used as a quality indicator (the flows should be equal, as-

suming there are no shunt lesions present). Similar approaches
can be used to confirm consistency between a feeding artery
and downstream branches. Pathlines are most commonly
used for visualization, but can also play an important role
for validation purposes. Pathlines accumulate error as they
are integrated over time and are therefore sensitive qual-
ity indicators. Ideally, a pathline should not leave the ves-
sel lumen. Validation methods can be achieved by applying
the conservation-of-mass principle to pathlines that describe
flow in and out of a specific region (64).
Quantitative Hemodynamic
Biomarkers
Visualization of complex cardiovascular flows is just one
aspect of 4D flow. The comprehensive data can also be
used to derive a wide range of quantitative hemodynamic
parameters, or biomarkers. The quantitative parameter Time
that is used most frequently in the clinical setting is flow.
As discussed above, it is typically measured using 2D cine
PC-MRI with through-plane VENC. A unique advantage Figure 33.6. Pulse-wave velocity (PWV) is a measure of vascular
of 4D flow is that flow can be retrospectively measured at stiffness that has been shown to be an important predictor of mortal-
ity. As depicted here, PWV is defined as the distance between two
any location. This can be particularly helpful for congenital
points divided by the time it takes the systolic pulse to travel the
heart disease where it can be challenging and time consum- distance.
ing to accurately study multiple vessels of interest with 2D
acquisitions (76). Good correlation between 4D flow and
conventional 2D data has been demonstrated for the mea-
surements needed for the estimation of collateral flow with improved robustness. The applicability of the cross-correlation
aortic coarctation (77). approach to aortic 4D flow data has recently been demon-
Four-dimensional flow permits applications well beyond strated (84). PWV calculated with PC-MRI using transit
the mapping and quantification of blood velocity and flow. time methods is accurate and reproducible (79,85).
In this section, we will discuss a selection of these parameters Aortic PWV is an important clinical parameter. Many
and their potential clinical significance. Many of these param- studies have shown that carotid–femoral PWV is an indepen-
eters can also be obtained from 2D cine PC-MRI or Fourier dent predictor of mortality in both the general population and
VENC MRI. Only with 4D flow, however, can all of these for important patient subgroups such as those with diabetes,
parameters be obtained retrospectively from a single dataset. hypertension, or renal disease (85–89). The generally accepted
mechanism for the harmful effects of the arterial stiffness
represented by increased PWV is that a premature return of
Pulse-Wave Velocity
reflected waves in late systole increases the central pulse pres-
The speed at which the systolic pulse propagates through sure and, in turn, increases the load on the left ventricle and
the cardiovascular tree, referred to as pulse-wave velocity myocardial oxygen demand (90). PWV correlates well with
(PWV), is a measure of vascular stiffness and vessel compli- aortic disease (91,92). It has been used to predict progressive
ance. PWV is simply defined as the distance traveled by the disease with Marfan syndrome (93) and assess treatment re-
pulse wave divided by the time it takes to travel that distance: sponse (94,95). An advantage of 4D flow over the commonly
PWV = distance/time (Fig. 33.6). Pressure catheter measure- used carotid–femoral PWV estimates is the ability to provide
ments are currently considered gold standard for determina- focused, regional assessment, which is more sensitive and pos-
tion of PWV, but are invasive and not commonly used. sibly clinically revealing than a global evaluation (93).
Estimation of central aortic PWV with PC-MRI is most
commonly done using so called transit time methods. PWV
Turbulence
is calculated from the temporal differences between specific
features of velocity-versus-time waveforms measured at two Turbulent flow is the antithesis of laminar flow. Whereas
or more locations in the vessel of interest (78–82). Flow–time laminar flow is comprised of orderly streaming of flow with
waveform features that have been used to track pulse-wave lower velocities toward the vessel wall, turbulent flow is
propagation include the foot, half maximum, peak, and characterized by apparent randomness in space and time.
second-order derivatives of the flow–time curve. Fielden et Many nonlaminar invivo flows do not meet all the criteria of
al. (83) extended this concept and introduced an approach fully developed turbulence, and are therefore often referred
that utilizes cross-correlation to estimate time shifts between to as “transitional” or “disturbed” flows. A distinct charac-
the waveforms. This approach has the advantage of tak- teristic of all of these nonlaminar flows, however, is the pres-
ing the complete flow waveform into account, resulting in ence of fluctuating velocities. The fluctuation of velocity, u´,
LWBK1209-ch33_p545-557.indd 551 16/05/13 9:14 PM

Figure 33.7. Left: A 3D isosurface rendering of a contrast-enhanced MR angiography in a patient with

aortic coarctation (arrow). Middle: The peak systolic velocity field visualized by color-coded 3D streamlines
shows the presence of a flow jet distal to the coarctation. Right: Visualization of flow turbulence intensity in
red/yellow (here, reported as turbulent kinetic energy, TKE) demonstrates elevated turbulence intensity in the
vicinity of the flow jet. Elevated turbulence intensity seen in the ascending aorta was caused by a subaortic
valve membrane. (From Dyverfeldt P, Kvitting JP, Sigfridsson A, et al. Assessment of fluctuating velocities in
disturbed cardiovascular blood flow: In vivo feasibility of generalized phase-contrast MRI. J Magn Reson
Imaging. 2008;28:655–663, with permission).
is defined as the deviation of the velocity, u, from its mean Turbulence intensity can be calculated from 4D flow data.
value U: u´ = U − u (m/s). The intensity of the velocity fluc- The approach has been used to visualize and quantify the
tuations can be measured in different directions, i, by their extent, degree, and timing of turbulence intensity at sites of val-
vular and vascular stenoses and valvular insufficiency (58,101)
standard deviation σ i = ui′ 2 (m/s), which is a common (Fig. 33.7). There are several potential applications for nonin-
measure of turbulence intensity. vasive turbulence intensity estimation. Turbulence is the major
Conventional PC-MRI velocity mapping measures the cause of pressure loss across a stenosis. Consequently, the
mean velocity, U, but is unable to resolve the fluctuating mapping of turbulence intensity could allow novel approaches
velocity u′. However, by exploiting the effects of flow velocity to assess the functional significance of a stenosis. More gener-
fluctuations on the amplitude of the complex-valued PC-MRI ally, exposure of blood constituents to abnormal stresses in
signal, turbulence intensity can be estimated from PC-MRI turbulent flow increased the risk for thrombus formation and
magnitude images (96–98): hemolysis (102,103). Another potential application relates to
the effects of fluctuating velocities on endothelial cells. In vitro
1
S − k2 σ 2 studies have indicated that such fluctuations significantly influ-
=e2 v
S0 ence endothelial turnover (104).
In this equation, S and S0 are PC-MRI magnitude images

Pressure Mapping
acquired with different motion sensitivity. Δkv = π/VENC
describes the net motion sensitivity. Turbulence intensity Estimation of pressure gradients is critical for evaluating the
mapping can be seen as a velocity analogue to diffusion MRI: severity of vascular and valvular stenoses. Four-dimensional
Both techniques characterize incoherent motion on a subvo- flow in combination with the modified Bernoulli equation
xel level by exploiting the fact that motion-induced signal can be used to estimate peak transvalvular pressure gradi-
attenuation is dependent on the first gradient moment. PC- ents. The multidimensional nature of 4D flow data, how-
MRI turbulence intensity mapping has been validated against ever, allows more advanced pressure evaluation. Assuming
laser Doppler (98) and particle image velocimetry (99), as nonturbulent flow, the pressure gradient field can be calcu-
well as computational fluid dynamics applied to both in vitro lated from the measured velocity field with the Navier–Stokes
(73) and in vivo (100) settings. equation. A straightforward application is to integrate pressure
LWBK1209-ch33_p545-557.indd 552 16/05/13 9:14 PM

gradients along arbitrary lines to estimate a pressure dif- Estimates of WSS using PC-MRI data underestimate
ference between different points in space (105,106). More actual WSS values (72,114,116). The degree of underes-
extensive information can be obtained by calculating rela- timation appears to be nonlinear and dependent on voxel
tive pressure fields in 3D domains and at several points in size (73,116,122). This implies that measures of the tem-
time (107–109); but this is associated with some challenges. poral variation of WSS, such as the oscillatory shear index
Choosing the correct computation methods and boundary (OSI), may not reflect the true temporal variation in WSS.
conditions is important to avoid systematic errors (110,111). Moreover, errors may be greater with nonisotropic data.
Since the current MR pressure field mapping technique ap- Nevertheless, evaluation of relative values with 4D flow can
plies only to nonturbulent flow, it cannot be applied to many reliably identify regions of abnormally high or low WSS.
cardiovascular disease processes including stenoses, where This allows characterization of gross abnormalities of WSS
marked turbulence is often present. MR pressure mapping, that may be useful clinically (120,121).
consequently, is primarily applied to intracardiac flow, but WSS is crucial for maintaining vascular homeostasis.
recent developments with PC-MRI allow the estimation of Alterations in WSS affect the endothelial cells that line the
turbulence intensity. This may enable the computation of pres- arterial lumen in predictable ways. Low and oscillatory
sure fields that take turbulence effects into account. The ap- WSS promote atherosclerosis through many well-studied
plication of MR pressure mapping could then be extended to pathways (12,123,124). Endothelial cells switch to an ath-
stenotic cardiovascular disease and potentially provide a non- erogenic phenotype that, among other things, recruits and
invasive measurement of the net pressure loss across a stenosis. activates monocytes, promotes vasoconstriction and plate-
let activation, and increases apoptosis and cellular turn-
over. As a consequence, atherosclerosis is found in regions
Wall Shear Stress
of disturbed, recirculating flow where there is low WSS
Wall shear stress (WSS) is the frictional force that flowing (<4 dyne/cm2), such as within the aortic arch and at vessel
blood exerts on the vessel wall. It is defined as the product bifurcations (125). Recent work demonstrates good correla-
of dynamic viscosity (m) and wall shear rate: tion between regions of low WSS, as estimated from 4D flow
datasets, and atherosclerotic plaque throughout the aorta
∂v (119). Low WSS has also been linked to the growth of intra-
WSS = µ
∂r r = 0 cranial aneurysms (126). For the aorta, however, WSS evalu-
ation of aneurysms associated with atherosclerosis (typically
where v is the velocity parallel to the wall and r is the radial in the infrarenal abdominal aorta) is confounded by the sub-
distance from the wall. By providing co-registered morpho- stantial intraluminal thrombus that is often present (127).
logic and velocity images, 4D flow is an appealing option for High WSS has been less extensively researched, but is
WSS estimation (Fig. 33.8). However, the lumen–wall con- linked to the pathologic remodeling of arteries. One study
trast can be poor and velocity measurements near the vessel of arteriovenous fistulas showed that radial arteries with in-
wall are limited by partial volume effects. These limitations creased flow and WSS dilated to a size at which WSS nor-
have been addressed by combining lumen–wall segmenta- malized (128). High WSS can lead to endothelial weaken-
tions with line or plane fitting of velocities in multiple voxels ing and injury. This was originally demonstrated in a study
(112–116). Most of these WSS methods are relatively easy that subjected dog aortae to markedly elevated WSS (129),
to use and are widely applied (117–121). and has more recently been incorporated into a low, then
high, WSS theory for why some atherosclerotic plaques rup-
ture (130). The first insult is low WSS, which drives plaque
growth and eventually leads to luminal narrowing. This
Wall Wall narrowing results in high WSS at the upstream margin of
the lesion, which weakens the vessel and predisposes to rup-
ture through a variety of mechanisms including extracellu-
lar matrix degradation and smooth muscle cell apoptosis.
V Gradients of WSS may be more important than absolute
WSS values in determining clinical significance. A recent cell
culture study suggests that positive WSS gradients are the
principal culprit for driving aneurysm growth (131).
High WSS has been investigated with 4D flow as a risk factor
for progressive thoracic aortic disease. Focally elevated WSS has
been correlated with characteristic sites of postoperative aortic
aneurysms and aneurysm rupture (132,133). Valve-related aor-
tic disease is another topic of active investigation. Aortic valve
disease is common, especially in the elderly, and is associated
with ascending aortic aneurysm. The mechanism of this post-
Figure 33.8. Wall shear stress (WSS) can be estimated from the stenotic dilatation is presumed to be flow related. Before 4D
near wall velocity gradients captured by 4D flow. This schematic figure flow, valve-related abnormalities of flow and hemodynamics
illustrates a 1D velocity profile for moderately eccentric flow, result- were not well characterized. Recent studies with the technique,
ing in relatively steep velocity gradient (red line) and high WSS at the however, reveal remarkably altered systolic flow patterns in the
right wall. ascending aorta with valve abnormalities (134–136). Many
LWBK1209-ch33_p545-557.indd 553 16/05/13 9:14 PM

B
Figure 33.9. Panel A exhibits normal blood flow in a healthy volunteer. From left to right, magnetic reso-
nance angiography (MRA), systolic streamlines in the ascending aorta, and cross-sectional analysis at the plane
depicted in the proximal ascending aorta are provided. The MRA shows normal aortic geometry, the stream-
lines normal laminar systolic flow, and the cross-sectional analysis central fast flow and an even distribution of
wall shear stress around the aortic lumen; the green bars represent the relative magnitude of wall shear stress.
Panel B is from a 34-year-old woman with bicuspid aortic valve (BAV), aortic stenosis and dilatation of the
ascending aorta up to 4.6 cm. Eccentric flow with a right-handed helix of systolic streamlines is demonstrated.
Shear stress is focally elevated where flow is marginalized against the aortic wall. (From Hope MD, Dyverfeldt
P, Acevedo-Bolton G, et al. Post-stenotic dilatation: Evaluation of ascending aortic dilatation with 4D flow
MR imaging. Int J Cardiol. 2012;156:e40–e42. (http://dx.doi.org/10.1016/j.ijcard.2011.08.018) [Epub Ahead
of print], with permission).
patients with bicuspid aortic valve demonstrate focally centrally (Fig. 33.9A). The common theme of the range
elevated systolic WSS at the convexity of the ascending aorta of abnormal systolic aortic flow patterns seen with aortic
(Fig. 33.9) (121,137). Intriguingly, the convexity is also the valve disease and/or aortic dilatation is eccentricity, with
site of asymmetric aneurysm formation with bicuspid aortic fastest flow peripheral rather than central (Fig. 33.9B).
valve (138,139) and where early and asymmetric extracellu- This off-center systolic flow results in an altered hemo-
lar matrix changes that are potentially flow related have been dynamic environment with asymmetric WSS profiles, ele-
reported in surgical specimen (140). More data are needed to vated in some areas and decreased or reversed in others.
better elucidate the relationship between abnormal flow and Such alterations in aortic blood flow are difficult to char-
valve-related aortic disease, but with the prevalence of aortic acterize by echocardiography, but they are well visualized
valve disease, it represents one of the more promising clinical with 4D flow. Eccentricity in the ascending aorta can be
applications of 4D flow. quantified with a flow jet angle approach (141), although
evaluating flow displacement from the aortic centerline
may be more reliable (Fig. 33.10). Recent work has shown
Flow Eccentricity
that semiautomatic analysis of 4D flow data can quan-
Skewed aortic blood flow is often referred to as eccentric. tify flow eccentricity with excellent agreement with visual
Normal systolic aortic flow is laminar with fastest flow observations (63).
LWBK1209-ch33_p545-557.indd 554 16/05/13 9:14 PM

intuitive and appealing to clinicians. A range of hemody-

namic biomarkers can be obtained and are being investi-
gated in various clinical scenarios. Preliminary data suggest
that some of these biomarkers may be able to predict pro-
gressive aortic disease. With larger prospective studies MRI
blood flow imaging may soon be used clinically to identify
cardiovascular risk well before life-threatening aortic com-
plications occur.
References
1. Elefteriades JA, Farkas EA. Thoracic aortic aneurysm clinically pertinent controversies
and uncertainties. J Am Coll Cardiol. 2010;55:841–857.
2. Scott RA, Ashton HA, Kay DN. Abdominal aortic aneurysm in 4237 screened
patients: Prevalence, development and management over 6 years. Br J Surg. 1991;78:
1122–1125.
3. Isselbacher EM. Thoracic and abdominal aortic aneurysms. Circulation. 2005;111:816–828.
4. Hiratzka LF, Bakris GL, Beckman JA, et al. 2010 ACCF/AHA/AATS/ACR/ASA/SCA/SCAI/
SIR/STS/SVM guidelines for the diagnosis and management of patients with Thoracic Aortic
Disease: A report of the American College of Cardiology Foundation/American Heart
Association Task Force on Practice Guidelines, American Association for Thoracic Surgery,
American College of Radiology, American Stroke Association, Society of Cardiovascular
Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society of
Interventional Radiology, Society of Thoracic Surgeons, and Society for Vascular Medicine.
Figure 33.10. Left: Schematic representation of investigated Circulation. 2010;121:e266–e369.
parameters: Displacement (Disp) of flow with respect to centerline 5. Elefteriades JA. Thoracic aortic aneurysm: Reading the enemy’s playbook. Curr Probl
(AoC), and angle of flow jet calculated between the normal to the Cardiol 2008;33:203–277.
cross-sectional plane (1) and the mean direction of peak systolic for- 6. Pape LA, Tsai TT, Isselbacher EM, et al. Aortic diameter >or = 5.5 cm is not a good pre-
dictor of type A aortic dissection: Observations from the International Registry of Acute
ward flow (2). Right: Cross-sectional views of the ascending aorta just Aortic Dissection (IRAD). Circulation. 2007;116:1120–1127.
distal to the sinotubular junction showing flow displacement from the 7. Fillinger MF, Marra SP, Raghavan ML, et al. Prediction of rupture risk in abdominal aortic
center of the aorta for normal, mildly, and markedly eccentric peak aneurysm during observation: Wall stress versus diameter. J Vasc Surg. 2003;37:724–732.
systolic flow patterns. Color coding represents the velocity magnitude 8. Nicholls SC, Gardner JB, Meissner MH, et al. Rupture in small abdominal aortic aneu-
rysms. J Vasc Surg. 1998;28:884–888.
of forward flow. (From Sigovan M, Hope MD, Dyverfeldt P, et al. 9. Brown LC, Powell JT. Risk factors for aneurysm rupture in patients kept under ultra-
Comparison of four-dimensional flow parameters for quantification sound surveillance. UK Small Aneurysm Trial Participants. Ann Surg. 1999;230:289–296;
of flow eccentricity in the ascending aorta. J Magn Reson Imaging. discussion 296–297.
2011;34:1226–1230, with permission). 10. Hope TA, Hope MD. Improved risk assessment for abdominal aortic aneurysm rupture:
Off-the-wall imaging. J Am Coll Cardiol. 2011;58:2531–2532.
11. Osborn EA, Jaffer FA. The year in molecular imaging. JACC Cardiovasc Imaging. 2010;3:
1181–1195.
12. Malek AM, Alper SL, Izumo S. Hemodynamic shear stress and its role in atherosclerosis.
JAMA. 1999;282:2035–2042.
A benefit of defining an accurate, quantitative approach 13. Wilton E, Jahangiri M. Post-stenotic aortic dilatation. J Cardiothorac Surg. 2006;1:7.
14. Kilner PJ, Firmin DN, Rees RS, et al. Valve and great vessel stenosis: Assessment with MR
for measuring flow eccentricity is that eccentricity may jet velocity mapping. Radiology. 1991;178:229–235.
be used as a surrogate for WSS. Since WSS is related to 15. Glockner JF, Johnston DL, McGee KP. Evaluation of cardiac valvular disease with MR
imaging: Qualitative and quantitative techniques. Radiographics. 2003;23:e9.
the near wall gradient of blood velocity, flow eccentric- 16. Dulce MC, Mostbeck GH, O’Sullivan M, et al. Severity of aortic regurgitation: Interstudy
ity is the principal factor driving changes in shear stress reproducibility of measurements with velocity-encoded cine MR imaging. Radiology. 1992;
(63). Calculation of eccentricity requires less computation 185:235–240.
17. Honda N, Machida K, Hashimoto M, et al. Aortic regurgitation: Quantitation with MR
and mathematical assumptions than are needed for WSS, imaging velocity mapping. Radiology. 1993;186:189–194.
making eccentricity a simple and attractive biomarker for 18. Pennell DJ, Sechtem UP, Higgins CB, et al. Clinical indications for cardiovascular mag-
regions of elevated WSS. Recent work suggests that eccen- netic resonance (CMR): Consensus Panel report. Eur Heart J. 2004;25:1940–1965.
19. Kilner PJ, Manzara CC, Mohiaddin RH, et al. Magnetic resonance jet velocity mapping in
tric flow does indeed drive growth in the ascending aorta. mitral and aortic valve stenosis. Circulation. 1993;87:1239–1248.
Restrictive aortic leaflet motion, which deflects systolic 20. Muzzarelli S, Meadows AK, Ordovas KG, et al. Usefulness of cardiovascular magnetic
resonance imaging to predict the need for intervention in patients with coarctation of
flow away from the aortic centerline, shows strong correla- the aorta. Am J Cardiol. 2012;109:861–865. (http://dx.doi.org/10.1016/j.amjcard.2011.
tion with increased aortic growth in a subset of patients 10.048).
with bicuspid aortic valve (142). In a separate study of a 21. Nielsen JC, Powell AJ, Gauvreau K, et al. Magnetic resonance imaging predictors of
coarctation severity. Circulation. 2005;111:622–628.
similar population of patients, flow displacement from cen- 22. Warnes CA, Williams RG, Bashore TM, et al. ACC/AHA 2008 Guidelines for the
terline measured with 4D flow exhibits good correlation Management of Adults with Congenital Heart Disease: A report of the American College
with increased growth rate (143). Flow eccentricity is thus of Cardiology/American Heart Association Task Force on Practice Guidelines (writing com-
mittee to develop guidelines on the management of adults with congenital heart disease).
a promising 4D flow hemodynamic biomarker of progres- Circulation. 2008;118:e714–e833.
sive aortic disease. 23. Steffens JC, Bourne MW, Sakuma H, et al. Quantification of collateral blood flow in
coarctation of the aorta by velocity encoded cine magnetic resonance imaging. Circulation.
1994;90:937–943.
24. Holmqvist C, Stahlberg F, Hanseus K, et al. Collateral flow in coarctation of the aorta
Summary with magnetic resonance velocity mapping: Correlation to morphological imaging of
collateral vessels. J Magn Reson Imaging. 2002;15:39–46.
25. Araoz PA, Reddy GP, Tarnoff H, et al. MR findings of collateral circulation are more
Aortic blood flow imaging with MRI is fast approaching clini- accurate measures of hemodynamic significance than arm-leg blood pressure gradient
cal medicine. Scan time has fallen to the clinical feasible thresh- after repair of coarctation of the aorta. J Magn Reson Imaging. 2003;17:177–183.
26. Muzzarelli S, Ordovas KG, Hope MD, et al. Diagnostic value of the flow profile in the
old of 15 minutes. Velocity data is accurate and reproduc- distal descending aorta by phase-contrast magnetic resonance for predicting severe coarc-
ible. Many visualization techniques are available that are tation of the aorta. J Magn Reson Imaging. 2011;33:1440–1446.
LWBK1209-ch33_p545-557.indd 555 16/05/13 9:14 PM

27. Pelc NJ, Bernstein MA, Shimakawa A, et al. Encoding strategies for three-direction phase- 63. Sigovan M, Hope MD, Dyverfeldt P, et al. Comparison of four-dimensional flow param-
contrast MR imaging of flow. J Magn Reson Imaging. 1991;1:405–413. eters for quantification of flow eccentricity in the ascending aorta. J Magn Reson Imaging.
28. Dumoulin CL, Souza SP, Walker MF, et al. Three-dimensional phase contrast angiogra- 2011;34:1226–1230.
phy. Magn Reson Med. 1989;9:139–149. 64. Eriksson J, Carlhall CJ, Dyverfeldt P, et al. Semi-automatic quantification of 4D left ven-
29. Lee AT, Pike GB, Pelc NJ. Three-point phase-contrast velocity measurements with increased tricular blood flow. J Cardiovasc Magn Reson. 2010;12:9.
velocity-to-noise ratio. Magn Reson Med. 1995;33:122–126. 65. Bernstein MA, Zhou XJ, Polzin JA, et al. Concomitant gradient terms in phase contrast
30. Johnson KM, Markl M. Improved SNR in phase contrast velocimetry with five-point bal- MR: Analysis and correction. Magn Reson Med. 1998;39:300–308.
anced flow encoding. Magn Reson Med. 2010;63:349–355. 66. Markl M, Bammer R, Alley MT, et al. Generalized reconstruction of phase contrast MRI:
31. Zhong X, Helm PA, Epstein FH. Balanced multipoint displacement encoding for DENSE Analysis and correction of the effect of gradient field distortions. Magn Reson Med.
MRI. Magn Reson Med. 2009;61:981–988. 2003;50:791–801.
32. Napel S, Lee DH, Frayne R, et al. Visualizing three-dimensional flow with simulated 67. Xiang QS. Temporal phase unwrapping for CINE velocity imaging. J Magn Reson
streamlines and three-dimensional phase-contrast MR imaging. J Magn Reson Imaging. Imaging. 1995;5:529–534.
1992;2:143–153. 68. Walker PG, Cranney GB, Scheidegger MB, et al. Semiautomated method for noise reduc-
33. Kilner PJ, Yang GZ, Mohiaddin RH, et al. Helical and retrograde secondary flow pat- tion and background phase error correction in MR phase velocity data. J Magn Reson
terns in the aortic arch studied by three-directional magnetic resonance velocity mapping. Imaging. 1993;3:521–530.
Circulation. 1993;88:2235–2247. 69. Lankhaar JW, Hofman MB, Marcus JT, et al. Correction of phase offset errors in main
34. Wigstrom L, Sjoqvist L, Wranne B. Temporally resolved 3D phase-contrast imaging. pulmonary artery flow quantification. J Magn Reson Imaging. 2005;22:73–79.
Magn Reson Med. 1996;36:800–803. 70. Rolf MP, Hofman MB, Gatehouse PD, et al. Sequence optimization to reduce velocity
35. Firmin DN GP, Konrad JP, Yang GZ, et al. Rapid 7-dimensional imaging of pulsatile flow. offsets in cardiovascular magnetic resonance volume flow quantification—a multi-vendor
Computers in Cardiology. London: IEEE Computer Society; 1993:353–356. study. J Cardiovasc Magn Reson. 2011;13:18.
36. Wigstrom L, Ebbers T, Fyrenius A, et al. Particle trace visualization of intracardiac flow 71. Hollnagel DI, Summers PE, Kollias SS, et al. Laser Doppler velocimetry (LDV) and 3D
using time-resolved 3D phase contrast MRI. Magn Reson Med. 1999;41:793–799. phase-contrast magnetic resonance angiography (PC-MRA) velocity measurements:
37. Kozerke S, Hasenkam JM, Pedersen EM, et al. Visualization of flow patterns distal to Validation in an anatomically accurate cerebral artery aneurysm model with steady flow.
aortic valve prostheses in humans using a fast approach for cine 3D velocity mapping. J J Magn Reson Imaging. 2007;26:1493–1505.
Magn Reson Imaging. 2001;13:690–698. 72. Boussel L, Rayz V, Martin A, et al. Phase-contrast magnetic resonance imaging measure-
38. Fyrenius A, Wigstrom L, Ebbers T, et al. Three dimensional flow in the human left atrium. ments in intracranial aneurysms in vivo of flow patterns, velocity fields, and wall shear stress:
Heart. 2001;86:448–455. Comparison with computational fluid dynamics. Magn Reson Med. 2009;61:409–417.
39. Markl M, Hennig J. Phase contrast MRI with improved temporal resolution by view shar- 73. Petersson S, Dyverfeldt P, Gardhagen R, et al. Simulation of phase contrast MRI of turbu-
ing: K-space related velocity mapping properties. Magn Reson Imaging. 2001;19:669–676. lent flow. Magn Reson Med. 2010;64:1039–1046.
40. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: Sensitivity encoding for fast 74. Jiang J, Johnson K, Valen-Sendstad K, et al. Flow characteristics in a canine aneurysm
MRI. Magn Reson Med. 1999;42:952–962. model: A comparison of 4D accelerated phase-contrast MR measurements and computa-
41. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially tional fluid dynamics simulations. Med Phys. 2011;38:6300–6312.
parallel acquisitions (GRAPPA). Magn Reson Med. 2002;47:1202–1210. 75. van Ooij P, Guedon A, Poelma C, et al. Complex flow patterns in a real-size intracranial
42. Tsao J, Boesiger P, Pruessmann KP. k-t BLAST and k-t SENSE: Dynamic MRI with high aneurysm phantom: Phase contrast MRI compared with particle image velocimetry and
frame rate exploiting spatiotemporal correlations. Magn Reson Med. 2003;50:1031–1042. computational fluid dynamics. NMR Biomed. 2012;25:14–26.
43. Huang F, Akao J, Vijayakumar S, et al. k-t GRAPPA: A k-space implementation for 76. Markl M, Geiger J, Kilner PJ, et al. Time-resolved three-dimensional magnetic resonance
dynamic MRI with high reduction factor. Magn Reson Med. 2005;54:1172–1184. velocity mapping of cardiovascular flow paths in volunteers and patients with Fontan
44. Lustig M, Donoho D, Pauly JM. Sparse MRI: The application of compressed sensing for circulation. Eur J Cardiothorac Surg. 2011;39:206–212.
rapid MR imaging. Magn Reson Med. 2007;58:1182–1195. 77. Hope MD, Meadows AK, Hope TA, et al. Clinical evaluation of aortic coarctation with
45. Jung H, Sung K, Nayak KS, et al. k-t FOCUSS: A general compressed sensing framework 4D flow MR imaging. J Magn Reson Imaging. 2010;31:711–718.
for high resolution dynamic MRI. Magn Reson Med. 2009;61:103–116. 78. Yu HY, Peng HH, Wang JL, et al. Quantification of the pulse wave velocity of the descend-
46. Jung B, Stalder AF, Bauer S, et al. On the undersampling strategies to accelerate time- ing aorta using axial velocity profiles from phase-contrast magnetic resonance imaging.
resolved 3D imaging using k-t-GRAPPA. Magn Reson Med. 2011;66:966–975. Magn Reson Med. 2006;56:876–883.
47. Carlsson M, Toger J, Kanski M, et al. Quantification and visualization of cardiovascular 79. Bradlow WM, Gatehouse PD, Hughes RL, et al. Assessing normal pulse wave velocity
4D velocity mapping accelerated with parallel imaging or k-t BLAST: Head to head com- in the proximal pulmonary arteries using transit time: A feasibility, repeatability, and
parison and validation at 1.5 T and 3 T. J Cardiovasc Magn Reson. 2011;13:55. observer reproducibility study by cardiovascular magnetic resonance. J Magn Reson
48. Marshall I. Feasibility of k-t BLAST technique for measuring “seven-dimensional” fluid Imaging. 2007;25:974–981.
flow. J Magn Reson Imaging. 2006;23:189–196. 80. Urchuk SN, Plewes DB. A velocity correlation method for measuring vascular compliance
49. Stadlbauer A, van der Riet W, Crelier G, et al. Accelerated time-resolved three-dimensional using MR imaging. J Magn Reson Imaging. 1995;5:628–634.
MR velocity mapping of blood flow patterns in the aorta using SENSE and k-t BLAST. Eur 81. Mohiaddin RH, Firmin DN, Longmore DB. Age-related changes of human aortic flow
J Radiol. 2010;75:e15–e21. wave velocity measured noninvasively by magnetic resonance imaging. J Appl Physiol.
50. Lauterbur PC. Image formation by induced local interactions: Examples employing 1993;74:492–497.
nuclear magnetic resonance. Nature. 1973;242:190–191. 82. Boese JM, Bock M, Schoenberg SO, et al. Estimation of aortic compliance using magnetic
51. Gu T, Korosec FR, Block WF, et al. PC VIPR: A high-speed 3D phase-contrast method for resonance pulse wave velocity measurement. Phys Med Biol. 2000;45:1703–1713.
flow quantification and high-resolution angiography. AJNR Am J Neuroradiol. 2005;26: 83. Fielden SW, Fornwalt BK, Jerosch-Herold M, et al. A new method for the determination
743–749. of aortic pulse wave velocity using cross-correlation on 2D PCMR velocity data. J Magn
52. Kecskemeti S, Johnson K, Wu Y, et al. High resolution three-dimensional cine phase con- Reson Imaging. 2008;27:1382–1387.
trast MRI of small intracranial aneurysms using a stack of stars k-space trajectory. J Magn 84. Markl M, Wallis W, Brendecke S, et al. Estimation of global aortic pulse wave velocity by
Reson Imaging. 2012;35:518–527. (doi: 10.1002/jmri.23501). flow-sensitive 4D MRI. Magn Reson Med. 2010;63:1575–1582.
53. Sigfridsson A, Petersson S, Carlhall CJ, et al. Four-dimensional flow MRI using spiral 85. Grotenhuis HB, Westenberg JJ, Steendijk P, et al. Validation and reproducibility of aor-
acquisition. Magn Reson Med. 2012;68:1065–1073. (doi: 10.1002/mrm.23297). tic pulse wave velocity as assessed with velocity-encoded MRI. J Magn Reson Imaging.
54. Bailes DR, Gilderdale DJ, Bydder GM, et al. Respiratory ordered phase encoding (ROPE): 2009;30:521–526.
A method for reducing respiratory motion artefacts in MR imaging. J Comput Assist 86. Laurent S, Boutouyrie P, Asmar R, et al. Aortic stiffness is an independent predic-
Tomogr. 1985;9:835–838. tor of all-cause and cardiovascular mortality in hypertensive patients. Hypertension.
55. Markl M, Chan FP, Alley MT, et al. Time-resolved three-dimensional phase-contrast MRI. 2001;37:1236–1241.
J Magn Reson Imaging. 2003;17:499–506. 87. Laurent S, Katsahian S, Fassot C, et al. Aortic stiffness is an independent predictor of fatal
56. Uribe S, Beerbaum P, Sorensen TS, et al. Four-dimensional (4D) flow of the whole heart stroke in essential hypertension. Stroke. 2003;34:1203–1206.
and great vessels using real-time respiratory self-gating. Magn Reson Med. 2009;62: 88. Cruickshank K, Riste L, Anderson SG, et al. Aortic pulse-wave velocity and its relation-
984–992. ship to mortality in diabetes and glucose intolerance: An integrated index of vascular
57. Buonocore MH. Visualizing blood flow patterns using streamlines, arrows, and particle function? Circulation. 2002;106:2085–2090.
paths. Magn Reson Med. 1998;40:210–226. 89. Blacher J, Guerin AP, Pannier B, et al. Impact of aortic stiffness on survival in end-stage
58. Dyverfeldt P, Kvitting JP, Carlhall CJ, et al. Hemodynamic aspects of mitral regurgitation renal disease. Circulation. 1999;99:2434–2439.
assessed by generalized phase-contrast MRI. J Magn Reson Imaging. 2011;33:582–588. 90. Laurent S, Cockcroft J, Van Bortel L, et al. Expert consensus document on arterial stiff-
59. Toger J, Carlsson M, Soderlind G, et al. Volume Tracking: A new method for quantitative ness: Methodological issues and clinical applications. Eur Heart J. 2006;27:2588–2605.
assessment and visualization of intracardiac blood flow from three-dimensional, time- 91. Tzemos N, Lyseggen E, Silversides C, et al. Endothelial function, carotid-femoral stiffness,
resolved, three-component magnetic resonance velocity mapping. BMC Med Imaging. and plasma matrix metalloproteinase-2 in men with bicuspid aortic valve and dilated
2011;11:10. aorta. J Am Coll Cardiol. 2010;55:660–668.
60. Heiberg E, Ebbers T, Wigström L, et al. Three dimensional flow characterization using 92. Cavalcante JL, Lima JA, Redheuil A, et al. Aortic stiffness: Current understanding and
vector pattern matching. IEEE Trans Vis Comp Graphics. 2003;9:313–319. future directions. J Am Coll Cardiol. 2011;57:1511–1522.
61. Harloff A, Simon J, Brendecke S, et al. Complex plaques in the proximal descending 93. Nollen GJ, Groenink M, Tijssen JG, et al. Aortic stiffness and diameter predict progressive
aorta: An underestimated embolic source of stroke. Stroke. 2010;41:1145–1150. aortic dilatation in patients with Marfan syndrome. Eur Heart J. 2004;25:1146–1152.
62. Bolger AF, Heiberg E, Karlsson M, et al. Transit of blood flow through the human left ven- 94. Ahimastos AA, Aggarwal A, D’Orsa KM, et al. Effect of perindopril on large artery stiff-
tricle mapped by cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2007;9: ness and aortic root diameter in patients with Marfan syndrome: A randomized con-
741–747. trolled trial. JAMA. 2007;298:1539–1547.
LWBK1209-ch33_p545-557.indd 556 16/05/13 9:14 PM

95. Groenink M, de Roos A, Mulder BJ, et al. Changes in aortic distensibility and pulse wave 119. Harloff A, Nussbaumer A, Bauer S, et al. In vivo assessment of wall shear stress in the ath-
velocity assessed with magnetic resonance imaging following beta-blocker therapy in the erosclerotic aorta using flow-sensitive 4D MRI. Magn Reson Med. 2010;63:1529–1536.
Marfan syndrome. Am J Cardiol. 1998;82:203–208. 120. Markl M, Wallis W, Harloff A. Reproducibility of flow and wall shear stress analysis
96. Gao JH, Gore JO. Turbulent flow effects on NMR imaging: Measurement of turbulent using flow-sensitive four-dimensional MRI. J Magn Reson Imaging. 2011;33:988–994.
intensity. Med Phys. 1991;18:1045–1051. 121. Hope MD, Hope TA, Crook SE, et al. 4D flow CMR in assessment of valve-related
97. Pipe JG. A simple measure of flow disorder and wall shear stress in phase contrast MRI. ascending aortic disease. JACC Cardiovasc Imaging. 2011;4:781–787.
Magn Reson Med. 2003;49:543–550. 122. Petersson S, Dyverfeldt P, Ebbers T. Assessment of the accuracy of MRI wall shear stress
98. Dyverfeldt P, Sigfridsson A, Kvitting JP, et al. Quantification of intravoxel velocity stan- estimation using numerical simulations. J Magn Reson Imaging. 2012;36:128–138.
dard deviation and turbulence intensity by generalizing phase-contrast MRI. Magn Reson 123. Cheng C, Tempel D, van Haperen R, et al. Atherosclerotic lesion size and vulnerability are
Med. 2006;56:850–858. determined by patterns of fluid shear stress. Circulation. 2006;113:2744–2753.
99. Elkins CJ, Alley MT, Saetran L, et al. Three-dimensional magnetic resonance velocimetry 124. Katritsis D, Kaiktsis L, Chaniotis A, et al. Wall shear stress: Theoretical considerations
measurements of turbulence quantities in complex flow. Exp Fluids. 2009;46:285–296. and methods of measurement. Prog Cardiovasc Dis. 2007;49:307–329.
100. Arzani A, Dyverfeldt P, Ebbers T, et al. In vivo validation of numerical prediction for 125. Cecchi E, Giglioli C, Valente S, et al. Role of hemodynamic shear stress in cardiovascular
turbulence intensity in an aortic coarctation. Ann Biomed Eng. 2012;40:860–870. (doi: disease. Atherosclerosis. 2011;214:249–256.
10.1007/s10439-011-0447-6). 126. Boussel L, Rayz V, McCulloch C, et al. Aneurysm growth occurs at region of low wall
101. Dyverfeldt P, Kvitting JP, Sigfridsson A, et al. Assessment of fluctuating velocities in dis- shear stress: Patient-specific correlation of hemodynamics and growth in a longitudinal
turbed cardiovascular blood flow: In vivo feasibility of generalized phase-contrast MRI. study. Stroke. 2008;39:2997–3002.
J Magn Reson Imaging. 2008;28:655–663. 127. Wang DH, Makaroun MS, Webster MW, et al. Effect of intraluminal thrombus on wall stress
102. Stein PD, Sabbah HN. Measured turbulence and its effect on thrombus formation. Circ in patient-specific models of abdominal aortic aneurysm. J Vasc Surg. 2002;36:598–604.
Res. 1974;35:608–614. 128. Girerd X, London G, Boutouyrie P, et al. Remodeling of the radial artery in response to a
103. Sallam AM, Hwang NH. Human red blood cell hemolysis in a turbulent shear flow: chronic increase in shear stress. Hypertension. 1996;27:799–803.
Contribution of Reynolds shear stresses. Biorheology. 1984;21:783–797. 129. Fry DL. Certain histological and chemical responses of the vascular interface to acutely
104. Davies PF, Remuzzi A, Gordon EJ, et al. Turbulent fluid shear stress induces vascular induced mechanical stress in the aorta of the dog. Circ Res. 1969;24:93–108.
endothelial cell turnover in vitro. Proc Natl Acad Sci U S A. 1986;83:2114–2117. 130. Slager CJ, Wentzel JJ, Gijsen FJ, et al. The role of shear stress in the destabilization of
105. Ebbers T, Wigstrom L, Bolger AF, et al. Estimation of relative cardiovascular pres- vulnerable plaques and related therapeutic implications. Nat Clin Pract Cardiovasc Med.
sures using time-resolved three-dimensional phase contrast MRI. Magn Reson Med. 2005;2:456–464.
2001;45:872–879. 131. Dolan JM, Meng H, Singh S, et al. High fluid shear stress and spatial shear stress gradi-
106. Thompson RB, McVeigh ER. Fast measurement of intracardiac pressure differences with ents affect endothelial proliferation, survival, and alignment. Ann Biomed Eng. 2011;39:
2D breath-hold phase-contrast MRI. Magn Reson Med. 2003;49:1056–1066. 1620–1631.
107. Tyszka JM, Laidlaw DH, Asa JW, et al. Three-dimensional, time-resolved (4D) relative 132. Crook SE, Hope MD. Arch hypoplasia and aneurysm after aortic coarctation repair:
pressure mapping using magnetic resonance imaging. J Magn Reson Imaging. 2000; Abnormal flow may be the link. JACC Cardiovasc Imaging. 2011;4:685–686.
12:321–329. 133. Hope MD, Hope TA, Urbania TH, et al. Four-dimensional flow magnetic resonance imag-
108. Ebbers T, Wigstrom L, Bolger AF, et al. Noninvasive measurement of time-varying ing with wall shear stress analysis before and after repair of aortopulmonary fistula. Circ
three-dimensional relative pressure fields within the human heart. J Biomech Eng. 2002; Cardiovasc Imaging. 2010;3:766–768.
124:288–293. 134. Hope MD, Meadows AK, Hope TA, et al. Images in cardiovascular medicine. Evaluation
109. Bock J, Frydrychowicz A, Lorenz R, et al. In vivo noninvasive 4D pressure difference of bicuspid aortic valve and aortic coarctation with 4D flow magnetic resonance imaging.
mapping in the human aorta: Phantom comparison and application in healthy volunteers Circulation. 2008;117:2818–2819.
and patients. Magn Reson Med. 2011;66:1079–1088. 135. Hope MD, Hope TA, Meadows AK, et al. Bicuspid aortic valve: Four-dimensional MR
110. Ebbers T, Farneback G. Improving computation of cardiovascular relative pressure fields evaluation of ascending aortic systolic flow patterns. Radiology. 2010;255:53–61.
from velocity MRI. J Magn Reson Imaging. 2009;30:54–61. 136. Hope MD, Dyverfeldt P, Acevedo-Bolton G, et al. Post-stenotic dilation: Evaluation of
111. Gresho PM SR. On pressure boundary conditions for the incompressible Navier-Stokes ascending aortic dilation with 4D flow MR imaging. Int J Cardiol. 2012;156:e40–e42.
equations. INT J NUMER METH FL. 1987;7:1111–1145. (http://dx.doi.org/10.1016/j.ijcard.2011.08.018).
112. Oshinski JN, Ku DN, Mukundan S Jr, et al. Determination of wall shear stress in the aorta 137. Viscardi F, Vergara C, Antiga L, et al. Comparative finite element model analysis
with the use of MR phase velocity mapping. J Magn Reson Imaging. 1995;5:640–647. of ascending aortic flow in bicuspid and tricuspid aortic valve. Artif Organs. 2010;34:
113. Oyre S, Ringgaard S, Kozerke S, et al. Quantitation of circumferential subpixel vessel wall 1114–1120.
position and wall shear stress by multiple sectored three-dimensional paraboloid model- 138. Bauer M, Gliech V, Siniawski H, et al. Configuration of the ascending aorta in patients
ing of velocity encoded cine MR. Magn Reson Med. 1998;40:645–655. with bicuspid and tricuspid aortic valve disease undergoing aortic valve replacement with
114. Masaryk AM, Frayne R, Unal O, et al. In vitro and in vivo comparison of three MR or without reduction aortoplasty. J Heart Valve Dis. 2006;15:594–600.
measurement methods for calculating vascular shear stress in the internal carotid artery. 139. Cotrufo M, Della Corte A. The association of bicuspid aortic valve disease with asym-
AJNR Am J Neuroradiol. 1999;20:237–245. metric dilatation of the tubular ascending aorta: Identification of a definite syndrome.
115. Cheng CP, Parker D, Taylor CA. Quantification of wall shear stress in large blood vessels J Cardiovasc Med (Hagerstown). 2009;10:291–297.
using Lagrangian interpolation functions with cine phase-contrast magnetic resonance 140. Della Corte A, Quarto C, Bancone C, et al. Spatiotemporal patterns of smooth muscle cell
imaging. Ann Biomed Eng. 2002;30:1020–1032. changes in ascending aortic dilatation with bicuspid and tricuspid aortic valve stenosis:
116. Stalder AF, Russe MF, Frydrychowicz A, et al. Quantitative 2D and 3D phase contrast Focus on cell-matrix signaling. J Thorac Cardiovasc Surg. 2008;135:8–18, 18.e1–18.e2.
MRI: Optimized analysis of blood flow and vessel wall parameters. Magn Reson Med. 141. den Reijer PM, Sallee D 3rd, van der Velden P, et al. Hemodynamic predictors of aortic
2008;60:1218–1231. dilatation in bicuspid aortic valve by velocity-encoded cardiovascular magnetic resonance.
117. Pedersen EM, Oyre S, Agerbaek M, et al. Distribution of early atherosclerotic lesions in J Cardiovasc Magn Reson. 2010;12:4.
the human abdominal aorta correlates with wall shear stresses measured in vivo. Eur J 142. Della Corte A, Bancone C, Conti CA, et al. Restricted cusp motion in right-left type
Vasc Endovasc Surg. 1999;18:328–333. of bicuspid aortic valves: A new risk marker for aortopathy. J Thorac Cardiovasc Surg.
118. Barker AJ, Lanning C, Shandas R. Quantification of hemodynamic wall shear stress in 2012;144:360–369. (http://dx.doi.org/10.1016/j.jtcvs.2011.10.014).
patients with bicuspid aortic valve using phase-contrast MRI. Ann Biomed Eng. 2010; 143. Hope MD, Wrenn J, Sigovan M, et al. Imaging biomarkers of aortic disease: Increased
38:788–800. growth rates with eccentric systolic flow. J Am Coll Cardiol. 2012;60:356–357.
LWBK1209-ch33_p545-557.indd 557 16/05/13 9:14 PM

Rutger W. van der Meer
Chapter Lucia J.M. Kroft
34
Jasper Florie
Martin N. Wasser
Albert de Roos
Computed Tomographic Angiography

of the Abdominal Arteries
■■ Abdominal Computed Tomographic resolution (1,7). Three-dimensional (3D) visualization can

Angiography Techniques facilitate image interpretation and might improve diagnostic
Contrast Administration accuracy.
MDCT is superior to single-slice CT for nearly all ap-
Scanning Technique
plications, especially for imaging the vascular system (1,8).
Postprocessing Compared with single-slice CT, there is a reduction in radiation
■■ Clinical Applications dose, except when thin sections of high quality are acquired or
Abdominal Aorta when multiphase imaging is performed (3). Compared with
Renal Applications single-slice CT, the injected volume of contrast medium can
Vascular Liver Imaging be reduced substantially by using more than 16-row MDCT
Mesenteric Ischemia technology.
MDCT has replaced x-ray angiography in many clinical
Intra-abdominal Arterial Bleeding
vascular applications (6). Furthermore, current MDCT
■■ Summary technology has a better spatial resolution than magnetic
resonance imaging (MRI). Another advantage of MDCT
for vascular imaging is that mural calcifications can be as-
Abdominal Computed sessed, which are usually not well recognized with MRI.
Tomographic Angiography Disadvantages of CT are the use of ionizing radiation and
Techniques the administration of a potentially nephrotoxic iodinated
contrast medium.
Computed tomography (CT) has undergone a revolution since In this chapter the abdominal CTA technique will be dis-
the introduction of multidetector-row computed tomography cussed. Practical applications of MDCT for assessing the
(MDCT) using systems currently up to 320 detector arrays abdominal aorta and renal and mesenteric vessels will be
(1–5). MDCT has three main advantages compared with sin- reviewed.
gle-slice CT: Faster imaging, the ability to cover larger body
volumes, and thinner section thickness with improved image
Contrast Administration
quality (6). Computed tomographic angiography (CTA) has
gained wide acceptance with the introduction of MDCT. Intravascular contrast material administration is essential in
Faster imaging results in substantial reduction of acquisition CTA (6). Both technical factors and patient-related factors
time (7). Imaging of the abdominal aorta and iliac arteries affect the process of contrast enhancement. It is important
can be performed in approximately 20 seconds with a 4-row to achieve homogeneous intravascular enhancement during
MDCT (depending on slice thickness and volume coverage) imaging. Arterial enhancement depends on iodine concen-
and in usually less than 10 seconds with a more than 4-row tration of the contrast medium used and the injection pro-
MDCT, thereby generating less motion artifacts because of tocol (9). More enhancement is achieved with higher iodine
imaging in a single breath-hold and less pulsation artifacts concentration. Faster injection rates rather than increased
because of reduced rotation time. The possibility to cover contrast medium volume also provide more enhancement,
larger anatomic areas allows for new applications, such as because a higher concentration of iodine is injected over
evaluation of the entire thoracoabdominal aorta in a single time (9).
breath-hold and imaging of the aortoiliac system and lower Several contrast injection protocols have been used for
extremities in one acquisition with high spatial resolution CTA. Uniphasic injection protocols with fixed scan delay, in-
(7). Thinner section thickness improves image quality and jection volume, and injection rate usually result in sufficient
allows acquisitions with isotropic or near isotropic spatial aortic enhancement (10). An interactive protocol, in which
558
LWBK1209-ch34_p558-574.indd 558 5/18/13 12:57 PM

Chapter 34 ■ Computed Tomographic Angiography of the Abdominal Arteries 559
of CT scanner (6). The second technique uses a test bolus to

determine patients’ contrast transit time. A region of interest
is placed in the distal descending aorta for abdominal CTA
and during injection of 15- to 20-mL contrast medium, the
test bolus, a series of low-dose nonincremental CT scans are
obtained. An enhancement curve is then automatically cal-
culated, and the time to maximum enhancement equals the
mean contrast transit time. The time to maximum enhance-
ment is set as the scan delay. However, scanning should be
started after an extra delay of approximately 10 seconds,
because with a large contrast bolus maximum enhancement
is achieved later compared with a test bolus because of the
accumulation of the contrast agent in the blood vessels (6).
With shorter acquisition times, the amount of contrast
agent to be injected can be reduced, but higher injection rates
are required for optimal enhancement. When parenchyma
imaging is also required (e.g., in liver imaging), the volume of
contrast medium cannot be reduced, because enhancement of
the parenchyma depends on the total iodine dose (9).
Oral contrast administration is not advisable for vascular
imaging; superposition of bowel loops filled with contrast
medium can hamper the evaluation of the contrast-enhanced
blood vessels, especially when reconstructions are obtained.
However, when oral contrast administration is required for
imaging, negative contrast medium (e.g., water) can be used,
which will not disturb evaluation of the contrast-enhanced
Figure 34.1. Coronal curved planar reformation (CPR) presenta- blood vessels. In addition, the use of negative oral contrast
tion in the abdominal aorta and right iliac artery in a patient with an agents facilitates assessment of the bowel wall when bowel
abdominal aortic aneurysm (AAA). Inhomogeneous enhancement in ischaemia is suspected.
the abdominal aorta and iliac arteries because of turbulent flow.
Scanning Technique
Vascular imaging of the abdominal aorta and its side
the injection of contrast medium is stopped manually, can branches requires a high-resolution protocol to produce a
be used to reduce the volume of contrast medium adminis- (near) isotropic dataset of thin overlapping cross-sectional
tered (11). Uniform, plateau-like arterial enhancement can images. Two-dimensional (2D) and 3D reconstructions of
only be achieved using a biphasic injection protocol (12). A high quality can be produced from this dataset. With 4- or
biphasic injection protocol starts with a small bolus with a 16-row MDCT, section thickness of 1 to 1.25 mm and 1 mm,
high flow rate, followed by a larger bolus with a lower flow respectively, can be obtained for high-resolution scanning.
rate (12). A saline flush is recommended to flush the veins With thinner section thickness, scanning is not fast enough
and push the contrast column into the circulation to reduce for covering the entire body volume of interest, whereas
the amount of contrast agent (6). Contrast injection for ab- with 64-row MDCT, section thickness of 0.5 mm can be
dominal imaging is preferably administered in a cubital vein. obtained routinely.
Major patient-related factors determining contrast en- Our standard scan protocol for the Aquilion 64 (Toshiba
hancement are body habitus and cardiac status. In case of Medical Systems, Toshiba cooperation, Tokyo, Japan) for
an abdominal aortic aneurysm (AAA), enhancement can be the abdominal aorta and mesenteric vessels is performed in
inhomogeneous because of turbulent flow (Fig. 34.1). the supine position, with the arms elevated. Scanning is per-
Because of these and other differences in circulation dynamics formed during a breath-hold. Section thickness is 0.5 mm,
between patients, scanning with a fixed delay is not recom- helical pitch is 53, and reconstruction index and interval are
mended (12). In addition contrast volume and injection rate 1 and 0.8 mm, respectively. The selected kV is 120, the mil-
need to be adjusted according to patient weight to achieve liampere varies per patient because tube current modulation
optimal vascular enhancement. Solutions to optimize scan is used, and the rotation time is 500 milliseconds. For an av-
delay for determination of the optimal enhancement phase in erage-weighted patient, arterial imaging is performed with an
a single patient are bolus triggering and the use of a test bolus. injection of 100-mL contrast agent (370 mg iodine/mL) at
With the first technique, a region of interest is placed in the 4 mL/s, followed by 40 mL of saline flush at the same in-
distal descending aorta for abdominal CTA. During contrast jection rate with at least 18G venous access, injected with
injection, low-dose nonincremental scans are obtained in a a power injector (Stellant injector, Medrad, Medrad Inc.,
dynamic fashion. The attenuation is monitored in the region Indianola, PA). Contrast agent volume and flow is adjusted
of interest and when a predefined enhancement threshold is according to patient weight. Delay is determined by bolus
reached, for example, 150 HU, the CT scan starts automati- triggering, and scanning starts with a delay of 10 seconds
cally with a delay of a few seconds, depending on the type when an enhancement increase of 100 HU is reached.
LWBK1209-ch34_p558-574.indd 559 17/05/13 5:17 PM

Figure 34.2. Multiplanar reformation (MPR) presentation in

the mid-sagittal pane. Patient with an AAA and penetrating ulcers
anteriorly (arrow). The focus of interest (penetrating ulcer) is clearly
depicted with this MPR technique, although the larger part of the
aorta is not visualized in this plane.
Figure 34.3. Same patient as in Figure 35.2. CPR presentation

in the sagittal plane through the aorta and right iliac artery. CPR
For the renal arteries slice thickness is 0.5 mm, helical presentation allows better evaluation of the vascular system through
pitch is 53, and reconstruction index and interval are 1 and the entire dataset, especially with tortuous atherosclerotic arteries.
0.8 mm, respectively. Tube current modulation is used result- Penetrating ulcer (arrow).
ing in different milliampere per patient; kV (120), and rota-
tion time (500 milliseconds) is unchanged. Parameters can be
changed individually depending on the patient’s weight.
Standard vascular protocols for the Aquilion 320 (Toshiba Imaging the peripheral arteries generates even more images,
Medical Systems) routinely use a slice thickness of 0.5 mm. because the scan volume is much larger. Workstation-based
Helical pitch is 65, and reconstruction index and interval are review is necessary not only to evaluate the axial images,
1 and 0.8 mm, respectively. which still remains the primary mode for evaluating the
abdominal aorta (14), but also to generate reconstructed and
3D images (13). Several reconstruction techniques can be
Postprocessing
used, which allows better evaluation of the entire vascula-
A great challenge of MDCT is in dealing with a large data ture. The main visualization techniques are multiplanar ref-
load (13), especially if an isotropic or near isotropic acquisi- ormation (MPR), maximum intensity projection (MIP), and
tion is performed (3). For example, routine multislice CTA of volume rendering (VR) (15).
the abdominal aorta with a scan range of 30 cm performed An MPR is a 2D reconstruction in any coronal, sagit-
with 1-mm collimation and a reconstruction index of 1 mm tal, or oblique plane. MPR presentation is easy to use, but
and reconstruction interval of 0.8 mm produces 375 images. restricted to a single 2D plane. Therefore it is of limited value
Using 64-slice CT scanners, collimations of 0.5 mm can be for vascular imaging, especially in tortuous arteries (Fig. 34.2).
achieved and thus a reconstruction index of 0.5 and a recon- A more sophisticated technique is curved planar reforma-
struction interval of 0.4 can be used. Using these param- tion (CPR). With CPR reformation, the display planes curve
eters, 750 images are generated for the same scan volume. along an anatomic structure through the entire dataset (16),
However, in our opinion, using slice thickness of 0.5 does for example, the abdominal aorta and a single iliac artery
not substantially add to the diagnosis in abdominal CTA (Fig. 34.3), projected in a 2D image. This technique is a
and we therefore use a reconstruction index and interval of valuable tool for vascular imaging. We use a Vitrea work-
1 and 0.8 mm as indicated earlier in 64- and 320-slice CT. station (Vital Images Incorporated, Plymouth, MN), which
LWBK1209-ch34_p558-574.indd 560 17/05/13 5:17 PM

Figure 34.4. Maximum intensity projection (MIP) presentation in

the anteroposterior plane in a patient with occlusion the left iliac arter-
ies (arrows) and left superficial femoral artery (dotted arrow). Editing
was performed to remove overlying bone. MIP presentation shows an
angio-like image.
Figure 34.5. Basic volume rendering (VR) presentation of abdomi-
nal computed tomographic angiogram (CTA) directly provided by
can generate CPRs semiautomatically. At the same time, ad- dedicated postprocessing software without editing. Note the infrarenal
ditional perpendicular images are displayed, allowing better aortic aneurysm.
evaluation of diameters, stenosis, and thrombus composi-
tion, thereby improving visualization of eccentric lesions.
MIP presentation visualizes only the brightest structures endoscopy to evaluate the internal surface of tubular struc-
in a selected plane (e.g., anteroposterior) and provides an tures (13,18). Mostly basic VR presentations are provided
angio-like image presentation (Fig. 34.4) (14). Thick- or thin- by dedicated post-processing software without prior editing.
slab MIP images can be obtained. Thin-slab images allow In addition, smaller regions of interest can be isolated from
better visualization of complicated anatomic structures. A the scan volume manually by using dedicated software.
limitation of the MIP technique, because it is a projection Additional visualization techniques like MIP and VR pre-
technique, is that more attenuated structures like bones ob- sentations are most often used in vascular imaging. The in-
scure the contrast-enhanced blood vessels (13). Editing to terpretation of an examination of the abdominal aorta takes
remove overlying structures from the MIP images is usually approximately 20 minutes, using a dedicated workstation.
time consuming. Another limitation of MIP presentation is Evaluating only 3D images is not an option because of the
the absence of the appreciation of depth relationships, es- limitations that have been discussed. It is always important
pecially in regions with a complex anatomy (13). MIP pre- to review the cross-sectional images to obtain additional
sentations are sometimes inadequate for visualizing small information (6).
vascular structures, for example, accessory renal arteries in
living-related kidney donors. Variants of MIP are minimum
intensity projections and curved-slab MIP. The latter can in-
clude multiple vessels in a single image with improvement of Clinical Applications
the interpretation efficiency (17).
VR is the most complex technique (Fig. 34.5). With VR, Abdominal Aorta
every voxel value is assigned an opacity level, ranging from
Abdominal Aortic Aneurysm
opacity to transparency. The opacification function can be
applied to a selected region in the histogram of voxel opac- Patients with AAA often remain without any symptoms
ity values, thereby visualizing the selected tissue of interest, until rupture occurs, with a high mortality rate. AAA occurs
for example, blood vessels (13). Colors can be applied to most often in middle-aged and elderly patients, and in more
the attenuation histogram, to differentiate voxel values and, men than women with a ratio of 4:1. The pathogenesis of an
for instance, to distinguish a calcified plaque from enhanc- AAA is complex. Risk factors for atherosclerosis and AAA
ing arterial lumen. VR presentations can be used for virtual are overlapping (19).
LWBK1209-ch34_p558-574.indd 561 17/05/13 5:17 PM

Aneurysms can be classified according to their cause (20). than with ultrasound (US) (26). With standardized measure-
Most aneurysms are degenerative atherosclerotic aneurysms ments on CT or US in routine follow-up, changes in diam-
(Fig. 34.5). Atherosclerotic penetrating ulcers can cause eter can be detected easily. However, reliable information for
small aneurysms (Figs. 34.2 and 34.3), with a high rupture optimal planning of endovascular intervention or surgery
rate up to 40% (21). Other types of aneurysms are infec- cannot be obtained by US alone.
tious, inflammatory, traumatic, congenital, and postopera- It has been demonstrated that the maximum diameter of
tive anastomotic aneurysms. It is important to differentiate the aneurysm is not enough for estimation of rupture risk,
between these causes, because therapy strategies differ with because small aneurysms may also rupture (27,28). Peak
the cause of the aneurysm. Aneurysms can also be classified wall stress distribution, which can be measured from the
according to their anatomic form (saccular, fusiform, true CT data in combination with patient-related factors such as
or false aneurysms) or their anatomic location (20). The blood pressure, seems to be superior for predicting rupture
latter allows classification of AAAs into thoracoabdominal risk compared with maximal aneurysm diameter measure-
aneurysms, suprarenal aneurysms (involvement of the aorta ments (29). Measuring peak wall stress might affect therapy
above the renal arteries), juxtarenal aneurysms (involvement strategy for small aneurysms with high wall stress and large
of the aorta at the level of the lowest renal artery or within aneurysms with low wall stress in patients at high risk for
5-mm distance from the lowest main renal artery), and in- surgery (30). Although research on aneurysm rupture and
frarenal aneurysms (with a neck of normal aorta between peak wall stress has been ongoing in the past years, the value
the renal arteries and the beginning of the aneurysm). of peak wall stress distribution for use in the routine clinical
The most important issue in evaluation of AAA is accu- practice has not yet been established.
rate diameter measurement, because the maximum diam- Contained rupture may be observed in patients pre-
eter of the aneurysm is an important predictor for rupture. senting with an acute symptomatic AAA if their condition
Aneurysms exceeding 5 cm in diameter have a higher rup- allows CT imaging (Fig. 34.6). CT imaging may show a
ture rate (22,23), whereas the risk of rupture of small (<5 cm) large retroperitoneal hematoma, and sometimes the location
AAA is quite low. Surveillance of AAA up to a diameter of of rupture (Fig. 34.7).
5.5 cm has been considered safe, unless rapid expansion
(>1 cm/y) or symptoms develop (24,25). In addition, there is
some uncertainty about the management of relatively small
aneurysms in females where aneurysm repair should be con-
sidered at a maximum aneurysm diameter of 5.2 cm (25).
Expansion rates of aneurysms are variable. Follow-up ex-
aminations are important for monitoring the expansion rate
(23). Diameter measurements are usually larger with CT
Figure 34.6. Axial computed tomography (CT) image in a patient

with an abdominal aneurysm with extension in both iliac arteries.
Image is taken at the level of the common iliac arteries. Contained Figure 34.7. Curved multiplanar reformation (MPR) of an acute
rupture of the left iliac aneurysm (arrowheads) with extension in the ruptured AAA. The site of aortic rupture is on the right (arrow) with
left psoas muscle. active bleeding (dotted arrow) and a large retroperitoneal hematoma (H).
LWBK1209-ch34_p558-574.indd 562 17/05/13 5:17 PM

and excludes the aneurysm from the circulation, thereby

preventing aneurysm rupture (32).
Randomized trials comparing conventional and endo-
vascular repair concluded that peri-operative mortality and
complications (the first 30 days after a procedure) were
lower with endovascular repair of AAA (37–39). However,
in recently published data, endovascular and open repair of
AAA resulted in similar rates of 6-year survival (40). In ad-
dition, the rate of secondary interventions was significantly
higher for endovascular repair (40). Therefore, the choice
between endovascular and open repair will mainly rely on
the preference of both patient and doctor and on aneurysm
anatomy.
Successful EVAR relies on accurate preoperative AAA im-
aging for patient selection (34). Detailed measurements must
be obtained, because endovascular stent-graft types and sizes
are selected per patient and based on these measurements. A
small error in stent-graft size can lead to major problems
such as endoleak, stent migration, or other complications
resulting in stent failure, the need for conversion to open
repair, or even aneurysm rupture (33).
Figure 34.8. CPR presentation through the aorta and right iliac
Most limitations for endovascular treatment of abdomi-
artery in a patient with an AAA. The line is perpendicular to the aor-
tic lumen. The distance displayed (80.1 mm) represents the distance
nal aneurysms are related to proximal neck anatomy of the
between the level of the lowest renal artery (small line at the top) and aneurysm (41–43). Evaluation of the neck requires measure-
the largest aneurysm diameter. Diameter and length measurements for ment of the neck length, tapering, degree of angulation, and
stent grafts are easy to perform with CPR presentations. evaluation of presence of mural thrombus. The length of the
neck is the distance between the most caudal renal artery
and the beginning of the aneurysm. When the neck length
is shorter than 15 mm, sealing problems are likely to occur
Exact evaluation of large aneurysms (>5 cm in diam- (44). Fenestrated stent grafts or stent grafts with suprarenal
eter) is important for clinical decision making and planning fixation can be used to solve this problem (45). Proximal
therapy, and should provide all anatomic information neces- fixation problems of the stent graft can occur when more
sary for planning surgery or endovascular aneurysm repair than a 2-mm reverse taper is present in the neck of the an-
(EVAR) (31). eurysm within the first centimeter below the renal arteries
Evaluation of aneurysms should include the location, (41). Angulation of the neck is defined as the angle between
length, and diameter of the aneurysm, the extension proxi- the longitudinal axis of the proximal aortic neck and the
mal and distal, the status of the neck of the aneurysm, and longitudinal axis of the aneurysm (Fig. 34.9). Increased risk
the diameter and length of the iliac arteries, and iliac aneu- for complications occurs with moderate (40 to 59 degrees)
rysms, when present (32). Abnormalities of renal, visceral, or severe (>59 degrees) angulation of the neck, compared
iliac, and femoral arteries, the number and size of the acces- with mild (<40 degrees) angulation, even if the neck length is
sory renal arteries and anatomic variants, and vessel con- more than 2 cm (46), although recently developed endopros-
figuration must also be evaluated (19,32). Furthermore, the theses might prevent these complications (47). Mural throm-
cause of the aneurysm should be sought (20). Previous tech- bus in the neck of the aneurysm may be another important
niques requiring both calibrated angiography and single- and possible contraindication for EVAR, because this can
slice CT, were invasive and did not visualize mural thrombus cause insufficient sealing resulting in endoleak or embo-
and the diameter of neck of the aneurysm and iliac arter- lism (48). Furthermore, nonsignificant aortic stenosis in the
ies adequately. The use of MDCT with 3D reconstructions region of the aneurysm neck may hamper expansion of the
allows all the information needed, thus eliminating x-ray stent graft and cause endoleak. A careful search for accessory
angiography as a preoperative examination method (33–35) renal arteries is required. Accessory renal arteries originating
(Fig. 34.8). from the aneurysm can cause endoleak after EVAR or lead
Unenhanced CT in the evaluation of AAA before EVAR to renal dysfunction when excluded from the circulation by
allows evaluation of calcifications. Contrast-enhanced CTA the stent graft. When small, these arteries may be embolized
is required for evaluating anatomic information and per- before EVAR (Fig. 34.10). Other relative contraindications
forming measurements, necessary for stent-graft selection. for EVAR are bilateral common iliac aneurysms that would
require stenting with exclusion of both internal iliac arteries
Endovascular Aortic Aneurysm Repair and Endoleak (44). Also, small vascular femoral access arteries may hinder
the use of EVAR. This problem occurs more commonly in
Since it was first used in the early 1990s (36), EVAR has women than in men (44).
become increasingly accepted for treatment of infrarenal AAA. Complications, especially late complications, are far more
The stent graft, composed of a wired metal frame and pros- frequent after EVAR compared with conventional abdomi-
thetic graft material, is inserted through the femoral arteries nal aneurysm surgery (40). Routine follow-up is therefore
LWBK1209-ch34_p558-574.indd 563 17/05/13 5:17 PM

30o
Figure 34.10. Coronal MIP presentation in a patient with a stent

graft for an infrarenal aortic aneurysm. Metallic artifacts next to the
stent graft (large arrows), caused by previous coiling of an accessory
renal artery arising from the aneurysm. Note the calcifications in the
excluded aortic aneurysm (small arrows).
It has therefore been suggested that using a single 60-sec-

Figure 34.9. Coronal MIP presentation (same patient as in Figure ond contrast material-enhanced acquisition may be all that
34.6). Mild angulation of the aortic neck (30 degrees). is required for surveillance after endovascular repair of AAA
(53).
In our hospital the first follow-up CT is usually performed
a few weeks after stent-graft placement; the second CT is
required, and CT is used most often (49). For the initial performed after 6 months and then annually. Knowledge of
post-EVAR CT we perform a triple-phase CT (unenhanced, normal morphologic changes and problems related to spe-
arterial phase, and delayed phase). The unenhanced scan cific types of stent grafts is important (54–57). The compli-
is essential to appreciate calcifications in the aneurysm sac cations most frequently reported are endoleak (Fig. 34.11),
which might otherwise be mistaken for an endoleak. The stent migration, stent-graft kinking (Fig. 34.12), stent-graft
arterial phase at initial follow-up is mainly used to detect thrombosis (Fig. 34.13), enlargement of the aneurysm, and
common femoral artery arteriovenous fistulas or pseudoa- infection (Fig. 34.14) that may cause pseudo-aneurysm
neurysms, which both are very short-term complications of formation (49,58). Signs of infection are soft tissue mass,
EVAR. On the delayed scans, endoleaks can reliably be in- stranding, fluid, and gas bubbles around the aorta (59). The
vestigated. For subsequent follow-up, only a delayed phase proximal neck of the aneurysm dilates after EVAR in up to
is acquired to assess possible endoleaks, stent migration or 30% of the patients. When this dilatation is severe, stent
other complications. Unless otherwise indicated, the use of migration can occur, resulting in endoleak (60). Other less
arterial phases in follow-up does not add to the detection frequent complications are mostly directly related to the
of complications, but does add substantially to the effective procedure, such as perforation of mural thrombus by the
dose delivered to the patient (50). delivery system, aortic dissection, shower embolism, and
Alternative scanning protocols can be used with recently hematoma at the arteriotomy site. Large bowel necrosis
clinically introduced dual-source CT systems. In these scan- can occur when both internal iliac arteries are overstented
ners, two separately operating tube and detector combina- (58).
tions are used for data acquisition. One of the advantages Early normal findings after EVAR are air in the aneurysm
of dual-source CT systems is that they allow to simulta- sac (Figs. 34.12 and 34.15), which is directly related to the
neously acquire CT datasets at two photon energy levels stent-graft placement. Mottled high attenuation around the
during a single acquisition (51). Typically energy levels of stent graft can be observed, probably caused by thrombo-
80 and 140 kVp are used, allowing reconstruction of pure sis of previous patent lumen, and occasionally calcifications
80 and 140 kVp datasets. Combining these datasets gener- are present in the thrombus, displaced from the aortic wall
ates a virtual 120 kVp dataset, which is comparable to a (61). The latter can be confusing, but comparison with pre-
standard 120 kVp dataset. In addition, using postprocessing al- operative CT examinations can be useful. These early find-
gorithms, virtual non-enhanced datasets can be generated (52). ings must not be confused with endoleak, which is also an
LWBK1209-ch34_p558-574.indd 564 17/05/13 5:17 PM

A B
Figure 34.11. Mid-sagittal MPR presentation. Patient with an AAA after stent-graft placement. In panel
(A) the initial CT after stent-graft placement is shown with correct position of the graft. During follow up
(panel B), dilatation of the neck of the aneurysm occurred, leading to insufficient sealing and migration of the
stent-graft. Endoleak type 1a (arrow) resulting from insufficient proximal sealing develloped.
early finding (61). Most patients show a decrease in their

aneurysm sac diameter over time after EVAR. Since the
aneurysm sac is excluded from the circulation, expanding
stops and the diameter of the aneurysm sac regresses (62).
By comparing the maximum diameter of different follow-
up examinations, shrinkage or enlargement of the aneurysm
can be noticed, but minor degrees of changes are difficult
to detect (63–65). The maximum diameter of the aneurysm,
measured perpendicular to the vascular axis on axial im-
ages, provides an easy measure for size follow-up. A more
precise approach to compare size changes of the aneurysm
is measuring volumes, which can be done easily with CT
(63,66) (Fig. 34.16). Sometimes the maximal diameter of
the aneurysm decreases or remains unchanged, whereas the
aneurysm volume increases, in the presence of endoleak or
endotension. In these cases the presence of expansion may
only be detected with volumetric changes. Volumetric mea-
surement is performed by measuring areas along the entire
aneurysm every 5 mm. Using dedicated software, 3D images
of the aneurysm can be reconstructed and volumes can be
calculated (Fig. 34.16).
Endoleak is defined as persistent blood flow outside the
Figure 34.12. Coronal CPR presentation through the distal aorta stent graft, but within the aneurysm sac. Endoleak is a fre-
and left iliac artery after stent-graft placement. Kinking of the left leg of quent complication (64). Classification is important, and
the prothesis (arrow) that resulted in sustained claudication of the left leg. four different categories can be distinguished.
LWBK1209-ch34_p558-574.indd 565 17/05/13 5:17 PM

Figure 34.14. Axial CT image in the delayed phase. Patient with

a stent graft for infrarenal AAA. Thickened aortic wall with periaortic
enhancing soft tissue mass (M) and obliteration of the fat pad (arrow)
against the psoas muscle (P). These signs, as well as clinical and labo-
ratory findings, were suggestive of infection. Patient underwent open
reconversion, after which the diagnosis infection was confirmed.
placement. When present, this is almost always identified at

the time of placement and occurs in patients who are fully
anticoagulated. Treatment is rarely required.
Endotension, defined as expansion of the aneurysm sac,
is sometimes classified as type 5 endoleak (67). Endotension
may occur with or without endoleak (64). The precise cause of
Figure 34.13. Frontal view of a 3D VR presentation of a patient

with an abdominal stent graft for an infrarenal AAA. Migration of the
left leg (that was a first-generation custom-made stent graft) caused
thrombosis of the left leg. Patient sustained left lower leg claudication
and was treated with an aortounilateral stent graft to the right common
iliac artery with surgical femoral–femoral crossover bypass to the left leg.
Type 1 endoleaks are attachment site endoleaks at the

proximal (1a) (Fig. 34.11) or distal (1b) site or from the iliac
artery occluder. These types of endoleak always require in-
tervention. Type 2 is the most frequent type of endoleak and
is caused by collateral flow to the excluded aneurysm sac,
typically from the lumbar arteries or inferior mesenteric artery.
Type 2 is further divided into type 2a, which is a “simple”
endoleak (Figs. 34.17 and 34.18) with one patent collateral
branch, and type 2b, which is a “complex” endoleak with
two or more patent branches. Most type 2 endoleaks resolve
spontaneously, but treatment is required when the aneurysm
expands. Type 3 endoleak is caused by structural stent-graft
failure, for example, stent fracture or junctional separation. Figure 34.15. Axial CT image. Small amount of air anterior in
Treatment is always required. Type 4 endoleak is caused the aneurysm sac (arrow), 2 days after stent-graft placement. This is a
by porosity of the stent graft, within 30 days after stent frequently observed finding immediately postoperative.
LWBK1209-ch34_p558-574.indd 566 17/05/13 5:17 PM

A B
Figure 34.16. Volume rendered 3D images of the aneurysm sac in a patient after endovascular repair. By
drawing contours around the borders of the aneurysm sac on axial CT images, dedicated software can recon-
struct a 3D image of the aneurysm sac (blue) and calculate the aneurysm sac volume. Note the increase in vol-
ume (from 358 mL in panel (A) to 425 mL in panel (B)) was much easier appreciated on the 3D than on the
axial images. The CT in panel (B) was performed 1 year after the CT in panel (A).
endotension without endoleak is unknown. If the aneurysm Magnetic resonance angiography (MRA) is an alterna-
sac does not decrease in size or volume, pressure measure- tive to CTA for evaluating patients after EVAR, especially in
ment directly in the aneurysm sac can be performed (Fig. patients with contraindications for CTA, such as renal dys-
34.19). When arterial pressure gradients are measured, in- function or allergy to iodinated contrast material. Several re-
tervention is required. cent studies even showed that MRI/MRA is superior to CTA
Figure 34.18. Three-dimensional reconstruction in a patient after

Figure 34.17. Axial CT image in a patient after endovascular AAA stent-graft placement. Large type 2a endoleak (E) caused by the infe-
repair. Endoleak type 2a (arrow) caused by inflow from a lumbar artery. rior mesenteric artery (arrow).
LWBK1209-ch34_p558-574.indd 567 17/05/13 5:18 PM

Preoperative anatomic screening starts with a US examina-

tion. When contraindications for donation are observed,
no other investigation is required. Further investigation can
be done with CT or MRI. The drawback of CT is the use
of ionizing radiation because most individuals assessed for
living-related kidney donation are healthy young adults and
ionizing radiation is not used in MRI. However imaging
assessment of potential renal donors demands high spatial
and temporal resolution of the imaging modality. It has not
been demonstrated if MR angiographic systems can compete
with the currently used multi-detector row CT units in terms
of these resolution parameters. Furthermore, it has been
shown that multi-detector row CT enables highly accurate
assessment of the renal anatomy including small accessory
renal arteries in living-related kidney donor candidates (74)
and correctly predicts the renal vascular anatomy in 96% of
patients (74). Therefore, at our institution we mainly use CT
and CTA for evaluating living-related kidney donors.
The donor must have two normal kidneys before donat-
ing one (73). Any renal disease forms a contraindication for
Figure 34.19. Same patient as in Figure 34.18. Axial posterior– donation. Additional relative contraindications are multiple
anterior CT-fluoroscopy image showing direct puncture of the aneu- renal arteries, early arterial branching, complex venous
rysm. This was performed because the volume of the aneurysm sac anatomy, or duplication anomalies (71,75). The left kidney
remained unchanged in size, to measure the pressure in the aneurysm is preferred for laparoscopic donation, because the left renal
sac. Arterial pressure gradients were measured (endotension). The infe- artery and vein are longer compared with the right, which is
rior mesenteric artery was embolized to prevent aneurysm enlargement
technically more convenient for surgery (75).
and rupture (not shown).
To limit radiation dose, unenhanced imaging for detecting
renal and ureteral calculi is usually not performed (73). When
reviewing cross-sectional images carefully in arterial phase im-
aging, most calculi are still detected easily (73,76). Dual-phase
in detecting endoleaks when dedicated imaging protocols
imaging, with an arterial phase (15- to 30-second delay) and a
are used (68–70). MRA is capable of demonstrating even
venous phase (55-second delay) or nephrogenic phase (75- to
very small endoleaks which can cause endotension. Since
100-second delay), is most often performed (73).
treatment of endotension without endoleak are limited, MR
Vascular anatomy is very important in laparoscopic
angiography or contrast-enhanced ultrasound should be
surgery (73,77). Venous bleeding during surgery can cause
performed as part of an exhaustive search for an endoleak in
serious complications, requiring conversion to an open pro-
patients with unexplained endotension (71).
cedure (75). Arterial phase imaging allows good visualiza-
However, with MRA, image quality and occurrence
tion of the renal arteries and the main renal veins generally.
of artifacts caused by the metallic stent grafts are highly
The adrenal, lumbar, and gonadal veins can be evaluated on
dependent on the composition of the stent graft (71) and
venous phase imaging. Anatomic variants of the renal arter-
therefore, in some patients MRA will not be helpful. X-ray
ies occur frequently. Approximately 70% of the potential
angiography is sometimes necessary for endoleak classifica-
donors have a bilateral single renal artery (73,75). The num-
tion and directly visualizes the source of endoleak. When an
ber of arteries, their size, and the location of origin from
endoleak is present, lumbar arteries or a patent inferior mes-
the aorta or iliac arteries and the length of the renal artery
enteric artery can enhance imaging secondarily because of out-
from its origin to its first bifurcation are important findings
flow from the aneurysm into these collateral vessels in the case
(Fig. 34.20). Up to four renal arteries may exist at each side.
of a type 1 or type 3 endoleak or directly because of inflow
Scanning should also include the celiac artery, mesenteric ar-
from these collateral vessels into the aneurysm (67). Recently,
teries, and iliac arteries, because occasionally accessory renal
contrast-enhanced ultrasound has been shown to be an effec-
arteries can arise from these arteries (73). Accessory renal
tive tool for surveillance after EVAR with similar accuracy for
arteries are usually small and may be missed when the spa-
detecting endoleaks as MRA. Ultrasound is fast and cheaper
tial resolution of the imaging tool is not sufficient (78). Early
than CTA or MRA. Limitations of CEUS are mainly due to its
pre-hilar branching must also be recognized preoperatively.
operator dependence and patients’ habitus (72).
When an artery branches within the first 2 cm from its ori-
gin, too little space is left for adequate surgical anastomosis
Renal Applications (Fig. 34.21) (73).
Variants in renal veins are less frequently observed
Living-related Kidney Donors
(73–76). The most frequent abnormality is a circumaortic
Laparoscopically assisted nephrectomy in a living-related left renal vein (5% to 7%) or a retroaortic left renal vein
kidney donor is currently the preferred method for kidney (2% to 3%) (Fig. 34.22) (73,76). The anomalous venous
donation (73). Preoperative anatomic evaluation is impor- drainage in the retroaortic component is important, which
tant, because the field of view is limited during laparoscopy. may cause severe bleeding during laparoscopic surgery. The
LWBK1209-ch34_p558-574.indd 568 17/05/13 5:18 PM

LRV
Figure 34.22. CPR presentation of the retroaortic part of a cir-

cumaortic left renal vein. A, aorta; LRV, left renal vein.
Figure 34.20. Coronal MIP presentation in a patient with polar

hypertension do not have renal artery stenosis, radiologic
arteries for the upper and lower pole of the left kidney (arrows).
Metallic artifacts are caused by previous coiling of an accessory renal
evaluation is only needed in patients with a suspected
artery (dotted arrows) for acute renal bleeding. Note a separate renal renovascular cause. Clinical symptoms that may indicate
vein for the upper and lower poles (arrowheads), running parallel with renovascular hypertension are hypertension at a young
the arteries at the left, which is an infrequent finding. age, sudden onset of hypertension, and rapid deterioration
of hypertension in patients treated with medication (80).
Renovascular hypertension is usually caused by athero-
sclerotic renal artery stenosis or fibromuscular dysplasia,
left gonadal vein and left adrenal veins usually drain into accounting for approximately 90% and less than 10%,
the left renal vein, whereas the right gonadal and adrenal respectively (79).
veins usually drain directly into the inferior vena cava (73). CTA has a high sensitivity and specificity for detecting
Lumbar veins can also drain into the renal vein (75). All of renal artery stenosis, with a high negative predictive value;
the arterial and venous structures should be evaluated care- normal renal arteries at MDCT angiography almost always
fully before the donation operation can take place. excludes renal artery stenosis (80). On the other hand when
The excretion system can be evaluated with delayed im- a stenosis is detected in the renal artery, it is not always the
aging with abdominal x-ray or a CT topogram, instead of a cause of hypertension (80). In patients with fibromuscular
delayed CT acquisition to limit radiation dose. A duplicated dysplasia, typically web-like stenosis may be visualized, but
collecting system can then be detected easily (73). abnormalities causing symptoms may also be missed com-
pletely with (CT) imaging (80). When fibromuscular dyspla-
Renal Artery Stenosis sia with hypertension is suspected but not visualized with
CT or MRI, x-ray angiography with pressure measurement
Hypertension can be caused by renovascular disease in a in the renal artery should be performed.
small number of patients (79). Since most patients with CT evaluation of renal artery stenosis starts with the
size and shape of the kidneys. All renal arteries must be
evaluated for the presence of stenosis. Axial images and
MPR, CPR, MIP, or VR presentations can be used. When a
stenosis is present it may be classified as nonsignificant ste-
nosis (with <50% diameter reduction of the vessel lumen)
or significant stenosis (with >50% diameter reduction)
(Fig. 34.23). Sometimes it is difficult to quantify a steno-
sis as either significant or nonsignificant. X-ray angiography
with pressure measurement can be used in case of doubt.
When a significant stenosis is thus detected, balloon an-
gioplasty can be performed immediately. Secondary signs,
for example, poststenotic dilatation, decreased size of the
ipsilateral kidney, and delayed enhancement can also be no-
ticed, and sustain the diagnosis of probably significant renal
artery stenosis (80).
MDCT is well suited for evaluation of renal artery steno-
sis. However, we usually prefer MRI and MRA for evaluating
patients with suspected renal artery stenosis. This is because
Figure 34.21. Three-dimensional image of the aorta and renal impaired renal function is often present in these patients,
arteries shows early branching of both renal arteries (arrows), a rela- which is a relative contraindication for administering iodin-
tive contraindication for laparoscopic donation. ated contrast agent.
LWBK1209-ch34_p558-574.indd 569 17/05/13 5:18 PM

Vascular Liver Imaging

Treatment of liver malignancies, either primary or second-
ary, is increasingly performed. Surgery, liver transplanta-
tion, radiofrequency ablation, intra-arterial chemotherapy,
and intra-arterial radioembolization are techniques applied.
Therapeutic strategies depend on radiologic findings. Dedicated
vascular CT protocols are necessary for adequate visualiza-
tion of the liver vasculature before surgical intervention.
This is especially true for preoperative assessment for liver
transplantation (85,86).
Living-donor liver transplantation was first developed to
overcome organ shortage in the pediatric population, but it
is also increasingly used in adults, mainly for the same reason
(87). The goal of CT for potential liver donors is to evaluate
the liver parenchyma and the hepatic and mesenteric anatomy
(87–89).
Characterization of focal liver lesions requires multiphase
imaging, combining different phases. Unenhanced phase, early
arterial phase, late arterial phase, portal venous phase, and
delayed phase can be used, depending on the type of lesion sus-
pected (85,86). The early arterial phase allows good evaluation
of the hepatic arteries and mesenteric circulation (85). The por-
tal venous phase is used for the portal venous anatomy (90).
Liver volumes should be measured before partial liver
resection to estimate postoperative success in case of a living
donor for transplantation, or for patients in whom partial
liver resection for malignancies is planned. Liver insufficiency
can occur postoperatively if the remaining liver volume is
not enough (90). In case of partial liver resection for malig-
nancies, embolization of the vena portae of the diseased part
of the liver can be performed to induce hypertrophy of the
remaining part of the liver (91). Assessment of the hepatic
Figure 34.23. CPR presentation through the proximal right renal arterial anatomy is of utmost importance for living-donor
artery.Significant stenosis at the origin (arrow). liver transplantation. Anatomic hepatic artery variants are
difficult to transplant because these arteries often are very
small and may cause anastomotic problems. This problem
does not occur during partial liver resection for malignancy.
Other Applications Classic vascular anatomy of the celiac trunk and mesen-
teric vessels is present only in approximately 50% of the
The most common primary tumor of the kidney is renal cell
population (Fig. 34.24) (87–89). The most common variants
carcinoma (81). CT imaging for renal tumors has a dual
are a right replaced hepatic artery, arising from the superior
role. First, CT can be used for diagnosis and characteriz-
mesenteric artery (Fig. 34.25), and a left replaced hepatic
ing renal masses suspect for malignancy. Second, CT allows
artery, arising from the left gastric artery (85). An important
evaluation of the relationship between the tumor and the
variant in potential living liver donors is the arterial supply
major vessels and collecting system, and comprises the ana-
of segment IV, which normally arises from the left hepatic
tomic information necessary to decide whether a patient is
artery (Fig. 34.26). When this artery arises from the right
suited for nephron-sparing surgery (81). Patients with a sin-
hepatic artery, left liver lobe donation cannot be performed
gle kidney (e.g., after prior nephrectomy for tumor surgery)
because adequate surgical anastomosis cannot be obtained.
obtain great benefit when partial nephrectomy is performed,
When there are doubts about the arterial anatomy, if one
because dialysis may be avoided (82).
or more hepatic arteries are not visualized by MDCT, or
Evaluation of renal masses requires multiphase CT,
when stenosis in the hepatic artery or their supplying arter-
combining an unenhanced phase, a corticomedullary
ies is suspected, x-ray angiography should be performed
phase, a nephrogenic phase, and an excretory phase (83).
(89). Variants may be present in the venous system as well.
Vascular analysis should reveal the arterial and venous
The most frequently reported venous variant is an accessory
anatomy and variants, and possible extension of tumor-
right inferior hepatic vein (87). The biliary anatomy should
thrombus into the renal veins. The vascular protocol is the
be assessed as well (87).
same as that used for living renal donors described earlier.
The surgical approach depends on the radiologic findings.
Mesenteric Ischemia
An alternative that is increasingly used, when surgery is
not preferable, is radiofrequency ablation or cryoablation CT examination of patients with suspected mesenteric isch-
(84). emia involves evaluation of both the mesenteric vessels and
LWBK1209-ch34_p558-574.indd 570 17/05/13 5:18 PM

Figure 34.26. Coronal maximum intensity projection (MIP)

presentation of the hepatic artery showing a separate artery for seg-
ment IV (arrow), where this artery usually arises from the left hepatic
artery. This variant anatomy is a contraindication for living-related
liver donation.
pneumatosis intestinalis (92–94). Acute mesenteric ischemia

is caused by severe stenosis or occlusion of the superior
mesenteric artery or vein, usually by an acute embolus or
thrombosis. The severity, size, and site of the bowel affected
Figure 34.24. Three-dimensional VR presentation of the normal depends on the cause of ischemia, the location and extent of
hepatic anatomy. The left gastric artery (LGA), common hepatic artery
the stenosis or occlusion, the duration of the process, and the
(CHA), and splenic artery (SA) arise from the celiac trunk. The common
hepatic artery divides into gastroduodenal artery (GDA) and proper presence of collateral vessels (93). Mesenteric venous throm-
hepatic artery (PHA). The latter divides into left hepatic artery (LHA) bosis may present with subacute symptomatology over 1 to
and right hepatic artery (RHA). SMA, superior mesenteric artery. 4 weeks (93,94).
Chronic mesenteric ischemia is known as abdominal an-
gina and usually caused by atherosclerosis (92). Symptoms
of abdominal epigastric pain typically occur postprandial,
the bowel walls (92,93). Mesenteric ischemia can be classified
when increased blood flow is required in the splanchnic ar-
as acute, subacute, or chronic. CT signs of acute mesenteric
teries (celiac trunk, superior mesenteric artery, and inferior
ischemia are bowel wall thickening and edema, resulting in
mesenteric artery). Chronic mesenteric ischemia generally
low attenuation, submucosal hemorrhage reflected by high
occurs only when two or all three arteries are significantly
attenuation, mesenteric stranding, mesenteric fluid, and
stenosed, because of the presence of extensive collateral ves-
sels between the splanchnic arteries (92). The diagnosis is
made after exclusion of other abnormalities.
CT allows excellent visualization of the bowel walls,
arterial anatomy with its collateral vessels, vascular vari-
ants, and anomalies (95–97). Arterial phase imaging allows
optimal visualization of the splanchnic arteries. CPR is very
useful for evaluating the degree of stenosis, because per-
pendicular CPR images centered on the central lumen line
SA
LHA drawn through the stenosis are displayed simultaneously
(Figs. 34.27 and 34.28). Most atherosclerotic stenoses are
located proximally in the affected artery (92). Stenosis can
be graded as nonsignificant or significant. Secondary signs
RHA of stenosis such as poststenotic dilatation are generally rec-
ognized well. An acute embolus can be diagnosed as a filling
defect in an enhanced artery.
GDA SMA
Intra-Abdominal Arterial Bleeding

MDCT is an excellent tool for assessing intra-abdominal
bleeding, for instance after trauma of abdominal surgery.
Reliable and timely localization of a bleeding source is impor-
tant for determining the therapeutic strategy. The presence
of active extravasation—which is contrast material leaking
from injured vessels or organs—is an indicator of morbid-
ity and mortality. Single-phase CT protocols might diagnose
Figure 34.25. Three-dimensional VR presentation image of a active extravasation, however dual-phase CT protocols
patient with a replaced RHA, arising from the superior mesenteric have higher sensitivity for the diagnosis of active extrava-
artery (SMA). See text of Figure 34.24 for abbreviations. sation. The classic pattern of active extravasation is a jet
LWBK1209-ch34_p558-574.indd 571 17/05/13 5:18 PM

Figure 34.27. Central lumen reformation through the celiac trunk.

This line is drawn manually when scrolling through the axial images.
The resulting sagittal CPR presentation is shown in Figure 34.28.
or focal area of hyperattenuation (usually >100 HU) within

a hematoma on initial images that fades into an enlarged,
enhanced hematoma on delayed images (98) (Fig. 34.29). It
is important to distinguish active extravasation from other
high-attenuating entities such as a pseudoaneurysm, bone
fragments, calcifications or (surgical) foreign bodies. Pseudo-
aneurysms, in contrast to extravasation, usually are well
defined, have a round or oval shape, and the neck that con-
nects the pseudoaneurysm with the accompanying artery can Figure 34.28. Sagittal CPR presentation shows 50% stenosis at
be visualized most of the times (98). Furthermore, the area the origin of the celiac trunk (arrow).
A B C
Figure 34.29. Coronal images of an arterial bleeding after high energetic trauma.In panel (A), a scan
without intravenous contrast shows a large retroperitoneal hematoma (arrow). In panel (B), the arterial phase
shows contrast extravasation (arrow), which further increases in the delayed phase in panel (C) (arrow).
LWBK1209-ch34_p558-574.indd 572 17/05/13 5:18 PM

of hyperattenuation in pseudoaneurysms does not enlarge 23. Glimaker H, Holmberg L, Elvin A, et al. Natural history of patients with abdominal aortic
aneurysm. Eur J Vasc Surg. 1991;5(2):125–130.
in delayed imaging. Likewise the hyperattenuating foreign 24. Lederle FA, Kane RL, MacDonald R, et al. Systematic review: Repair of unruptured
bodies, calcifications or bone fragments do also appear simi- abdominal aortic aneurysm. Ann Intern Med. 2007;146(10):735–741.
25. Brewster DC, Cronenwett JL, Hallett JW Jr, et al. Guidelines for the treatment of abdomi-
lar in the delayed phase as compared to the early phase. As nal aortic aneurysms. Report of a subcommittee of the Joint Council of the American
mentioned earlier, in active extravasation, the hyperattenu- Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg. 2003;
ating area is enlarging in the delayed phase. 37(5):1106–1117.
26. Sprouse LR, Meier GH III, Lesar CJ, et al. Comparison of abdominal aortic aneurysm
diameter measurements obtained with ultrasound and computed tomography: Is there a
difference? J Vasc Surg. 2003;38(3):466–471.
Summary 27. Fillinger MF, Racusin J, Baker RK, et al. Anatomic characteristics of ruptured abdominal
aortic aneurysm on conventional CT scans: Implications for rupture risk. J Vasc Surg.
2004;39(6):1243–1252.
CTA has rapidly gained wide acceptance for imaging of 28. Nicholls SC, Gardner JB, Meissner MH, et al. Rupture in small abdominal aortic aneu-
rysms. J Vasc Surg. 1998;28(5):884–888.
the aorta and its abdominal side branches because of its 29. Fillinger MF, Marra SP, Raghavan ML, et al. Prediction of rupture risk in abdominal aortic
fast imaging with high spatial resolution. The introduction aneurysm during observation: Wall stress versus diameter. J Vasc Surg. 2003;37(4):724–732.
of 16-row and 64-row MDCT has especially advanced the 30. Fillinger MF, Raghavan ML, Marra SP, et al. In vivo analysis of mechanical wall stress and
abdominal aortic aneurysm rupture risk. J Vasc Surg. 2002;36(3):589–597.
technology and provides an alternative to x-ray angiogra- 31. Qanadli SD, Mesurolle B, Coggia M, et al. Abdominal aortic aneurysm: Pretherapy assess-
phy and MRA for several indications. More recent introduc- ment with dual-slice helical CT angiography. AJR Am J Roentgenol. 2000;174(1):181–187.
tion of 320 slice and dual-source CT scanners even further 32. Rydberg J, Kopecky KK, Lalka SG, et al. Stent grafting of abdominal aortic aneurysms: Pre-
and postoperative evaluation with multislice helical CT. J Comput Assist Tomogr. 2001;
expanded the possibilities of CT-scanning. Further studies 25(4):580–586.
should evaluate the cost-effectiveness of MDCT for specific 33. Wyers MC, Fillinger MF, Schermerhorn ML, et al. Endovascular repair of abdominal aortic
aneurysm without preoperative arteriography. J Vasc Surg. 2003;38(4):730–738.
abdominal indications, as has been performed for peripheral 34. Blum U, Voshage G, Lammer J, et al. Endoluminal stent-grafts for infrarenal abdominal
vascular imaging (99). Optimized injection protocols and aortic aneurysms. N Engl J Med. 1997;336(1):13–20.
specific contrast medium concentrations are required for 35. Willmann JK, Wildermuth S, Pfammatter T, et al. Aortoiliac and renal arteries: Prospective
intraindividual comparison of contrast-enhanced three-dimensional MR angiography and
further optimization. multi-detector row CT angiography. Radiology. 2003;226(3):798–811.
36. Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for
abdominal aortic aneurysms. Ann Vasc Surg. 1991;5(6):491–499.
37. Prinssen M, Verhoeven EL, Buth J, et al. A randomized trial comparing conventional and endo-
References vascular repair of abdominal aortic aneurysms. N Engl J Med. 2004;351(16):1607–1618.
38. Greenhalgh RM, Brown LC, Kwong GP, et al. Comparison of endovascular aneurysm repair
1. Rydberg J, Buckwalter KA, Caldemeyer KS, et al. Multisection CT: Scanning techniques with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day opera-
and clinical applications. Radiographics. 2000;20(6):1787–1806. tive mortality results: Randomised controlled trial. Lancet. 2004;364(9437):843–848.
2. Hu H, He HD, Foley WD, et al. Four multidetector-row helical CT: Image quality and 39. Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs open
volume coverage speed. Radiology. 2000;215(1):55–62. repair of abdominal aortic aneurysm: A randomized trial. JAMA. 2009;302(14):1535–1542.
3. Prokop M. General principles of MDCT. Eur J Radiol. 2003;45:S4–S10. 40. De Bruin JL, Baas AF, Buth J, et al. Long-term outcome of open or endovascular repair of
4. Saini S. Multi-detector row CT: Principles and practice for abdominal applications. abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1881–1889.
Radiology. 2004;233(2):323–327. 41. Dillavou ED, Muluk SC, Rhee RY, et al. Does hostile neck anatomy preclude successful
5. Kroft LJ, Roelofs JJ, Geleijns J. Scan time and patient dose for thoracic imaging in neonates endovascular aortic aneurysm repair? J Vasc Surg. 2003;38(4):657–663.
and small children using axial volumetric 320-detector row CT compared to helical 64-, 42. Mohan IV, Laheij RJ, Harris PL. Risk factors for endoleak and the evidence for stent-graft
32-, and 16-detector row CT acquisitions. Pediatr Radiol. 2010;40(3):294–300. oversizing in patients undergoing endovascular aneurysm repair. Eur J Vasc Endovasc Surg.
6. Napoli A, Fleischmann D, Chan FP, et al. Computed tomography angiography: State-of- 2001;21(4):344–349.
the-art imaging using multidetector-row technology. J Comput Assist Tomogr. 2004;28 43. Stanley BM, Semmens JB, Mai Q, et al. Evaluation of patient selection guidelines for endo-
(suppl 1):S32–S45. luminal AAA repair with the Zenith Stent-Graft: The Australasian experience. J Endovasc
7. Rubin GD. MDCT imaging of the aorta and peripheral vessels. Eur J Radiol. 2003;45(suppl Ther. 2001;8(5):457–464.
1):S42–S49. 44. Carpenter JP, Baum RA, Barker CF, et al. Impact of exclusion criteria on patient selection
8. Rubin GD, Shiau MC, Leung AN, et al. Aorta and iliac arteries: Single versus multiple for endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2001;34(6):1050–1054.
detector-row helical CT angiography. Radiology. 2000;215(3):670–676. 45. Verhoeven EL, Prins TR, Tielliu IF, et al. Treatment of short-necked infrarenal aortic
9. Fleischmann D. Use of high concentration contrast media: Principles and rationale-vascular aneurysms with fenestrated stent-grafts: Short-term results. Eur J Vasc Endovasc Surg.
district. Eur J Radiol. 2003;45(suppl 1):S88–S93. 2004;27(5):477–483.
10. Macari M, Israel GM, Berman P, et al. Infrarenal abdominal aortic aneurysms at multi- 46. Sternbergh WC III, Carter G, York JW, et al. Aortic neck angulation predicts adverse outcome
detector row CT angiography: Intravascular enhancement without a timing acquisition. with endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2002;35(3):482–486.
Radiology. 2001;220(2):519–523. 47. Rouwet EV, Torsello G, de Vries JP, et al. Final results of the prospective European trial
11. Ho LM, Nelson RC, Thomas J, et al. Abdominal aortic aneurysms at multi-detector row of the Endurant stent graft for endovascular abdominal aortic aneurysm repair. Eur J Vasc
helical CT: Optimization with interactive determination of scanning delay and contrast Endovasc Surg. 2011;42(4):489–497.
medium dose. Radiology. 2004;232(3):854–859. 48. Gitlitz DB, Ramaswami G, Kaplan D, et al. Endovascular stent grafting in the presence of
12. Fleischmann D, Rubin GD, Bankier AA, et al. Improved uniformity of aortic enhance- aortic neck filling defects: Early clinical experience. J Vasc Surg. 2001;33(2):340–344.
ment with customized contrast medium injection protocols at CT angiography. Radiology. 49. Armerding MD, Rubin GD, Beaulieu CF, et al. Aortic aneurysmal disease: Assessment of stent-
2000;214(2):363–371. graft treatment-CT versus conventional angiography. Radiology. 2000;215(1):138–146.
13. Rubin GD. 3-D imaging with MDCT. Eur J Radiol. 2003;45(suppl 1):S37–S41. 50. Macari M, Chandarana H, Schmidt B, et al. Abdominal aortic aneurysm: Can the arterial
14. Prokop M. Multislice CT angiography. Eur J Radiol. 2000;36:86–96. phase at CT evaluation after endovascular repair be eliminated to reduce radiation dose?
15. Cody DD. AAPM/RSNA physics tutorial for residents: Topics in CT. Image processing in Radiology. 2006;241(3):908–914.
CT. Radiographics. 2002;22(5):1255–1268. 51. Johnson TR, Krauss B, Sedlmair M, et al. Material differentiation by dual energy CT:
16. Nino-Murcia M, Jeffrey RB Jr, Beaulieu CF, et al. Multidetector CT of the pancreas and Initial experience. Eur Radiol. 2007;17(6):1510–1517.
bile duct system: Value of curved planar reformations. AJR Am J Roentgenol. 2001;176(3): 52. Scheffel H, Stolzmann P, Frauenfelder T, et al. Dual-energy contrast-enhanced computed
689–693. tomography for the detection of urinary stone disease. Invest Radiol. 2007;42(12):823–829.
17. Raman R, Napel S, Rubin GD. Curved-slab maximum intensity projection: Method and 53. Chandarana H, Godoy MC, Vlahos I, et al. Abdominal aorta: Evaluation with dual-source
evaluation. Radiology. 2003;229(1):255–260. dual-energy multidetector CT after endovascular repair of aneurysms–initial observations.
18. Sun Z, Gallagher E. Multislice CT virtual intravascular endoscopy for abdominal aortic Radiology. 2008;249(2):692–700.
aneurysm stent grafts. J Vasc Interv Radiol. 2004;15(9):961–970. 54. Tillich M, Hausegger KA, Tiesenhausen K, et al. Helical CT angiography of stent-grafts
19. Singh K, Bonaa KH, Jacobsen BK, et al. Prevalence of and risk factors for abdominal aortic in abdominal aortic aneurysms: Morphologic changes and complications. Radiographics.
aneurysms in a population-based study: The Tromso Study. Am J Epidemiol. 2001;154 1999;19(6):1573–1583.
(3):236–244. 55. Mita T, Arita T, Matsunaga N, et al. Complications of endovascular repair for thoracic and
20. Chaikof EL, Blankensteijn JD, Harris PL, et al. Reporting standards for endovascular aor- abdominal aortic aneurysm: An imaging spectrum. Radiographics. 2000;20(5):1263–1278.
tic aneurysm repair. J Vasc Surg. 2002;35(5):1048–1060. 56. Dattilo JB, Brewster DC, Fan CM, et al. Clinical failures of endovascular abdominal aortic
21. Eggebrecht H, Baumgart D, Schmermund A, et al. Penetrating atherosclerotic ulcer of the aorta: aneurysm repair: Incidence, causes, and management. J Vasc Surg. 2002;35(6):1137–1144.
Treatment by endovascular stent-graft placement. Curr Opin Cardiol. 2003;18(6):431–435. 57. Haulon S, Devos P, Willoteaux S, et al. Risk factors of early and late complications in
22. Nevitt MP, Ballard DJ, Hallett JW Jr. Prognosis of abdominal aortic aneurysms. A population- patients undergoing endovascular aneurysm repair. Eur J Vasc Endovasc Surg. 2003;25(2):
based study. N Engl J Med. 1989;321(15):1009–1014. 118–124.
LWBK1209-ch34_p558-574.indd 573 17/05/13 5:18 PM

58. Rydberg J, Kopecky KK, Lalka SG, et al. Stent grafting of abdominal aortic aneurysms: 78. Rankin SC, Jan W, Koffman CG. Noninvasive imaging of living related kidney donors:
Pre-and postoperative evaluation with multislice helical CT. J Comput Assist Tomogr. Evaluation with CT angiography and gadolinium-enhanced MR angiography. AJR Am J
2001;25(4):580–586. Roentgenol. 2001;177(2):349–355.
59. Macedo TA, Stanson AW, Oderich GS, et al. Infected aortic aneurysms: Imaging findings. 79. Safian RD, Textor SC. Renal-artery stenosis. N Engl J Med. 2001;344(6):431–442.
Radiology. 2004;231(1):250–257. 80. Fleischmann D. Multiple detector-row CT angiography of the renal and mesenteric vessels.
60. Napoli V, Sardella SG, Bargellini I, et al. Evaluation of the proximal aortic neck enlarge- Eur J Radiol. 2003;45(suppl 1):S79–S87.
ment following endovascular repair of abdominal aortic aneurysm: 3-years experience. Eur 81. Sheth S, Scatarige JC, Horton KM, et al. Current concepts in the diagnosis and man-
Radiol. 2003;13(8):1962–1971. agement of renal cell carcinoma: Role of multidetector CT and three-dimensional CT.
61. Sawhney R, Kerlan RK, Wall SD, et al. Analysis of initial CT findings after endovascular Radiographics. 2001;21 Spec No:S237–S254.
repair of abdominal aortic aneurysm. Radiology. 2001;220(1):157–160. 82. Catalano C, Fraioli F, Laghi A, et al. High-resolution multidetector CT in the preoperative eval-
62. Farner MC, Carpenter JP, Baum RA, et al. Early changes in abdominal aortic aneurysm uation of patients with renal cell carcinoma. AJR Am J Roentgenol. 2003;180(5):1271–1277.
diameter after endovascular repair. J Vasc Interv Radiol. 2003;14(2 Pt 1):205–210. 83. Foley WD. Renal MDCT. Eur J Radiol. 2003;45(suppl 1):S73–S78.
63. Kritpracha B, Beebe HG, Comerota AJ. Aortic diameter is an insensitive measurement of 84. Gervais DA, McGovern FJ, Arellano RS, et al. Renal cell carcinoma: Clinical experience
early aneurysm expansion after endografting. J Endovasc Ther. 2004;11(2):184–190. and technical success with radiofrequency ablation of 42 tumors. Radiology. 2003;226(2):
64. Veith FJ, Baum RA, Ohki T, et al. Nature and significance of endoleaks and endoten- 417–424.
sion: Summary of opinions expressed at an international conference. J Vasc Surg. 2002;35 85. Foley WD, Mallisee TA, Hohenwalter MD, et al. Multiphase hepatic CT with a multirow
(5):1029–1035. detector CT scanner. AJR Am J Roentgenol. 2000;175(3):679–685.
65. Elkouri S, Panneton JM, Andrews JC, et al. Computed tomography and ultrasound in 86. Foley WD. Special focus session: Multidetector CT: Abdominal visceral imaging. Radiographics.
follow-up of patients after endovascular repair of abdominal aortic aneurysm. Ann Vasc 2002;22(3):701–719.
Surg. 2004;18(3):271–279. 87. Erbay N, Raptopoulos V, Pomfret EA, et al. Living donor liver transplantation in adults:
66. Pollock JG, Travis SJ, Whitaker SC, et al. Endovascular AAA repair: Classification of aneu- Vascular variants important in surgical planning for donors and recipients. AJR Am J
rysm sac volumetric change using spiral computed tomographic angiography. J Endovasc Roentgenol. 2003;181(1):109–114.
Ther. 2002;9(2):185–193. 88. Lee SS, Kim TK, Byun JH, et al. Hepatic arteries in potential donors for living related
67. Stavropoulos SW, Baum RA. Imaging modalities for the detection and management of liver transplantation: Evaluation with multi-detector row CT angiography. Radiology.
endoleaks. Semin Vasc Surg. 2004;17(2):154–160. 2003;227(2):391–399.
68. Alerci M, Oberson M, Fogliata A, et al. Prospective, intraindividual comparison of MRI 89. Byun JH, Kim TK, Lee SS, et al. Evaluation of the hepatic artery in potential donors for living
versus MDCT for endoleak detection after endovascular repair of abdominal aortic aneu- donor liver transplantation by computed tomography angiography using multidetector-row
rysms. Eur Radiol. 2009;19(5):1223–1231. computed tomography: Comparison of volume rendering and maximum intensity projec-
69. Cornelissen SA, Prokop M, Verhagen HJ, et al. Detection of occult endoleaks after endo- tion techniques. J Comput Assist Tomogr. 2003;27(2):125–131.
vascular treatment of abdominal aortic aneurysm using magnetic resonance imaging 90. Onodera Y, Omatsu T, Nakayama J, et al. Peripheral anatomic evaluation using 3D CT
with a blood pool contrast agent: Preliminary observations. Invest Radiol. 2010;45(9): hepatic venography in donors: Significance of peripheral venous visualization in living-
548–953. donor liver transplantation. AJR Am J Roentgenol. 2004;183(4):1065–1070.
70. van der Laan MJ, Bartels LW, Viergever MA, et al. Computed tomography versus mag- 91. van Gulik TM, van den Esschert JW, de Graaf W, et al. Controversies in the use of portal
netic resonance imaging of endoleaks after EVAR. Eur J Vasc Endovasc Surg. 2006;32(4): vein embolization. Dig Surg. 2008;25(6):436–444.
361–365. 92. Cademartiri F, Raaijmakers RH, Kuiper JW, et al. Multi-detector row CT angiography in
71. Stavropoulos SW, Charagundla SR. Imaging techniques for detection and management of patients with abdominal angina. Radiographics. 2004;24(4):969–984.
endoleaks after endovascular aortic aneurysm repair. Radiology. 2007;243(3):641–655. 93. Horton KM, Fishman EK. Multi-detector row CT of mesenteric ischemia: Can it be done?
72. Cantisani V, Ricci P, Grazhdani H, et al. Prospective comparative analysis of colour- Radiographics. 2001;21(6):1463–1473.
Doppler ultrasound, contrast-enhanced ultrasound, computed tomography and magnetic 94. Bradbury MS, Kavanagh PV, Bechtold RE, et al. Mesenteric venous thrombosis: Diagnosis
resonance in detecting endoleak after endovascular abdominal aortic aneurysm repair. Eur and noninvasive imaging. Radiographics. 2002;22(3):527–541.
J Vasc Endovasc Surg. 2011;41(2):186–192. 95. Laghi A, Iannaccone R, Catalano C, et al. Multislice spiral computed tomography angiog-
73. Kawamoto S, Montgomery RA, Lawler LP, et al. Multi-detector row CT evaluation of liv- raphy of mesenteric arteries. Lancet. 2001;358(9282):638–639.
ing renal donors prior to laparoscopic nephrectomy. Radiographics. 2004;24(2):453–466. 96. Lawler LP, Fishman EK. Celiomesenteric anomaly demonstration by multidetector CT and
74. Holden A, Smith A, Dukes P, et al. Assessment of 100 live potential renal donors for laparo- volume rendering. J Comput Assist Tomogr. 2001;25(5):802–804.
scopic nephrectomy with multi-detector row helical CT. Radiology. 2005;237(3):973–980. 97. Horton KM, Fishman EK. 3D CT angiography of the celiac and superior mesenteric arteries
75. Rydberg J, Kopecky KK, Tann M, et al. Evaluation of prospective living renal donors for with multidetector CT data sets: Preliminary observations. Abdom Imaging. 2000;25(5):
laparoscopic nephrectomy with multisection CT: The marriage of minimally invasive imag- 523–525.
ing with minimally invasive surgery. Radiographics. 2001;21 Spec No:S223–S236. 98. Hamilton JD, Kumaravel M, Censullo ML, et al. Multidetector CT evaluation of active
76. Kawamoto S, Montgomery RA, Lawler LP, et al. Multidetector CT angiography for pre- extravasation in blunt abdominal and pelvic trauma patients. Radiographics. 2008;28(6):
operative evaluation of living laparoscopic kidney donors. AJR Am J Roentgenol. 2003; 1603–1616.
180(6):1633–1638. 99. Visser K, Kock MC, Kuntz KM, et al. Cost-effectiveness targets for multi-detector row CT
77. Halpern EJ, Mitchell DG, Wechsler RJ, et al. Preoperative evaluation of living renal donors: angiography in the work-up of patients with intermittent claudication. Radiology. 2003;
Comparison of CT angiography and MR angiography. Radiology. 2000;216(2):434–439. 227(3):647–656.
LWBK1209-ch34_p558-574.indd 574 17/05/13 5:18 PM

Chapter
Martin N. Wasser
Harrie C.M. van den Bosch 35
Magnetic Resonance and Computed
Tomography Angiography of
Peripheral Arteries
■■ Magnetic Resonance Angiography of affected in 90% of cases. Presently, atherosclerosis in the

Peripheral Arteries lower limbs accounts for 50,000 to 60,000 percutaneous
Enhanced versus unenhanced MRA angioplasties, implantation of 110,000 vascular prostheses,
Techniques for Peripheral Contrast-Enhanced and 100,000 amputations annually in the United States (1).
Before the planning of therapeutic actions, detailed infor-
mation on localization, extent, and severity of the disease is
Pitfalls essential. Conventional angiography has long served as the
■■ Computed Tomography Angiography of imaging modality of choice in this respect. However, x-ray
Peripheral Arteries angiography has definite risks and limitations, including the
General possibility of a severe contrast medium reaction, even when
Technique for Peripheral Computed Tomography nonionic agents are used (2). Also, because approximately
Angiography 70% of patients with severe occlusive peripheral disease have
evidence of impairment of renal function (3), contrast-induced
Visualization Techniques
renal insufficiency is a matter of concern. Therefore, nonin-
■■ Clinical Applications of Peripheral Magnetic vasive alternatives, such as duplex ultrasound and magnetic
Resonance Angiography and Computed resonance (MR) angiography, are increasingly used in the
Tomography Angiography diagnostic workup of patients with peripheral arterial disease.
Atherosclerosis Contrast-enhanced magnetic resonance angiography (CE-
Aneurysms MRA) has rapidly emerged as an attractive alternative to
Diabetes conventional angiography. The reason for this rapid accep-
tance in the “vascular community” is the close resemblance
Surveillance of Peripheral Bypass Grafts of CE-MRA images to conventional angiography. The con-
■■ Comparison Between Peripheral Magnetic trast agent gadolinium diethylene triamine pentaacetic acid
Resonance Angiography and Computed (DTPA) in the dosages used for MRA has long been con-
Tomography Angiography sidered non-nephrotoxic (4). The recent association between
high doses of gadolinium-based contrast material and neph-
■■ New Developments rogenic systemic fibrosis (NSF) (5,6), however, has made it
Magnetic Resonance Angiography Technique imperative to consider nonenhanced alternatives for angiog-
Magnetic Resonance Contrast Agents raphy in patients with moderate to severe renal insufficiency
Computed Tomography Angiography and vascular disease. Fortunately, it appears that the occur-
Technique rence of NSF is related to the formulation of the gadolinium
compound being applied and the contrast agents that have
■■ Summary been associated with NSF have been taken of the market.
With the advent of multidetector row or multislice com-
With the increasing average age of the population in the puted tomography (MSCT), another modality has emerged
industrialized world, atherosclerotic disease of the periph- for the noninvasive evaluation of the peripheral vasculature.
eral vessels is gaining importance. Although atherosclerotic Although this technique involves the use of radiation and
lesions occur in all vascular territories, the lower limbs are iodinated contrast agents, the spatial resolution that can
575
LWBK1209-ch35_p575-598.indd 575 17/05/13 5:18 PM

be obtained with multislice CT angiography (CTA) is su- high-resistance flow pattern with identical artifacts on con-
perior to that of MR angiography. CTA has therefore rap- ventional MRA.
idly become a serious competitor for MRA in evaluation of The discovery of association between Gd-compounds and
peripheral vascular disease in patients with nonimpaired NSF has led to a renewed interest in nonenhanced MRA.
renal function. Improvements in MR hardware and software, including the
This chapter will provide an overview of the techniques widespread availability of parallel imaging have helped to
and indications for CE-MRA and CE-CTA of the peripheral reduce acquisition times and have made some methods clini-
arteries of the lower limbs. Although atherosclerosis is the cally practical. Nonenhanced MR angiography may play a
major cause of peripheral arterial disease, other abnormali- useful role as a supplement to gadolinium-enhanced MR
ties will also be addressed. angiography, particularly when the contrast-enhanced meth-
ods fail or have serious artifacts.
While long acquisition times of TOF-MRA and PC-MRA
Magnetic Resonance have limited their routine use in patients with vascular
Angiography of Peripheral disease recent research has focused on TSE- or FSE-based
approaches (8). ECG-gated FSE-based MRAs are acquired
Arteries
along the axis of the run-off vessels, which allow for larger
anatomical coverage in z-axis. FSE-based techniques are
Enhanced Versus Unenhanced MRA
typically based on the subtraction of two datasets obtained
The older unenhanced MRA techniques [phase-contrast during systole and diastole with ECG gating. Cine sequences
MRA (PCA) and time-of-flight (TOF) MRA] suffered from are generally applied to determine the appropriate timing for
artifacts caused by turbulence and in-plane saturation. These each acquisition. FSE-based techniques have a high sensitiv-
techniques were also highly time consuming. An important ity and negative predictive value for detection of stenosis of
reason for the long acquisition times in conventional MRA is the lower extremities (9). As this approach is based on sub-
the requirement of obtaining images in the transverse plane traction of a systolic acquisition from a diastolic acquisition
to circumvent in-plane saturation. Due to the triphasic flow for selective arterial angiograms, it is susceptible to motion
pattern in the peripheral vessels with diastolic flow reversal, artifacts, which may lead to nondiagnostic image quality in
cardiac triggering has to be performed in conventional MRA up to 50% of patients (9,10). Recent improvements with the
in order to prevent erroneous signal loss. This cardiac trig- development of a variable flip angle with spatially nonselec-
gering causes additional lengthening of the study. tive RF refocusing pulses (SPACE, VISTA, and Cube) can
In 2000, a meta-analysis was published in which the re- dramatically reduce the data sampling duration because of
sults of 21 studies on peripheral MRA published in the litera- reduced echo spacing. This technique can also provide more
ture from January 1991 through June 1999 were evaluated detailed angiography of distal arteries (11).
(7). The sensitivity for detection of hemodynamically signi Until now contrast-enhanced MRA remains the best
ficant stenosis (>50% luminal reduction) ranged from 64% MRA technique to evaluate arterial disease in the lower
to 100% for two-dimensional (2D) TOF and 92% to 100% limbs, in terms of time effectiveness and accuracy.
for three-dimensional (3D) CE-MRA. Specificity ranged from
68% to 96% for 2D TOF and 91% to 99% for 3D CE-MRA.
This indicates that 3D CE-MRA is superior to 2D TOF for the Techniques for Peripheral Contrast-
detection of peripheral arterial disease. Especially for imaging Enhanced Magnetic Resonance
of the proximal vessels in the legs 2D TOF appears to be less Angiography
reliable (sensitivity 83% to 100%, specificity 23% to 98%). Basics
In tortuous, elongated iliac arteries, 2D TOF may show false-
positive results due to in-plane saturation or saturation of ar- The basic principle behind CE-MRA is imaging the arterial first
terial signal by the venous presaturation slab. Short segments pass of paramagnetic contrast agent in the vessels after intra-
of signal void may be caused by either a short occlusion or a venous injection (12). The injected contrast agent produces
severe stenosis. Also, retrograde flow may not be recognized a strong shortening of the T1-relaxation time of blood. This
due to the venous presaturation slab. results in high signal intensity of the arteries on T1-weighted,
Also in the femoropopliteal and infrapopliteal regions spoiled gradient-echo images. The delay between arterial and
CE-MRA performed better than 2D TOF. Sensitivity ranged venous enhancement provides a time window for pure arte-
between 88% and 92% and between 94% and 100% for rial imaging. This time window depends on the rate of con-
2D TOF and CE-MRA, respectively (specificity 82% to trast agent injection, but also on the anatomic region being
98% vs. 97% to 99%). With respect to the infrapopliteal imaged. In the upper limbs, venous return is much faster than
vessels sensitivity/specificity were 89% to 98%/91% to in the lower limbs, requiring a much shorter measurement
95% versus 91% to 94%/100% for 2D TOF and CE-MRA, time. Since the technique depends on imaging of the first pass
respectively. of contrast agent, timing the start of image acquisition after
No figures exist for the accuracy of MRA for arterial injection is essential. Various techniques have been developed
disease in the upper limbs. Although conventional MRA to calculate this so-called scan delay, as will be explained
and CE-MRA have not been compared, there is no rea- later.
son to doubt that CE-MRA is also the best technique to The quality of the MRA images with regard to selective
evaluate arteries in this territory. Similar to arteries in the arterial visualization, resolution, and volume of interest de-
lower extremities, arteries in the arms show a triphasic, pends on both sequence parameters used and the contrast
LWBK1209-ch35_p575-598.indd 576 17/05/13 5:18 PM

Chapter 35 ■ Magnetic Resonance and Computed Tomography Angiography of Peripheral Arteries 577
medium bolus geometry. Arterial enhancement depends on Timing of Image Acquisition

individual physiologic parameters, such as cardiac output
and blood volume, but also on contrast agent application In CE-MRA, adequate timing of the start of data acquisi-
parameters, which can be manipulated (flow rate, dose, and tion, with arrival of the contrast agent in the vessels of inter-
saline flush volume). The T1-shortening of blood depends on est, is essential. Arterial enhancement must coincide with
the intravascular concentration of the contrast agent, which, acquisition of central k-lines, and for improved arterial/
in turn, depends on the rate of injection. In general, the venous differentiation, the central k-lines have to be sampled
higher the injection rate, the higher the concentration will before venous return. Essentially, three methods have been
be, although rates exceeding 5 mL/s do not result in further developed to ensure proper timing.
increase in signal intensity. In fact, intravascular signal may
become lower at rates higher than 6 mL/s (13). On the other Test Bolus
hand, a higher infusion rate results in faster venous return Adequate timing can be performed by measuring the circu-
and a shorter bolus length, requiring a shorter measurement lation time after intravenous injection of a small test bolus
time. Always a compromise between imaging time, resolu- (1 to 2 mL) of contrast agent, using a rapid dynamic imaging
tion and arterial/venous time window must be sought. Since, sequence. With this dynamic series, the arrival time of the
for examination of the lower limbs, the entire vascular tree test bolus after injection in the region of interest (ROI) can
from renal arteries to the ankles has to be imaged with high be imaged and, therefore, the scan delay can be calculated.
resolution (especially in the lower legs), a long arteriovenous To avoid pooling of the test bolus in veins, the tubing has to
time interval is necessary. Therefore, in the lower extremities be flushed with saline. Both the test bolus and the final infu-
injection rates of 0.5 to 1.5 mL/s are used. After injection sion should be injected at the same rate. This timing method
of contrast agent, the venous system should be flushed with is fairly robust and easy to perform. A disadvantage is that
saline to push the contrast medium column from the arm the method requires additional administration of contrast
veins into the circulation. agent. It lengthens the total procedure time with 2 to 3 min-
A heavily T1-weighted, spoiled gradient-echo sequence utes. Also, in case of irregular heart action, the circulation
[short repetition time/echo time (TR/TE); flip angle, 25 to time calculated from the test bolus may sometimes differ
50 degrees] is used in which the achievable field of view, from that during injection of the final bolus.
matrix size, scan volume, and number of partitions are de-
termined by the arterial/venous time interval. Subtraction of
Fluoroscopic Triggering
pre- and postcontrast images is performed to visualize only
the arteries for greater ease of image interpretation and to Fast imaging sequences can be used to visualize the influx of
reduce artifacts (Fig. 35.1) (14). The obtained volumetric contrast into the ROI. MRA data acquisition is then started
dataset containing high intensity voxels corresponding to automatically or manually [MR Smart-Prep (15), bolus track
arteries can be postprocessed using the maximum intensity (16)]. A disadvantage of this method is the delay of 2 to
projection (MIP) algorithm. For improved diagnostic perfor- 4 seconds that occurs between visualizing the arrival of the
mance, review of the individual source images or reformat- contrast bolus and the actual start of acquisition. Also, time
ting images in the transverse plane is required. for breathing instructions, if required, may be too short.
A B
Figure 35.1. Nonsubtracted (A) and subtracted (B) CE-MRA images of the aortoiliac region. The sub-
tracted image shows higher contrast/background ratio. More vessels can be seen on the subtracted image.
LWBK1209-ch35_p575-598.indd 577 17/05/13 5:18 PM

Time-resolved Imaging
With ongoing improvements in hardware and gradient
technology, the time required to image a large volume of
Enhancement ratio
interest with high resolution has greatly reduced. With a 2D
slab, the entire anteroposterior diameter of the legs can be
covered in 2 to 3 seconds (17). Therefore, by performing a
dynamic volumetric series after intravenous injection of the
contrast agent, the arterial phase will always be included
in one of the datasets, thus obviating the need for bolus
timing (18).
1 2 3 4 5 6
Imaging Strategies No. of bolus injections
For imaging of the leg vessels, a large anatomic region Figure 35.2. Enhancement ratios of subtracted images after
(approximately 100 cm) has to be covered. Since the lon- repeated injection of gadolinium-DTPA in a flowing phantom model.
gitudinal field of view of MR systems is limited to 45 to Repeated injection results in substantially reduced contrast-to-noise
50 cm, complete imaging of the lower extremity requires ratios. In vivo, however, extravasation of contrast agent also occurs.
imaging of several regions (or “stations”). This means repo- The effect of intravascular accumulation of contrast will therefore be
sitioning of the patient to the isocenter of the magnet. Two lower in vivo. (From: Westenberg JJ, Wasser MN, van der Geest RJ,
methods for imaging of the entire legs have been developed: et al. Scan optimization of gadolinium-enhanced three-dimensional
So-called multistation/multi-injection imaging and the bolus MRA of peripheral arteries with multiple bolus injections and in
vitro validation of stenosis quantification. Magn Reson Imaging.
chase method. In both techniques one has to ensure suffi-
cient overlap between the separate stations. No studies have
been performed comparing the two methods directly; how-
ever, because of its short examination time and more effi- chased using a floating table. This technique requires appli-
cient use of contrast agent, the bolus-chase technique is the cation of special hardware to move the table to predefined
current standard of practice. positions. This can be done using homemade table-stopping
devices or special MR-table hardware. The advantage of the
Multistation/Multi-injection Imaging or moving-table technique is that it is very rapid (coverage of
Step-by-step Technique the entire legs in less than 4 minutes), requiring a relatively
low dose of contrast agent (approximately 30 cc).
With this technique, the three regions of the legs (aortoil- Initially, a disadvantage of the stepping-table technique
iac, femoropopliteal, and infrapopliteal vessels) are imaged was that the three stations were imaged sequentially while
sequentially with two or three separate bolus injections of the scan preparation had been optimized for only a single
contrast medium (19). Imaging of each station involves station. Later on, the method has become more flexible (16),
acquisition of a mask dataset, which is subtracted from and even dedicated surface coils can be used (25,26).
the subsequent images obtained after contrast injection. Although with the bolus-chase technique, imaging times
Contrast agent can be injected at moderately high rates of of the various segments are short, early venous enhancement
1 to 1.5 mL/s, ensuring high signal intensity in the vessels. may sometimes cause difficulties in evaluation of the arter-
The advantage of this technique is that it can be applied ies of the lower leg (Fig. 35.3). This venous overlay may
on any clinical scanner without use of special hardware, be especially prominent in patients with diabetes, venous
and imaging parameters can be tailored to every single sta- ulcers, or cellulitis. Enhancement of the veins in the third
tion. The multistation/multi-injection technique, however, station, the lower legs, can be minimized by using an opti-
is not very time efficient (30 to 45 min/patient); a total con- mized imaging protocol: Use of a biphasic injection protocol
trast agent dose of 0.2 to 0.3 mmol/kg/patient is required, (e.g., 1 mL/s during 10 seconds and the rest with 0.5 mL/s),
and the contrast-to-noise ratio of the subtracted images use of “centric” k-space acquisition, in which the contrast-
becomes less at the second and third stations as a result of determining k-lines are acquired first, and the use of parallel
circulating gadolinium (Fig. 35.2) (20). Using larger dos- imaging. With the use of parallel imaging techniques such as
ages in the subsequent stages can partly circumvent this simultaneous acquisition of spatial harmonics (SMASH) (27)
decrease in signal intensity. Surface coils can be used for or sensitivity encoding (SENSE) (28) techniques, combina-
imaging of the different stations. Huber et al. (21) achieved tion of coils can be used to compensate for omitted gradient
100% sensitivity and 94% specificity for identifying 80 steps. The increase in imaging speed can be used to reduce
hemodynamically significant stenoses and 39 occlusions in acquisition times or to increase resolution of the scan (29).
24 patients. Binkert et al. (30) used a dual-bolus technique, in which they
started with a dedicated calf MRA and—after a 15-minute
interval—proceeded with the moving-table MRA. This led
Bolus Chase or Moving-table Technique
to better visualization of the infrapopliteal arteries with less
In this technique, mask images of all three stations of the legs venous overlay and therefore a significantly increased di-
are made first (16,22–24). Then, during slow constant infu- agnostic accuracy, compared to moving-table MRA alone.
sion of contrast agent, the flow of contrast through the legs is Also, venous compression by means of a tourniquet, either
LWBK1209-ch35_p575-598.indd 578 17/05/13 5:18 PM

A B C
Figure 35.3. MRA of the lower limbs in a patient with a kidney transplant in the left iliac fossa. Note the
venous enhancement in the lower legs. Aorta-iliac region (A), upper legs (B), and, lower legs (C).
at the midfemoral (31) or at infragenual level (32,33), can be Computed Tomography

used to slow down the flow in the vessels of the legs. Angiography of Peripheral
Arteries
Image Presentation
To encourage acceptance of the technique by referring clini- General
cians, one has to provide them with images to which they are The introduction of multidetector row or MSCT represented
accustomed and which appear similar to conventional angiog- a revolutionary advance in CT technology. Compared with
raphy. The number of images should be limited as a practical single-slice helical CT, MSCT offers faster rotation times
concern for vascular surgeons. The images should be large to and simultaneous acquisition of multiple helices per rota-
be viewed from a distance, displaying standard angiographic tion, resulting in shorter acquisition times, increased longi-
views. Also, bony landmarks should be provided for better tudinal volume coverage, lower-dose contrast medium, and
orientation, similar to images of conventional angiography improved spatial resolution for assessing smaller arterial
(Fig. 35.4). branches. With every new generation of multislice scanners,
these advantages became more apparent, enabling increased
Pitfalls
table feed (and therefore shorter examination times) and
Although signal loss owing to turbulence and in-plane sat- increased spatial resolution.
uration is usually not present in CE-MRA, overestimation Radiation dose is an issue in MSCT, but in the age group of
of the length of a stenosis can still occur, especially at high patients with intermittent claudication, it is probably not very
velocity rates at the stenotic area and a low concentration of important. Moreover, the radiation dose in conventional angio
contrast material (34). graphy is approximately four times as high as in MSCT (37).
If the contrast agent arrives in the vessel of interest after
sampling of the central k-lines as a result of improper tim- Technique for Peripheral Computed
ing, a central dark line in the vessel can be seen, the so-called Tomography Angiography
pseudodissection (35).
As in CE-MRA, the principle behind peripheral multislice
Care has to be taken to include all vessels of interest in
CTA is imaging the arterial phase of contrast enhancement
the imaging volume. Inappropriate image coverage may
before substantial interstitial or venous enhancement occurs.
mimic occlusion of tortuous arteries, such as the external iliac
The advent of multislice scanning has enabled chasing of the
arteries (36). Therefore, it is recommended to present also the
contrast bolus while maintaining high spatial resolution.
sagittal or transverse views to demonstrate the entire imaging
Three important factors determine the quality of images
volume (Fig. 35.5).
in CTA (38).
Sometimes a ghost or phase artifact can be seen parallel
to vessels with very high signal intensity (36) (Fig. 35.6). •• High spatial resolution, especially required for imaging of
Also, subtraction misregistration artifact may occur. Clips the infrapopliteal vessels, depends largely on the detector
and metallic stents may cause signal voids as a result of sus- collimation and reconstruction kernel. Although inferior
ceptibility artifacts (Fig. 35.7). to conventional catheter angiography, CTA can assess
LWBK1209-ch35_p575-598.indd 579 17/05/13 5:18 PM

A B C
D E F
Figure 35.4. MRA of the lower limbs in a patient with occlusion of the distal aorta and proximal common
iliac arteries (A–C). The nonsubtracted source images provide the bony land marks, especially of the joints (D–F).
arteries smaller than 1 mm in diameter. Since it is volu- The timing of a CTA scan is of critical importance so that
metric, CTA allows 3D visualization of the vasculature to the data is acquired during peak enhancement and not
separate superimposed structures. before or after the contrast has arrived. Table feed should
•• High temporal resolution is achieved from fast gantry therefore not exceed 50 mm/s (39).
rotation. Modern scanners have a gantry rotation time of
280 to 350 milliseconds. Therefore, a volume of data (3 to As in CE-MRA, the delay time between contrast injec-
4 cm on 64-detector row scanners) can be acquired with tion and start of the scan may be determined by a test
full scan mode in less than a half second. bolus (20 cc). An ROI is drawn in the distal aorta, and,
•• The z-axis coverage per gantry rotation and table speed after injection of the test bolus, low-dose scans of the ROI
plays a large role in the CTA acquisition. With 16- or are obtained in a dynamic (nonincremental) fashion. The
greater detector row systems the tradeoff between acquisi- time to peak enhancement is then chosen as scan-delay time.
tion speed and slice thickness is no longer an issue because Most often, automated bolus timing is performed, in which
of larger volume coverage. In fact, 4 cm or greater cover- scanning is started after the enhancement in the ROI has
age per gantry rotation from larger detectors introduces exceeded a predefined enhancement level (e.g., 100 HU).
the problem of imaging too quickly and thus outrun- After this threshold is reached, scanning is started following
ning the iodinated contrast bolus. Care must be taken to a delay of 4 to 12 seconds. This delay is required to allow
match the helical pitch and table speed with the first cir- the scanner to move from the ROI-measurement level to the
culation of contrast material through the anatomic ROI. level at which scanning is to begin.
LWBK1209-ch35_p575-598.indd 580 17/05/13 5:18 PM

duration. Total duration of injection can be 5 to 10 seconds

shorter than the actual scanning time (40). After injection
of contrast agent, the venous system should be flushed with
saline. Saline flushing prolongs arterial enhancement.
Visualization Techniques
In CTA, postprocessing of data is usually done using two
projection algorithms: MIPs and volume rendering (VR).
Since MIP reformation is a projection technique, structures
with high attenuation, such as bones, are superimposed over
the enhancing vessels. Therefore, before applying the MIP
algorithm, segmentation of bony structures is necessary to
visualize the vessels in their entirety (Fig. 35.8). Calcified
plaques are clearly visible on MIPs (Fig. 35.9). An important
disadvantage of MIP reformation, however, is that calcified
plaques and stents can obscure visualization of the lumen,
especially when the calcium has a significant circumferential
extent (Fig. 35.10). For a better overview of these difficult
areas with calcifications and stents, a curved multiplanar
reconstruction or central luminal line can be made (Fig.
35.11). Large calcifications in the MIP can be removed by
segmentation using the algorithm for bone removal (Figs.
Figure 35.5. Coronal (left) and sagittal (right) CE-MRA MIP 35.11–35.13). This segmentation method, however, is too
images of the aortoiliac region. Especially the sagittal view shows the crude to be used in segments with small calcifications, that
elongated external iliac arteries in this patient. is, in the lower legs (Figs. 35.14 and 35.15). Another disad-
vantage of the MIP algorithm is that vessel edges and ste-
noses may not be accurately mapped because of decreased
Since CT is less sensitive than MR for contrast enhance- attenuation or signal secondary to partial volume effects.
ment, a larger volume of contrast agent (120 to 160 mL) Another algorithm for reconstruction of the 3D data is
has to be injected in CTA, compared to MRA at a higher VR. In VR, no data are discarded. Opacification of a voxel,
rate (3 to 4 mL/s). The level of enhancement of the arteries ranging from transparency to total opacity, is assigned based
depends on the infusion rate and the type of contrast agent on the attenuation coefficient. Since the image is inclusive off
used. If a contrast agent with an iodine concentration of 370 all voxels acquired, there are no surfaces. Since both opaci-
to 400 mg/cc is used, the injection rate can be lower than fication and lighting effects are coded in gray scale, color is
if a contrast medium with 300 mg iodine/cc is used. The often added to clarify the image features. Color is assigned on
total amount of contrast agent needed depends on the scan the basis of the attenuation coefficient of a voxel (Fig. 35.16).
Figure 35.6. Example of ghost arti-

facts in images of the lower limbs (arrows):
MIP (left) and source image (right).
LWBK1209-ch35_p575-598.indd 581 17/05/13 5:19 PM

Stent
Stent
No pressure gradient
Figure 35.7. Patient with aortoiliac

Stenosis bifurcation prosthesis and implantation
Stent of stents in the distal aorta and right
iliac leg of the prosthesis. At MRA (left)
susceptibility artifacts falsely suggest
significant stenoses in the stents; no pres-
sure gradient was found on angiography
(right).
Clinical Applications of Steno-occlusive Disease

Peripheral Magnetic Resonance In patients with acute limb threatening ischemia, angio-
Angiography and Computed graphic imaging is performed as soon as possible, followed
Tomography Angiography by thrombolytic therapy in the same session. MRA or CTA
has no role in the management of this condition.
Atherosclerosis In patients with chronic ischemia, however, MRA has
emerged as a valuable alternative pretreatment procedure to
Atherosclerosis is a systemic process that may cause stenosis,
catheter angiography. In analyzing MRA images of the lower
occlusion, or aneurysmal dilatation of the affected arteries.
limbs of patients with intermittent claudication, one has to
understand what the treating physician wants to know. In
other words, one has to “think as a vascular radiologist,”
and one has to be familiar with the clinical issues and the
potential treatment options (41). Analyzing the proximal
inflow to a lesion, the distal outflow from the lesion, and
the lesion itself is important. If the inflow is impaired, a dis-
tal arterial reconstruction may be endangered. On the other
hand, if the outflow is limited, a proximal reconstruction
may be compromised. A general rule in vascular surgery is
to first treat the most proximal lesions, because they usually
have the greatest impact on flow to the limbs.
Regarding the lesion itself, it is important to report the
localization, severity, and length of the stenosis. In general,
hemodynamically significant stenoses, usually taken as >50%
luminal narrowing, require treatment. In case of borderline
stenosis (40% to 60%), it may be difficult to determine the
hemodynamic significance of the lesion. This, however, also
holds for conventional angiography, and, in these cases, intra-
vascular pressure measurements are performed to assess the
hemodynamic implications of a lesion (42). In almost all stud-
ies evaluating the value of MRA, conventional DSA is taken
as the standard of reference. This standard, however, may not
be as “gold” as one is inclined to believe. Eccentric plaques in
the large proximal vessels may go undetected with DSA be-
cause of the limited number of projections (Fig. 35.17). In the
renal arteries we found a sensitivity of only 70% for DSA
in detecting hemodynamically significant stenoses when
compared with intravascular pressure measurements (43).
In the iliac arteries, Wikstrom et al. (44) found sensitivities/
Figure 35.8. Volume rendering projection (left) and MIP (right) of specificities of 86%/88% for DSA, 81%/75% for MRA, and
a CTA study in a 62-year-old patient with diabetes mellitus (see also 72%/88% for duplex, in detection of significant stenoses
Figure 35.15). To visualize the entire arterial tree in the MIP image, using an aortofemoral pressure gradient of 20 mm Hg in-
skeletal structures were removed by segmentation. dicating hemodynamic significance. Probably, conventional
LWBK1209-ch35_p575-598.indd 582 17/05/13 5:19 PM

Figure 35.9. Volume rendering pro-

jection (left) and MIP (right) of a CTA
examination in a 63-year-old male with
extensive atherosclerotic disease. The
calcified plaques are better appreciated
on the MIP.
Figure 35.10. MIP images of a CTA examination in a 56-year-old man with atherosclerosis after stent
placement in the right external iliac artery. Large calcified atherosclerotic lesions were present in the distal
aorta and both common iliac arteries (left). The calcified plaques could be removed by segmentation, and
evaluation of patency of the stent was not possible (right).
LWBK1209-ch35_p575-598.indd 583 17/05/13 5:19 PM

B
Figure 35.11. CTA images of a 56-year-old man with multiple atherosclerotic lesions in the distal aorta
and iliac arteries. Curved multiplanar reformatted image or center luminal line reconstruction (A) shows pat-
ent right iliac arteries, despite extensive calcified plaques as visualized on the MIP reconstruction (B, left).
After segmentation of the disfications, patency of the right iliac arteries is also visible on MIP (B, right).
LWBK1209-ch35_p575-598.indd 584 17/05/13 5:19 PM

Figure 35.12. CTA-MIP images of multiple calcified plaques in the aorta (A, top). These calcifications could
be removed from the images by segmentation (A, bottom). On the segmented images, a stenosis was appreci-
ated in the distal aorta. This stenosis was also visible on DSA (B).
LWBK1209-ch35_p575-598.indd 585 17/05/13 5:19 PM

Figure 35.13. CTA of a patient with multiple calcifications in the distal aorta and iliac arteries (A).
An aneurysm of the right common iliac artery is seen, also visible on DSA (B). Because of a large calcified
plaque, patency of the left common iliac artery cannot be assessed (A, left). After segmentation, the left
common iliac artery appears patent (A, right), which was confirmed at DSA (B).
LWBK1209-ch35_p575-598.indd 586 17/05/13 5:19 PM

Figure 35.14. CTA of a 52-year-old patient with diabetes. A large calcified plaque in the distal aorta does
not result in stenosis of the aorta (A), before (left) and after (right) segmentation of the calcification. Smaller cal-
cifications in the femoral and infrapopliteal arteries could not be removed (B).
LWBK1209-ch35_p575-598.indd 587 17/05/13 5:19 PM

Figure 35.15. CTA-MIP images of a 53-year-old woman with diabetes. Extensive calcifications, especially
in the infrapopliteal vessels, were not removable by segmentation.
Figure 35.16. Volume rendering images of

a CTA examination. Color is assigned on the
basis of the attenuation coefficient of a voxel
(A). Subsequently, the opacified vessels can be
selectively displayed, rotated and viewed from
A B
any angle (B).
LWBK1209-ch35_p575-598.indd 588 17/05/13 5:19 PM

Figure 35.17. Patient with athero-

sclerotic lesion in the left common iliac
artery, as visualized on MRA (left). The
stenosis, located on the dorsal wall of the
vessel, was not appreciated at angiogra-
phy (right), although a pressure gradient
of 25 mm Hg was measured.
angiography is also not an adequate standard of reference reserved to determine the hemodynamic significance of ste-
in analyzing runoff vessels, because CE-MRA demonstrated noses found on CE-MRA.
more patent vessels than did conventional angiography In a meta-analysis in 2009, Met et al. (50) found an over-
(Fig. 35.18) (45,46). all sensitivity of CTA for detecting more than 50% stenosis
Duplex Doppler ultrasonography (US) is an accurate or occlusion of 95% (95% CI: 92% to 97%) and a specific-
method to detect stenoses in the carotid arteries. Also in the ity of 96% (95% CI: 93% to 97%) . Overstaging of stenosis
lower limbs, duplex ultrasound has been shown to be a reli- occurred in 8% of segments and understaging in 15%. CTA
able noninvasive diagnostic modality (47,48). Duplex US, therefore appears as an accurate modality to assess presence
however, is operator-dependent, laborious, and it does not and extent of arterial disease in patients with intermittent
provide a complete road map for treatment planning. claudication.
In 2000, a meta-analysis was done based on 9 papers on In another meta-analysis in 2009, Heijenbrok-Kal et al. (51)
CE-MRA (216 patients) and 18 on color-guided duplex US found a pooled sensitivity and specificity for detecting a steno-
(1,059 patients) (49). In all of these papers, conventional sis of at least 50% per segment of 92% (95% CI: 89%, 95%)
angiography was used as the gold standard. The pooled sen- and 93% (95% CI: 91%, 95%), respectively. The diagnostic
sitivity of CE-MRA for assessing arterial disease (i.e., >50% performance of multidetector CT angiography in the infrapop-
luminal reduction) in these studies was 97% [95% confi- liteal tract was lower but not significantly different from that
dence interval (CI): 95.7% to 99.3%] and for duplex US, in the aortoiliac (p >.11) and femoropopliteal (P >.40) tracts.
87.6% (95% CI: 84.4% to 90.8%). The pooled specificities The diagnostic accuracy of CTA seems to compare well
were similar (96.2% for CE-MRA and 94.7% for duplex with MRA (7,52), although studies directly comparing these
US). These figures for CE-MRA correspond to the ones re- imaging modalities are lacking.
ported in the meta-analysis mentioned earlier, comparing 2D Intermittent claudication is commonly caused by stenosis
TOF with CE-MRA. or occlusion of vessels in the legs. Sometimes, however, the
In almost all duplex US studies, results for the different symptoms may be caused by severe stenosis or occlusion of
stations were reported, but this was done in only five of nine the aorta (Leriche syndrome). Also, this entity can be visual-
CE-MRA studies. Stratification per anatomic site was there- ized with CE-MRA (53) (Figs. 35.4 and 35.22).
fore not possible. In general, the interval between CE-MRA
and conventional angiography was shorter (mean: 5 days)
Aneurysms
than between duplex US and angiography (mean: 17 days).
One might expect that as the interval between the studied CE-MRA has been found to be adequate in demonstrating
examination and the reference examination increases, the the extent of aortoiliac aneurysmal disease (54). Aneurysms
discriminatory power of the studied examination decreases. may also occur in the femoral and popliteal arteries. Popliteal
This was indeed found for duplex US. The lower sensitiv- aneurysm is an important entity, because up to one-third of
ity of duplex US in the reported studies may therefore be patients have either distal embolization or severe acute limb
partly attributed to the longer interval between duplex and ischemia from an unrecognized and nonpalpable popliteal
angiography. aneurysm (55). Femoral aneurysms are less important and
It appears that, because of its high sensitivity and speci- they may be followed conservatively (56,57). In a study in
ficity, CE-MRA may replace conventional angiography 313 patients with abdominal aortic aneurysms (AAAs) who
in the workup of patients with intermittent claudication all underwent ultrasound examination of the femoropopliteal
(Figs. 35.19–35.21). Color-guided duplex US can then be region, Diwan et al. (58) encountered femoral and popliteal
LWBK1209-ch35_p575-598.indd 589 17/05/13 5:19 PM

Figure 35.18. MRA (A) and DSA (B) of a

patient with bilateral occlusion of the superficial
femoral artery. Infrapopliteal runoff vessels are
much better visualized on MRA than on DSA. B
LWBK1209-ch35_p575-598.indd 590 17/05/13 5:19 PM

B
Figure 35.19. Comparison between MRA and DSA in a patient with stenosis (arrow) in the right iliac
artery (A). The more distal vessels show no stenosis (B).
A B C
Figure 35.20. MRA-MIP images of a patient with a stent in the right external iliac artery, a long occlusion
in the left superficial femoral artery and a short occlusion in the right superficial femoral artery. Two arteries in
both lower legs.
LWBK1209-ch35_p575-598.indd 591 17/05/13 5:19 PM

A B C
Figure 35.21. MRA of a patient with a stent in the right superficial artery, occlusion of the left superficial
artery, poor vascularization of the left lower leg.
aneurysms in 12% of patients (36 patients with 51 peri world. Diabetic patients often suffer from long segments of
pheral aneurysms, of which 31 were popliteal aneurysms). arterial occlusions, especially of the infrapopliteal vessels
The peripheral aneurysms occurred only in men with AAA (Fig. 35.24). Surgical reconstitution of blood flow remains
(incidence 14%), no aneurysms were found in the 62 women the most important therapeutic option for limb salvage in
with AAA. The reason for this sex difference is not known. these patients. Improvements in bypass graft technology
In 14 (39%) of the 36 men with peripheral aneurysms, have enabled the use of the pedal arteries as target sites. For
peripheral occlusive disease was present as opposed to 20 planning of distal revascularizations, information on distal
(9%) of the 215 men without peripheral aneurysm. vessels in the foot (including the pedal arch) is necessary.
Also, popliteal aneurysms can be detected on CE-MRA The evaluation of the arteries in the calf with CE-MRA often
(Fig. 35.23). is hampered because of venous contamination due to either
bolus-timing problems or inflammation of soft tissues (59).
Both limitations are accentuated for patients with diabetes,
Diabetes
because the associated peripheral vascular disease typically
Diabetes-related angiopathy is an important cause of non- spares the proximal vessels and mainly affects the more
traumatic lower extremity amputations in the industrialized distal arteries in the calf and, in a later stage, the foot (60).
A B C
Figure 35.22. MRA of a patient with occlusion of the distal aorta (Leriche syndrome).
LWBK1209-ch35_p575-598.indd 592 17/05/13 5:20 PM

A B C
Figure 35.23. MRA of a patient with aneurysm of the left popliteal artery.
Some studies have indicated that even 2D TOF-MRA disease, CE-MRA should be used for pedal bypass planning,
can demonstrate patent distal vessels, not visible on conven- instead of DSA.
tional angiography (45,61). Kreitner et al. (62) evaluated the
value of CE-MRA in 24 diabetic patients, using the head coil
Surveillance of Peripheral Bypass Grafts
for signal reception. All patients underwent angiography for
comparison within 5 days. Seven vascular segments were CE-MRA may also be used in the postoperative follow-
evaluated in each extremity: Distal anterior tibial artery, up of arterial bypass grafts. Peripheral bypass grafts may
distal posterior tibial artery, distal peroneal artery, dorsal consist of autologous saphenous vein, expanded polytetra-
pedal artery, lateral plantar artery, medial plantar artery, and fluoroethylene (PTFE) or Dacron. Autologous venous grafts
plantar arch. Of possible 168 segments, 74 were seen to be have a superior patency rate compared to the other grafts,
patent on DSA and 104 on CE-MRA. Thirty vessel segments but still a 12% incidence of bypass graft stenosis in the first
that were suitable for distal grafting in 9 (38%) patients were postoperative year. Eighty percent of stenosis occurs during
seen exclusively on MRA. In seven of the nine patients this the first postoperative year, and almost all occur in the first
resulted in a change of treatment plans. Therefore, this study 18 months (63). In 60% of cases of stenosis this occurs with
shows that in diabetic patients with severe peripheral occlusive no initial clinical symptoms (63). Therefore, these bypasses
A B C
Figure 35.24. MRA of a patient with diabetes. The iliac and femoral vessels are normal, but there are
multiple stenoses in the arteries of the lower legs.
LWBK1209-ch35_p575-598.indd 593 17/05/13 5:20 PM

A B C
Figure 35.25. MRA-MIP images of a patient with an aortic bifurcation prosthesis. A stenosis has occurred
at the distal anastomosis in the right common femoral artery.
are carefully monitored to detect stenosis in time to enable area of stenosis of sufficient severity to require treatment.
reconstitution of flow (secondary patency). Surveillance cri- These data strongly suggest that MRA is well suited as a
teria, such as the ankle-brachial index, have been successfully screening method in the follow-up of lower limb bypass
tested. Also, duplex US is routinely performed for surveil- grafts (Fig. 35.25). Further study to analyze the cost–benefit
lance. Midgraft peak systolic velocity (PSV), PSV ratio before ratio is required to assess whether CE-MRA can be used as
and after stenosis can be calculated, or the entire graft can the single follow-up examination and road map for addi-
be visualized. Conventional angiography is performed when tional reconstruction.
an abnormality at duplex US is found and when a vascular Willmann et al. (67) found no statistically significant dif-
intervention is needed. ference in sensitivity or specificity between CTA and duplex
Duplex US is adequate in assessing graft patency, and it US for detection of hemodynamically significant bypass
can detect stenoses adequately. However, duplex US is lim- stenosis or occlusion, aneurismal changes, or arteriovenous
ited in detecting collateral flow, and also proximal lesions fistulas.
that may endanger future patency of the graft may be missed CE-MRA also appears useful in the surveillance of extra-
at duplex US. anatomic bypass grafts. These grafts are made of PTFE or
Three studies evaluated the use of CE-MRA in surveil- Dacron; veins cannot be used. Possible reconstructions in
lance of lower limb grafts. Bertschinger et al. (64) evaluated the lower limbs are axillofemoral grafts, femorofemoral
30 patients with 31 grafts who underwent both IA-DSA and crossover bypasses (Fig. 35.26), and thoracic aortofemoral
CE-MRA. All abnormalities (ten stenoses, nine occlusions, artery bypass grafts (68).
and eight aneurysms or ectasias) could be visualized by CE-
MRA (sensitivity, 100%). Since six segments could not be
evaluated, specificity was 90.3%. Bendib et al. (65) selected
23 patients with 40 vascular grafts with either clinical symp- Comparison Between Peripheral
toms or abnormal duplex findings. All patients also under- Magnetic Resonance
went x-ray angiography. MRA depicted 38 grafts (95%) Angiography and Computed
with 28 abnormalities. Two stenoses were overestimated. Tomography Angiography
Some grafts had more than one abnormality. Sensitivity
and specificity for detection of stenoses and occlusions of Thus far, no studies have been published comparing mul-
CE-MRA were 91% and 95%, respectively. In five cases, tislice CTA and MRA in assessment of patency of the
CE-MRA showed extra complications: Four nonthrom- peripheral vessel directly. Willman et al. (69) compared the
botic ectasias not seen on ultrasound and one thrombotic performance of CTA and MRA in evaluation of stenoses in
ectasia, overlooked at conventional angiography. Meissner the renal and aortoiliac arteries and found no significant
et al. (66) performed MRA in 24 patients with 26 femoro- differences in detecting hemodynamically significant ste-
tibial or femoropedal grafts. The degree of stenosis at MRA noses: Sensitivity/specificity figures (two observers) were
was compared to findings at duplex US, and, in case of dis- 92%/93% and 100%/99% for MRA and 91%/92% and
crepancy, to DSA findings. In 109 of 117 evaluated segments 99%/99% for CTA, respectively. However, a significant
(93%), MRA and duplex US showed concordant findings. difference in examination time occurred: On an average,
In eight discordant segments in seven patients, duplex US 24 minutes for CTA and 35 minutes for MRA. On the other
overlooked four high-grade stenoses, correctly identified at hand, times required for 3D reconstruction and image inter-
MRA and confirmed by DSA. In no case did MRA miss an pretation were much longer for CTA (20 to 45 minutes)
LWBK1209-ch35_p575-598.indd 594 17/05/13 5:20 PM

A B C
Figure 35.26. MRA-MIP images of a patient with a patent femorofemoral crossover bypass because of
occluded right iliac arteries.
than for MRA (5 to 9 minutes). The most time-consuming Magnetic Resonance Contrast Agents
procedure in CTA is segmentation of surrounding nonvas-
In MR arteriography of the lower limbs, when extracellular
cular structures.
contrast agents, such as gadolinium-DTPA, are administered
They also examined patient acceptance of the various
at a low infusion rate, venous enhancement occurs relatively
examination procedures. Patients found MS-CTA the least
late as a result of diffusion into the interstitium. The disad-
uncomfortable procedure and DSA the most uncomfortable
vantage of using low infusion rates is the dilution of contrast
procedure. Surprisingly, no statistically significant difference
agent in the bloodstream, with resulting relatively moderate
occurred in patient acceptance between DSA and MRA.
enhancement. Using higher infusion rates results in higher
The most disturbing factors in MRA are noise, having to
local concentration of the contrast agent in the vessels of
keep still, and the lengthy procedure and arterial puncture
interest, but at the cost of faster venous return and a shorter
and postprocedural application of pressure in DSA.
arterial/venous time window for imaging. A possible solu-
tion to this problem may be the use of higher concentration
formulations, such as gadobutrol 1M (Gadovist, Schering,
New Developments Berlin, Germany) (72,73). With these extracellular contrast
agents, however, extravasation into the interstitium results
in reduced contrast-to-background ratios.
Technique
Although rapid first-pass image acquisition with conven-
Ongoing hardware and gradient improvements result in tional gadolinium agents can help prevent artifacts due to
shorter scan times, enabling time-resolved imaging with high patient movement and flow artifacts in nonaxial arteries,
resolution. Time-resolved imaging will continue to improve MRA suffers from reduced spatial resolution compared with
with higher resolution, real-time reconstruction, and view- alternative procedures such as CTA and diagnostic DSA and
ing capabilities (70). Besides obviating the need for con- is susceptible to poor synchronization between bolus arrival
trast bolus timing, time-resolved imaging also adds an extra in distal arterial branches and acquisition especially in pa-
dimension to imaging of the peripheral arteries. Because of tients with heavily diseased vessels and slow arterial flow
the dynamic nature of the series, visualizing differences in below the knee. This has been highlighted by the results of
influx of contrast agent in the two limbs is possible, similar studies with dedicated intravascular (blood pool) contrast
to conventional angiography. Also, the use of magnets with agents that have demonstrated improved sensitivity for ste-
higher field strengths (3 T) results in improved visualization nosis depiction and quantification with the additional avail-
of small infrapopliteal vessels and better delineation of ste- ability of steady-state (SS) images (74).
noses (71). The advent of intravascular “blood pool” gadolinium-
In addition, dedicated coils for peripheral vascular im- based MR contrast agents which bind strongly to serum
aging have become available which result in images with albumin offers the possibility to obtain SS images with in-
higher resolution and signal-to-noise ratios, but—in case of creased in-plane and slice resolution during the equilibrium
phased array coils—also aid in reducing imaging times by phase of circulation following the acquisition of conventional
allowing acquisition of data in parallel. FP images (75).
LWBK1209-ch35_p575-598.indd 595 17/05/13 5:20 PM

Combining first-pass and SS k-space segmentation after adequate treatment planning (37). CTA on four-slice scan-
injection of a blood pool contrast agent is feasible and pro- ners was therefore not a cost-effective alternative to MRA.
vides high-resolution data with sharp delineation of the arte- The results of CTA with the “64-plus” generation scanners
rial vessel tree while reducing venous contamination (76). have dramatically improved and are at least comparable to
Unfortunately, however, blood pool agents are costly and of the results of MRA.
limited value for applications other than MRA; as a result, the Which of the two modalities—CTA or MRA—will even-
solitary approved intravascular agent in Europe (Gadofosveset tually prevail for imaging of the peripheral vessels is diffi-
or Vasovist) is no longer commercially available. cult to say. Considering the disadvantages of CTA (use of
A separate contrast agent, Gadobenate Dimeglumine or iodinated contrast agents and difficulties in assessment of
MultiHance (Bracco), binds weakly to albumin (77,78). patency of vessels, in case of calcified plaques), an indication
Because of this weak binding to protein, interstitial diffusion for MRA will remain in patients with diabetes and intoler-
is diminished but not absent. MultiHance, therefore, is not a ance to iodinated contrast agents. In other patients, it will
blood pool agent in sensu strictu. SS imaging of the peri depend on local facilities, infrastructure, and local experi-
pheral arteries with MultiHance is possible and may result ence to determine which of the two modalities will be used
in improved depiction of the arteries in calf and foot (79). for noninvasive imaging of the peripheral arteries.
The major advantage of MultiHance is the higher relax-
ivity as compared to the extracellular agents. This property
can be used to obtain higher signal in the vessels or to lower References
the dose and injection volume requirement.
1. Martin EC. Transcatheter therapies in peripheral and nonvascular disease. Circulation.
1991;83:1–5.
2. Waugh JR, Sacharias N. Arteriographic complications in the DSA era. Radiology. 1992;138:
Computed Tomography 237–281.
Angiography Technique 3. Goldman K, Salvesen S, Hegedus V. Acute renal failure after contrast medium injection.
Invest Radiol. 1984;S19–S125.
Until recently, state of the art CT was defined by “64- 4. Niendorf HP, Haustein J, Louton T, et al. Safety and tolerance after intravenous adminis-
tration of 0.3 mmol/kg Gd-DTPA: Results of a randomized, controlled clinical trial. Invest
generation” scanners that produced 64-slices, either 64-detector Radiol. 1991;26(suppl 1):221S–225S.
rows, or 32-detector rows and a strategy to double the slice 5. Sadowski EA, Bennett LK, Chan MR, et al. Nephrogenic systemic fibrosis: Risk factors and
number by alternate deflection of the focal spot of the x-ray incidence estimation. Radiology. 2007;243:148–157.
6. Grobner T. Gadolinium: A specific trigger for the development of nephrogenic fibrosing der-
source. More modern CT scanners make up the “post-64 era” mopathy and nephrogenic systemic fibrosis? Nephrol Dial Transplant. 2006;21:1104–1108.
and improve on the fundamental CT parameters. 7. Nelemans PJ, Leiner T, de Vet HCW, et al. Peripheral arterial disease: Meta-analysis of the
Wide area detector CT extends craniocaudal coverage to diagnostic performance of MR angiography. Radiology. 2000;217:105–114.
8. Miyazaki M, Takai H, Sugiura S, et al. Peripheral MR angiography: Separation of arter-
16 cm and enables simultaneous large volume craniocaudal ies from veins with flow-spoiled gradient pulses in electrocardiography-triggered three-
imaging either with no table motion (axial imaging) or via dimensional half-Fourier fast spin-echo imaging. Radiology. 2003;227:890–896.
9. Lim R, Hecht E, Xu J, et al. 3D nongadolinium-enhanced ECG-gated MRA of the distal lower
helical scanning. extremities: Preliminary clinical experience. J Magn Reson Imaging. 2008;28(1):181–189.
Introduction of automatic segmentation software has en- 10. Haneder S, Attenberger UI, Riffel P, et al. Magnetic resonance angiography (MRA) of the calf
abled much faster processing of the large dataset from peri station at 3.0 T: Intraindividual comparison of non-enhanced ECG-gated flow-dependent MRA,
continuous table movement MRA and time-resolved MRA. Eur Radiol. 2011;21:1452–1461.
pheral CTA so that postprocessing of CT is now comparable 11. Lim R, Storey P, Atanasova I, et al. Three-dimensional electrocardiographically gated vari-
to that of MRA. able flip angle FSE imaging for MR angiography of the hands at 3.0 T: Initial experience.
In dual source CT, high temporal resolution is achieved Radiology. 2009;252(3):874–881.
12. Prince MR, Yucel EK, Kaufman JA, et al. Dynamic gadolinium-enhanced three-dimen-
from two coplanar x-ray tubes and detector arrays. The sional abdominal MR arteriography. J Magn Reson Imaging. 1993;3:677–881.
temporal resolution can be reduced to one-quarter of a gan- 13. Kopka L, Vosshenrich R, Rodenwaldt J, et al. Differences in injecting rates on contrast-
enhanced breath-hold three-dimensional MR angiography. AJR Am J Roentgenol. 1998;170:
try rotation, with the greatest advantage being coronary im- 345–348.
aging with reduced motion artifact. The second advantage 14. Ho KY, de Haan MW, Kessels AG, et al. Peripheral vascular tree stenosis: Detection with
is that the system can operate at two energies, one at a low subtracted and nonsubtracted MR angiography. Radiology. 1998;206:673–681.
15. Foo TKF, Saranathan M, Prince MR, et al. Automated detection of bolus arrival and initia-
kVp and the other at a high kVp to achieve a dual-energy tion of data acquisition in fast, three-dimensional MR angiography image quality. Radiology.
acquisition. The ability to provide material decomposition 1997;203:275–280.
imaging can be applied to CTA with removal of vessel wall 16. Leiner T, Ho KY, Nelemans PJ, et al. Three-dimensional contrast-enhanced moving-bed
infusion-tracking (MoBi-track) peripheral MR angiography with flexible choice of imaging
calcium and decrease of blooming artifact (80). parameters for each field of view. J Magn Reson Imaging. 2000;11:368–377.
However, while correlating well with DSA above the 17. Frayne R, Grist TM, Swan JS, et al. 3D MR DSA: Effects of injection protocol and image
masking. J Magn Reson Imaging. 2000;12:476–487.
knee, accuracy of DE-CTA MIPs appeared to be moderate 18. Andreisek G, Pfammatter T, Goepfert K, et al. Peripheral arteries in diabetic patients:
in the calf and largely insufficient in calcified pedal arteries, Standard bolus-chase and time-resolved MR angiography. Radiology. 2007;242:610–620.
especially in patients with critical limb ischemia (81). 19. Watanabe Y, Dohke M, Okumura A, et al. Dynamic subtraction contrast-enhanced MR angi-
ography: Technique, clinical applications, and pitfalls. Radiographics. 2000;20:135–152.
20. Westenberg JJ, Wasser MN, van der Geest RJ, et al. Scan optimization of gadolinium-
enhanced three-dimensional MRA of peripheral arteries with multiple bolus injections and
Summary in vitro validation of stenosis quantification. Magn Reson Imaging. 1999;17:47–57.
21. Huber A, Heuck A, Bauer A, et al. Dynamic contrast-enhanced MR angiography from the
distal aorta to the ankle joint with a step-by-step techniques. AJR Am J Roentgenol. 2000;175:
In 2003, Visser et al. (82) concluded that CTA, compared 1291–1298.
22. Ho KY, Leiner T, De Haan MW, et al. Peripheral vascular tree stenoses: Evaluation with
to MRA, will be cost-effective if the costs are $300 or less, moving-bed infusion-tracking MR angiography. Radiology. 1998;206:683–692.
if sensitivity for detection of significant stenoses is higher 23. Wang Y, Lee HM, Khilnani NM, et al. Bolus-chase MR digital subtraction angiography in
than 94%, and if 20% or fewer patients require additional the lower extremity. Radiology. 1998;207:263–269.
24. Meaney JF, Ridgway JP, Chakraverty S, et al. Stepping-table gadolinium-enhanced digital
workup. Using a four-slice CT, it has been found that after subtraction MR angiography of the aorta and lower extremity arteries: Preliminary experi-
CTA 35% of patients needed additional imaging before ence. Radiology. 1999;211:59–67.
LWBK1209-ch35_p575-598.indd 596 17/05/13 5:20 PM

25. Alley MT, Grist TM, Swan JS. Development of a phased-array coil for the lower extremi- 55. Whitehouse WM, Wakefield TW, Graham LM, et al. Limb-threatening potential of arterio-
ties. Magn Reson Med. 1995;34:260–267. sclerotic popliteal aneurysms. Surgery. 1983;93:694–699.
26. Leiner T, Nijenhuis RJ, Maki JH, et al. Use of a three-station phased array coil to improve 56. Adiseshiah M, Bailey DA. Aneurysms of the femoral artery. Br J Surg. 1977;64:174–176.
peripheral contrast-enhanced magnetic resonance angiography. J Magn Reson Imaging. 57. Graham LM, Zelenock GB, Whitehouse WM, et al. Clinical significance of arteriosclerotic
2004;20:417–425. femoral artery aneurysms. Arch Surg. 1980;115:502–507.
27. Sodickson DK, McKenzie CA, Li W, et al. Contrast-enhanced 3D MR angiography with 58. Diwan A, Sarkar R, Stanley JC, et al. Incidence of femoral and popliteal artery aneurysms
simultaneous acquisition of spatial harmonics: A pilot study. Radiology. 2000;217:284–289. in patients with abdominal aneurysms. J Vasc Surg. 2000;31:863–869.
28. Weiger M, Pruessmann KP, Kassner A, et al. Contrast-enhanced 3D MRA using SENSE. 59. Zhang HL, Kent KC, Bush HL, et al. Soft tissue enhancement on time-resolved peripheral
J Magn Reson Imaging. 2000;12:671–677. magnetic resonance angiography. J Magn Reson Imaging. 2004;19:590–597.
29. Bezooijen R, van den Bosch HC, Tielbeek AV, et al. Peripheral arterial disease: Sensitivity- 60. Dyet JF, Nicholson AA, Ettles DF. Vascular imaging and intervention in peripheral arteries
encoded multiposition MR angiography compared with intraarterial angiography and con- in the diabetic patient. Diabetes Metab Res Rev. 2000;16(suppl 1):S16–S22.
ventional multiposition MR angiography. Radiology. 2004;231:263–271. 61. Baum RA, Rutter CM, Sunshine JH, et al. Multicenter trial to evaluate vascular magnetic
30. Binkert CA, Baker PD, Petersen BD, et al. Peripheral vascular disease: Blinded study of resonance angiography of the lower extremity. JAMA. 1995;274:875–880.
dedicated calf MR angiography versus standard bolus-chase MR angiography and film 62. Kreitner KF, Kalden P, Neufang A, et al. Diabetes and peripheral arterial occlusive disease.
hard-copy angiography. Radiology. 2004;232:860–866. Prospective comparison of contrast-enhanced three-dimensional MR angiography with
31. Vogt FM, Ajaj W, Hunold P, et al. Venous compression at high-spatial-resolution three- conventional digital subtraction. AJR Am J Roentgenol. 2000;174:171–179.
dimensional MR angiography of peripheral arteries. Radiology. 2004;233:913–920. 63. Berkowitz HD, Hobbs CL, Roberts B, et al. Value of routine vascular laboratory studies to
32. Bilecen D, Schulte AC, Bongartz G, et al. Infragenual cuff-compression reduces venous identify vein graft stenosis. Surgery. 1981;90:971–979.
contamination in contrast-enhanced MR angiography of the calf. J Magn Reson Imaging. 64. Bertschinger K, Cassina PC, Debatin J, et al. Surveillance of peripheral arterial bypass
2004;20:347–351. grafts with three-dimensional MR angiography: Comparison with digital subtraction angi-
33. Bilecen D, Jager KA, Aschwanden M, et al. Cuff-compression of the proximal calf to ography. AJR Am J Roentgenol. 2001;176:215–220.
reduce venous contamination in contrast-enhanced stepping-table magnetic resonance 65. Bendib K, Berthezene Y, Croisille P, et al. Assessment of complicated arterial bypass grafts:
angiography. Acta Radiol. 2004;45:510–515. Value of contrast-enhanced subtraction magnetic resonance angiography. J Vasc Surg. 1997;
34. Mitsuzaki K, Yamashita Y, Onomichi M, et al. Delineation of simulated vascular steno- 26:1036–1042.
sis with Gd-DTPA-enhanced 3D gradient echo MR angiography: An experimental study. 66. Meissner OA, Verrel F, Tato F, et al. Magnetic resonance angiography in the follow-up
J Comput Assist Tomogr. 2000;24:77–82. of distal lower-extremity bypass surgery: Comparison with duplex ultrasound and digital
35. Maki JH, Prince MR, Londy FJ, et al. The effects of time-varying intravascular signal inten- subtraction angiography. J Vasc Interv Radiol. 2004;15:1269–1277.
sity and k-space acquisition order on three-dimensional MR angiography image quality. 67. Willmann JK, Mayer D, Banyai M, et al. Evaluation of peripheral arterial bypass grafts
J Magn Reson Imaging. 1996;6:642–651. with multi-detector row CT angiography: Comparison with duplex US and digital subtrac-
36. Korosec FR, Mistretta CA. MR angiography: Basic principles and theory. Magn Reson tion angiography. Radiology. 2003;229:465–474.
Imaging Clin N Am. 1998;6:223–256. 68. Krinsky G, Jacobowitz G, Rofsky N. Gadolinium-enhanced MR angiography of extraana-
37. Rubin GD, Schmidt AJ, Logan LJ, et al. Multi-detector row CT angiography of lower tomic arterial bypass grafts. AJR Am J Roentgenol. 1998;170:735–741.
extremity arterial inflow and runoff: Initial experience. Radiology. 2001;221:146–158. 69. Willmann JK, Wildermuth S, Pfammatter T, et al. Aortoiliac and renal arteries: Prospective
38. Kumamaru KK, Hoppel BE, Mather RT, et al. CT angiography: Current technology and intraindividual comparison of contrast-enhanced three-dimensional MR angiography and
clinical use. Radiol Clin North Am. 2010;48:213–235. multi-detector row CT angiography. Radiology. 2003;226:798–811.
39. Fleischmann D. Aorto-popliteal bolus transit times in peripheral CT angiography: Can fast 70. Du J, Korosec FR, Thornton FJ, et al. High-resolution multistation peripheral MR angiography
acquisitions outrun the bolus? Eur Radiol. 2003;13:S218. using undersampled projection reconstruction imaging. Magn Reson Med. 2004;52:204–208.
40. Fleischmann D. Present and future trends in multislice detector-row CT applications: CT 71. Leiner T, de Vries M, Hoogeveen R, et al. Contrast-enhanced peripheral MR angiography
angiography. Eur Radiol. 2002;2(suppl 2):S11–S15. at 3.0 Tesla: Initial experience with a whole-body scanner in healthy volunteers. J Magn
41. Rofsky NM, Adelman MA. MR angiography in the evaluation of atherosclerotic periph- Reson Imaging. 2003;17:609–614.
eral vascular disease: What the clinician wants to know. Radiology. 2000;214:325–338. 72. Herborn CU. Peripheral contrast-enhanced 3D MRA with 1.0 M gadobutrol. Eur Radiol.
42. Kinney TB, Rose SC. Intraarterial pressure measurements during angiographic evaluation 2004;14(suppl 5):M23–M25.
of peripheral vascular disease: Techniques, interpretation, applications, and limitations. 73. Hentsch A, Aschauer MA, Balzer jO, et al. Gadobutrol-enhanced moving-table mag-
AJR Am J Roentgenol. 1996;116:277–284. netic resonance angiography in patients with peripheral vascular disease: A prospec-
43. Wasser MN, Westenberg J, van der Hulst VP, et al. Hemodynamic significance of renal tive, multi-centre blinded comparison with digital subtraction angiography. Eur Radiol.
artery stenosis: Digital subtraction angiography versus systolically gated three-dimensional 2003;13:2103–2114.
phase-contrast MR angiography. Radiology. 1997;202:333–338. 74. Hadizadeh DR, Gieseke J, Stefan H, et al. Peripheral MR angiography with blood pool
44. Wikstrom J, Holmberg A, Johansson L, et al. Gadolinium-enhanced magnetic resonance contrast agent: Prospective intraindividual comparative study of high-spatial-resolution
angiography, digital subtraction angiography and duplex of the iliac arteries compared with steady-state MR angiography versus standard-resolution first-pass MR angiography and
intra-arterial pressure gradient measurements. Eur J Vasc Endovasc Surg. 2000;19:516–523. DSA. Radiology. 2008;249:701–711.
45. Owen RS, Carpenter JP, Baum RA, et al. Magnetic resonance imaging of angiographi- 75. Bonel HM, Bettina Saar B, Hanno Hoppe H, et al. MR angiography of infrapopliteal arter-
cally occult runoff vessels in peripheral arterial occlusive disease. N Engl J Med. 1992;326: ies in patients with peripheral arterial occlusive disease by using Gadofosveset at 3.0 T:
1577–1581. Diagnostic accuracy compared with selective DSA. Radiology. 2009;253:879–890.
46. Carpenter JP, Owen RS, Baum RA, et al. Magnetic resonance angiography of peripheral 76. Kinner S, Zenge MO, Heilmaier C, et al. Peripheral MRA with k-space segmentation and
vessels. J Vasc Surg. 1992;16:807–813. blood-pool contrast agent. Acad Radiol. 2011;18:113–119.
47. Koelemay MJ, den Hartog D, Prins MH, et al. Diagnosis of arterial disease of the lower 77. Lohr HA, Froehlich JM, Pfyffer M, et al. Comparison of Gd-BOPTA and Gd-DOTA
extremities with duplex ultrasonography. Br J Surg. 1996;83:404–409. for peripheral CE-MRA: A double-blind clinical study. Acad Radiol. 2002;9(suppl 2):
48. Legemate DA, Teeuwen C, Hoeneveld H, et al. Value of duplex scanning compared with S421–S424.
angiography and pressure measurement in the assessment of aortoiliac arterial lesions. Br J 78. Wyttenbach R, Gianella S, Alerci M, et al. Prospective blinded evaluation of Gd-DOTA-
Surg. 1991;78:1003–1008. versus Gd-BOPTA-enhanced peripheral MR angiography, as compared with digital sub-
49. Visser K, Hunink MG. Peripheral arterial disease: Gadolinium-enhanced MR angiography traction angiography. Radiology. 2003;227:261–269.
versus color-guided duplex US—a meta-analysis. Radiology. 2000;216:67–77. 79. Anzidei M, Napoli A, Zaccagna F, et al. First-pass and high-resolution steady-state mag-
50. Met R, Bipat S, Legemate DA, et al. Diagnostic performance of computed tomography netic resonance angiography of the peripheral arteries with gadobenate dimeglumine: An
angiography in peripheral arterial disease: A systematic review and meta-analysis. JAMA. assessment of feasibility and diagnostic performance. Invest Radiol. 2011;46:307–316.
2009;301:415–424. 80. Meyer BC, Werncke T, Hopfenmuller W, et al. Dual energy CT of peripheral arteries: Effect
51. Heijenbrok-Kal MH, Kock MC, Hunink MG. Lower extremity arterial disease: Multidetector of automatic bone and plaque removal on image quality and grading of stenoses. Eur J
CT angiography meta-analysis. Radiology. 2007;245:433–439. Radiol. 2008;68:414–422.
52. Koelemay MJ, Lijmer JG, Stoker J, et al. Magnetic resonance angiography for the evaluation 81. Kau T, Eicher W, Reiterer C, et al. Dual-energy CT angiography in peripheral arterial
of lower extremity arterial disease: A meta-analysis. JAMA. 2001;285(10):1338–1345. occlusive disease-accuracy of maximum intensity projections in clinical routine and sub-
53. Ruehm SG, Weishaupt D, Debatin JF. Contrast-enhanced MR angiography in patients with group analysis. Eur Radiol. 2011;21:1677–1686.
aortic occlusion (Leriche syndrome). J Magn Reson Imaging. 2000;11:401–410. 82. Visser K, Kock MC, Kuntz KM, et al. Cost-effectiveness targets for multi-detector row
54. Prince M, Narasimham D, Stanley J, et al. Gadolinium-enhanced magnetic resonance angio CT angiography in the work-up of patients with intermittent claudication.Radiology.
graphy of abdominal aortic aneurysms. J Vasc Surg. 1995;21:656–669. 2003;227:647–656.
LWBK1209-ch35_p575-598.indd 597 17/05/13 5:20 PM

LWBK1209-ch35_p575-598.indd 598 17/05/13 5:20 PM
PART
6
Interventional MRI
LWBK1209-ch36_p599-617.indd 599 17/05/13 5:20 PM

LWBK1209-ch36_p599-617.indd 600 17/05/13 5:20 PM
C h apt e r
Harald H. Quick 36
Catheter Tracking and Devices
■■ Interventional Device Visualization cardiovascular interventions such as MR-guided guidewire

Passive Instrument Visualization and catheter tracking, aortic stent grafting, and MR-guided
Active Instrument Visualization transarterial aortic valve implantation (TAVI). Issues of MR
Tracking of Resonant Radiofrequency Markers safety related to interventional devices in an MR environ-
ment are discussed.
Wireless Active Catheter Visualization
Combination with Conventional X-ray
Fluoroscopy Interventional Device
■■ Intravascular Magnetic Resonance Imaging Visualization
Vascular Stents in Magnetic Resonance Imaging
Prerequisite to the safe and successful performance of vas-
■■ Interventional Magnetic Resonance Scanner
cular interventions with MRI is not only the collection of
Features relevant anatomic information, but also the reliable visual-
Interactive User Interface ization of catheters and guidewires in relation to the sur-
■■ Safety rounding tissue morphology. In contrast to ultrasound, x-ray
fluoroscopy, or computed tomography (CT), visualization of
■■ Future Perspectives interventional instruments in MRI has proven to be difficult.
The technique used to render vascular instruments visible
Several attributes make magnetic resonance imaging (MRI) in MRI would ideally be characterized by high spatial and
attractive for guidance of intravascular therapeutic proce- temporal resolution. It should also provide a high-contrast
dures, including high soft tissue contrast, imaging in arbitrary instrument signature, making it easy to pick out the instru-
oblique planes, lack of ionizing radiation, and the ability ment in the MR image. Several approaches have been de-
to provide functional information, such as flow velocity or veloped for depicting vascular instruments in an MR envi-
flow volume per unit time, in conjunction with morphologic ronment. They can be broadly grouped into two categories:
information. For MR guidance of vascular interventions to Passive and active visualization. The passive techniques are
be safe, the interventionist must be able to visualize catheters familiar from ultrasound, x-ray fluoroscopy, and CT. The
and guidewires relative to the vascular system and surround- material properties of the instrument are manipulated so
ing tissues. Several approaches for rendering instruments vis- that the instrument appears with sufficient contrast in the
ible in an MR environment have been developed, including image itself. No additional hardware or instrument modi-
both passive and active techniques. Passive techniques depend fications are required. The active techniques rely on addi-
on contrast agents or susceptibility artifacts that enhance the tional hardware and postprocessing to achieve instrument
appearance of the catheter in the image itself, whereas active localization.
techniques rely on supplemental hardware built into the cath-
eter, such as a radiofrequency (RF) coil. In addition, the abil-
Passive Instrument Visualization
ity to introduce an RF coil mounted on a catheter presents
the opportunity to obtain high-resolution images of the vessel Two approaches to passive catheter guidance exist. The first
wall. These images can provide the capability to distinguish is to alter the blood signal relative to the catheter signal. This
and identify various plaque components. The additional capa- can be achieved with administration of a vascular contrast
bilities of MRI could potentially open up new applications agent. Gadolinium diethylenetriaminepentaacetate (DTPA)
within the purview of vascular interventions beyond those allows the acquisition of high-resolution MR angiograms
currently performed under x-ray fluoroscopic guidance. that could be used for tracking vascular instruments relative
This chapter reviews the technical and clinical require- to arterial morphologic background “road maps.” However,
ments for MR-compatible interventional instruments. The commercially available contrast agents rapidly leak out of
basic techniques for MR-guided instrument visualization the vascular space, resulting in increased signal in the back-
are demonstrated with current examples from preclinical ground tissues. This alters the signal characteristics of the
601
LWBK1209-ch36_p599-617.indd 601 17/05/13 5:20 PM

602 Part 6 ■ Interventional MRI
target vessel and potentially reduces catheter visualization. shown in Figure 36.1. Passive guidewire visualization is often
Intravascular MR contrast agents have a prolonged intra- hampered by the rather small diameter (below 1 mm) of the
vascular presence but equally opacify both arteries and veins flexible vascular instruments. To increase instrument vis-
(1,2). Strategies for reducing venous opacification have been ibility, magnetic susceptibility markers have been integrated
explored by using contrast agents based on superparamag- into the design at defined positions relative to the instrument
netic iron oxides with both T1- and T2-shortening effects (3). tip. Such markers provide small signal voids due to suscep-
By filling the catheter with a Gd-based contrast agent, both tibility artifacts at defined positions that facilitate instru-
vascular system and instruments can be visualized separately ment visualization (Figs. 36.1 and 36.2). MR-compatible
with a double-echo gradient-echo sequence. The image guidewires have successfully been tested in preclinical animal
based on the short echo renders both the vasculature and the experiments (11–15) as well as most recently in the first
catheter bright, whereas the image based on the long echo- applications in humans (16). It remains to be tested and it
time (TE) renders only the catheter bright. The catheter-only will depend on the specific applications whether such pas-
image (second echo) can be threshold and overlaid in color sively visualized MR-compatible guidewires will provide
on the vascular image (first echo) (3). sufficient mechanical stiffness, long-term stability, and image
The second approach to passive catheter guidance visibility to substitute standard guidewires that are used in
involves altering the catheter signal relative to the blood sig- x-ray guided interventions.
nal. Instruments can be filled with contrast-doped solution,
shortening the relaxation time. Imaging is accomplished by
Aortic Stent Grafting
using a short repetition time (TR)/short (TE) pulse sequence
along with a high flip angle, achieving a catheter image that Beyond the development of vascular guidewires, more
is bright relative to the background. The slice thickness is application-specific instrumentation has been investigated
generally limited, however, because the instrument rapidly for its potential use in MR-guided interventions. In a study
disappears as a result of partial voluming as the thickness is in 2006 (17), commercially available thoracic aortic stent
increased (4). Rather than filling the lumen of the catheter grafts have been systematically investigated regarding their
with liquid contrast solution, another approach is to treat MR compatibility and artifact behavior. Furthermore, the
the surface of the catheter with Gd3+ ions (5). As a conse- associated delivery catheters also have been tested toward
quence, the T1 of blood in the immediate vicinity of the cath- MR compatibility and MR safety. From an initially tested
eter is shortened, rendering the catheter visible. group of six different stent grafts and delivery devices, just
A different approach to achieving adequate catheter con- one combination of stent graft and device, the GORE TAG
trast is based on enhancing the inherent signal void (i.e., neg- aortic stent graft has been tested MR-compatible for further
ative contrast) of an instrument as it displaces spins during experimental evaluation in an in vivo study on pigs (18) (Fig.
insertion. Differences in magnetic susceptibility can be used 36.3). The stent graft with its nitinol mesh showed only mild
to create large, local losses in signal as a result of intravoxel susceptibility artifacts which indicates that no ferromagnetic
dephasing (6–8). Unfortunately, these signal losses are most wires or components have been used in the production.
often accompanied by geometric distortion of the underlying Furthermore, the large-caliber (18F) catheter delivery device
vascular anatomy. In addition, the effect is highly dependent does not contain any metal braidings or metallic compo-
on several factors, including field strength, pulse sequence nents that would render this specific device potentially MR
parameters, and device orientation within the magnetic field. unsafe. Aortic stent grafting provides an excellent example
These dependences prevent a consistent portrayal of instru- for the strength of MR guidance as well as for the advan-
ments. A useful approach has been to incorporate multiple tages of passive instrument visualization. Large-diameter
rings of paramagnetic dysprosium oxide (Dy2O3) along the instruments guided within the large-diameter aorta can
catheter tip, allowing the catheter to be consistently visual- robustly be visualized using a fast steady-state free preces-
ized independent of orientation (8). sion (SSFP, TrueFISP, balanced FFE, Fiesta, etc.) sequence.
The following sections provide three current examples With this sequence type, passive instrument visualization
for the development of MR-compatible instrumentation that provides good instrument-to-background contrast featuring
make use of passive visualization techniques. bright blood and dark instrument profiles with a high frame
rate (Fig. 36.4) (18).
Guidewires
Transarterial Aortic Valve Implantation
In the recent past, a number of MR-guided vascular inter-
ventions have been performed, evaluating MR-compatible Building on similar features of passive instrument visualiza-
and MR-safe instruments that make use of passive visualization, a most recent study has investigated the possibility to
tion techniques. As one of the most important basic instru- guide the rather new clinical procedure of TAVI under MRI
ments, vascular guidewires have been developed by different (19). Again, MR compatibility of the associated devices is
groups that are based on glass fiber or polyetheretherketone one of the preconditions for a potential translation from ani-
(PEEK)-fiber reinforced plastics (9–15). The rationale of such mal models into the clinical arena. While the tested aortic
fiber reinforcement is to provide mechanical stiffness of the valve prosthesis (CoreValve, Medtronic) has been tested MR
rather small-diameter instruments, and at the same time to compatible, the associated delivery device showed ferromag-
obviate the need for metallic cores and wires that are com- netic attraction as well as severe artifacts due to ferromag-
mon in x-ray guidewires and which pose a potential safety netic components and metallic braidings integrated along
risk in MRI. An example of a PEEK-reinforced guidewire is the delivery catheter shaft (19) (Fig. 36.5). Accordingly, the
LWBK1209-ch36_p599-617.indd 602 17/05/13 5:20 PM

Chapter 36 ■ Catheter Tracking and Devices 603
Figure 36.1. Schematic illustration, photography, and cross-section of a PEEK-based MR-safe 0.035-inch guidewire. The schematic illustration
depicts the structure and measures of the 160-cm long guidewire (A). The 0.57-mm core fiber-reinforced PEEK compound (white central zone) is
tapered toward the 120-mm long flexible distal part with a minimal diameter of 0.15 mm at the 20-mm long floppy and thereby atraumatic tip
of the wire. The guidewire has a polyurethane polymer jacket (gray peripheral zone) and a hydrophilic coating, yielding a final diameter of 0.035
inches along the whole guidewire axis. Passive MR markings are positioned at the distal wire (white circles). Photos show the flexible distal part of
the guidewire (B) and a magnification of the most distal two MR-markers (arrows, C). A photograph of a bright field microscopy (D) demonstrates
the PEEK/fiberglass core compound in cross-section. (Courtesy of Sebastian Kos, MD, and Deniz Bilecen, PhD, University of Basel, Switzerland.)
A B C
Figure 36.2. Passive instrument visualization of a PEEK-based MR-safe guidewire (see also Fig. 36.1). MR image-guided advancement of
the guidewire into the abdominal aorta of a pig. The distal end of the guidewire can be depicted in the real-time MR images due to its markings
along the guidewire structure showing as a line of dark signal spots. The guidewire is advanced from the left iliac artery (A) into the abdominal
aorta of the pig (B,C). Arrows in (A–C) point toward the guidewire tip. The larger signal void in (A–C) (marked with arrowhead in (A)) shows the
image artifact resulting from a metallic stent placed in the right renal artery of the pig. (Courtesy of Sebastian Kos, MD, and Deniz Bilecen, PhD,
University of Basel, Switzerland.)
LWBK1209-ch36_p599-617.indd 603 17/05/13 5:21 PM

A B C D
Figure 36.3. Commercially available GORE TAG thoracic aortic stent graft for catheter-based therapy of
thoracic aortic dissection. A: The stent graft loaded on the distal end of the 18F polyurethane delivery catheter.
B: Partly released stent graft. C: Fully released nitinol-based, membrane-covered stent graft. D: Corresponding
minimum intensity projection from a high-resolution T1-weighted 3D FLASH sequence. The FLASH sequence
allows for detailed evaluation of the lumen and surroundings of the stent graft indicating no severe imaging
artifacts thus serving as reference for MR-compatibility testing.
A B C D E F G
Figure 36.4. MR image-guided thoracic aortic stent graft deployment in a pig model of aortic dissec-
tion. A: Preinterventional high-resolution MRI (cine-TrueFISP retro) in parasagittal orientation showing the
dissection flap (arrowheads) in the proximal descending thoracic aorta. One of twenty acquired frames per
RR-interval is shown. B–F: Safe advancement of the stent graft delivery system (arrows) up to the level of the
dissection under rtMRI guidance. G: The correct stent graft position (arrowheads) is confirmed immediately
after stent graft deployment showing complete coverage of the dissection.
LWBK1209-ch36_p599-617.indd 604 17/05/13 5:21 PM

A B C D F
Figure 36.5. Photograph of the commercially available self-expandable nitinol-based Medtronic CoreValve
bioprosthesis and its original delivery catheter for image-guided transarterial aortic valve implantation (TAVI).
Photography (A) and high-resolution 3D FLASH image (B) of the original delivery catheter showing a side-
by-side correlation of individual device details and according imaging artifacts in MRI. The original device
here (B) shows severe imaging artifacts due to ferromagnetic catheter components as well as due to mechanical
stability reinforcing metal braiding in the catheter shaft. The redesigned delivery catheter (C) was developed
while not using metallic components or metallic braiding, and thus shows absolutely no artifacts in the high-
resolution 3D FLASH images (D). The nitinol-based aortic valve and redesigned delivery device can thus be
considered MR safe. E: Photography of the nitinol stent frame with the aortic valve. F: High-resolution 3D
FLASH sequence revealing slight MR signal attenuation within the metallic stent frame and excellent correla-
tion in showing the eyelet (arrows) at the outflow tract of the stents.
catheter delivery device has been redesigned without ferro-

Active Instrument Visualization
magnetic components and without metal braiding to render
the instrument MR compatible. Figure 36.5 shows a side- Several of the active tracking techniques that have been dem-
by-side comparison of the original delivery catheter pro- onstrated to be suitable for vascular interventions involve
viding strong artifacts and the redesigned MR-compatible the incorporation of an RF coil into the instrument itself.
delivery catheter showing no artifacts. This catheter and MR tracking relies on the incorporation of a miniature sole-
valve combination has then been further investigated in an noidal coil into the instrument (21–25). The coil is connected
in vivo study in pigs (20). Passive instrument visualization to the scanner via a thin coaxial cable passing through the
with real-time TrueFISP sequence here provided excellent catheter and provides a robust signal, identifying the instru-
MR guidance of the TAVI procedure (Fig. 36.6). ment location with high contrast. Early tracking catheter
LWBK1209-ch36_p599-617.indd 605 17/05/13 5:21 PM

A B C D E
Figure 36.6. A: Distal end of the TAVI catheter delivery device with the loaded and covered stent valve.
B: Partly released nitinol stent valve that can mechanically be released by slowly retracting the catheter sheath.
C–E: Real-time TrueFISP images of MR-guided Core Valve deployment in vivo in a pig. Arrowheads in (C–E)
show successive stent release.
designs incorporated an RF coil on the tip of interventional is directed toward the outside of the antenna, providing a
instruments (Fig. 36.7A,B), thus the tip could be visualized signal beyond the constraints of the instrument. This might
with high contrast and high temporal resolution in three be advantageous for simultaneously displaying the imme-
dimensions (22). The actively available three-dimensional diate anatomic surrounding when tracking the instrument;
(3D) spatial coordinates could also be used to steer the however, this signal characteristic hampers sharp delinea-
actual imaging plane with the instrument tip, allowing for tion of the instrument. In addition, the signal in early pro-
two-dimensional (2D) imaging updates at the exact location totypes of dipole antennas has been inherently faint at the
of the coil, with corresponding depiction of the surrounding antenna tip, which has led to insufficient instrument tip
anatomy. More current setups for catheter tip tracking now visibility.
combine fast updates of the catheter position with real-time In previous experimental studies that explored the poten-
imaging sequences, such that real-time imaging is always tial for performing MR-guided cardiovascular interventions
performed at the current catheter position (26–28). Up to with actively visualized instruments, the instruments were
three tracking coils implemented into the catheter over sev- tracked on preacquired, high-quality contrast-enhanced
eral centimeters enable to link the slice position and ori- vascular road maps (24,25,30). With this strategy, the time-
entation to the distal end of the instrument (26,27). The consuming acquisition of vascular road maps could be per-
interventionist can thus assess the current interventional formed before moving on to the acquisition of real-time
situation in real time. Practical considerations, however, updates for continuous display of the instrument position.
suggest that the entire instrument needs to be visualized, Although vascular road map and real-time instrument track-
rather than only the instrument tip. ing were thus independent from each other, such strategies
One way to obtain the curvature information missing had obvious drawbacks: The anatomic road map did not
with the MR tracking technique is to elongate the RF coil always represent the current anatomic situation during the
in the instrument. Magnetically coupled antennas with course of an intervention. Gross movements of the patient
reduced signal penetration depth can be used. These are or cardiac and breathing motion sometimes led to misregis-
the traditional looped antennas of MR, familiar from all trations between the anatomic road map and the instrument
surface coils: The coils are simply wound very thin and ex- position.
tended over a length of several centimeters (29–31). These The drawbacks of preinterventional road map acquisi-
antennas generate an outline of very limited extent, which tion strategies can be overcome by the use of fast imaging
sharply delineates the instrument. The acquisition of a con- techniques with steady-state free precession (SSFP), also
ventional MR image with these antennas leads to an outline referred to as TrueFISP, balanced FFE, Fiesta, and so forth.
or “profile” of the instrument as a result of the localized sen- Such sequences, in general, offer high image acquisition
sitivity of the coils, thus the designation “MR profiling” (Fig. speed, signal-to-noise ratio (SNR), contrast-to-noise ratio
36.7C,D) (30,31). (CNR), and furthermore, contrast characteristics that ren-
Another group of RF antennas suitable for integration der vessels hyperintense, even without the administration of
into small-diameter vascular instruments are the electri- a contrast agent, making them attractive for the guidance
cally coupled loopless antennas (dipoles or stubs) (32,33). of interventional MRA. Current setups for performing MR-
Such antennas provide a relatively homogeneous signal guided cardiovascular interventions are based on the com-
profile along the whole instrument. Signal sensitivity here bination of actively visualized instruments with real-time
LWBK1209-ch36_p599-617.indd 606 17/05/13 5:21 PM

A B
C D
Figure 36.7. Active instrument visualization through implementation of radiofrequency (RF) coils into
vascular instruments. A: “Tip-tracking” guidewire. The integrated solenoid RF coil at the instrument tip
provides localized signal that can be used to superimpose the tip position in color (red cross) onto a previ-
ously acquired road map (arrow) (B), or to steer the position of the imaging plane in real time, respectively.
C: “Profiling” guidewire. The integrated elongated helical wound RF coil provides signal over the distal end
of the instrument, resulting in a sharp signal “profile” of the tip and curvature of the instrument (red line
between arrow and arrowhead) (D). B,D: In vivo results in pig experiments
TrueFISP imaging (27,34,35). Figure 36.8 shows a sche- guidewire here is equipped with a loopless dipole antenna,
matic example of an interventional real-time imaging setup while the catheter is visualized with three independent loop
that combines active instrument visualization and TrueFISP tracking coils and additionally with a loopless antenna along
imaging to enable real-time image acquisition, reconstruc- the catheter shaft. This active visualization setup provides three
tion, fusion, and display with multiple RF receiver channels bright spots in MR images along the distal end of the catheter,
for the simultaneous and independent display of actively in combination with a bright instrument profile provided by
visualized instruments and vascular morphology (34). the loopless antenna (Fig. 36.10E) (35,36).
Figure 36.9A shows an example of a 6F 1-lumen catheter
that has been equipped with a coiled dipole antenna in order
Tracking of Resonant
to render the catheter visible during advancement in the aortic
Radiofrequency Markers
arch of a pig (Fig. 36.9B–D). Figure 36.10 shows a combination
of actively visualized guidewire and guiding catheter, featuring Thus far, most of the proposed designs for active instrument
a more complex design for active instrument visualization. The visualization in MRI have necessitated some kind of electrically
LWBK1209-ch36_p599-617.indd 607 17/05/13 5:21 PM

Images Images Fusion image
Windowing Overlay
Leveling
Color tables Contours
Catheter
Guidewire
Figure 36.8. Schematic of a setup for real-time instrument visualization. The signals of up to six phased-
array surface coil elements are fed to separate RF receiver channels. Image reconstruction results in an anatomic
image A. The individual signals of the guidewire and the catheter are fed into separate receiver channels and
reconstructed independently (G, guidewire image; C, catheter image). The reconstructed images A, C, and G are
transferred via a 100 Mbit/s. Ethernet connection to a stand-alone PC, where a real-time software application
enables individual windowing, leveling, and color coding of the individual images. An image fusion function
allows the use of an overlay technique, RGB signal mix, or contour visualization in real time. The resulting
composite image, ACG, is displayed on an in-room monitor to be viewed by the interventionist. (From: Quick
HH, Kuehl H, Kaiser G, et al. Interventional MRA using actively visualized catheters, trueFISP, and real-time
image fusion. Magn Reson Med. 2003;49:129–137, with permission.)
A B C D
Figure 36.9. Active catheter-visualization in MRI. A: A 6F catheter into which a coiled RF dipole antenna
was integrated. This RF antenna detects and amplifies the local MR signal in the close vicinity of the catheter.
B–D: Image fusion (see also Fig. 36.8) allows for color-coded overlay of the catheter signal on the TrueFISP
real-time images of the beating heart (6 images per second) while advancing the catheter. (From: Quick HH,
Kuehl H, Kaiser G, et al. Interventional MRA using actively visualized catheters, trueFISP, and real-time image
fusion. Magn Reson Med. 2003;49:129–137, with permission.)
LWBK1209-ch36_p599-617.indd 608 17/05/13 5:21 PM

A B
C D E
Figure 36.10. A: Photographs of a 0.035-inch active guidewire with the distal portion of a 7F active guid-
ing catheter also shown in (B). C: A 0.035-inch J-curved active guidewire (150-cm long). D: A 7F active guiding
catheter that has three separate channels with 0.035-inch active guidewire advanced to the tip. Note the three
small-diameter RF plugs for connecting the guiding catheter to the RF surface coil port of the MR system.
Image (E) shows these actively visualized instruments during an MR-guided in vivo procedure in a pig. The
guiding catheter at its distal end features three loop tracking coils colored in green and red. A loopless RF dipole
antenna that is embedded into the device shaft using the braiding layer displays in blue. (Courtesy of Ozgur
Kocaturk, PhD, and Robert Lederman, MD, National Heart, Lung and Blood Institute, Bethesda, Maryland.)
conducting wire connection between the instrument and structures exposed to RF fields, can be successfully elimi-
the MR scanner. This wire connects the RF coil or dipole nated with the approach described previously. However,
antenna through the instrument body to remote external tun- this approach still requires a mechanical connection (laser
ing, matching, and decoupling electronics. These electronics fiber) between the instrument and the scanner. Such connec-
are typically contained within an RF-shielded box that is tions (electrical or optical) greatly hamper handling of the
connected with an interface plug to the surface coil port of interventional instruments. Unlike in conventional x-ray
the scanner (21–37). fluoroscopy, instruments cannot be freely manipulated and
An alternative strategy pursues omission of the electri- rotated. In addition, catheters and other instruments can-
cally conducting cable. This strategy is based on self-resonant not be easily exchanged over an already-positioned guide-
RF circuits that initially have been successfully used as high- wire, which is a standard maneuver in conventional x-ray
contrast markers for localization purposes in MRI (38). Such fluoroscopy.
markers consist of a miniature high-quality RF coil tuned to The principle of inductive coupling of RF coils (43–47)
the Larmor frequency of the scanner and surrounding a small has recently successfully been applied to catheters to enable
container filled with a short T1 solution. The application of a new instrument visualization strategy: Wireless active
low flip angle excitation pulses in a fast imaging sequence catheter visualization (48). Here, catheters are designed to
allows bright depiction of the coil’s interior, because the ef- contain longitudinal single-loop RF-resonant circuits. The
fective excitation angle inside the coil is increased as a result catheter thus acts as an intravascular RF receiver whose
of the coil resonance. The background will give relatively lit- signal can be coupled to outside surface coils (Fig. 36.11).
tle signal at these low excitation angles, resulting in a positive This active instrument visualization strategy aims to (a)
contrast between marker and background. This technique provide reliable and robust high-contrast visualization of
has recently been adapted to instrument tip tracking (39–42), an instrument portion larger than just the tip, (b) avoid
where miniaturized resonant circuits with solenoidal coils an electrically conducting wire connection to avoid RF
have been mounted on vascular catheters. An optical fiber heating, and (c) avoid hampering instrument handling by
running through the catheter shaft supplies laser light pulses eliminating any mechanical (electrical wire or laser fiber)
from the scanner to a photodiode at the instrument tip to connection. The concept has been evaluated in phantom
intermittently tune and detune the resonant circuit. This experiments (Fig. 36.12A,B) and in numerous vascular ma-
technique enables high-contrast visibility and thus real-time nipulations performed on pigs (Figs. 36.12C–E and 36.13).
tracking of the instrument tip (40,41). In these interventions, wireless active catheters were guided
under real-time visualization into several different arterial
segments. Selective, time-resolved contrast-enhanced MR
Wireless Active Catheter Visualization
angiography (49–53) has been performed subsequently
By omitting the electrically conducting wire connection, at each selected location to verify the catheter position
potential RF heating issues, associated with long conducting (Fig. 36.13).
LWBK1209-ch36_p599-617.indd 609 17/05/13 5:21 PM

Inductive coupling of an elongated, resonant structure

provides high-contrast visualization of both the instrument
to receiver
tip and the curvature of the distal segment of the catheter
A Loop surface coil
body. The method completely eliminates the necessity of a
mechanical connection between the catheter and the MR
scanner, thereby simplifying instrument handling. Finally,
B1
the instrument-to-background contrast can be influenced
Single-loop coil
Distal end of
Tuning capacitor
by adjusting the flip angle of the guidance sequence. These
catheter characteristics make the technique of wireless active cath-
B Catheter-mounted
eter visualization an appealing new addition to the palette
inductively coupled single-loop coil of available visualization strategies for MR-guided interven-
tions (48).
Figure 36.11. Schematic of wireless signal coupling between two
RF coils and its application to wireless active catheter visualization.
A: Loop surface coil that is connected to the RF receiver of the scan-
Combination with Conventional X-Ray
ner. B: Distal end of a catheter that is equipped with a closed-loop RF
resonator that is tuned to resonance with a capacitor. In body coil (not
Fluoroscopy
shown) RF transmit mode, the resonant catheter coil locally multiplies One approach to bridging the current difficulties with cathe-
the excitation flip angle. In RF receive mode, the resonant catheter coil ter visualization in MR is to combine an MR magnet with an
picks up the MR signal in its immediate vicinity, resulting in a B1-field
alternative guidance modality, usually an x-ray fluoroscopy
vector that can be inductively coupled to that of the loop surface coil
(A). (From: Quick HH, Zenge MO, Kuehl H, et al. Interventional
system. One system (56,57) couples a conventional C-arm
MRA with no strings attached: Wireless active catheter visualization. fluoroscopic unit, complete with digital subtraction angiog-
Magn Reson Med. 2005;53:446–455, with permission.) raphy, road mapping, and so forth, with a cylindrical-bore
1.5 T magnet. The patient is placed on a common table that
is both MR compatible and x-ray transparent. The table can
be moved back and forth between the two systems without
This principle of wireless active instrument visualization the need for repositioning the patient. For guidewire and
has recently been investigated further and extended toward catheter placement, the fluoroscopic system can be used,
a method called reverse signal polarization. This feature pro- whereas the MR can be used for immediate postdilatation
vides a means for separation of the catheter visualization sig- flow measurements. Such combined systems can potentially
nal from the anatomical background signal, thus providing play an increasingly important role when transferring the
enhanced flexibility when it comes to catheter-to-background MR-guided interventional concepts from preclinical animal
signal isolation (54,55). studies to human applications.
A B C D E
Figure 36.12. Active wireless catheter visualization. Photograph of the distal end of a 6F catheter with
integrated self-resonant RF circuit (A), and the corresponding signal characteristics in a “high-resolution”
TrueFISP image (B). Despite the small excitation flip angle of 5 degrees, the distal end of the catheter is dis-
played with high and homogenous signal and thus with high contrast compared to the background (water-
filled phantom). The MR signal profile precisely matches the shape of the catheter up to the very instrument
tip. Arrows in (A) and (B) mark the length of the resonant RF loop coil and of the position of the tuning
capacitors at the proximal end. C–E: Sagittal plane in vivo catheter guidance through the abdominal aorta
of a pig into the celiac trunk. Image sequence was a projection reconstruction TrueFISP with a frame rate of
6 frames per second. Note the flip angle of 20 degrees allows for a bright instrument signal while the back-
ground signal remains low, thus allowing for high instrument-to-background contrast. The position and the
curvature of the distal end of the catheter, including the tip, are always exactly determinable. (From: Quick
HH, Zenge MO, Kuehl H, et al. Interventional MRA with no strings attached: Wireless active catheter visual-
ization. Magn Reson Med. 2005;53:446–455, with permission.)
LWBK1209-ch36_p599-617.indd 610 17/05/13 5:21 PM

A C
Figure 36.13. A: Maximum intensity projection (MIP) of an MR angiogram (MRA) acquired in the
abdominal aorta of a pig following intravenous contrast injection gives a comprehensive view of the vascular
morphology. Images (B) and (C) show MIPs of selective time-resolved high-resolution 3D FLASH contrast-
enhanced MRA of the celiac trunk with the superior mesenteric artery (SMA) in coronal (B) and sagittal
(C) projections. Contrast agent was administered through the tip of the wireless active catheter, which had been
previously tracked into the celiac trunk, as shown in Figure 34.7. The distal active end of the catheter was col-
ored in red. Note excellent correlation between the intravenous MRA in (A) and the intra-arterial MRA in
(B) and (C). (From: Quick HH, Zenge MO, Kuehl H, et al. Interventional MRA with no strings attached:
Wireless active catheter visualization. Magn Reson Med. 2005;53:446–455, with permission.)
Intravascular Magnetic safety and image quality requirements: (a) High local SNR
Resonance Imaging for high-resolution imaging; (b) minimized radial sensitivity
falloff for sufficient signal penetration depth into the vessel
Beyond active catheter visualization for tracking purposes, wall; (c) homogeneous signal response over several centime-
catheter-mounted RF coils potentially offer the ability to ters for multislice imaging; and (d) insensitivity to the orien-
acquire high-resolution intravascular MR images of the ves- tation of the coil, with respect to the main magnetic field B0.
sel wall. Since such coils are introduced into and manipu- Furthermore, features, such as the suppression of flow arti-
lated through the vessels, they can be placed at virtually every facts, a small diameter of the RF coil system, and a nonrigid,
vascular region of interest. Such an approach for vessel wall flexible catheter-based design for easy insertion into small
imaging inherently provides very high SNR localized at the vessels, have to be considered essential prerequisites (64).
lesion of interest, which often outperforms the SNR that sur- A significant number of different coil designs have been
face coils would provide at an identical position (47,58). The developed with the challenging objective to fulfill most of
concept of intravascular MRI has been explored for some these design requirements that are often conflicting goals
time now (59–65). For an intravascular MRI concept to suc- and have been evaluated in vitro as well as in vivo in ani-
ceed, the design of the imaging coil needs to fulfill various mal experiments. Among these intravascular imaging coil
LWBK1209-ch36_p599-617.indd 611 17/05/13 5:21 PM

A C D E F
B G H 1 year I 2 years J 3 years
Figure 36.14. A: Close-up of the tip of a prototype catheter with a balloon-mounted single-loop RF coil
for high-resolution intravascular MRI. B: Coronal high-resolution view acquired with the balloon imaging
catheter inside an arterial segment. Dark structures inside the vessel wall represent calcifications (three arrows,
B). Images (C–F) show in vivo high-resolution (117 × 156 mm) intravascular T2-weighted images, acquired
with the balloon-mounted intravascular coil inside the abdominal aorta of atherosclerotic Watanabe rabbits.
Images (G–J) represent histopathologic correlations (hematoxylin–eosin stain) transecting the abdominal aorta
of four different animals aged 6 months (C,G), 1 (D,H), 2 (E,I), and 3 (F,J) year in identical locations. Wall
thickness and plaque area increase with increasing age. Calcified plaque characterized by reduced signal intensi-
ties on MR images (arrow, F) was confirmed in the histologic hematoxylin–eosin stain (arrow, J). The 0.035-inch
guidewire lumen is visible inside the inflated balloon (arrow).
designs were opposed solenoid coils (59,60,62,65), flexible requirement for such metallic implants to be suited for use
single-loop coils for intra-arterial (63,64) and intravenous in MR-guided procedures is that they are MR compatible.
(66) insertion, balloon-mounted loop coils (Fig. 36.14A) This implies that no hazardous forces or torques that could
(64,67), and “loopless” dipole antennas (33) that inherently dislodge the implant occur within the MRI scanner, and
provide less SNR in the near-field when compared to con- that the artifacts produced by the implants are acceptable.
ventional looped coils. Many authors have studied the artifacts produced by metal-
Since the placement of such endoluminal coils requires lic stents (73–78), with the conclusion that for the majority
arterial puncture, their use must be coupled with an endo- of stents a detailed evaluation of the stent lumen is difficult,
vascular intervention, and is thus not considered to be purely or even impossible. The extent of the artifact depends on
diagnostic. The outstanding resolution of the intravascular the stent type, the material, the geometry, and the stent ori-
images might be used to characterize various plaque compo- entation to the main magnetic field, B0. The consequence is
nents, thereby providing guidance for choice of the optimal that stents that can be used in interventional MRI proce-
therapeutic strategy to pursue (Fig. 36.14) (68–70). Although dures should preferably be made of so-called low-artifact
intravascular ultrasound can be used currently to character- materials, such as nitinol or tantalum, which will not cause
ize plaques, the advantages of MRI lie in its superior soft extensive susceptibility artifacts. However, even stents that
tissue contrast (71), its ability to see behind calcified plaques are made of such materials, in general, allow for only quali-
and the struts of intravascular stents (72), and the ease with tative patency assessment but not quantification of in-stent
which tomographic or 3D data, with a known relation to re-stenosis (77).
other anatomic landmarks, can be generated. In a more re-
cent study, the concept of opposed solenoid RF coils has been
Active Stent Visualization
revisited to combine active catheter tracking with subsequent
high-resolution intravascular MRI (Fig. 36.15A) (65). One To eliminate the aforementioned MRI limitations associated
demonstration revealed that an intravascular RF coil in con- with conventional stents, investigators have pursued the idea
junction with fast TrueFISP sequences can provide sufficient to use vascular stents as intravascular RF antennas. In an initial
SNR to achieve an in-plane image resolution of around 300 mm study (72), catheter-mounted self-expandable Wallstents were
within 6 seconds per slice (Fig. 36.15C). Image acquisition connected via coaxial cable to tuning and matching electron-
speed will be a critical factor when transferring the concept ics. Thus, the stent acted as an active intravascular antenna
of intravascular MRI to human applications. that provided a high signal for MR-guided stent placement.
Furthermore, that technique provided high-resolution imaging
of the vessel wall outside the stent and high-resolution imag-
Vascular Stents in Magnetic
ing inside of the stent lumen. Such a cable-based approach
Resonance Imaging
for active stent visualization is invasive and thus is limited to
Many interventional vascular procedures use arterial stents the time of the intervention. Follow-up studies are potentially
to reduce re-stenosis rates following angioplasty. A major difficult to achieve, because the approach requires electrical
LWBK1209-ch36_p599-617.indd 612 17/05/13 5:21 PM

B C
Figure 36.15. Opposed solenoid prototype tracking and intravascular imaging catheter (A). The catheter
contains three independent loops (L1, L2, L3) that can be individually connected to independent RF receivers.
B,C: Intravascular imaging of the abdominal aorta in a porcine animal model in vivo. Imaging sequence in
(B) was a double inversion recovery (IR) T2-weighted turbo spin-echo with TE 96 milliseconds, TR two car-
diac cycles, slice thickness 3 mm, in-plane resolution 195 × 195 mm, and an image acquisition time of 6 minutes per
slice. Imaging sequence in (C) was a TrueFISP with TE 3.8 milliseconds, TR 7.5 milliseconds, slice thickness
3 mm, and in-plane resolution 313 × 313 mm. Image acquisition time was 6 seconds per slice only. Although
in-plane resolution in (C) was less than in (B), imaging with the TrueFISP sequence in (C) provided enough
SNR to significantly reduce the image acquisition time. (Courtesy of Hillenbrand CM and Duerk JL.
University Hospitals of Cleveland, Cleveland, OH.)
contact between the stent and the MR scanner, and hence often conflicting goals. For optimal performance, a stent
repeated invasive access. antenna should ideally: (a) Resonate at the Larmor frequency
A novel study applies the principle of wireless inductive of the MR system; (b) provide high, local signal amplifica-
coupling between two RF coils to wireless active visualization, that is, provide a high-quality factor, Q, of the resonant
tion of stents (Fig. 36.16) (79,80). Here, the stent mesh itself circuit; (c) provide high signal homogeneity, both radially
acts as a closed loop inductor that is tuned with a chip capac- and axially inside the stent; (d) function largely independent
itor to resonance. The stent thus functions as an intravascu- of orientation relative to the B1-field of the RF transmit coil;
lar RF resonator that inductively and thus wirelessly couples (e) provide optimal coupling to an outside surface coil, with
its RF signal to an outside RF surface coil that is connected the coupling being largely independent of the stent orienta-
to the MR scanner. The receiver coil system consisting of tion relative to the surface coil; and (f) be insensitive to dif-
stent resonator and surface coil performs as an intravascular ferent loading conditions once the stent is implanted, that is,
signal-amplifier and enables high-contrast stent visualization in-stent tissue remodeling, the state of stent expansion, and
and, furthermore, high-resolution MRI of the stent lumen as the radius of stent bending should not influence the reso-
a result of its high, local SNR (Figs. 36.17 and 36.18) (80). nance frequency or Q. Finally, to facilitate implantation, the
By using the stent prototypes as resonant structures in stent mesh has to be designed balloon- or self-expandable
an MR environment, the stent mesh has to fulfill at least for catheter-based stent delivery. The proposed method for
two basic requirements (80): The mechanical stent function active stent visualization could provide a powerful diagnos-
itself, providing optimal vessel wall support following treat- tic means for the noninvasive long-term follow-up of stent
ment of the lesion, and the electrical RF antenna function. patency, thereby enhancing the understanding of the mecha-
The latter can be further characterized by the following, nisms of in-stent re-stenosis.
LWBK1209-ch36_p599-617.indd 613 17/05/13 5:21 PM

B1 features than just instrument and device visualization. For

an interventional imaging concept to succeed, the setup must
account for imaging flexibility and interactivity to stream-
A Loop surface coil to receiver
line the workflow required for performing image-guided
interventions. Moving patient tables and flexible, interac-
tive user interfaces with in-room monitors or displays are
important prerequisites to successfully perform MR-guided
B1 vascular interventions.
B Inductively coupled stent Tuning capacitor Interactive User Interface

The guidance of vascular interventions with MRI requires
Figure 36.16. Schematic of the principle of inductive coupling interactive and flexible adaptation of sequence parameters
between two coils and its application to active wireless visualization (e.g., slice position, slice orientation, slice thickness, flip angle,
of stents. A: Loop surface coil that is connected to the RF receiver of field-of-view (FOV), imaging matrix, and so forth) during
the scanner. B: Schematic of an active stent RF resonator. The mesh continuous real-time data acquisition (27,81). Simultaneous
of the stent forms a loop coil that is tuned to the Larmor frequency of multiplanar data acquisition and display of two or more
the MR scanner with a chip tuning capacitor. In body coil (not shown)
independent imaging slices facilitates operator orientation
RF transmit mode, the stent resonator locally multiplies the excita-
and instrument navigation in complex vascular territories
tion flip angle. In RF receive mode, the stent resonator picks up the
MR signal in its immediate vicinity, resulting in a B1-field vector that (34,81). The visual information needs to be presented to the
can be inductively coupled to that of the loop surface coil (A). The RF interventionist inside of the scanner room. Large in-room
receiver coil system, consisting of implanted stent and surface coil, is monitors or high-contrast back projection screens, in con-
thus acting as an intravascular signal amplifier and potentially allows junction with RF-shielded video beamers, have turned out to
high-resolution MRI of deep-sited regions of interest. be a practical and convenient solution.
Interventional Magnetic Safety

Resonance Scanner Features
The primary consideration when performing any interven-
The transition of MR from a purely diagnostic imaging tion is ensuring the safety of the patient. A major concern
modality into an imaging platform to guide therapeutic vas- in interventional MR is the possibility of localized increases
cular interventions requires realization of more technical in the RF specific absorption rate (SAR) near interventional
A B C
D E F
Figure 36.17. Photographs of a balloon-expandable stent resonator prototype, length 28 mm, inner
diameter before/after expansion 1.8/4 mm. Stent resonator in the expanded state, tuned to resonance at 63.8
MHz (A), stent in the folded state (B), folded stent mounted on a 5F balloon catheter (balloon 40 × 4 mm)
(C), unfolded stent after full inflation of balloon (D), fully deployed stent (E), and MR image acquired with the
stent immersed in an NaCl phantom (F).
LWBK1209-ch36_p599-617.indd 614 17/05/13 5:21 PM

A C
Figure 36.18. In vivo images of a solenoidal stent (length 25 mm, inner diameter 2 mm) implanted into
the right iliac artery of a pig. The outside surface receive coil was placed coaxially above the position of the
stent. The distance from the middle of the stent to the center of the loop coil was approximately 12 cm. The
interior of the stent displays with high signal on the MR image (A). Image parameters were 2D FLASH, TR/TE
300/7 ms, field-of-view (FOV) 200 × 200 mm, matrix 256 × 128, slice 2 mm, flip angle 10 degrees, time 4 hours
and 18 minutes. Further reduction of the FOV to 100 × 100 mm while maintaining all other imaging parameters
enables full assessment of the stent lumen over its full length, parallel (B), and orthogonal to the axis of the stent
(C). The signal void in the middle of the stent lumen (arrows, B,C) was identified as thrombus after explan-
tation. Image portions of 30 × 25 mm (B) and 25 × 25 mm (C) are shown, respectively. (From: Quick HH,
Kuehl H, Kaiser G, et al. Inductively coupled stent antennas in MRI. Magn Reson Med. 2002;48:781–790, with
permission.)
instruments. The local electric field can be amplified, espe- coil at the tip. Incorporation of coaxial chokes can reduce
cially if the instruments are composed of long, conducting the electric-field coupling and prevent excessive heating
structures, making the peak SAR difficult to predict. Most (83,86). Another strategy is to eliminate the coaxial cable
MR tracking, MR profiling, and intravascular imaging tech- with its potential risk for RF heating, on the basis of the
niques currently involve incorporation of a long, electrically implementation of self-resonant structures into the instru-
conducting cable and a small coil. The small coil, if prop- ments that are detuned with optical photoresistors driven by
erly detuned, will not couple significantly with the transmit optic fibers (39–42) or to completely omit any conducting
energy of the body coil. The coupling to the coaxial cable is cable connection, as described in the wireless active catheter
more difficult to reduce. The cable is basically a long antenna visualization section (48).
sensitive to the transmit electric field of the body coil.
Significant temperature increases have been demonstrated
in high-field imagers (1.5 T) near the tips of MR track- Future Perspectives
ing and MR profiling instruments, when using RF-intense
imaging sequences, such as fast spin-echo (82,83). This type Interventional cardiovascular MRI has the potential to pro-
of heating has also been shown for conventional vascular foundly alter intravascular therapy and is currently a thriv-
guidewires with conducting cores (84,85), indicating that ing area of research. Beyond duplicating the success of pro-
the problem is truly related to the long cable and not the cedures currently performed with x-ray fluoroscopy in an
LWBK1209-ch36_p599-617.indd 615 17/05/13 5:21 PM

environment free of ionizing radiation, new applications 16. Tzifa A, Krombach GA, Krämer N, et al. Magnetic resonance-guided cardiac interven-
tions using magnetic resonance-compatible devices: A preclinical study and first-in-
may emerge. Vessel wall characterization has already been man congenital interventions. Circ Cardiovasc Interv. 2010;3(6):585–592. Epub 2010
demonstrated in proof-of-concept form. Another exciting Nov 23.
17. Eggebrecht H, Zenge M, Ladd ME, et al. In vitro evaluation of current thoracic aortic
field of endeavor on the horizon is the transvascular delivery stent-grafts for real-time MR-guided placement. J Endovasc Ther. 2006;13(1):62–71.
of therapeutic drugs and genes. MRI offers a unique oppor- 18. Eggebrecht H, Kühl H, Kaiser GM, et al. Feasibility of real-time magnetic resonance-
tunity for both guidance of gene delivery and monitoring of guided stent-graft placement in a swine model of descending aortic dissection. Eur Heart J.
2006;27(5):613–620. Epub 2006 Jan 23.
gene expression. Beyond the guidance of pure vascular inter- 19. Kahlert P, Eggebrecht H, Plicht B, et al. Towards real-time cardiovascular magnetic reso-
ventions, cardiac interventions benefit from the inherent 2D nance-guided transarterial aortic valve implantation: In vitro evaluation and modification
and 3D natures of MRI and its unique ability to simulta- of existing devices. J Cardiovasc Magn Reson. 2010;12:58.
20. Kahlert P, Parohl N, Albert J, et al. Real-time magnetic resonance imaging-guided transar-
neously visualize cardiac chambers and myocardium, with terial aortic valve implantation: In vivo evaluation in swine. J Am Coll Cardiol. 2012;59
excellent contrast and sufficient temporal resolution. (2):192–193.
21. Dumoulin CL, Souza SP, Darrow RD. Real-time position monitoring of invasive devices
In this context, passive and active instrument visualiza- using magnetic resonance. Magn Reson Med. 1993;29:411–415.
tion techniques both certainly have their right to coexist. 22. Ladd ME, Zimmermann GG, McKinnon GC, et al. Visualization of vascular guidewires
Although active techniques offer a broad range of inherent using MR tracking. J Magn Reson Imaging. 1998;8:251–253.
23. Wendt M, Busch M, Wetzler R, et al. Shifted rotated keyhole imaging and active
advantages (e.g., high positive instrument contrast, instru- tip-tracking for interventional procedure guidance. J Magn Reson Imaging. 1998;8:
ment color coding, instrument overlay onto road maps, co- 258–261.
ordinates available for active steering of the imaging plane, 24. Leung DA, Debatin JF, Wildermuth S, et al. Intravascular MR tracking catheter:
Preliminary experimental evaluation. Am J Roentgenol. 1995;164:1265–1270.
and so forth), their widespread use is currently limited as a 25. Wildermuth S, Debatin JF, Leung DA, et al. MR imaging-guided intravascular procedures:
result of more complex instrument design and related RF Initial demonstration in a pig model. Radiology. 1997;202:578–583.
26. Zhang Q, Wendt M, Aschoff AJ, et al. A multielement RF coil for MRI guidance of inter-
safety issues. Here, the passive instrument visualization tech- ventional devices. J Magn Reson Imaging. 2001;14:56–62.
niques are valuable in performing initial basic but safe MR- 27. Elgort DR, Wong EY, Hillenbrand CM, et al. Real-time catheter tracking and adaptive
guided vascular procedures (16,18,19), and thus to generate imaging. J Magn Reson Imaging. 2003;18:621–626.
28. Zuehlsdorff S, Umathum R, Volz S, et al. MR coil design for simultaneous tip tracking and
confidence into the method. Also, they define the necessary curvature delineation of a catheter. Magn Reson Med. 2004;52:214–218.
workflow while moving the field from its current research 29. Ladd ME, Erhart P, Debatin JF, et al. Guidewire antennas for MR fluoroscopy. Magn
platform with animal feasibility studies into widespread Reson Med. 1997;37:891–897.
30. Ladd ME, Zimmermann GG, Quick HH, et al. Active MR visualization of a vascular
clinical applications for the benefit of patients and investiga- guidewire in vivo. J Magn Reson Imaging. 1998;8:220–225.
tors. Certainly, these promises warrant further investment in 31. Burl M, Coutts GA, Herlihy DJ, et al. Twisted-pair RF coil suitable for locating the track
of a catheter. Magn Reson Med. 1999;41:636–638.
developing this burgeoning field. 32. McKinnon GC, Debatin JF, Leung DA, et al. Towards active guidewire visualization in
interventional magnetic resonance imaging. MAGMA. 1996;4:13–18.
33. Ocali O, Atalar E. Intravascular magnetic resonance imaging using a loopless catheter
References antenna. Magn Reson Med. 1997;37:112–118.
34. Quick HH, Kuehl H, Kaiser G, et al. Interventional MRA using actively visualized cath-
1. Bluemke DA, Stillman AE, Bis KG, et al. Carotid MR angiography: Phase II study of safety eters, TrueFISP, and real-time image fusion. Magn Reson Med. 2003;49:129–137.
and efficacy for MS-325. Radiology. 2001;219:114–122. 35. Kocaturk O, Saikus CE, Guttman MA, et al. Whole shaft visibility and mechanical perfor-
2. Grist TM, Korosec FR, Peters DC, et al. Steady-state and dynamic MR angiography with mance for active MR catheters using copper-nitinol braided polymer tubes. J Cardiovasc
MS-325: Initial experience in humans. Radiology. 1998;207:539–544. Magn Reson. 2009;11:29.
3. Nanz D, Weishaupt D, Quick HH, et al. TE-switched double-contrast enhanced visualiza- 36. Kocaturk O, Kim AH, Saikus CE, et al. Active two-channel 0.035’’ guidewire for interven-
tion of vascular system and instruments for MR-guided interventions. Magn Reson Med. tional cardiovascular MRI. J Magn Reson Imaging. 2009;30(2):461–465.
2000;43:645–648. 37. Lederman RJ, Guttman MA, Peters DC, et al. Catheter-based endomyocardial injection
4. Omary RA, Unal O, Koscielski DS, et al. Real-time MR imaging-guided passive catheter with real-time magnetic resonance imaging. Circulation. 2002;105:1282–1284.
tracking with use of gadolinium-filled catheters. J Vasc Interv Radiol. 2000;11(8):1079–1085. 38. Burl M, Coutts GA, Young IA. Tuned fiducial markers to identify body locations with
5. Unal O, Li J, Cheng W, et al. MR-visible coatings for endovascular device visualization. minimal perturbation of tissue magnetization. Magn Reson Med. 1996;36:491–493.
J Magn Reson Imaging. 2006;23(5):763–769. 39. Wong EY, Zhang Q, Duerk JL, et al. An optical system for wireless detuning of parallel
6. Rubin DL, Ratner AV, Young SW. Magnetic susceptibility effects and their application in resonant circuits. J Magn Reson Imaging. 2000;12:632–638.
the development of new ferromagnetic catheters for magnetic resonance imaging. Invest 40. Eggers H, Weiss S, Boernert P, et al. Image-based tracking of optically detunable parallel
Radiol. 1990;25:1325–1332. resonant circuits. Magn Reson Med. 2003;49(6):1163–1174.
7. Kochli VD, McKinnon GC, Hofmann E, et al. Vascular interventions guided by ultrafast 41. Weiss S, Kuehne T, Brinkert F, et al. In vivo safe catheter visualization and slice tracking
MR imaging: Evaluation of different materials. Magn Reson Med. 1994;31:309–314. using an optically detunable resonant marker. Magn Reson Med. 2004;52:860–868.
8. Bakker CJ, Hoogeveen RM, Hurtak WF, et al. MR-guided endovascular interven- 42. Kuehne T, Fahrig R, Butts K. Pair of resonant fiducial markers for localization of
tions: Susceptibility-based catheter and near-real-time imaging technique. Radiology. endovascular catheters at all catheter orientations. J Magn Reson Imaging. 2003;17:
1997;202:273–276. 620–624.
9. Buecker A, Spuentrup E, Schmitz-Rode T, et al. Use of a nonmetallic guide wire for magnetic 43. Schnall MD, Barlow C, Subramanian VH, et al. Wireless implanted magnetic resonance
resonance-guided coronary artery catheterization. Invest Radiol. 2004;39(11):656–660. probes for in vivo NMR. J Magn Reson. 1986;68:161–167.
10. Mekle R, Hofmann E, Scheffler K, et al. A polymer-based MR-compatible guidewire: A 44. Kuhns PL. Inductive coupling and tuning in NMR probes: Applications. J Magn Reson.
study to explore new prospects for interventional peripheral magnetic resonance angiography 1988;78:69–76.
(ipMRA). J Magn Reson Imaging. 2006;23(2):145–155. 45. Farmer TH, Cofer GP, Johnson GA. Maximizing contrast to noise with inductively coupled
11. Krueger S, Schmitz S, Weiss S, et al. An MR guidewire based on micropultruded fiber- implanted coils. Invest Radiol. 1990;25:552–558.
reinforced material. Magn Reson Med. 2008;60(5):1190–1196. 46. Wirth ED III, Mareci TH, Beck BL, et al. A comparison of an inductively coupled implanted
12. Mekle R, Zenge MO, Ladd ME, et al. Initial in vivo studies with a polymer-based coil with optimized surface coils for in vivo NMR imaging of the spinal cord. Magn Reson
MR-compatible guide wire. J Vasc Interv Radiol. 2009;20(10):1384–1389. Epub 2009 Med. 1993;30:626–633.
Aug 21. 47. Arnder LL, Shattuck MD, Black RD. Signal-to-noise ratio comparison between surface
13. Kos S, Huegli R, Hofmann E, et al. MR-compatible polyetheretherketone-based guide wire coils and implanted coils. Magn Reson Med. 1996;35:727–733.
assisting MR-guided stenting of iliac and supraaortic arteries in swine: Feasibility study. 48. Quick HH, Zenge MO, Kuehl H, et al. Interventional MR angiography with no strings
Minim Invasive Ther Allied Technol. 2009;18(3):181–188. attached: Wireless active catheter visualization. Magn Reson Med. 2005;53:446–455.
14. Kos S, Huegli R, Hofmann E, et al. Feasibility of real-time magnetic resonance-guided 49. Serfaty JM, Atalar E, Declerck J, et al. Real-time projection MR angiography: Feasibility
angioplasty and stenting of renal arteries in vitro and in Swine, using a new polyetherether- study. Radiology. 2000;217:290–295.
ketone-based magnetic resonance-compatible guidewire. Invest Radiol. 2009;44(4): 50. Bos C, Smits HF, Bakker CJ, et al. Selective contrast-enhanced MR angiography. Magn
234–241. Reson Med. 2000;44:575–582.
15. Kos S, Huegli R, Hofmann E, et al. First magnetic resonance imaging-guided aortic stent- 51. Serfaty JM, Yang X, Foo TK, et al. MRI-guided coronary catheterization and PTCA: A
ing and cava filter placement using a polyetheretherketone-based magnetic resonance feasibility study on a dog model. Magn Reson Med. 2003;49:258–263.
imaging-compatible guidewire in swine: Proof of concept. Cardiovasc Intervent Radiol. 52. Omary RA, Green JD, Schirf BE, et al. Real-time magnetic resonance imaging-guided coro-
2009;32(3):514–521. Epub 2008 Dec 30. nary catheterization in swine. Circulation. 2003;107:2656–2659.
LWBK1209-ch36_p599-617.indd 616 17/05/13 5:21 PM

53. Green JD, Omary RA, Schirf BE, et al. Catheter-directed contrast-enhanced coronary 70. Correia LC, Atalar E, Kelemen MD, et al. Intravascular magnetic resonance imaging
MR angiography in swine using magnetization-prepared True-FISP. Magn Reson Med. of aortic atherosclerotic plaque composition. Arterioscler Thromb Vasc Biol. 1997;17:
2003;50:1317–1321. 3626–3632.
54. Celik H, Ulutürk A, Tali T, et al. A catheter tracking method using reverse polarization for 71. Martin AJ, Ryan LK, Gotlieb AI, et al. Arterial imaging: Comparison of high-resolution US
MR-guided interventions. Magn Reson Med. 2007;58(6):1224–1231. and MR imaging with histologic correlation. Radiographics. 1997;17:189–202.
55. Celik H, Atalar E. Reverse polarized inductive coupling to transmit and receive radiofre- 72. Quick HH, Ladd ME, Nanz D, et al. Vascular stents as RF antennas for intravascular MR
quency coil arrays. Magn Reson Med. 2012;67(2):446–456. guidance and imaging. Magn Reson Med. 1999;42:738–745.
56. Van Vaals JJ. Interventional MRI with a hybrid high-field system. In: Debatin JF, Adam G, 73. Teitelbaum GP, Bradley WG Jr, Klein BD. MR imaging artifacts, ferromagnetism, and mag-
eds. Interventional Magnetic Resonance Imaging. Berlin: Springer-Verlag; 1997:19–32. netic torque of intravascular filters, stents, and coils. Radiology. 1988;166:657–664.
57. Adam G, Neuerburg J, Bucker A, et al. Interventional magnetic resonance. Initial clinical 74. Hilfiker PR, Quick HH, Debatin JF. Plain and covered stent-grafts: In vitro evaluation of
experience with a 1.5-tesla magnetic resonance system combined with c-arm fluoroscopy. characteristics at three-dimensional MR angiography. Radiology. 1999;211:693–697.
Invest Radiol. 1997;32:191–197. 75. Schurmann K, Vorwerk D, Bucker A, et al. Magnetic resonance angiography of nonfer-
58. Ocali O, Atalar E. Ultimate intrinsic signal-to-noise ratio in MRI. Magn Reson Med. 1998; romagnetic iliac artery stents and stent-grafts: A comparative study in sheep. Cardiovasc
39:462–473. Intervent Radiol. 1999;22:394–402.
59. Hurst GC, Hua J, Duerk JL, et al. Intravascular (catheter) NMR receiver probe: 76. Klemm T, Duda S, Machann J, et al. MR imaging in the presence of vascular stents: A
Preliminary design analysis and application to canine iliofemoral imaging. Magn Reson systematic assessment of artifacts for various stent orientations, sequence types, and field
Med. 1992;24:343–357. strengths. Magn Reson Imaging. 2000;12:606–615.
60. Martin AJ, Plewes DB, Henkelman RM. MR imaging of blood vessels with an intravascu- 77. Bartels LW, Smits HF, Bakker CJ, et al. MR imaging of vascular stents: effects of suscepti-
lar coil. J Magn Reson Imaging. 1992;2:421–429. bility, flow, and radio frequency eddy currents. J Vasc Interv Radiol. 2001;12:365–371.
61. Kandarpa K, Jakab P, Patz S, et al. Prototype miniature endoluminal MR imaging catheter. 78. Wall A, Kugel H, Bachman R, et al. 3.0 T vs. 1.5 T MR angiography: In vitro comparison
J Vasc Interv Radiol. 1993;4:419–427. of intravascular stent artifacts. J Magn Reson Imaging. 2005;22(6):772–779.
62. Martin AJ, Henkelman RM. Intravascular MR imaging in a porcine animal model. Magn 79. Kivelitz D, Wagner S, Hansel J, et al. The active magnetic resonance imaging stent
Reson Med. 1994;32:224–229. (AMRIS): Initial experimental in vivo results with locally amplified MR angiography and
63. Atalar E, Bottomley PA, Ocali O, et al. High resolution intravascular MRI and MRA by flow measurements. Invest Radiol. 2001;36:625–631.
using a catheter receiver coil. Magn Reson Med. 1996;36:596–605. 80. Quick HH, Kuehl H, Kaiser G, et al. Inductively coupled stent antennas in MRI. Magn
64. Quick HH, Ladd ME, Zimmermann-Paul GG, et al. Single-loop coil concepts for intravas- Reson Med. 2002;48:781–790.
cular MR imaging. Magn Reson Med. 1999;41:751–758. 81. Guttman MA, Lederman RJ, Sorger JM, et al. Real-time volume rendered MRI for inter-
65. Hillenbrand CM, Elgort DR, Wong EY, et al. Active device tracking and high-resolution ventional guidance. J Cardiovasc Magn Reson. 2002;4:431–442.
intravascular MRI using a novel catheter-based, opposed-solenoid phased array coil. Magn 82. Wildermuth S, Dumoulin CL, Pfammatter T, et al. MR-guided percutaneous angioplasty:
Reson Med. 2004;51:668–675. Assessment of tracking safety, catheter handling and functionality.Cardiovasc Intervent
66. Martin AJ, McLoughlin RF, Chu KC, et al. An expandable intravenous RF coil for arterial Radiol. 1998;21:404–410.
wall imaging. J Magn Reson Imaging. 1998;8:226–234. 83. Ladd ME, Quick HH. Reduction of resonant RF heating in intravascular catheters using
67. Quick HH, Ladd ME, Hilfiker PR, et al. Autoperfused balloon catheter for intravascular coaxial chokes. Magn Reson Med. 2000;43:615–619.
MR imaging. J Magn Reson Imaging. 1999;9:428–434. 84. Nitz WR, Oppelt A, Renz W, et al. On the heating of linear conductive structures as guide
68. Zimmermann-Paul GG, Quick HH, Vogt P, et al. High-resolution intravascular magnetic wires and catheters in interventional MRI. J Magn Reson Imaging. 2001;13:105–114.
resonance imaging: Monitoring of plaque formation in heritable hyperlipidemic rabbits. 85. Konings MK, Bartels LW, Smits HF, et al. Heating around intravascular guidewires by
Circulation. 1999;99:1054–1061. resonating RF waves. J Magn Reson Imaging. 2000;12:79–85.
69. Martin AJ, Gotlieb AI, Henkelman RM. High-resolution MR imaging of human arteries. 86. Atalar E. Safe coaxial cables. In: Proc., ISMRM, 7th Scientific Meeting and Exhibition,
J Magn Reson Imaging. 1995;5:93–100. 1999. Philadelphia, PA; 1999:1006.
LWBK1209-ch36_p599-617.indd 617 17/05/13 5:21 PM

Chapter
37 Arno Bücker
Endovascular Interventional
■■ Endovascular Interventional Mri of the background anatomy as well as of the interventional

Introduction device. Of course, both the vascular anatomy and the inter-
Techniques ventional device would have to be depicted with high spatial
resolution. The application of contrast media should not be
Safety Aspects
necessary in order to avoid toxic concentrations of nowa-
■■ Endovascular Applications of Interventional days commercially available contrast agents in case of mul-
Mri tiple injections. While the application of blood pool agents
MR-guided Percutaneous Transluminal can circumvent this problem, the resulting overlap of arterial
Dilatation and venous anatomy is not desirable.
MR-guided Stent Placement So far, we are still struggling to get close to the ideal situ-
MR-guided Placement of Vena Cava Filters ation of real-time 3D imaging for vascular MR interventions.
The acquisition of 3D datasets is still too slow to allow real-
MR-guided Tips Procedure
time imaging. But first attempts for vascular interventions
MR-guided Radiofrequency Ablation have been successfully performed relying on roadmaps of 3D
of the Heart datasets in conjunction with real-time projection of a cathe-
MR-guided Coil Embolization ter tip onto previously reconstructed maximum intensity pro-
MR-guided Cardiac Interventions jections (MIPs) (1). In addition to high temporal, high spatial
■■ Future Perspectives resolution is also needed, in order to visualize small vessels
as well as interventional instruments. These instruments have
to be made of non-ferromagnetic material in order to avoid
Endovascular large artifacts. Furthermore, safety aspects have to be con-
Interventional MRI sidered, which do not allow the use of metallic devices like
guidewires, which could act as an antenna (2).
Introduction Many of the above-mentioned “challenges” have been, at
least partially, met by now, although the ideal solution has
Magnetic resonance angiography (MRA) has supplanted not been found yet. Further research is needed to optimize
x-ray angiography as the first diagnostic step after ultra- currently applied techniques and instruments before they will
sound. A competitor to MRA is computed tomography be integrated in a clinical routine setup. This is reflected by
angiography (CTA), despite its disadvantages of ionizing the fact that almost all of the applications reported have only
radiation, the need for nephrotoxic contrast agents, and been performed in animal studies. Despite the unexpected
deteriorating accuracy in heavily calcified vessels. The suc- long development time, I am convinced that the technical
cess of MRA in daily routine was made possible by technical development of real-time control for endovascular interven-
progress on the front of sequence development as well as tional MRI will greatly influence the clinical routine as well
through hardware improvement. While coronary angiogra- and therefore, has its clinical impact even for those who are
phy is still a challenge for MRA, the advent of fast gradients “merely” applying Magnetic Resonance Imaging (MRI) for
with short repetition time (TR) and echo time (TE) enabling diagnostic purposes.
breath-held three-dimensional (3D) contrast-enhanced angi-
ography was a breakthrough for the clinical application of
Techniques
MRA in the region of the aorta, its branches, and even for
run-off vessels. The different requirements for endovascular interventional
The ideal MR sequence for vascular interventions would MR mentioned above have to be fulfilled all at the same time.
be the acquisition of a 3D dataset with real-time upgrades A real-time imaging sequence will not be acceptable for the
618
LWBK1209-ch37_p618-633.indd 618 16/05/13 9:17 PM

Chapter 37 ■ Endovascular Interventional Magnetic Resonance Imaging 619
control of interventions if it displays the vascular anatomy guidance of vascular interventions, although mid-field systems
without depicting interventional devices at the same time. have been successfully applied (1). Up-to-date high gradients
For didactical reasons the different technical aspects will be are also mandatory for state of the art imaging, because they
discussed separately. But it has to be kept in mind that the provide the short TR and TE not only needed for speed but
need for coordinated functionality and simultaneous fulfill- also for high image quality of so-called trueFISP (fast imag-
ment of all the major prerequisites makes MR control of ing with steady state precession) sequences, which already play
vascular interventions such a demanding and challenging a major role for diagnostic imaging of the heart and can be
field of research. expected to do so for interventions as well. Despite the fact that
many articles have been written about possible applications of
even low-field magnets for interventional MR in general—like
Real-time MR Imaging
biopsies and thermoablation (6,7)—there have been no reports
As early as 1984, real-time MR imaging has been described published about vascular interventions on low-field systems
in the literature by means of echo planar imaging (EPI) (3). showing a reasonable imaging speed and quality.
But the spatial resolution of these images has been quite low. Vascular interventions under MR guidance require
This is due to the basic physics of MRI causing a trade-off additional hardware modifications. An in-room monitor is
between spatial and temporal resolutions for the nowadays needed beside the magnet in order to be able to directly con-
routinely used Cartesian acquisition of MR images. As each trol the intervention. It should be possible to change the slice
imaging point has to be measured individually, the spatial res- position and at least some of the scan parameters on the fly
olution is directly proportional to the acquisition time of an in order to be flexible enough to optimize image quality
MR image. Therefore, standard MR techniques yield either by changing TR, TE, and flip angle. At least currently, an
images of high spatial and low temporal resolution or vice angiographic backup system is needed to be able to control
versa. New imaging strategies like radial and spiral scanning the success of MR-guided vascular interventions and to be
have become essential for vascular MR-guided interventions prepared for possible complications and emergencies. In ad-
and will be discussed as far as understanding of the technical dition, dedicated real-time MR sequences will require special
details is necessary for performing interventions and as far as hardware for the real-time reconstruction of the raw data.
the new techniques can be expected to become important for Different interventional setups fulfill these requirements at
diagnostic real-time imaging, especially of the heart. least partially and are described in the literature (8–10).
Hardware Considerations Sequence Design

The need for speed favors high-field systems for interventional MR images are created by collecting raw data in the so-
MR. On the other hand, high-field systems offer less patient called k-space. The standard acquisition scheme nowadays
access compared to open-, mid-, and low-field MR scanners is Cartesian imaging filling k-space with lines, where each
(4). A further disadvantage of high-field systems is the increase line represents one phase-encoding step (Fig. 37.1). For stan-
of susceptibility artifacts, which can be a problem for the use dard gradient- or spin-echo sequences, the acquisition of one
of metallic devices (5). Considering the need for high-quality k-space line will require one TR. Consequently, an MR gradi-
diagnostic MR angiographies as well as the prerequisite of ent-echo image with a 256 matrix (256 phase-encoding steps)
real-time imaging capabilities, it can be concluded that and a TR of 4 milliseconds will still require a total acqui-
high-field scanners of up to 1.5 T will be preferable for the sition time of 1 second. The application of fast spin-echo,
Figure 37.1. The filling of k-space can be accomplished by different strategies. Cartesian (left), spiral
(middle), and radial scanning are schematically demonstrated. In order to reconstruct an MR modulus image,
central as well as outer parts of k-space have to be covered. The central parts being mainly responsible for
contrast and signal. In case of spiral or radial scanning the number of spirals or radials, respectively, are not
directly related to spatial resolution.
LWBK1209-ch37_p618-633.indd 619 16/05/13 9:17 PM

gradient-echo techniques, or EPI allows to sample more than no problems arise due to irregular heart rhythm. The speed
one phase-encoding step during one TR period, speeding up of spiral scanning is combined with flow sensitivity which
the image acquisition. Although imaging in subseconds is yields good contrast of the cardiac chambers as well as ves-
possible with either technique, we found neither the image sels, allowing the application of this technique for vascular
contrast nor the acquisition speed sufficient for the guidance interventions. One drawback of spiral scanning is the need
of vascular interventions, at least when compared to other for a homogenous magnetic field to avoid major artifacts.
k-space strategies as described below. But Bakker et al. (11) This can cause problems when metallic devices are used,
exploit Cartesian imaging for MR guidance of interventions. which usually will disturb the required homogeneity.
They showed that on the fly background subtraction greatly Another sampling technique for k-space is radial scan-
improves the visibility of interventional instruments. This ning (16). Similar to Cartesian scanning, multiple lines are
requires special hardware modifications in order to calcu- acquired, but these lines fill k-space in a radial manner (Fig.
late and display the images fast enough (12). The frame rate 37.1). Compared to spiral scanning this technique is much
achieved was about one image every 2 seconds consisting of less time efficient, because a relatively high number of radi-
an image acquisition time of 0.5 to 1 second and additional als has to be collected in order to avoid streaking artifacts
latency due to image calculation and display. Despite this due to severe undersampling and in order to achieve a suf-
relatively low imaging speed, Bakker et al. were able to per- ficient signal-to-noise ratio. As the acquisition time of each
form balloon dilatations in vitro as well as in vivo under MR radial line is comparable to the above-mentioned Cartesian
control (13). scanning, relatively long imaging times result. Nonetheless,
In terms of fast k-space coverage, spiral scanning is one radial scanning is fast enough and can be successfully ex-
of the most efficient ways of data sampling. As the name im- ploited for the guidance of vascular interventions (17). This
plies, one or more spirals are acquired in order to fill k-space is achieved by two means. Some degree of undersampling
(Fig. 37.1). In theory, one spiral is sufficient to reconstruct can be accepted for radial scanning, because the spatial reso-
an MR image. Usually more than one spiral is acquired and lution is not directly related to the number of radial lines.
the different spirals are interleaved. The effects of this as Therefore, the spatial resolution is not essentially reduced
well as more refined techniques like partial k-space coverage by undersampling (18), and only a few radials are needed
are beyond the scope of this text and the interested reader to be able to depict a high contrast structure (Fig. 37.3). In
is referred to the literature (14,15). Depending on the exact addition, radial scanning can be ideally combined with the
form of the spiral and its length, the acquisition time for one sliding window reconstruction technique also called view
spiral will vary. Spiral scanning does allow acquisition times sharing (19). This reconstruction technique uses partially
for one image in the range of 100 milliseconds, allowing, for new data in conjunction with old data to calculate new MR
example, real-time visualization of the beating heart, which images with a frame rate higher than the acquisition speed
makes this technique also interesting for diagnostic imaging for a single complete MR image (Fig. 37.4). This results in
(Fig. 37.2). No cardiac triggering is required and therefore, an MR movie where fast movements are partially frozen due
to averaging. While this feature makes the technique at least
theoretically less well suited for diagnostic imaging of heart
function, it excellently satisfies the need to visualize a mov-
ing interventional device on a basically unchanging back-
ground anatomy. One further advantage of this technique is
the lack of its susceptibility to motion artifacts (20) allow-
ing, for example, its diagnostic application for the detection
of pulmonary embolism without the need for breath-holding
(21). Many images in this chapter are acquired by radial
scanning combined with the sliding window reconstruction
technique, yielding a frame rate of 20 images per second. As
the imaging time for one complete MR image is in the range
of seconds, there is a minor latency between the movement
of the interventional instrument as seen on the real-time im-
ages and the actual movement.
Currently, real-time imaging sequences used for MR
guidance consist of two-dimensional imaging repetitively
acquiring a single slice. Three-dimensional imaging is still
too slow to accomplish even near real-time imaging, not
to mention the difficulties of maintaining good contrast
between vessels and background anatomy. The need for
contrast does not allow to apply the principle of projec-
tion imaging to interventional MRI. Compared to x-ray, the
tomographic nature of MR is clearly a disadvantage in this
Figure 37.2. A real-time spiral image acquired in 120 milliseconds field. In order to contain the vascular anatomy of interest
nicely demonstrates a left atrial myxoma (arrow) on this long-axis view. including tortuous vessels, the acquired single slice should
The dark areas in the left atrium due to turbulent flow (curved arrow) be as thick as possible, while maintaining a good contrast
could be easily distinguished from a mass on the real-time movie. between the vascular anatomy and the background. At the
LWBK1209-ch37_p618-633.indd 620 16/05/13 9:17 PM

Figure 37.3. Images acquired with radial k-space filling of a catheter with dysprosium markers (see also
Figure 37.7) in a water bath. The numbers in the images give the number of radials acquired to calculate
the MR image, demonstrating how few radials are needed to depict the structure of the catheter markers. In
addition, the effect of increasing the number of radials can be seen yielding a higher signal-to-noise ratio and
diminishing streaking artifacts, which are caused by undersampling.
same time, the depiction of interventional devices has to thereby simulating a blood pool-agent–filled vessel. The
be considered. While the application of blood pool agents growing loss of contrast between contrast agent containing
has been shown to be advantageous (22), there are limits structures and the background with increasing slice thick-
to the slice thickness that can be used. Figure 37.5 shows ness is a principle that holds true for all MR sequences, lim-
radial images of the ureter, which is of high signal inten- iting the maximum slice thickness. Furthermore, the use of
sity because of the earlier application of gadolinium-DTPA, blood pool agents causes an overlap of venous and arterial
anatomy, which can significantly degrade the image quality.
This is illustrated by Figure 37.6 showing the intrinsic con-
trast of radial images. One image displays veins as well as
Oldest part of MR data Newest part of MR data
MR Image n
arteries. The other image is acquired with a saturation band
saturating the venous flow and thereby allowing for a much
clearer depiction of arterial anatomy.
Parallel imaging techniques like SENSE and SMASH
Continuous Data acquisition
gained fast acceptance for clinical MR imaging. After incor-
poration of these techniques into interventional MR scan-
ners, additional imaging speed will be available to further
improve image quality (23).
MR Image n + 1 Instrument Visualization

Despite the progress made in real-time imaging, the image
Time quality of the different techniques is still poor compared to
x-ray fluoroscopy. This becomes especially apparent, when a
Figure 37.4. The sliding window reconstruction technique calcu-
small catheter is to be visualized on an MR image acquired
lates images with a speed higher than the acquisition speed for a single
image. This is achieved by using old and new data to calculate the next
in real-time. Plastic catheters usually are black on MR
image. While this technique can, in principle, be applied to all k-space images due to their lack of protons. Because of the smallness
filling strategies, it is especially useful for radial scanning, because a of interventional devices, they can only be seen if the spatial
single radial line acquires inner and outer parts of k-space adding fresh resolution of the MR image is high enough. In addition, the
low- and high-frequency components to the newly calculated image. background has to be bright in order to provide a reasonable
LWBK1209-ch37_p618-633.indd 621 16/05/13 9:17 PM

Figure 37.5. Images acquired

with radial k-space filling and of
increasing slice thickness are shown.
The gadolinium-filled ureter simulates
a vessel, which is filled with a blood
pool agent, thereby constantly reduc-
ing the intraluminal T1. Despite the
contrast agent, thicker slices yield a
constantly diminishing contrast of the
ureter against the background. The
ring-like artifacts are due to the radial
filling of k-space.
contrast against the low signal intensity of instruments. So for reliable detection on the MR images. Therefore, they
far, the spatial resolution and contrast characteristics of real- homogeneously doted polyethylene catheters with ferro-
time imaging sequences are not sufficient to reliably depict magnetic material. Depending on the concentration of the
an interventional instrument. Therefore, special means have ferromagnetic material a susceptibility artifact was caused,
been introduced to visualize devices. These are divided into which exceeded the actual catheter size, thereby making the
passive and active techniques. Passive visualization is defined catheter easily visible on MR images (24). But the artifact
as direct visualization of the instrument in the acquired MR behavior of those ferromagnetic catheters largely depends
image. No changes are made on the side of the MR sequence on their orientation to the main magnetic field B0. As this
or scanner hardware for passive visualization. Active visu- parameter cannot be controlled during angiography but is
alization, on the other hand, requires special hardware and given by the vascular anatomy, those catheters are not well
usually modifications of the MR imaging sequence in order suited for MR interventions. The problem of dependency of
to exploit the abilities of MR to localize a dedicated coil in susceptibility artifacts on the orientation to B0 was solved
3D space. This information is used to actively project the by Bakker et al. (25), who composed ring-like markers of
position of this coil onto the MR image. dysprosium oxide around a catheter (Fig. 37.7). Dysprosium
belongs to the group of lanthanides (rare earth metals) as
does the well-known element gadolinium. High concentra-
Passive Visualization with Susceptibility Markers
tions of lanthanides cause signal voids due to differences
In 1990, Rubin et al. found the artifact behavior of stan- in the ability to become magnetized compared to the sur-
dard conventional radiographic catheters on MR images rounding human tissue. This ability to become magnetized
to be characterized by a small signal void, not allowing is called susceptibility and signal voids, which are due to
LWBK1209-ch37_p618-633.indd 622 16/05/13 9:17 PM

Figure 37.6. Radial scanning of the

iliac arteries demonstrates the intrinsic
contrast of this gradient-echo tech-
nique. For the first image, arteries and
veins are superimposed. The second
image was acquired with a saturation
band distally to the imaged area, result-
ing in a complete suppression of the
venous signal, thereby allowing clear
depiction of the iliac arteries.
differences in the ability to become magnetized, are therefore insulated copper wire is wound around the instrument going
termed susceptibility artifacts. The ability to reliably visualize from the hub to the tip and back (Fig. 37.7). A small energy
these catheters was proven in vitro as well as in a human vol- source is connected to the copper wire loop to allow the
unteer (26). While the dependency of susceptibility markers flow of a small amount of direct current through the wire.
on the orientation to B0 was solved, the size of the marker in According to the rule of thumb the flowing current creates a
the MR image is still dependent on sequence parameters, the local magnetic field which causes a local field inhomogeneity
TE being the most important one. Considering the difficulty and, thereby, a signal loss. Depending on the strength of the
of real-time MR imaging, it is not desirable to have to adapt current, the artifact size can be varied (Fig. 37.9). The signal
imaging parameters in order to optimally visualize interven- void can be shaped by different wire configurations. This
tional instruments. Furthermore, the tomographic nature of allows the production of only local markers at the begin-
MRI means that a large signal void is advantageous as long ning and the end of a balloon or the marking of the whole
as the instrument is located in a large vessel like the aorta. catheter length. As with the passive method first proposed
This will ensure that the device can be localized even if it is by Bakker, the wire can be wound in order to cause signal
not positioned in the middle of the imaged slice. On the other voids independent of the catheter orientation to B0 (29).
hand, for smaller vessels large susceptibility artifacts will The small currents of up to 150 mA—which were needed
obscure the vascular anatomy (27). Many dedicated inter- to produce sufficiently large areas of signal void during in
ventional scanners allow on-the-fly adaptation of sequence vivo experiments (30)—are too small to be a safety problem,
parameters, and thereby can at least somewhat influence the especially when the insulated copper wires can be positioned
size of susceptibility markers, but considering reasonable TEs inside the catheter walls as it is done with standard braiding.
for real-time scanning the effect is relatively small (Fig. 37.8). Nonetheless, the wires can act as antennas and in case of
resonance, they could heat up significantly (2). This safety
aspect will be discussed in more detail in a separate section
Passive Visualization with the
below.
Field Inhomogeneity Concept
The field inhomogeneity concept has been evaluated in
In order to be able to change the size of an instrument in animal experiments showing its advantage for the visualiza-
the MR image independently of the sequence parameters the tion of the interventional device in vessels of different size
field inhomogeneity concept was developed (28). A loop of (Fig. 37.10) (31). Applying a current around 150 mA it was
LWBK1209-ch37_p618-633.indd 623 16/05/13 9:17 PM

Figure 37.7. Passive visualization can be achieved by susceptibility markers (left image, arrows) or by the
field inhomogeneity principle (middle image), where a copper wire is forming a loop along the catheter shaft.
If wound like a double helix, as in this example, a small current sent through the insulated copper wire will
cause a signal drop around the whole length of the catheter shaft. Active visualization needs a microcoil at the
catheter tip to localize its position (right image, arrows).
Figure 37.8. The size of passive susceptibility markers can be changed by modifying the echo time. Longer
echo times yielding bigger signal voids around the markers, as can be seen on the radial gradient-echo images
acquired with increasing TE. As the increase in echo time lengthens the image acquisition and degrades the
image quality, there is a limit to the possibility of changing the effect of susceptibility markers.
LWBK1209-ch37_p618-633.indd 624 16/05/13 9:17 PM

Figure 37.9. A catheter with the

field inhomogeneity principle is imaged
in a flow phantom by radial scanning.
The artifact around the catheter can
be increased by sending a current—in
this example of 150 mA—through the
copper wire running along the catheter
shaft.
A B
Figure 37.10. Real-time radial images of a field inhomogeneity

catheter with current switched on (A,B) and off (C). While the catheter
is positioned in the aorta the proximal part of the left renal artery can
be depicted (A, white straight arrow). Artifacts caused by bowel gas
obscure the ostium of the left renal artery and the distal right renal
artery (A, curved arrows). After steering the catheter (B, long arrow)
into the left renal artery, the vessel cannot be seen, with the current
switched on (B, short arrow). Switching the current off allows visu-
alization of the left renal artery again, although it is still not possible
to see the catheter tip on this frozen image. In this image, the contrast
of the catheter shaft in the aorta was better when the current was
C switched off (C, long arrow). This is only the case if the catheter is
positioned close to the middle of the imaged slice.
LWBK1209-ch37_p618-633.indd 625 16/05/13 9:17 PM

possible to constantly visualize a catheter over its full length

in the aorta of a pig. As soon as the catheter was steered
into the renal artery the signal void obscured the vascular
anatomy. This could be simply changed by switching off the
current, while the MR imaging sequence parameters were
kept constant (17).
Although the catheter could be visualized in the proxi-
mal part of the renal artery, it was very difficult to find the
correct imaging plane in order to depict the full course of
the renal artery. As described above, the tomographic nature
of MR is a disadvantage for finding smaller vessels and/or
depicting a tortuous vascular anatomy. Active visualization
can help to solve these problems.
Active Visualization
Active visualization uses a small microcoil which is mounted
on an instrument (Fig. 37.7). The sensitivity of an MR coil
is related to its size. Therefore, a small microcoil used for
receiving MR signal only yields a signal in close proximity
Figure 37.12. The position of an active tip tracking catheter is
to the microcoil. Applying an adequate MR sequence, the projected onto this real-time radial image, being indicated by a blink-
frequency of this signal can be used to localize the position ing cross (straight arrow). The plain gradient-echo image demonstrates
of the microcoil in two-dimensional (2D) (Fig. 37.11) or nicely a stenosis in the proximal iliac artery (curved arrow), which
even in 3D space (32). If the microcoil is placed on the tip could be successfully dilated by real-time tip tracking of the active
of a catheter, the position of the catheter tip can be calcu- balloon catheter.
lated and projected onto an MR image (Fig. 37.12) (1,33).
In addition, the knowledge of the position of the microcoil
in 3D space can be exploited to position the MR imag-
ing plane in real-time to contain the microcoil, so-called profiling (36) and can be combined with the use of other
slice tracking (34). It has been shown that the number of microcoils as well.
microcoils can be increased albeit at the cost of reducing the Initially, the technique of MR tracking was used to
speed for updating the coil position (35). Furthermore, the superimpose the position of the microcoil onto a previously
shape of the microcoil can be changed to cover the length acquired MIP image (1). The 3D dataset was acquired in a
of an instrument in order to visualize a longer distance of a breath-hold after the bolus application of contrast material.
guidewire for example. This technique has been called MR Therefore, a real-time update of the background anatomy
was not possible and any movement of the patient or even
bending of a vessel due to catheter manipulation could lead
to the false impression that the microcoil was positioned
outside the vascular tree. Nonetheless, balloon occlusion,
embolization, and puncture of the portal venous system have
been successfully performed with this technique (1). An in
vitro comparison between active MR-catheter tracking and
x-ray-guided catheter steering showed a similar time needed
for both techniques (37). If this can be expected to hold true
for the more complex anatomy, in vivo conditions has to be
proven still. In order to avoid the above-mentioned move-
ment artifacts for the active tip tracking technique, real-time
imaging of the vascular anatomy as well as the real-time tip
tracking is desirable. Simultaneous real-time depiction of the
anatomy as well as the catheter tip was performed applying
a 2D radial scanning technique in vitro (34) and in vivo (38).
Multiple MR receiver channels are necessary to simultane-
ously collect the data from the microcoil as well as from the
standard MR coil. For this 2D imaging technique it is neces-
sary to position the vessel of interest in the imaged slice. This
can be ensured by using the slice tracking technique which
Figure 37.11. A microcoil receives signal only from its very exploits the knowledge of the position of the microcoil to
proximity, thereby creating an image which consists of a small-, high- change the position of the imaged slice. The general slice
intensity region. The analysis of this signal allows to calculate the orientation has to be known in order to ensure that a longer
position of the microcoil, which can be projected onto a “standard” part of the vessel of interest is included in the scan plane.
MR image. An even more sophisticated technique uses three microcoils
LWBK1209-ch37_p618-633.indd 626 16/05/13 9:17 PM

mounted on the catheter tip (39). The microcoils are needed

to calculate two points near the catheter tip, thereby giv-
ing information about the orientation of the interventional
instrument. In addition, a software was programmed which
allowed to define a point of interest, for example, a stenosis.
Those three points could be used to automatically define the
optimal imaging plane for an intervention which contained
the instrument tip and the target region. The additional
hardware and software modifications and the implementa-
tion on a 0.2 T scanner did not allow for real-time imaging.
Nonetheless, if this sophisticated approach can be fully au-
tomated and combined with a real-time imaging sequence,
the whole procedure would be applicable to clinical routine
despite its complexity.
Other Techniques
The difficulty and complexity of visualizing interventional Figure 37.13. A 40-kg pig was placed as far off center in the mag-
instruments by fast MR imaging techniques is demon- net bore as possible. The tip of the nitinol guidewire is positioned in
strated by the number of different approaches. Besides the the aorta and the distal end is bend backward in the magnet to touch
above-described main categories, there are other techniques, the at this time dead animal. Sparks, which caused skin burns, could
which vary greatly concerning their potential imaging speed, be seen at the distal end of the guidewire.
the available spatial resolution, and other general advan-
tages and disadvantages. Small, tuned antennas at the tip
of catheters have been described as fiducial markers (40),
electron spin resonance (41) and the Overhauser effect (42) (49). Another in vitro study examined guidewires and found
have been exploited for instrument visualization as well. only a temperature increase of 15° at 1.5 T and no signifi-
Catheters filled with contrast material are used in different cant heating at 0.2 T (50). Experiments performed with
imaging strategies (43–45). Intravascular MR imaging was guidewires at 1.5 T observed a heating of almost 50° in
applied for MR-guided dilatation of rabbit aortas (46). This 30 seconds reaching a maximum of 74°C (2). Touching a
list is far from complete and the interested reader is referred standard nitinol guidewire did lead to skin burns in this
to the respective literature. study in one case. During first in vivo animal experiments,
we measured a temperature increase around the tip of a stan-
dard nitinol guidewire placed in the aorta of a living pig of
35°. The pig itself was placed as far off center in the magnet
Safety Aspects
bore as possible (Fig. 37.13). Furthermore, we could repeat-
Respecting the general contraindications, MR is a safe edly produce sparks at the distal end of a commercially avail-
examination technique. Of course, no ferromagnetic mate- able nitinol guidewire simply by bending the distal wire tip to
rials can be used close to an MR scanner. But even non- touch the animal (Fig. 37.13) (27). Despite these data, there
ferromagnetic metals can cause safety problems due to heating are studies showing no heating of interventional instruments
(47) or induced electrical currents (48). For interventional when special conditions regarding the devices and the MR
instruments, the danger lies in the fact that the devices can sequences are being met; the authors draw the conclusion
act as antennas as soon as a critical length of the instrument that low SAR and limited imaging durations allow perfor-
is reached. The heating is caused by resonance of the instru- mance of interventional MRI without any substantial risk for
ment which leads to a constant feeding of radiofrequency thermal injury (51).
energy into and, thereby, heating of the device. Like a radio First steps to solve the safety problems of metallic wires
antenna which is tuned in to a radio station, the interven- or other metallic instruments have been undertaken. The
tional device may be tuned in (resonant) to the radiofre- insertion of chokes along wires has been proposed and its
quency transmitted by the body coil of the MR scanner. The feasibility to prevent heating has been shown (52). Another
conditions for resonance to occur are very difficult to predict method applies photoresistors and photo-optical methods in
under clinical conditions. It is even impossible to simulate a order to abolish the need for an electrical connection be-
worst-case scenario and measure the maximum heating for tween the MR scanner and a microcoil, thereby making the
one sequence. Besides the position of the instrument inside occurrence of heating due to resonance impossible (53,54).
the magnet bore, its orientation and also the shape of the Another approach is the use of a laser fiber to deliver en-
patient examined play a major role. The amount of radiofre- ergy to a small coil at the catheter tip, causing intravoxel
quency energy applied by an MR sequence is also important, dephasing as described by the field inhomogeneity concept
but in case of resonance even a sequence with a low specific but without the need for a conducting wire (55). Despite nu-
absorption rate (SAR) can, in theory, cause significant and merous announcements, there are no commercially available
potentially harmful heating. First in vitro experiments with MR-safe guidewires available. This is still one of the main
an active tip tracking catheter found no significant heating reasons why so few MR-guided vascular interventions have
at 0.5 T, but a temperature increase of up to 20° at 1.5 T been performed clinically.
LWBK1209-ch37_p618-633.indd 627 16/05/13 9:17 PM

Endovascular Applications of behavior of those two stent types indicated a similar artifact
Interventional MRI behavior. This nicely documents the progress that has been
made for real-time image quality (Fig. 37.14), although the
Due to the unresolved safety problems for standard metallic hardware and software demands are quite high (69). This is
instruments, almost all of the MR-guided interventions have also demonstrated by other studies, which exploited active
been performed in vitro or in animal experiments. The con- stent visualization for stent placement, using the stent as
tinuously growing number of applications will surely stimu- a receiver antenna (70) or placed stents in the pulmonary
late further research for safe instruments and the commercial valve and main pulmonary artery (71), the aorta (72–74).
production of instruments with the currently available safe
techniques. MR-Guided Placement of Vena Cava Filters
The placement of vena cava filters is a relatively simple inter-
MR-Guided Percutaneous TRANSLUMINAL vention, which, therefore, could be performed under MR
Dilatation guidance by applying relatively slow MR imaging techniques
(73,75,76,77). All studies were performed relying on passive
Percutaneous transluminal angioplasties (PTA) were among visualization of the instruments including the different vena
the first interventions performed under MR-guidance. Initially cava filters. As with stents, the artifact behavior of these filters
active tip tracking was used projecting the catheter tip onto an varies among different filter types and depends on the imaging
MIP of a previously acquired contrast-enhanced 3D dataset sequence used. Real-time radial imaging can be also used for
(1). Despite the fact that this method lacks real-time updates vena cava filter placement being able to show the renal veins
of the vascular anatomy, the technique was successfully and the inferior vena cava allowing for fast and exact position-
applied in one patient in the clinical setup of an open 0.5 T ing of the filter (Fig. 37.15) (78). The concept of active vena
scanner for PTA of an iliac artery. Simultaneous real-time cava filters allows for clear depiction of the filter and possible
active tip tracking and real-time visualization of the vascular thrombus material within (79).
anatomy was performed to dilate iliac artery stenoses in a pig Other studies proved the feasibility of filter retrieval in
model (56). Passive visualization was applied for dilatations vitro and in animal experiments (80).
of the aorta, iliac arteries, and for dialysis shunts (11,13,57).
Human studies relied on the visualization of standard nitinol
guidewires, stents, and gadolinium-filled balloons for PTA in MR-Guided TIPS Procedure
iliac (13 patients) (58), femoral, and popliteal arteries (three In 1994, MR was used in order to alleviate the planning
patients) (59). No side effects due to the use of metallic guide- of transjugular intrahepatic portosystemic shunts (TIPS)
wires were observed during these studies. Nonetheless, the before the procedure itself was performed under x-ray guid-
risk of heating of nitinol guidewires, especially, at high-field ance (81). By now there are some reports about MR-guided
systems should not be neglected. Potential radiofrequency TIPS procedures (82,83). According to my own experience,
heating of guidewires in vitro has been described by several MR guidance is of help only for the puncture of the por-
groups (2,27,50,60,61) (see also Figure 37.13). Another tal vein. Especially, the stent placement is quite difficult and
clinical study involved patients with stenoses of hemodialysis time consuming under MR guidance, because the slice has
shunts (62). Passive visualization with dysprosium markers to be oriented along the plane of the TIPS tract (83). In this
was successfully applied in conjunction with a subtraction regard, the tomographic nature of MR is a disadvantage for
technique. Renal artery dilatation under MR guidance has visualization of the TIPS tract, the portal vein, and the infe-
been performed in an animal model as well (63). rior vena cava in one image plane. Special plan scan tools,
active tip tracking, and slice tracking are needed to allow
MR-Guided Stent Placement
localization of the ideal MR imaging plane for control of the
TIPS procedure in a reasonable time.
One drawback of contrast-enhanced 3D MRA is the poor
visualization of stent lumina and the following inability to
MR-Guided Radiofrequency Ablation
quantify in stent restenosis (64). Nonetheless, there are stents
of the Heart
causing only minor artifacts in the MR images, therefore
allowing stent placement under MR guidance (65). First, The heart is an especially difficult challenge for MR-guided
the feasibility of real-time control of MR-guided stent place- interventions, because of its constant movement and the com-
ment by means of radial scanning and passive visualization plex anatomy. One article claims the possibility for MR-guided
of the stents was shown in animal experiments (66). Passive catheterization of the coronary arteries, indirectly deducing
visualization was also applied for the first MR-guided stent this possibility from successful steering of a catheter through
placement in humans (67). No real-time MR control was the aorta (84). Radiofrequency ablation of the heart has been
used in this study to perform the intervention. The artifact performed under MR guidance by means of passive catheter
behavior of stents depends on the type of stent as well as visualization (85). Besides control of the intervention MR
on the imaging sequences applied. Comparing the real- imaging offered the possibility to directly visualize the success
time radial images of ZA stents (68) to the slower standard of transmural ablation. The difficulty of defining the correct
gradient-echo images and Memotherm stents used in the scan plane, which contains the catheter, was also apparent
only clinical study (67), there is a favorable image quality during this study and the authors suggest the introduction of
of the real-time images. In vitro comparisons of the artifact active tip tracking to solve this problem (85).
LWBK1209-ch37_p618-633.indd 628 16/05/13 9:17 PM

Figure 37.14. Real-time radial images acquired during the MR-guided placement of a ZA stent (62). The
first image shows a guidewire with dysprosium markers (arrows) in the aorta of a living pig. In the second
image, the stent is introduced into the aorta (arrows). After withdrawing the stent (third image) it is partially
(fourth and fifth images), and then completely deployed (last image).
LWBK1209-ch37_p618-633.indd 629 16/05/13 9:17 PM

A B
C D
Figure 37.15. Real-time radial images with a cranially positioned saturation slab, yielding a black aorta
(A, straight arrow). The introducer sheath of an inferior vena cava filter is advanced (A, curved arrow) up to
the renal veins (B, arrows). The filter head (C, arrow) is positioned at the ostium of the left renal vein. After
deployment the filter legs (D, arrows) can be seen as well as the position of the filter head.
MR-Guided Coil Embolization (∼3 images per second) was achieved by incorporation of
time-resolved imaging contrast kinetic elements and a projec-
Besides the MR-guided dilatation of renal arteries, coil
tion dephaser-–but let it suffice to call this technique by its
embolization has been performed in pigs (86). The tortu-
apt abbreviation TRICKS (45). This technique did not allow
ous anatomy of the renal arteries did hinder the depiction
for direct visualization of the Guglielmi detachable coils,
of the whole length of the renal arteries by the applied real-
which was achieved by repeated acquisition of new MR an-
time radial scanning technique. The application of blood
giograms and, therefore, could not be controlled in real-time.
pool contrast agents might allow the acquisition of thicker
slices while maintaining the vessel-to-background contrast,
but were not used in this study. The passive visualization of
MR-Guided Cardiac Interventions
platinum and nitinol coils was possible allowing their cor-
rect placement and the flow-sensitive technique of radial Improved spatial and temporal resolution together with
scanning made it possible to directly judge the success of the excellent contrast achieved by steady-state imaging has
embolization. But the artifacts of the coils and the relatively opened MR guidance for different cardiac interventions.
low spatial resolution as compared to x-ray angiography Despite their smallness and constant movement, the coro-
made it impossible to exactly visualize the coil shape by real- nary arteries were catheterized under MR guidance (27,78)
time MRI. and even stent placement was carried out successfully in
Successful closure of surgically created canine carotid animal experiments (Figs. 37.16 and 37.17) (88). Placement
artery aneurysms was reported applying the superposition of atrial septal occluders was performed (89,90) as well as
of a gadolinium-filled catheter onto an MIP of a previously MR-guided radiofrequency ablation (85). The cardiac cham-
acquired MR angiogram (87). Near real-time visualization bers were successfully negotiated and pressure measurements
LWBK1209-ch37_p618-633.indd 630 16/05/13 9:17 PM

Figure 37.16. Real-time MR radial images showing a stainless steel stent (curved arrow), which is placed
in the LAD (arrow) under MR guidance.
Figure 37.17. MR-compatible septal

occluder placement under real-time spiral MR
imaging. After deployment of the first half of
the septal occluder in the left atrium (arrow),
the occluder is withdrawn against the septum
and its second half is deployed in the right
atrium.
LWBK1209-ch37_p618-633.indd 631 16/05/13 9:17 PM

taken (91), pulmonary (71), and aortic valve stents (92) were 6. Duerk JL, Lewin JS, Wendt M, et al. Remember true FISP? A high SNR, near 1-second
imaging method for T2-like contrast in interventional MRI at .2 T. J Magn Reson Imaging.
placed under MR guidance. One of the most promising future 1998;8:203–208.
applications for interventional cardiac MR is the transcath- 7. Busch M, Bornstedt A, Wendt M, et al. Fast “real time” imaging with different k-space
update strategies for interventional procedures. J Magn Reson Imaging. 1998;8:944–954.
eter application of substances into infarcted myocardium. 8. Adam G, Neuerburg J, Bücker A, et al. Interventional magnetic resonance. Initial clinical
Feasibility studies on intramyocardial injections were done experience with a 1.5-tesla magnetic resonance system combined with c-arm fluoroscopy.
(93,94) and the possibility to correctly demonstrate the dis- Invest Radiol. 1997;32:191–197.
9. Silverman SG, Jolesz FA, Newman RW, et al. Design and implementation of an interven-
tribution of the injected substances by MRI was proved (95). tional MR imaging suite. AJR Am J Roentgenol. 1997;168:1465–1471.
MR-guided transfemoral aortic valve placement was exam- 10. Paley M, Mayhew JE, Martindale AJ, et al. Design and initial evaluation of a low-cost
ined in vitro (96). 3-Tesla research system for combined optical and functional MR imaging with interven-
tional capability. J Magn Reson Imaging. 2001;13:87–92.
Due to the lack of safe interventional instruments, all of 11. Bakker CJ, Smits HF, Bos C, et al. MR-guided balloon angioplasty: In vitro demonstration
the above described cardiac interventions were done in ani- of the potential of MRI for guiding, monitoring, and evaluating endovascular interven-
tions. J Magn Reson Imaging. 1998;8:245–250.
mal experiments only. But a first approach using standard, 12. van der Weide R, Zuiderveld KJ, Bakker CJ, et al. Image guidance of endovascular inter-
purely nonmetallic instruments for the interventions has ventions on a clinical MR scanner. IEEE Trans Med Imaging. 1998;17:779–785.
been made in 16 patients (97). Especially, the children were 13. Smits HF, Bos C, van der Weide R, et al. Endovascular interventional MR: Balloon angio-
plasty in a hemodialysis access flow phantom [corrected] [published erratum appears in
profiting from these cardiac interventions due to the reduced J Vasc Interv Radiol. 1998;9(6):1024]. J Vasc Interv Radiol. 1998;9:840–845.
radiation exposure. 14. Spielman DM, Pauly JM, Meyer CH. Magnetic resonance fluoroscopy using spirals with
variable sampling densities. Magn Reson Med. 1995;34:388–394.
15. Pipe JG, Ahunbay E, Menon P. Effects of interleaf order for spiral MRI of dynamic pro-
cesses. Magn Reson Med. 1999;41:417–422.
Future Perspectives 16. Rasche V, de Boer RW, Holz D, et al. Continuous radial data acquisition for dynamic MRI.
Magn Reson Med. 1995;34:754–761.
17. Bücker A, Adam G, Neuerburg JM, et al. [Real-time MRI with radial k-space scanning
The most obvious advantage of MR is the lack of radia- technique for control of angiographic interventions]. Rofo. 1998;169:542–546.
18. Peters DC, Korosec FR, Grist TM, et al. Undersampled projection reconstruction applied
tion and consequently, it will have to compare against x-ray to MR angiography. Magn Reson Med. 2000;43:91–101.
guidance for interventions. But in addition to the lack of 19. Riederer SJ, Tasciyan T, Farzaneh F, et al. MR fluoroscopy: Technical feasibility. Magn
radiation, there are a number of further potential benefits Reson Med. 1988;8:1–15.
20. Glover GH, Pauly JM. Projection reconstruction techniques for reduction of motion effects
for MR: The ability to visualize vascular anatomy without in MRI. Magn Reson Med. 1992;28:275–289.
contrast material and to depict the vessel walls and differen- 21. Haage P, Bücker A, Kruger S, et al. [Radial k-scanning for real-time MR imaging of central
tiate between lipid-rich and calcified plaques, the possibility and peripheral pulmonary vasculature]. Rofo. 2000;172:203–206.
22. Bakker CJ, Bos C, Weinmann HJ. Passive tracking of catheters and guide wires by contrast-
to measure blood flow to name only a few. Of course, these enhanced MR fluoroscopy. Magn Reson Med. 2001;45:17–23.
advantages are balanced by difficulties due to the tomo- 23. Guttman MA, Kellman P, Dick AJ, et al. Real-time accelerated interactive MRI with adap-
tive TSENSE and UNFOLD. Magn Reson Med. 2003;50:315–321.
graphic nature of MRI, the problems of using and visual- 24. Rubin DL, Ratner AV, Young SW. Magnetic susceptibility effects and their application in
izing standard interventional instruments, the complexity of the development of new ferromagnetic catheters for Magnetic Resonance Imaging. Invest
the MR technique, and the currently still time-consuming Radiol. 1990;25:1325–1332.
25. Bakker CJ, Hoogeveen RM, Weber J, et al. Visualization of dedicated catheters using fast
acquisition of image data compared to high-resolution real- scanning techniques with potential for MR-guided vascular interventions. Magn Reson
time x-ray imaging. But all of these still-existing problems Med. 1996;36:816–820.
have been solved or constantly reduced over the last years. 26. Bakker CJ, Hoogeveen RM, Hurtak WF, et al. MR-guided endovascular interventions:
Susceptibility-based catheter and near-real-time imaging technique. Radiology. 1997;202:
The most important next step will be to develop MR-safe 273–276.
interventional instruments. The high number of different 27. Bücker A, Spuentrup E, Schmitz-Rode T, et al. Use of a nonmetallic guide wire for Magnetic
Resonance-guided coronary artery catheterization. Invest Radiol. 2004;39:656–660.
approaches mirrors the complexity of this task. On the other 28. Glowinski A, Adam G, Bücker A, et al. Catheter visualization using locally induced,
hand, the manifold applications successfully performed actively controlled field inhomogeneities. Magn Reson Med. 1997;38:253–258.
under MR guidance in animal experiments and a few even 29. Glowinski A, Kursch J, Adam G, et al. Device visualization for interventional MRI using
local magnetic fields: Basic theory and its application to catheter visualization. IEEE Trans
in a clinical setting clearly demonstrate the potential of MR. Med Imaging. 1998;17:786–793.
In the field of pediatric cardiology and radiofrequency abla- 30. Adam G, Glowinski A, Neuerburg J, et al. [Catheter visualization in MR-tomography:
tion, lack of radiation exposure makes it especially worth- Initial experimental results with field-inhomogeneity catheters]. Rofo. 1997;166:324–328.
31. Adam G, Glowinski A, Neuerburg J, et al. Visualization of MR-compatible catheters by
while to explore MR-guided interventions (98). electrically induced local field inhomogeneities: Evaluation in vivo. J Magn Reson Imaging.
Furthermore, MR offers unique imaging abilities for the 1998;8:209–213.
32. Ackerman JL, Offutt MC, Buxton RB, et al. Rapid 3-D tracking of small RF coils. In:
depiction of myocardial infarction and guidance of intra- Proceedings of 5th Annual Meeting of SMR, Montreal, Canada 1986:1131.
myocardial injections. This might make MR not only an 33. Dumoulin CL, Souza SP, Darrow RD. Real time position monitoring of invasive devices
alternative but the imaging modality of choice for some using magnetic resonance imaging. Magn Reson Med. 1993;29:411–415.
34. Rasche V, Holz D, Kohler J, et al. Catheter tracking using continuous radial MRI. Magn
interventions. Reson Med. 1997;37:963–968.
35. Ladd ME, Zimmermann GG, McKinnon GC, et al. Visualization of vascular guide wires
using MR tracking. J Magn Reson Imaging. 1998;8:251–253.
36. Ladd ME, Erhart P, Debatin JF, et al. Guide wire antennas for MR fluoroscopy. Magn
References Reson Med. 1997;37:891–897.
37. Zimmermann-Paul GG, Ladd ME, Pfammatter T, et al. MR versus fluoroscopic guidance
1. Wildermuth S, Debatin JF, Leung DA, et al. MR imaging-guided intravascular procedures: of a catheter/guide wire system: In vitro comparison of steerability. J Magn Reson Imaging.
Initial demonstration in a pig model. Radiology. 1997;202:578–583. 1998;8:1177–1181.
2. Konings MK, Bartels LW, Smits HF, et al. Heating around intravascular guide wires by 38. Bücker A, Adam G, Neuerburg JM, et al. Simultaneous real-time visualization of the cath-
resonating RF waves. J Magn Reson Imaging. 2000;12:79–85. eter tip and vascular anatomy for MR-guided PTA of iliac arteries in an animal model.
3. Mansfield P. Real-time echo-planar imaging by NMR. Br Med Bull. 1984;40:187–190. J Magn Reson Imaging. 2002;16:201–208.
4. Adam G, Bücker A, Glowinski A, et al. [Interventional MR tomography: Equipment con- 39. Zhang Q, Wendt M, Aschoff AJ, et al. Active MR guidance of interventional devices with
cepts]. Radiologe. 1998;38:168–172. target-navigation. Magn Reson Med. 2000;44:56–65.
5. Lewin JS, Duerk JL, Jain VR, et al. Needle localization in MR-guided biopsy and aspira- 40. Coutts GA, Gilderdale DJ, Chui M, et al. Integrated and interactive position tracking and
tion: Effect of field strength, sequence design, and magnetic field orientation. AJR Am J imaging of interventional tools and internal devices using small fiducial receiver coils.
Roentgenol. 1996;166:1337–1345. Magn Reson Med. 1998;40:908–913.
LWBK1209-ch37_p618-633.indd 632 16/05/13 9:17 PM

41. Ehnholm GJ, Vahala ET, Kinnunen J, et al. Electron spin resonance (ESR) probe for inter- 71. Kuehne T, Saeed M, Higgins CB, et al. Endovascular stents in pulmonary valve and artery
ventional MRI instrument localization. J Magn Reson Imaging. 1999;10:216–219. in swine: Feasibility study of MR imaging-guided deployment and postinterventional
42. Joensuu RP, Sepponen RE, Lamminen AE, et al. A shielded Overhauser marker for MR assessment. Radiology. 2003;226:475–481.
tracking of interventional devices. Magn Reson Med. 2000;43:139–145. 72. Mahnken AH, Chalabi K, Jalali F, et al. Magnetic resonance-guided placement of aor-
43. Nanz D, Weishaupt D, Quick HH, et al. TE-switched double-contrast enhanced visualiza- tic stents grafts: Feasibility with real-time magnetic resonance fluoroscopy. J Vasc Interv
tion of vascular system and instruments for MR-guided interventions. Magn Reson Med. Radiol. 2004;15:189–195.
2000;43:645–648. 73. Kos S, Huegli R, Hofmann E, et al. First magnetic resonance imaging-guided aortic stent-
44. Omary RA, Unal O, Koscielski DS, et al. Real-time MR imaging-guided passive catheter ing and cava filter placement using a polyetheretherketone-based magnetic resonance
tracking with use of gadolinium-filled catheters. J Vasc Interv Radiol. 2000;11:1079–1085. imaging-compatible guide wire in swine: Proof of concept. Cardiovasc Intervent Radiol.
45. Unal O, Korosec FR, Frayne R, et al. A rapid 2D time-resolved variable-rate k-space sam- 2009;32:514–521.
pling MR technique for passive catheter tracking during endovascular procedures. Magn 74. Frericks BB, Elgort DR, Hillenbrand C, et al. Magnetic resonance imaging-guided renal
Reson Med. 1998;40:356–362. artery stent placement in a Swine model: Comparison of two tracking techniques. Acta
46. Yang X, Atalar E. Intravascular MR imaging-guided balloon angioplasty with an MR Radiol. 2009;50:21–27.
imaging guide wire: Feasibility study in rabbits. Radiology. 2000;217:501–506. 75. Neuerburg J, Bücker A, Adam G, et al. Kavafilterimplantation unter MRT-Kontrolle:
47. Buchli R, Boesiger P, Meier D. Heating effects of metallic implants by MRI examinations. Experimentelle in vitro- und in vivo-Untersuchungen. Fortschr Röntgenstr. 1997;167:
Magn Reson Med. 1988;7:255–261. 418–424.
48. Peden CJ, Collins AG, Butson PC, et al. Induction of microcurrents in critically ill patients 76. Bartels LW, Bos C, van der Weide R, et al. Placement of an inferior vena cava filter in a pig
in magnetic resonance systems. Crit Care Med. 1993;21:1923–1928. guided by high-resolution MR fluoroscopy at 1.5 T. J Magn Reson Imaging. 2000;12:599–605.
49. Wildermuth S, Erhart P, Leung DA, et al. [Active instrumental guidance in interventional 77. Frahm C, Gehl HB, Lorch H, et al. MR-guided placement of a temporary vena cava filter:
MR tomography: Introduction to a new concept]. Rofo. 1998;169:77–84. Technique and feasibility. J Magn Reson Imaging. 1998;8:105–109.
50. Liu CY, Farahani K, Lu DS, et al. Safety of MRI-guided endovascular guide wire applica- 78. Bücker A, Neuerburg JM, Adam G, et al. Real-time MR guidance for inferior vena cava
tions. J Magn Reson Imaging. 2000;12:75–78. filter placement. J Vasc Interv Radiol. 2001;12:753–756.
51. Martin AJ, Baek B, Acevedo-Bolton G, et al. MR imaging during endovascular procedures: 79. Kraemer NA, Immel E, Donker HC, et al. Evaluation of an active vena cava filter for MR
An evaluation of the potential for catheter heating. Magn Reson Med. 2009;61:45–53. imaging in a swine model. Radiology. 2011;258:446–454.
52. Ladd ME, Quick HH. Reduction of resonant RF heating in intravascular catheters using 80. Shih MC, Rogers WJ, Bonatti H, et al. Real-time MR-guided retrieval of inferior vena cava
coaxial chokes. Magn Reson Med. 2000;43:615–619. filters: An in vitro and animal model study. J Vasc Interv Radiol. 2011;22:843–850.
53. Wong ME, Zhang DQ, Duerk DJ, et al. An optical system for wireless detuning of parallel 81. Muller MF, Siewert B, Stokes KR, et al. MR angiographic guidance for transjugular intra-
resonant circuits. J Magn Reson Imaging. 2000;12:632–638. hepatic portosystemic shunt procedures. J Magn Reson Imaging. 1994;4:145–150.
54. Weiss S, Kuehne T, Brinkert F, et al. In-vivo safe catheter visualization and slice tracking 82. Kee ST, Rhee JS, Butts K, et al. MR-guided transjugular portosystemic shunt placement in
using an optically detunable resonant marker. Magn Reson Med. 2004;52:860–868. a swine model. J Vasc Interv Radiol. 1999;10:529–535.
55. Konings MK, Bartels LW, van Swol CF, et al. Development of an MR-safe tracking catheter 83. Bücker A, Neuerburg JM, Adam GB, et al. MR-guidance of TIPS procedures performed at
with a laser-driven tip coil. J Magn Reson Imaging. 2001;13:131–135. 1.5 T. Eur Radiol. 2000;10(suppl 1):171.
56. Bücker A, Adam G, Neuerburg J, et al. MR-guided PTA applying radial k-space filling 84. Serfaty JM, Yang X, Aksit P, et al. Toward MRI-guided coronary catheterization:
and active tip tracking: Simultaneous real-time visualization of the catheter tip and the Visualization of guiding catheters, guidewires, and anatomy in real time. J Magn Reson
anatomy. In: Proceedings of the International Society of Magnetic Resonance in Medicine. Imaging. 2000;12:590–594.
1999:575. 85. Lardo AC, McVeigh ER, Jumrussirikul P, et al. Visualization and temporal/spatial char-
57. Godart F, Beregi JP, Nicol L, et al. MR-guided balloon angioplasty of stenosed aorta: acterization of cardiac radiofrequency ablation lesions using magnetic resonance imaging.
In vivo evaluation using near-standard instruments and a passive tracking technique. J Magn Circulation. 2000;102:698–705.
Reson Imaging. 2000;12:639–644. 86. Bücker A, Neuerburg JM, Adam G, et al. MR-gesteurte Spiralembolisation von Nierenarterien
58. Manke C, Nitz WR, Djavidani B, et al. MR imaging-guided stent placement in iliac arterial in einem Tiermodell. Rofo. 2003;175(2):271–274.
stenoses: A feasibility study. Radiology. 2001;219:527–534. 87. Strother CM, Unal O, Frayne R, et al. Endovascular treatment of experimental canine
59. Paetzel C, Zorger N, Bachthaler M, et al. Feasibility of MR-guided angioplasty of femoral aneurysms: Feasibility with MR imaging guidance. Radiology. 2000;215:516–519.
artery stenoses using real-time imaging and intraarterial contrast-enhanced MR angiogra- 88. Spuentrup E, Ruebben A, Schaeffter T, et al. MR-guided coronary stent placement in a pig
phy. Rofo. 2004;176:1232–1236. model. Circulation. 2002;105:874–879.
60. Nitz WR, Oppelt A, Renz W, et al. On the heating of linear conductive structures as guide 89. Bücker A, Spuentrup E, Grabitz R, et al. Magnetic resonance-guided placement of atrial sep-
wires and catheters in interventional MRI. J Magn Reson Imaging. 2001;13:105–114. tal closure device in animal model of patent foramen ovale. Circulation. 2002;106:511–515.
61. Wildermuth S, Dumoulin CL, Pfammatter T, et al. MR-guided percutaneous angioplasty: 90. Rickers C, Jerosch-Herold M, Hu X, et al. Magnetic resonance image-guided transcatheter
Assessment of tracking safety, catheter handling and functionality. Cardiovasc Intervent closure of atrial septal defects. Circulation. 2003;107:132–138.
Radiol. 1998;21:404–410. 91. Razavi R, Hill DL, Keevil SF, et al. Cardiac catheterisation guided by MRI in children and
62. Smits HF, Bos C, van der Weide R, et al. Interventional MR: Vascular applications. Eur adults with congenital heart disease. Lancet. 2003;362:1877–1882.
Radiol. 1999;9:1488–1495. 92. Kuehne T, Yilmaz S, Meinus C, et al. Magnetic resonance imaging-guided transcatheter
63. Omary RA, Frayne R, Unal O, et al. MR-guided angioplasty of renal artery stenosis in a implantation of a prosthetic valve in aortic valve position: Feasibility study in swine. J Am
pig model: A feasibility study. J Vasc Interv Radiol. 2000;11:373–381. Coll Cardiol. 2004;44:2247–2249.
64. Meyer JM, Bücker A, Schuermann K, et al. MR evaluation of stent patency: In vitro tests 93. Lederman RJ, Guttman MA, Peters DC, et al. Catheter-based endomyocardial injection
of 22 metallic stents and the possibility of determining their patency by MR angiography. with real-time magnetic resonance imaging. Circulation. 2002;105:1282–1284.
Invest Radiol. 2000;35:739–746. 94. Saeed M, Lee R, Martin A, et al. Transendocardial delivery of extracellular myocardial
65. Manke C, Nitz WR, Lenhart M, et al. Magnetic resonance monitoring of stent deployment: markers by using combination X-ray/MR fluoroscopic guidance: Feasibility study in dogs.
In vitro evaluation of different stent designs and stent delivery systems. Invest Radiol. Radiology. 2004;231:689–696.
2000;35:343–351. 95. Krombach GA, Baireuther R, Higgins CB, et al. Distribution of intramyocardially injected
66. Bücker A, Neuerburg JM, Adam G, et al. [Stent placement with real time MRI guidance: extracellular MR contrast medium: Effects of concentration and volume. Eur Radiol. 2004;
Initial animal experiment experiences]. Rofo. 1998;169:655–657. 14:334–340.
67. Manke C, Nitz WR, Lenhart M, et al. [Stent angioplasty of pelvic artery stenosis with MRI 96. Kahlert P, Eggebrecht H, Plicht B, et al. Towards real-time cardiovascular magnetic reso-
control: Initial clinical results]. Rofo. 2000;172:92–97. nance-guided transarterial aortic valve implantation: In vitro evaluation and modification
68. Bücker A, Neuerburg JM, Adam GB, et al. Real-time MR fluoroscopy for MR-guided iliac of existing devices. J Cardiovasc Magn Reson. 2010;12:58–62.
artery stent placement. J Magn Reson Imaging. 2000;12:616–622. 97. Schalla S, Saeed M, Higgins CB, et al. Magnetic resonance–guided cardiac catheterization
69. Bücker A, Adam G, Neuerburg JM, et al. [Interventional magnetic resonance imaging–nonin a swine model of atrial septal defect. Circulation. 2003;108:1865–1870.
invasive imaging for interventions]. Rofo. 2000;172:105–114. 98. Tzifa A, Krombach GA, Krämer N, et al. Magnetic resonance-guided cardiac interventions
70. Quick HH, Ladd ME, Nanz D, et al. Vascular stents as RF antennas for intravascular MR using magnetic resonance-compatible devices: A preclinical study and first-in-man congenital
guidance and imaging. Magn Reson Med. 1999;42:738–745. interventions. Circ Cardiovasc Interv. 2010;3:585–592.
LWBK1209-ch37_p618-633.indd 633 16/05/13 9:17 PM

Chapter
38 Titus Kuehne
Endovascular Interventions—
■■ Techniques in the future the work of clinicians caring for such patients is
Interventional MRI Unit: General Considerations likely to focus more and more on maintaining low morbid-
Imaging Techniques ity and higher quality of life. This requires the development
of optimized diagnostic tools to better guide therapy; the
MRI Characteristics of Medical Implants
establishment of non-ionizing and less invasive interven-
MRI Tracking of Endovascular Catheters and tional methods; and the progressive replacement of surgical
Guidewires by transcatheter techniques. Moreover, advances in image
■■ Applications postprocessing and numerical modeling have opened new
MRI-guided Cardiac Catheterization perspectives using MRI for virtual surgery and catheteriza-
MRI-guided Catheter-based Intervention tion and for simulating and compare the effects of different
treatment strategies.
■■ Future Perspectives
The ability of MRI to acquire images with detailed infor-

mation about anatomy and function has set new diagnos-
Techniques
tic standards for the assessment of patients with congenital
Interventional MRI Unit:
heart disease. Fast imaging techniques have emerged that
General Considerations
have extended MRI from a purely diagnostic to a dynamic
modality that is suited to guide endovascular interventions Most MRI-guided cardiovascular interventions are con-
and to provide immediate information about the physiologic ducted in 1.5-T scanners with a short wide bore. Three-tesla
and anatomic responses to catheter-based treatment. Latest and low-field open systems are theoretically also suited, but
concepts even aim at predicting the hemodynamic outcome are limited by higher affinity to susceptibility or lower reso-
of intervention by using MRI data for numerical modeling. lution for real-time cardiac imaging, respectively. The short
The desire to replace x-ray fluoroscopy by alternative im- bore of modern scanners allows reasonable access to the
aging modalities for the guidance of catheter-based interven- groin or neck of the patient, although catheter manipulation
tion is motivated by the association of x-ray with significant can be difficult in small infants or children (Fig. 38.1).
exposure to ionizing radiation (1) and contrast media and In the clinical setting, x-ray fluoroscopy is used for almost
also its lack of soft tissue visualization and functional data all endovascular interventions, due to excellent spatial and
acquisition. As an alternative to x-ray transesophageal echo- temporal resolutions in combination with high contrast be-
cardiography (TEE) was successfully introduced for trans- tween vessels and background. Recent developments, such as
catheter closure of atrial septal defects (ASD) and evolved contrast-enhanced 3D rotational angiography, exploit these
in a short time into an important tool that can substitute features further and open up new possibilities. However, x-ray
fluoroscopy completely (2). The capability of MRI, however, is still limited in terms of soft tissue visualization and acquisi-
to acquire at any arbitrary orientation in three-dimensional tion of quantitative functional parameters. Combined x-ray/
(3D) space real-time images with high soft tissue contrast MRI units were proposed to take advantage of both imaging
makes this technique attractive for technically even more de- modalities. Fahrig et al. (3) reported simultaneous x-ray and
manding procedures. In addition, MRI allows accurate pre- MRI by integrating, similar to PET-MRI systems, an x-ray
interventional and immediate postinterventional assessment camera into an open 0.5-T scanner. A different strategy is to
of cardiovascular morphology and function, and thus may connect the MRI scanner over a mobile table with C-arms or
simplify catheter-based procedures. even fully equipped x-ray angiography laboratories (4,5). The
Mortality rates in patients with congenital heart disease use of such combined units makes sense until MRI catheter-
have decreased significantly over the past decades. Therefore, tracking techniques have reached a higher standard in terms
634
LWBK1209-ch38_p634-647.indd 634 16/05/13 9:16 PM

Chapter 38 ■ Endovascular Interventions—Congenital Heart Disease 635
B C
Figure 38.1. MRI catheterization of patients. Panel A: It shows angioplasty balloon inflated with saline
solution in a patient with Fontan circulation. Panels B,C: They show flow directional catheters in the pul-
monary artery and right ventricle. The catheter tipped balloon is inflated by CO2. The arrows point on flow-
directed catheters in the pulmonary artery and right ventricle (lower panels) and an angioplasty balloon cath-
eter in the vena cava (upper right panel).
of reliability and safety. Until then an x-ray safety back-up can be well adapted to intermittent noise by training of the staff
system should be available. involved in the procedure.
The overlay of MRI-derived anatomic roadmaps to x-ray Safety aspects for the use of an interventional MRI unit
fluoroscopy is used in an increasing number of applications. require that general contraindications for MRI of patients
Such an approach can be applied in combined x-ray/MRI are respected. In addition, the use of interventional instru-
units or, in a sequential approach by importing previously ments must be chosen with their interaction with the mag-
acquired MR images into a conventional catheterization netic field in mind (8–11). Metallic materials that yield
laboratory. Any of such multi-modality imaging requires significant magnetic torque, such as introducer needles or
methods for image registration. For having a kind of over- wires, must not be brought into close proximity to the mag-
view, registration can be done by anatomical landmarks. netic field. In addition, all elongated electrical conductors,
For procedures that require more precise information, more metallic or not, are prone to heating effects if they reach
sophisticated registration methods need to be applied. Such the critical length of 20 cm at 1.5 T (10,12,13). Therefore,
methods include among others the use of fiducial markers such interventional instruments must be excluded from en-
(Fig. 38.2) (6,7). dovascular manipulation. Further details on the MRI com-
MRI-compatible equipment for anesthesia and monitor- patibility of catheters, guidewires, and medical implants are
ing vital parameters are offered by different manufacturers. provided in other chapters.
In-room display monitors and operation consoles allow
image acquisition by an interventionalist and imaging expert
Imaging Techniques
standing alongside. An aspect that must be considered when
performing MRI-guided intervention is high noise levels during Real-time 3D imaging is a task that remains to be realized.
imaging. However, our experience shows that communication However, MRI already offers an exceptional variety of
LWBK1209-ch38_p634-647.indd 635 16/05/13 9:16 PM

A B C
Figure 38.2. Multi-modality imaging by x-ray fluoroscopy and MRI for closure of membranous ventricu-
lar septal defects. Image registration is realized by fiducial markers (panel A), the fused images are shown
(panel B). MRI provides detailed anatomical roadmaps (panel C). Arrow (panels A, B) position of the ven-
tricular septal defect. In addition, arrow show position of the aortic root (panel C) and guidewire (panel B).
two-dimensional (2D) real-time and ultra-fast acquisition iology studies. In a more experimental level, such methods
modalities for the assessment of anatomy and function. Up-to- were described for guiding myocardial biopsy and closure of
date high-gradient systems provide the short TR and TE needed VSD (17). These approaches will be described in more detail
for fast imaging with steady state free precession (SSFP). This in the chapters below.
acquisition technique in conjunction with spiral or radial filling
of the k-space is currently considered one of the most reliable
MRI Characteristics of Medical Implants
tools for high-quality and robust real-time imaging (14,15).
Nevertheless, during real-time MRI the interventionalist To date most medical implants used in cardiovascular medi-
has to find a compromise between temporal and spatial res- cine comprise at least in part metal alloys that can cause
olutions or signal-to-noise ratio (SNR). In-plane resolutions artifacts on MR images. Such devices include endovascular
of 1 mm can be obtained if lower acquisition rates of ap- stents or stent-prostheses, septal or duct occluders, embo-
proximately 5 frames per second are accepted. In contrast, lization coils, and prosthetic heart valves. These implants
at lower spatial resolutions, frame rates of up to 20 frames are generally not a contraindication for MRI studies: They
per second can be achieved. yield only minor magnetic torque and heating and therefore
It is important to note, that x-ray fluoroscopic frame rates can be considered MRI compatible in terms of patient safety
of less than 10 frames per second are often applied in order to (8,19,20). However, susceptibility artifacts of the metallic
minimize the radiation exposure of the patient. In addition, device can cause pitfalls or even make assessment of cardio-
MRI allows interleaved high-resolution imaging if needed. vascular morphology and function impossible (20).
The rationale for MRI would be the transition between dif- The extent of imaging artifacts depends mainly on the
ferent imaging sequences that provide either high-spatial/ paramagnetic properties of the device and the field strength
low-temporal or low-spatial/high-temporal resolution. of the magnet. Less important factors are material thickness,
Accurate assessment of cardiovascular morphology prior the orientation of metallic struts toward the B0 field direc-
to the intervention can facilitate the procedure and reduce tion, and the sequence parameters used during MRI.
its duration. The acquisition of high-quality whole-heart im- Metals with weak paramagnetic properties, such as niti-
aging provides detailed 3D information that can be used as nol, yield susceptibility values that are relatively close to those
an anatomical roadmap when overlaid to real-time images of human tissue and therefore cause only slight local imaging
or x-ray in combined units (7,16). Such multi-modality im- artifacts (21–24). By contrast, metals with strong paramag-
aging approaches are known for electrophysiology studies netic properties, such as stainless steel, produce severe image
but are also promising for closure of ventricular septal defects distortion (22–24). Furthermore, the extent of susceptibility
or performing myocardial biopsy (17). is directly related to the magnetic field strength. MR scanners
High-quality MRI of anatomy in combination with with high field strengths cause more susceptibility artifacts
MRI-derived blood flow provides valuable information that than scanners with low field strengths (8,25). Finally, parallel
permit simulating different interventional or surgical treat- alignment of a metallic strut to the B0 field direction causes
ment strategies prior to the actual procedures. Thus such ap- fewer artifacts than orthogonal alignment (25,26). However,
proaches enable to determine for the individual patient the the orientation of a metallic implant toward the field direc-
type of treatment that offers optimum outcome (Figs. 38.3 tion is obviously difficult to predict in the clinical setting.
and 38.4) (16,18). The MR artifact behavior of endovascular stents is further
Multi-modality imaging that fuses MRI or CT images complicated by radiofrequency shielding effects. Shielding
with x-ray fluoroscopy is routinously applied in electrophys- effects are caused by eddy currents on the surface of the stent
LWBK1209-ch38_p634-647.indd 636 16/05/13 9:16 PM

Figure 38.3. MRI 3D whole-heart -based cast models of the heart of a patient with complex double outlet
right ventricle (which means that the pulmonary artery and the aorta arise from the right ventricle (RV). The upper
panels shows the cast before (left) and after surgery (right panel).The asterix indicates the position of the ventricular
septal defect through which a surgical conduit is positioned (dotted line, right upper panel) in order to connect the
left ventricle (LV) with the aorta.The upper mid panel is a magnification with a view through the ventricular septal
defect into the RV. Computer-assisted methods allow for the simulation of the surgical procedures. In this patient
the size, form, and position of the conduit form is determined prior surgery. The results of the simulation were con-
firmed after surgery (upper right panel). MPA, main pulmonary artery with banding (arrow); RVOT, right ventricu-
lar outflow tract; LVOT, left ventricular outflow tract; LPA, left pulmonary artery; Ao, aorta.
that prevent radiofrequency pulses from fully penetrating the Endovascular Stents
wire mesh. This results in decreased signal intensities within
the lumen of the stent. The obscured lumen can be misinter- MRI has been proven to be an effective noninvasive tool to
preted as in-stent stenosis or obstruction (21,26–28). assess vascular stenosis (32,33). However, the evaluation of
The impact of pulse sequences for minimizing susceptibility in-stent stenosis, a frequently encountered problem, is difficult
or radiofrequency shielding was investigated in several studies using MRI due to susceptibility artifacts and radiofrequency
(23,24,27,29,30). The use of large flip-angle can reduce the shielding effects of the stent.
amount of signal loss by radiofrequency shielding (23,31). Numerous in-vitro and in-vivo studies have investigated the
Short TE and read recording direction parallel to B0 were re- artifact behavior of clinically available stents and aimed to grade
ported to reduce dephasing and thus susceptibility (23,27). A them in terms of their MRI suitability (22,25,28,29,31,34).
further, quite effective sequence implement is 180 degrees refo- Although several findings of these studies were not congruent,
cusing pulses, as used in spin-echo imaging. However, artifacts all studies demonstrated that stainless steel stents are generally
on spin-echo images must be carefully interpreted because re- not suited for MRI examination due to severe susceptibility ar-
maining radiofrequency shielding and residual susceptibility tifacts. In terms of susceptibility, platinum and tantalum stents
artifacts cannot be easily differentiated from the dark blood showed intermediate and nitinol stents good results. In addi-
pool, and thus in-stent stenosis could be falsely overlooked. tion, the degree of observed radiofrequency shielding varied
LWBK1209-ch38_p634-647.indd 637 16/05/13 9:16 PM

A B
Figure 38.4. MR angiogram of a patient with right pulmonary stenosis (panel A, arrow) with moderate
pressure gradient but substantial flow mismatch between the right and the left pulmonary artery. Current
guidelines show indication for stent placement. Anatomical information and quantitative flow are used for
numerical simulation of the hemodynamic effects of an intervention (panel B). Simulation revealed that stent
placement would not improve perfusion mismatch which was later confirmed.
greatly between different stent types because specific design not visible on MR images at all. Current studies investigate
properties can either favor or reduce eddy currents and thus such stents for use in the coronary arteries or as carrier for
RF shielding. In summary, large nitinol and platinum stents next generation valved stent devices.
of the aorta or pulmonary arteries can be assessed with MRI
(Figs. 38.5 and 38.6). In contrast, smaller stents or stents Embolization Coils and Duct Occluders
made from stainless steel are not assessable by MRI.
Most embolization coils are made of platinum or nitinol
In some circumstances MR-flow measurements have the
alloys that produce only slight local susceptibility artifacts
potential to overcome, at least in part, the problem of RF
(20). On the other hand, some occluders for the closure of
shielding for the assessment of stent lumen patency, because
patent ductus arteriosus contain stainless steel components
phase images do not depend on the amplitude of flip angle
that can severely distort MR images due to susceptibility
excitation (35–37). Therefore, stenosed stents will not con-
and make assessment of the adjacent pulmonary arteries and
tain any phase information in areas with tissue in-growth
thoracic aorta problematic.
or thrombus formation. However, phase images are very
sensitive to susceptibility effects and therefore this technique
Atrial and Ventricular Septal Occluders
requires stents with both low-paramagnetic properties and
relatively large diameter (35–37). In contrast to metallic Metallic components of atrial and ventricular septal occlud-
stent, bioabsorbable stents have the drawback that they are ers are generally manufactured from a nitinol mesh that
LWBK1209-ch38_p634-647.indd 638 16/05/13 9:16 PM

A B
C D E
Figure 38.5. Real-time MRI for placement of custom made nitinol valved stents in the pulmonary posi-
tion. Panels A, D: Material properties of the delivery system and of the stent allow for reasonable visualiza-
tion by MRI. Many commercially available systems for transcatheter aortic valve position are also made from
nitinol and thus suited for MRI-guided intervention. Arrows in panels A–D show the position of the delivery
system and the valved stent after implantation. Panel E: It shows the Medtronic Corevalve. The arrows in the
panle E show the eylets at the outflow tract of the stent. (Kahlert P, Eggebrecht H, Plicht B, et al. Towards real-
time cardiovascular magnetic resonance-guided transarterial aortic valve implantation: In vitro evaluation and
modification of existing devices. J Cardiovasc Magn Reson. 2010;12:58.)
produces only moderate susceptibility artifacts (Fig. 38.7) transcatheter techniques into the pulmonary or aortic valve po-
(38,39). Due to the relatively large size of the device, the sition. The Medtronic Melody, Medtronic CoreValve, Symetis
artifact can sometimes extend to the adjacent anatomy and Acurate, and Jena Valve, have carrier stents made from Nitinol
make analysis of atrial or ventricular volumes problematic. alloys or, for the Melody system, from platinum. Such devices
In contrast, the magnetic torque on such devices is minor are MRI compatible and they can be well visualized during and
and patients can be safely examined with MRI. after MRI-guided procedures (Figs. 38.5 and 38.6). In addition,
after implantation valve function, anatomic position and ad-
Mechanical Heart Valve Prostheses verse effects of valved stent can be determined by MRI (42–44).
Only few mechanical heart valves are known to have con-
traindication for MRI, including some early models of the
MRI Tracking of Endovascular Catheters
Starr-Edwards and Jomed Monodisc prosthesis (40,41).
and Guidewires
Other mechanical valves are safe with MRI, however, mor-
phologic and to some extent functional assessment is difficult There are a broad range of catheters that are manufactured
because of susceptibility artifacts (41). for specific cardiovascular applications. Some of these cath-
New generations of heart valves are incorporated into eters are braided with iron oxides or wire meshes in order to
carrier stents that can be implanted in selected patients by make them radio-opaque or to add flexibility. Such catheters
LWBK1209-ch38_p634-647.indd 639 16/05/13 9:16 PM

Figure 38.6. MRI-guided

transcatheter placement of a
valved stent in the aortic valve
position in swine. MRI enables
continuous visualization of
anatomy and interventional instru-
A B ments. Passive catheter tracking of
the delivery system is achieved by
the use of two small susceptibil-
ity markers that are positioned
at the distal and proximal end of
the loaded valved stent (arrows,
panels A,B). Important anatomical
landmarks can be identified during
MRI including the coronary arter-
ies or mitral valve leaflet (arrow-
heads, panels B–D). The position
of the implanted valved stent can
be clearly identified by slight local
susceptibility artifacts and radio-
frequency shielding. Panel D: It
shows the prosthetic valve leaflets
within the valved stent (arrow),
the unobstructed orifice, and first
segments of the left coronary
C D
artery (arrowheads).
Figure 38.7. MRI-guided closure of apical ventricular septal defect. A transthoracic access sheath is guided
under real time and by active tracking through the chest directly into the heart. Arrows show the position of
the occluder and of the delivery system.
LWBK1209-ch38_p634-647.indd 640 16/05/13 9:17 PM

A B C
Figure 38.8. Interactive real-time images of a patient with aortic coarctation. Panels A and B show MR
and x-ray fluoroscopy angiograms of the aorta before and after intervention. A balloon angioplasty catheter
is advanced from the ascending aorta through the coarctation. Angioplasty is performed with an iron oxide–
based contrast media (panel B). The contrast medium produces good contrast between the catheter shaft and
the balloon and the blood pool of the aorta. The balloon catheter was splinted by a long sheath for stabiliza-
tion. The vessel lumen was secured by a custom made polymer guidewire.
are generally not suited for use during MRI because they addition, carbon dioxide (CO2), iron oxide or Gadolinium-
can produce severe image distortion due to susceptibility. based contrast media have been used for the visualization of
On the other hand, catheters made from polyesters only do balloon tipped catheters (52,53,55–57) (Figs. 38.1B,C and
not generate any signal at all and are therefore invisible dur- 38.8). Filling of balloons with isotonic saline solutions also
ing MRI. However, appropriate catheter-tracking methods produces good signal contrast in large vessels (Fig. 38.1A).
are essential to perform successful MRI-guided endovascu- Whereas CO2 or iron oxide contrast media again produce
lar procedures. Ideally, tracking methods generate high con- susceptibility, a solution of 10% Gd DPTA offers maximum
trast between the catheter and the background anatomy and T1 enhancement on T1-weighted images (52,53,55,57,58).
readily allow automated slice and tip detection. In addition, Other recent studies used 19F or hyperpolarized 13C con-
the method should visualize both the tip of the catheter and trast media (59,60).
its shaft in order to be able to identify possible looping of
the catheter. The development of catheter-tracking methods
Hybrid Tracking
that fulfill all clinical needs is technically very challenging.
These difficulties are one of the major barriers that delayed Small wireless resonance circuits (RCs) were recently
the transition of preclinical interventional MRI into the clin- proposed as fiducial markers of endovascular catheters
ical settings. (61–64). This hybrid of active and passive tracking tech-
niques generates an intense local signal enhancement that
is readily perceptible to the observer, allows automated
Active Tracking
slice tracking and is safe for the patient because no elon-
Active tracking is commonly accomplished by incorporat- gated conductors are used (65). RCs are based on small
ing a small receiver coil or antennas into the tip of catheters copper coils that are tuned to the Larmor frequency. Single
or guidewires (45–47). The coil picks up signal during slice or multiple coils mounted on endovascular catheters can
excitation and generates frequency-encoded recall echoes. serve for catheter tip and shaft detection. However, RCs
These echoes can be detected in 3D space in real-time and are technically difficult to manufacture, relatively large
at a spatial resolution of approximately 1 mm (Fig. 38.7). and therefore increase substantially the profile of any cath-
The major drawback of this technique is the need to inte- eter system.
grate the electronic components into the catheters. This has
direct impact on the mechanical properties and the profile
Tracking of Guidewires
of the catheters. Progress was made in the development of
safe transmission lines that overcome potential heating of The availability of appropriate guidewires remains one of the
the catheter tip (48–50). However, all preclinical devices still main challenges for MRI-guided cardiovascular interventions.
require further safety evaluation for obtaining approval for Although good visualization of the shaft of metallic guidewires
clinical use. can be achieved by the concept of loopless antennas, visualiza-
tion of the tip of the antenna is ambiguous due to significant
signal decreases toward its ends (66,67). Even more important
Passive Tracking
are unsolved safety issues. All electrical conductors, metallic
Passive tracking techniques aim to directly visualize the or not, are prone to heating effects (10,68,69). Thus, x-ray
interventional instrument within the imaging plane and metallic guidewires are not suited for clinical use. To over-
therefore do not require any additional data postprocess- come this problem, guidewires with active bioelectrically safe
ing. Passive techniques are based either on local signal voids transmission lines were proposed. They are tested in preclini-
due to field inhomogeneities/susceptibility or signal enhance- cal studies and the results are very promising (50,70). As an
ment by the use of specific MRI-contrast media (51–54). A alternative, non-metallic guidewires with adequate mechanical
broad variety of passive markers have been reported for properties were used in preclinical studies (71,72). However,
the visualization of catheters that generally consist of metal their mechanical and thus stirability properties as well as their
alloys, dysprosium oxide or that are simply exploiting the robustness concerning stress that could potentially induce
material properties of x-ray opaque catheters (Fig. 38.6). In fractures require further evaluation.
LWBK1209-ch38_p634-647.indd 641 16/05/13 9:17 PM

myocardial contractility, ventricular–arterial coupling or

Applications diastolic performance in-vivo (Fig. 38.9) (58,86–88). In
the clinical setting there are several important advantages
MRI-Guided Cardiac Catheterization
of MRI over other available methods for the assessment of
MRI-guided catheterization for the assessment of pulmo- ventricular pressure–volume relations or the measurement
nary vascular function and myocardial performance has of pulmonary vascular resistance. MRI-guided catheteriza-
found its way into the clinical setting at fast pace since it was tion is technically easy to perform, measurements are very
first reported (73,74). This method will become part of the reproducible and data acquisition can be accomplished
guidelines of the AHA/ATS for the assessment of children in short imaging times (57,73,74). In contrast, combined
with pulmonary hypertension (published in 2012). pressure-conductance catheter techniques, yet still the gold
One important characteristic of MRI is its ability to standard for measuring ventricular pressure–volume relations,
assess both, cardiovascular anatomy and function. To date are more complicated and often require multiple catheter ma-
MRI is widely considered the gold standard for quantifica- nipulation (89).
tion of ventricular volumes and blood flow. Measurements Pulmonary vascular resistance: The combination of pul-
have been demonstrated to be accurate and reproducible monary flow volumes with invasive pressures permits to
and to have relatively low interobserver and interstudy vari- calculate pulmonary vascular resistance. This is a clinically
ability (75–78). In addition, the advent of fast imaging tech- valuable and often used parameter of vascular function that
niques enables a full set of data to be acquired in short imag- is conventionally obtained by thermodilution techniques
ing times. Fast-imaging methods allow ventricular volumes or Oxymetry (Fick’s principle). Both latter techniques have
to be determined at good quality within only few breath- several drawbacks; thermodilution is known to have limited
hold periods. Quantitative blood flow can be measured reproducibility in the presence of RV dysfunction; Oxymetry
even in real-time or with four-dimensional (4D) methods is time consuming and quite invasive as repeated catheter
that measure phasic flow velocities in a 3D volume (79–81). manipulations are required. In addition, the presence of
MRI-derived flow volume data provide quantitative infor- shunt and collateral flow can hamper its accuracy. As an al-
mation about pulmonary perfusion, left–right perfusion ternative, MRI-guided catheterization was introduced and
mismatch, intracardiac shunt volumes or collateral flow validated for simultaneous assessment of MRI flow and in-
through aortopulmonary or venovenous collaterals (82–85). vasive pressures (73,74).
Whereas conventional x-ray angiography is limited in the For MRI catheterization, typically liquid-filled catheters
quantitation of ventricular volumes and vascular blood flow, are used for measuring invasive pressures. The balloon at
MRI is restricted in its ability to determine cardiovascular the tip directional catheters can be visualized during MRI by
pressures. However, increased understanding and experience susceptibility due to CO2 inflation (Fig. 38.1). The contrast
in the field of intervention assisting imaging and catheter generated between the inflated balloon and the bright blood
tracking has enabled successful MRI-guided cardiac cath- pool of the heart chambers and the main segments of the
eterization studies in patients with congenital heart disease pulmonary arterial system was sufficient to safely control
(73,74,86,87). catheter advancement.
Ventricular pressure–volume relations: The combined as- Normally, measurement of pressures is realized by con-
sessment of cardiovascular pressures and quantitative ven- necting the liquid-filled catheters to conventional pressure
tricular volumes or blood flow permits to determine several transducers (56,58). However, in the MRI environment the
important parameters. From these data the pressure–volume transducer has to be effectively shielded to avoid electro-
relation can be constructed for measuring load-independent magnetic interaction with the RF pulses. As an alternative to
Right ventricle Left ventricle

120 Rest
40 Rest 110 Dobutamine
Dobutamine
100
90
Pressure (mm Hg)
Pressure (mm Hg)
30
80
70
20 60
50
40
10 30
20
10
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Volume (mL) Volume (mL)
Figure 38.9. Pressure–volume loops acquired by MRI catheterization. Measurements allow deriving loops
simultaneously in the right and the left ventricle. Integrated parameters of global pump function, myocontrac-
tility, and diastolic performance can be determined form the loops.
LWBK1209-ch38_p634-647.indd 642 16/05/13 9:17 PM

liquid-filled catheters, MRI-compatible pressure-tipped mi- Noninvasive sizing of ASD would be desirable. However,
cromanometer catheters are available but are not suited for similar to the findings of echocardiography studies, sizing
human application because of safety issues due to potential with MRI is described as difficult (38,93,95). Nevertheless,
heating effects. MRI-guided transcatheter closure of ASD is of potential
Current directions go to combine MRI flow with sequen- clinical interest because there would be no need to sedate the
tially obtained pressures from right heart catheterization patient as for closure under TEE guidance.
(90). The time window should be small and physiologic con- Another application of clinical interest would be the
dition like volume loading similar between the two study one-stop assessment and treatment of elderly patients with
parts. There are also promising work estimating vascular ASD. These patients are at risk of developing pulmonary
pressures and pulmonary resistance non-invasively from edema after ASD closure due to LV diastolic dysfunction (2).
MRI flow–derived pressure maps. Such future approach Conditioning these patients with diuretics and positive ino-
might be accomplished fully noninvasive (91,92). tropic medication can help to prevent the onset of diastolic
and systolic LV dysfunction. At our institution we assess
biventricular function with MRI during transient balloon
MRI-Guided Catheter-Based Intervention
occlusion of the ASD in order to monitor the effectiveness of
The desire to replace x-ray fluoroscopy and angiography by medical conditioning prior to final ASD closure.
alternative imaging methods for the guidance of endovascu- Because of the high level of experiences that accumulated
lar intervention was fueled by the disadvantages of repeated during the past decade concerning the manufacturing of ASD
exposure to often significant doses of ionizing radiation in delivery catheters and the conduction of the procedure, the
infants and children (1), possible renal dysfunction due to barrier for a transition from standard TEE to MRI-guided
considerable quantities of contrast media, and the lack of soft methods is quite high. However, remarkable progress of
tissue visualization and incomplete functional data. Although MRI-guided closure of VSD was made in the past years.
MRI has considerable potential for monitoring endovascu- Transcatheter closure of membranous subaortic VSDs is a
lar interventions, most MRI-guided procedures described in complex procedure that is often complicated due to poor
the literature are still performed in animal models. This is visualization of the defect. In addition, VSDs with a mus-
mainly because of unsolved safety or compatibility issues in cular apical location often cannot be closed by transcatheter
connection with the currently available implants, catheters or techniques. However, also surgical closure is challenging and
delivery systems. A further limiting factor is that more expe- show in a considerable number of cases residual defects (96).
rience needs to be accumulated in order to make the move The overlay of previously acquired MR images with x-ray
to an entirely new environment of interventional MRI units, fluoroscopy is a promising multi-modality approach that can
with their limited patient access, high noise levels, and tomo- facilitate closure of subaortic VSD (Fig. 38.2). MRI provides
graphic images. Finally, the costs of MRI-guided procedures detailed anatomic information of the VSD that is helpful for
have to be weighed against the potential advantages and ben- the guidance of catheters when images are fused to x-ray
efits. Therefore, in future it seems to be important to define fluoroscopy. The creation of such anatomical roadmaps by
specific applications that emphasize the strengths of MRI fiducial markers requires a technical infrastructure and op-
rather than attempting to transfer one-to-one interventional erator experiences (97). Both can be found in an increasing
procedures from x-ray to MRI. MRI-guided intervention number of medical centers.
would add considerable benefit for procedures that rely on Lederman et al. showed in a pig study of closure of mem-
good visualization of soft-tissue anatomy or immediate pre- branous VSD, that x-ray combined with MRI significantly
and postinterventional assessment of physiologic parameters reduces fluoroscopy time (Fig. 38.2) (97). For transcatheter
of cardiovascular function. closure of apical VSD, Lederman et al. also reported a novel
strategy using real-time MRI for guiding a transthoracic ac-
cess sheath through the chest directly into the heart (98).
Transcatheter Closure of Atrial and Ventricular Septal
The sheath was visualized by active tracking method, the
Defects and Patent Ductus Arteriosus
closure was done by a commercially available closure de-
Many large centers have partially or completely abandoned vice (Fig. 38.7).
x-ray fluoroscopy for closure of ASD. Transcatheter clo-
sure of ASD has became a routine procedure and is carried
Intervention in Patent Ductus Arteriosus,
out mainly under guidance with TEE (93). Several animal
Collaterals and Shunts
studies and a human landmark report have been published
that describe successful MRI-guided transcatheter closure To our knowledge no reports of successful MRI-guided
of ASD (5,38,39,94). Septal closure devices that are man- transcatheter closure of patent ductus arteriosus collater-
ufactured from nitinol mesh are not prone to significant als or shunts have yet been published. The closure of pat-
magnetic torque or susceptibility (Fig. 38.7). Therefore, the ent ductus arteriosus or collaters is typically done by coils
devices can be considered safe and suited for the use within or plug devices that are manufactured from platinum or
the MRI environments. In contrast, commercially available nitinol. These devices produce only slight susceptibility arti-
delivery systems for ASD closure devices are generally not facts. MRI-guided coil embolization was described for the
suited for MRI, because their release system either contains renal arteries in pigs (99). However, for application in con-
ferromagnetic materials that cause significant susceptibility genital heart disease the determination of the final device
artifacts and/or they comprise elongated conductors that are position and its 3D shape is important but cannot be defined
prone to heating effects. after delivery with MRI because of too small device artifacts
LWBK1209-ch38_p634-647.indd 643 16/05/13 9:17 PM

and limited spatial resolution of MRI compared to x-ray metal alloys with low paramagnetic properties and only
fluoroscopy (29,100). minor RF shielding.
Though, MRI appears to be not suited for directly guiding While coronary artery MRI-guided stenting is complicated
intervention in “small” tortuous vessels, it is of increasing im- by the small, moving and tortuous vessels, larger vasculature
portance for the planning and assistance of procedures. MRI seem to be more attractive. Several studies, mostly in animal
flow studies permit to quantify blood flow through aortopul- models, which reported that MRI-guided endovascular stent-
monary shunts or collaterals (83–85,101). These information ing is feasible. In these studies the stents of different manu-
can be put into biofluid dynamic models for computing the facturers were positioned in the aorta, including stenting of
hemodynamic effects that different catheter-based or surgical aortic coarctation and placement of tube endographs in aor-
procedures would have on blood flow and oxygen saturation tic aneurysm and the pulmonary artery (116–118). Recent
(102,103). studies also demonstrated that vessel injury in the form of
dissection or perforation can be detected by MRI (117,119).
Other studies describe MRI-guided placement of stents in pe-
Balloon Angioplasty for Treatment of
ripheral vasculature, including the carotid, renal, and iliac
Aortic Coarctation and Pulmonary Stenosis
artery (56,63,120,121).
MRI-guided percutaneous balloon angioplasty is a promis-
ing technique for the treatment of a variety of congenital
Valved Stent Placement
heart diseases. MR-guided angioplasty of the pulmonary,
iliac, and renal arteries and the aorta have been conducted Percutaneous transcatheter placement of valved stents in the
in animals (104–107). A first human study of angioplasty pulmonary, aortic, and atrioventricular valve position has
of the iliac artery has been reported but safety issues were revolutionized valve replacement in the past decade (122–
raised because metallic guidewires caused hazards to the 124). These innovative devices are based on the visionary
patients due to potential heating (108). work of Dr H. Andersen from Denmark in the late 1980s
MRI can provide detailed information on 3D anatomy but were introduced in first human studies some 20 years
and blood flow profiles and insight into the morphology of later. Recently, follow-up data of pulmonary valved stents
the vascular wall. This information is of benefit in planning became available and show overall promising performance
the intervention, monitoring its success, and keeping cath- at midterm (125). Transcatheter placement of valved stents
eter manipulation limited. For angioplasty, the balloon has is a technically demanding procedure and requires careful
to be inflated with contrast media that are suited for MRI. planning. MRI assessment of function and anatomy plays an
The use of isotonic saline solution in T2/T2*-weighted SSFP important role for timing of the procedure, to determine the
images or of saline with approximately 10% Gadolinium optimum size of the device and to assess whether transcath-
in T1-weighted pulse sequences produces bright signal (Fig. eter techniques can be applied at all.
38.1A) (87,109). As an alternative iron oxide–based MR Transcatheter placement of valved stents appears very at-
contrast media can be used to produce dark susceptibility tractive being conducted under MRI guidance. The relatively
effect on SSFP images (Fig. 38.8). large delivery systems can be well visualized by passive track-
A large problem associated with MRI-guided balloon an- ing methods. MRI offers the advantage of x-ray of soft tissue
gioplasty is the fact that no metallic guidewires can be used visualization. The visualization of pulmonary or aortic valve
due to potential heating in the MRI environment. The use of leaflets, mitral valve and coronary orifices is crucial for the suc-
guidewires is, however, mandatory for “securing” the lumen cess of the procedure. Successful preclinical MRI-guided im-
of the vessel in the event of vessel dissection and to stabilize plantation of a valved stent in aortic position of pigs was first
the balloon catheter during inflation. In a pilot study that we described by Kuehne et al. (42) using a transarterial approach
conducted in patients with aortic coarctation we used only a and an entirely custom made valved stent system (Fig. 38.6).
thin custom-made guidewire manufactured from polyesters In further pig studies, aortic stent valves were implanted under
that secured the vessel lumen (110). Stabilization of the bal- MRI guidance by a surgical transapical approach with balloon
loon catheter was achieved by a long sheath that splinted the expandable stent-valve and by transarterial access using a self-
shaft of the catheter (Fig. 38.8). Of course, active MRI guide- expanding Core Valve system. The materials used in commer-
wires with safe transmission lines and approved mechanical– cially available valved stent systems are MRI compatible and
elastic properties would be the first choice because such a allow for assessment of morphology and function after implan-
set-up can be applied only in selected procedure. tation. These issues are described in the above chapter.
In the future, MRI will certainly also play an important
role for producing more personalized valved stent devices.
Endovascular Stent Placement
Particularly for the pulmonary position, devices that do not
Modern stents are characterized by high radial strength come out of the shelf but that adapt to the anatomical con-
and surface properties that minimize tissue in-growth and dition of individual patient are promising to provide better
thrombus formation (111). In addition, low profiles and long-term performance and to reduce the risk of device failure
minimal shortening enables the stents to be deployed from by stent fractures (126,127).
small delivery systems with high precision (111–113). There
are several studies that aimed to add MRI-compatible prop-
MRI-guided Electrophysiology and Catheter Ablation
erties to stents in order to enable MR assessment of in-stent
stenosis (114,115). Such stents must either be nonmetallic, The desire to use MRI for electrophysiology (EP) studies
like the generation of bioabsorbable stents, or made from goes back to the 1990s (128). Today, MRI is increasingly
LWBK1209-ch38_p634-647.indd 644 16/05/13 9:17 PM

used for providing anatomical roadmaps that are imported 9. Schenck JF. The role of magnetic susceptibility in magnetic resonance imaging: MRI mag-
netic compatibility of the first and second kinds. Med Phys. 1996;23:815–850.
into EP mapping systems (129). In addition, the develop- 10. Yeung CJ, Susil RC, Atalar E. RF safety of wires in interventional MRI: Using a safety
ment of MRI compatible instruments enabled to acquire index. Magn Reson Med. 2002;47:187–193.
11. Ordidge RJ, Shellock FG, Kanal E. A y2000 update of current safety issues related to
intracardiac electrograms and to demonstrate the feasibility MRI. J Magn Reson Imaging. 2000;12:1.
of radiofrequency ablation in animal studies and a first pilot 12. Nitz WR, Oppelt A, Renz W, et al. On the heating of linear conductive structures as guide
study with patients (130–133). This ability of MRI for direct wires and catheters in interventional MRI. J Magn Reson Imaging. 2001;13:105–114.
13. Shellock FG. Radiofrequency energy-induced heating during MR procedures: A review.
visualization of ablation lesions opens the perspectives to J Magn Reson Imaging. 2000;12:30–36.
improve outcome by relating the lesions with therapy results 14. Martin AJ, Weber OM, Saeed M, et al. Steady-state imaging for visualization of endovas-
(134,135). cular interventions. Magn Reson Med. 2003;50:434–438.
15. Spuentrup E, Schroeder J, Mahnken AH, et al. Quantitative assessment of left ventricular
function with interactive real-time spiral and radial MR imaging. Radiology. 2003;227:
870–876.
16. Riesenkampff E, Rietdorf U, Wolf I, et al. The practical clinical value of three-dimensional
Future Perspectives models of complex congenitally malformed hearts. J Thorac Cardiovasc Surg. 2009;
138:571–580.
17. Gutierrez LF, Silva R, Ozturk C, et al. Technology preview: X-ray fused with mag-
MRI provides like no other imaging modality a large range netic resonance during invasive cardiovascular procedures. Catheter Cardiovasc Interv.
of high-quality information about anatomy and function. 2007;70:773–782.
18. Rietdorf U, Riesenkampff E, Schwarz T, et al. Planning of vessel grafts for reconstructive
These data can be used as input variables for numerical surgery in congenital heart disease. In: SPIE Medical Imaging 2010: Visualization, Image-
models and computational biofluid dynamics that provide Guided Procedures, and Modeling. 2010:7625–7105.
an interesting alternative to experimental investigations. 19. Shellock FG, Shellock VJ. Metallic stents: Evaluation of MR imaging safety. AJR Am J
Roentgenol. 1999;173:543–547.
These methods allow simulation of surgical techniques and 20. Teitelbaum GP, Bradley WG Jr, Klein BD. MR imaging artifacts, ferromagnetism, and
to approximate the hemodynamic effects of surgical or inter- magnetic torque of intravascular filters, stents, and coils. Radiology. 1988;166:657–664.
ventional procedures, respectively. Such models are based on 21. Wang Y, Truong TN, Yen C, et al. Quantitative evaluation of susceptibility and shield-
ing effects of nitinol, platinum, cobalt-alloy, and stainless steel stents. Magn Reson Med.
the patient’s individual characteristics and thus pave the way 2003;49:972–976.
for more personalized treatment strategies in cardiovascular 22. Meyer JM, Buecker A, Schuermann K, et al. MR evaluation of stent patency: In vitro test
of 22 metallic stents and the possibility of determining their patency by MR angiography.
disease (Figs. 38.3 and 38.4). Invest Radiol. 2000;35:739–746.
The focus of early studies was on predicting blood flow 23. Meyer JM, Buecker A, Spuentrup E, et al. Improved in-stent magnetic resonance angiog-
in different types of the Fontan operation (136,137). Other raphy with high flip angle excitation. Invest Radiol. 2001;36:677–681.
24. Lenhart M, Volk M, Manke C, et al. Stent appearance at contrast-enhanced MR angiog-
investigations modeled blood flow in aortic coarctation, raphy: In vitro examination with 14 stents. Radiology. 2000;217:173–178.
collaterals or complex procedures like the staged Norwood 25. Klemm T, Duda S, Machann J, et al. MR imaging in the presence of vascular stents: A
(102,103,138). These studies accomplished important ex- systematic assessment of artifacts for various stent orientations, sequence types, and field
strengths. J Magn Reson Imaging. 2000;12:606–615.
perimental groundwork in the past years. Current work 26. Bartels LW, Smits HF, Bakker CJ, et al. MR imaging of vascular stents: Effects of suscep-
focus on closing the gap between the experimental stage of tibility, flow, and radiofrequency eddy currents. J Vasc Interv Radiol. 2001;12:365–371.
27. Bartels LW, Bakker CJ, Viergever MA. Improved lumen visualization in metallic vascular
research and its clinical introduction. For CHD, potential implants by reducing RF artifacts. Magn Reson Med. 2002;47:171–180.
clinical applications are manifold and it can be expected that 28. Schurmann K, Vorwerk D, Bucker A, et al. Magnetic resonance angiography of nonfer-
the indication and planning of intervention will be substan- romagnetic iliac artery stents and stent-grafts: A comparative study in sheep. Cardiovasc
Intervent Radiol. 1999;22:394–402.
tially improved by these emerging techniques. 29. Graf H, Klemm T, Lauer UA, et al. [Systematics of imaging artifacts in MRT caused by
Methods of MRI-guided cardiac catheterization and en- metallic vascular implants (stents)]. Rofo. 2003;175:1711–1719.
dovascular interventions are still in the process of transition 30. Kuehne T, Saeed M, Moore P, et al. Influence of blood-pool contrast media on MR imaging and
flow measurements in the presence of pulmonary arterial stents in swine. Radiology. 2002;
from a preclinical stage of research to the clinical setting. First 223:439–445.
application, like MRI-guided catheterization for the assess- 31. Nordmeyer J, Gaudin R, Tann OR, et al. MRI may be sufficient for noninvasive assess-
ment of great vessel stents: An in vitro comparison of MRI, CT, and conventional angiog-
ment of pulmonary hypertension has become part of clinical raphy. AJR Am J Roentgenol. 2011;195:865–871.
routine in specialized centers. Other applications are expected 32. Loewe C, Schillinger M, Haumer M, et al. MRA versus DSA in the assessment of occlu-
to be introduced soon after safety-approved active MRI sive disease in the aortic arch vessels: Accuracy in detecting the severity, number, and
length of stenoses. J Endovasc Ther. 2004;11:152–160.
guidewires and catheters will become clinically available. 33. Sundgren PC, Sunden P, Lindgren A, et al. Carotid artery stenosis: Contrast-enhanced MR
angiography with two different scan times compared with digital subtraction angiography.
Neuroradiology. 2002;44:592–599.
34. Maintz D, Kugel H, Schellhammer F, et al. In vitro evaluation of intravascular stent arti-
References facts in three-dimensional MR angiography. Invest Radiol. 2001;36:218–224.
35. Lethimonnier F, Bouligand B, Thouveny F, et al. Error assessment due to coronary stents in flow-
1. Modan B, Keinan L, Blumstein T, et al. Cancer following cardiac catheterization in child- encoded phase contrast MR angiography: A phantom study. J Magn Reson Imaging. 1999;
hood. Int J Epidemiol. 2000;29:424–428. 10:899–902.
2. Ewert P, Berger F, Nagdyman N, et al. Masked left ventricular restriction in elderly patients 36. van Holten J, Kunz P, Mulder PG, et al. MR-velocity mapping in vascular stents to assess
with atrial septal defects: A contraindication for closure? Catheter Cardiovasc Interv. 2001; peak systolic velocity. In vitro comparison of various stent designs made of stainless steel
52:177–180. and nitinol. Magma. 2002;15:52–57.
3. Fahrig R, Butts K, Wen Z, et al. Truly hybrid interventional MR/X-ray system: Investigation 37. Kuehne T, Saeed M, Reddy G, et al. Sequential magnetic resonance monitoring of pulmo-
of in vivo applications. Acad Radiol. 2001;8:1200–1207. nary flow with endovascular stents placed across the pulmonary valve in growing swine.
4. Saeed M, Lee R, Martin A, et al. Transendocardial delivery of extracellular myocardial Circulation. 2001;104:2363–2368.
markers by using combination X-ray/MR fluoroscopic guidance: Feasibility study in dogs. 38. Rickers C, Jerosch-Herold M, Hu X, et al. Magnetic resonance image-guided transcath-
Radiology. 2004;231:689–696. eter closure of atrial septal defects. Circulation. 2003;107:132–138.
5. Razavi R, Hill DL, Keevil SF, et al. Cardiac catheterisation guided by MRI in children and 39. Schalla S, Saeed M, Higgins CB, et al. Magnetic resonance–guided cardiac catheterization
adults with congenital heart disease. Lancet. 2003;362:1877–1882. in a swine model of atrial septal defect. Circulation. 2003;108:1865–1870.
6. George AK, Sonmez M, Lederman RJ, et al. Robust automatic rigid registration of MRI 40. Hassler M, Le Bas JF, Wolf JE, et al. [Effects of the magnetic field in magnetic resonance
and X-ray using external fiducial markers for XFM-guided interventional procedures. imaging on 15 tested cardiac valve prostheses]. J Radiol. 1986;67:661–666.
Med Phys. 2011;38:125–141. 41. Edwards MB, Taylor KM, Shellock FG. Prosthetic heart valves: Evaluation of magnetic field
7. Gao G, Penney G, Ma Y, et al. Registration of 3D trans-esophageal echocardiography to interactions, heating, and artifacts at 1.5 T. J Magn Reson Imaging. 2000;12:363–369.
X-ray fluoroscopy using image-based probe tracking. Med Image Anal. 2012;16:38–49. 42. Kuehne T, Yilmaz S, Meinus C, et al. Magnetic resonance imaging-guided transcatheter
8. Shellock FG. Biomedical implants and devices: Assessment of magnetic field interactions implantation of a prosthetic valve in aortic valve position: Feasibility study in swine. J Am
with a 3.0-tesla MR system. J Magn Reson Imaging. 2002;16:721–732. Coll Cardiol. 2004;44:2247–2249.
LWBK1209-ch38_p634-647.indd 645 16/05/13 9:17 PM

43. Kahlert P, Eggebrecht H, Plicht B, et al. Towards real-time cardiovascular magnetic 77. Beerbaum P, Barth P, Kropf S, et al. Cardiac function by MRI in congenital heart dis-
resonance-guided transarterial aortic valve implantation: In vitro evaluation and modifi- ease: Impact of consensus training on interinstitutional variance. J Magn Reson Imaging.
cation of existing devices. J Cardiovasc Magn Reson. 2010;12:58. 2009;30:956–966.
44. Kahlert P, Parohl N, Albert J, et al. Real-time magnetic resonance imaging-guided transar- 78. Grothues F, Moon JC, Bellenger NG, et al. Interstudy reproducibility of right ventricu-
terial aortic valve implantation: In vivo evaluation in swine. J Am Coll Cardiol. 2012;59: lar volumes, function, and mass with cardiovascular magnetic resonance. Am Heart J.
192–193. 2004;147:218–223.
45. Bock M, Volz S, Zuhlsdorff S, et al. MR-guided intravascular procedures: Real-time 79. Markl M, Draney MT, Hope MD, et al. Time-resolved 3-dimensional velocity mapping in
parameter control and automated slice positioning with active tracking coils. J Magn the thoracic aorta: Visualization of 3-directional blood flow patterns in healthy volunteers
Reson Imaging. 2004;19:580–589. and patients. J Comput Assist Tomogr. 2004;28:459–468.
46. Hillenbrand CM, Elgort DR, Wong EY, et al. Active device tracking and high-resolution 80. Markl M, Harloff A, Bley TA, et al. Time-resolved 3D MR velocity mapping at 3T:
intravascular MRI using a novel catheter-based, opposed-solenoid phased array coil. Improved navigator-gated assessment of vascular anatomy and blood flow. J Magn Reson
Magn Reson Med. 2004;51:668–675. Imaging. 2007;25:824–831.
47. Zhang Q, Wendt M, Aschoff AJ, et al. Active MR guidance of interventional devices with 81. Stalder AF, Russe MF, Frydrychowicz A, et al. Quantitative 2D and 3D phase contrast
target-navigation. Magn Reson Med. 2000;44:56–65. MRI: Optimized analysis of blood flow and vessel wall parameters. Magn Reson Med.
48. Weiss S, Wirtz D, David B, et al. In vivo evaluation and proof of radiofrequency safety of 2008;60:1218–1231.
a novel diagnostic MR-electrophysiology catheter. Magn Reson Med. 2011;65:770–777. 82. Beerbaum P, Korperich H, Barth P, et al. Noninvasive quantification of left-to-right shunt
49. Kocaturk O, Kim AH, Saikus CE, et al. Active two-channel 0.035’’ guidewire for interven- in pediatric patients: Phase-contrast cine magnetic resonance imaging compared with
tional cardiovascular MRI. J Magn Reson Imaging. 2009;30:461–465. invasive oximetry. Circulation. 2001;103:2476–2482.
50. Fandrey S, Weiss S, Muller J. A novel active MR probe using a miniaturized optical link 83. Beerbaum P, Korperich H, Gieseke J, et al. Rapid left-to-right shunt quantification in chil-
for a 1.5-T MRI scanner. Magn Reson Med. 2012;67:148–155. dren by phase-contrast magnetic resonance imaging combined with sensitivity encoding
51. Bakker CJ, Hoogeveen RM, Weber J, et al. Visualization of dedicated catheters using fast (sense). Circulation. 2003;108:1355–1361.
scanning techniques with potential for MR-guided vascular interventions. Magn Reson 84. Grosse-Wortmann L, Al-Otay A, Yoo SJ. Aortopulmonary collaterals after bidirec-
Med. 1996;36:816–820. tional cavopulmonary connection or Fontan completion: Quantification with MRI. Circ
52. Wacker FK, Reither K, Ebert W, et al. MR image-guided endovascular procedures with Cardiovasc Imaging. 2009;2:219–225.
the ultrasmall superparamagnetic iron oxide SH U 555 C as an intravascular contrast 85. Valverde I, Rachel C, Kuehne T, et al. Comprehensive four-dimensional phase-contrast
agent: Study in pigs. Radiology. 2003;226:459–464. flow assessment in hemi-fontan circulation: Systemic-to-pulmonary collateral flow quan-
53. Wacker FK, Maes RM, Jesberger JA, et al. MR imaging-guided vascular procedures using tification. Cardiol Young. 2011;21:116–119.
CO2 as a contrast agent. AJR Am J Roentgenol. 2003;181:485–489. 86. Schmitt B, Steendijk P, Lunze K, et al. Integrated assessment of diastolic and systolic ventric-
54. Glowinski A, Adam G, Bucker A, et al. Catheter visualization using locally induced, ular function using diagnostic cardiac magnetic resonance catheterization: Validation in pigs
actively controlled field inhomogeneities. Magn Reson Med. 1997;38:253–258. and application in a clinical pilot study. JACC Cardiovasc Imaging. 2009;2:1271–1281.
55. Miquel ME, Hegde S, Muthurangu V, et al. Visualization and tracking of an inflatable bal- 87. Schmitt B, Steendijk P, Ovroutski S, et al. Pulmonary vascular resistance, collateral flow,
loon catheter using SSFP in a flow phantom and in the heart and great vessels of patients. and ventricular function in patients with a Fontan circulation at rest and during dobuta-
Magn Reson Med. 2004;51:988–995. mine stress. Circ Cardiovasc Imaging. 2010;3:623–631.
56. Kuehne T, Saeed M, Higgins CB, et al. Endovascular stents in pulmonary valve and artery 88. Brimioulle S, Wauthy P, Ewalenko P, et al. Single-beat estimation of right ventricular
in swine: Feasibility study of MR imaging-guided deployment and postinterventional end-systolic pressure-volume relationship. Am J Physiol Heart Circ Physiol. 2003;284:
assessment. Radiology. 2003;226:475–481. H1625–H1630.
57. Omary RA, Unal O, Koscielski DS, et al. Real-time MR imaging-guided passive catheter 89. Bishop A, White P, Oldershaw P, et al. Clinical application of the conductance catheter
tracking with use of gadolinium-filled catheters. J Vasc Interv Radiol. 2000;11:1079–1085. technique in the adult human right ventricle. Int J Cardiol. 1997;58:211–221.
58. Kuehne T, Yilmaz S, Steendijk P, et al. Magnetic resonance imaging analysis of right ven- 90. Sanz J, Kariisa M, Dellegrottaglie S, et al. Evaluation of pulmonary artery stiffness in
tricular pressure-volume loops: In vivo validation and clinical application in patients with pulmonary hypertension with cardiac magnetic resonance. JACC Cardiovasc Imaging.
pulmonary hypertension. Circulation. 2004;110:2010–2016. 2009;2:286–295.
59. Kozerke S, Hegde S, Schaeffter T, et al. Catheter tracking and visualization using 19f 91. Frydrychowicz A, Stalder AF, Russe MF, et al. Three-dimensional analysis of segmen-
nuclear magnetic resonance. Magn Reson Med. 2004;52:693–697. tal wall shear stress in the aorta by flow-sensitive four-dimensional-MRI. J Magn Reson
60. Magnusson P, Johansson E, Mansson S, et al. Passive catheter tracking during interven- Imaging. 2009;30:77–84.
tional MRI using hyperpolarized 13C. Magn Reson Med. 2007;57:1140–1147. 92. Garcia-Alvarez A, Fernandez-Friera L, Mirelis JG, et al. Non-invasive estimation of
61. Burl M, Coutts GA, Young IR. Tuned fiducial markers to identify body locations with pulmonary vascular resistance with cardiac magnetic resonance. Eur Heart J. 2011;32:
minimal perturbation of tissue magnetization. Magn Reson Med. 1996;36:491–493. 2438–2445.
62. Kuehne T, Fahrig R, Butts K. Pair of resonant fiducial markers for localization of endovas- 93. Ewert P, Berger F, Daehnert I, et al. Transcatheter closure of atrial septal defects without
cular catheters at all catheter orientations. J Magn Reson Imaging. 2003;17:620–624. fluoroscopy: Feasibility of a new method. Circulation. 2000;101:847–849.
63. Kuehne T, Weiss S, Brinkert F, et al. Catheter visualization with resonant markers at MR 94. Buecker A, Spuentrup E, Grabitz R, et al. Magnetic resonance-guided placement of atrial
imaging-guided deployment of endovascular stents in swine. Radiology. 2004;233:774–780. septal closure device in animal model of patent foramen ovale. Circulation. 2002;106:
64. Flask C, Elgort D, Wong E, et al. A method for fast 3D tracking using tuned fiducial 511–515.
markers and a limited projection reconstruction FISP (LPR-FISP) sequence. J Magn Reson 95. Beerbaum P, Korperich H, Esdorn H, et al. Atrial septal defects in pediatric patients:
Imaging. 2001;14:617–627. Noninvasive sizing with cardiovascular MR imaging. Radiology. 2003;228:361–369.
65. Weiss S, Kuehne T, Brinkert F, et al. In vivo safe catheter visualization and slice tracking 96. Pedra CA, Pedra SR, Chaccur P, et al. Perventricular device closure of congenital muscular
using an optically detunable resonant marker. Magn Reson Med. 2004;52:860–868. ventricular septal defects. Expert Rev Cardiovasc Ther. 2010;8:663–674.
66. Ladd ME, Erhart P, Debatin JF, et al. Guidewire antennas for MR fluoroscopy. Magn 97. Ratnayaka K, Raman VK, Faranesh AZ, et al. Antegrade percutaneous closure of mem-
Reson Med. 1997;37:891–897. branous ventricular septal defect using X-ray fused with magnetic resonance imaging.
67. Ocali O, Atalar E. Intravascular magnetic resonance imaging using a loop less catheter JACC Cardiovasc Interv. 2009;2:224–230.
antenna. Magn Reson Med. 1997;37:112–118. 98. Ratnayaka K, Saikus CE, Faranesh AZ, et al. Closed-chest transthoracic magnetic reso-
68. Yang X, Yeung CJ, Ji H, et al. Thermal effect of intravascular MR imaging using an nance imaging (MRI)-guided ventricular septal defect closure in swine. JACC Cardiovasc
MR imaging-guidewire: An in vivo laboratory and histopathological evaluation. Med Sci Interv. 2011;4(12):1326–1334.
Monit. 2002;8:MT113–MT117. 99. Bucker A, Neuerburg JM, Adam G, et al. [MR-guided coil embolisation of renal arteries
69. Liu CY, Farahani K, Lu DS, et al. Safety of MRI-guided endovascular guidewire applica- in an animal model]. Rofo. 2003;175:271–274.
tions. J Magn Reson Imaging. 2000;12:75–78. 100. Marshall MW, Teitelbaum GP, Kim HS, et al. Ferromagnetism and magnetic resonance
70. Weiss S, Vernickel P, Schaeffter T, et al. Transmission line for improved RF safety of inter- artifacts of platinum embolization microcoils. Cardiovasc Intervent Radiol. 1991;14:
ventional devices. Magn Reson Med. 2005;54:182–189. 163–166.
71. Mekle R, Hofmann E, Scheffler K, et al. A polymer-based MR-compatible guidewire: A 101. Fratz S, Hess J, Schwaiger M, et al. More accurate quantification of pulmonary blood
study to explore new prospects for interventional peripheral magnetic resonance angiog- flow by magnetic resonance imaging than by lung perfusion scintigraphy in patients with
raphy (IPMRA). J Magn Reson Imaging. 2006;23:145–155. Fontan circulation. Circulation. 2002;106:1510–1513.
72. Krueger S, Schmitz S, Weiss S, et al. An MR guidewire based on micropultruded fiber- 102. Bove EL, Migliavacca F, de Leval MR, et al. Use of mathematic modeling to compare
reinforced material. Magn Reson Med. 2008;60:1190–1196. and predict hemodynamic effects of the modified Blalock–Taussig and right ventricle-
73. Kuehne T, Yilmaz S, Schulze-Neick I, et al. Magnetic resonance imaging guided cath- pulmonary artery shunts for hypoplastic left heart syndrome. J Thorac Cardiovasc Surg.
eterisation for assessment of pulmonary vascular resistance: In vivo validation and clinical 2008;136:312–320 e2.
application in patients with pulmonary hypertension. Heart. 2005;91:1064–1069. 103. Hsia TY, Cosentino D, Corsini C, et al. Use of mathematical modeling to compare and
74. Muthurangu V, Taylor A, Andriantsimiavona R, et al. Novel method of quantifying predict hemodynamic effects between hybrid and surgical norwood palliations for hypo-
pulmonary vascular resistance by use of simultaneous invasive pressure monitoring and plastic left heart syndrome. Circulation. 2011;124:S204–S210.
phase-contrast magnetic resonance flow. Circulation. 2004;110:826–834. 104. Yang X, Atalar E. Intravascular MR imaging-guided balloon angioplasty with an MR
75. Luijnenburg SE, Robbers-Visser D, Moelker A, et al. Intra-observer and interobserver imaging guide wire: Feasibility study in rabbits. Radiology. 2000;217:501–506.
variability of biventricular function, volumes and mass in patients with congenital heart 105. Yang X, Bolster BD Jr, Kraitchman DL, et al. Intravascular MR-monitored balloon angio-
disease measured by CMR imaging. Int J Cardiovasc Imaging. 2010;26:57–64. plasty: An in vivo feasibility study. J Vasc Interv Radiol. 1998;9:953–959.
76. Winter MM, Bernink FJ, Groenink M, et al. Evaluating the systemic right ventricle by 106. Rickers C, Seethamraju RT, Jerosch-Herold M, et al. Magnetic resonance imaging guided
CMR: The importance of consistent and reproducible delineation of the cavity. J Cardiovasc cardiovascular interventions in congenital heart diseases. J Interv Cardiol. 2003;16:
Magn Reson. 2008;10:40. 143–147.
LWBK1209-ch38_p634-647.indd 646 16/05/13 9:17 PM

107. Godart F, Beregi JP, Nicol L, et al. MR-guided balloon angioplasty of stenosed aorta: In 124. Cribier A, Eltchaninoff H, Bash A, et al. Percutaneous transcatheter implantation of
vivo evaluation using near-standard instruments and a passive tracking technique. J Magn an aortic valve prosthesis for calcific aortic stenosis: First human case description.
Reson Imaging. 2000;12:639–644. Circulation. 2002;106:3006–3008.
108. Manke C, Nitz WR, Djavidani B, et al. MR imaging-guided stent placement in iliac arte- 125. Eicken A, Ewert P, Hager A, et al. Percutaneous pulmonary valve implantation: Two-
rial stenoses: A feasibility study. Radiology. 2001;219:527–534. centre experience with more than 100 patients. Eur Heart J. 2011;32:1260–1265.
109. Godart F, Labrot G, Devos P, et al. Coarctation of the aorta: Comparison of aortic dimen- 126. Nordmeyer J, Tsang V, Gaudin R, et al. Quantitative assessment of homograft function
sions between conventional MR imaging, 3D MR angiography, and conventional angiog- 1 year after insertion into the pulmonary position: Impact of in situ homograft geometry
raphy. Eur Radiol. 2002;12:2034–2039. on valve competence. Eur Heart J. 2009;30:2147–2154.
110. Krueger JJ, Ewert P, Yilmaz S, et al. Magnetic resonance imaging-guided balloon angio- 127. Lurz P, Nordmeyer J, Muthurangu V, et al. Comparison of bare metal stenting and per-
plasty of coarctation of the aorta: A pilot study. Circulation. 2006;113:1093–1100. cutaneous pulmonary valve implantation for treatment of right ventricular outflow tract
111. Okubo M, Benson LN. Intravascular and intracardiac stents used in congenital heart dis- obstruction: Use of an X-ray/magnetic resonance hybrid laboratory for acute physiologi-
ease. Curr Opin Cardiol. 2001;16:84–91. cal assessment. Circulation. 2009;119:2995–3001.
112. Ovaert C, Caldarone CA, McCrindle BW, et al. Endovascular stent implantation for the 128. Jerzewski A, Pattynama PM, Steendijk P, et al. Development of an MRI-compatible cath-
management of postoperative right ventricular outflow tract obstruction: Clinical efficacy. eter for pacing the heart: Initial in vitro and in vivo results. J Magn Reson Imaging. 1996;
J Thorac Cardiovasc Surg. 1999;118:886–893. 6:948–949.
113. Redington AN, Weil J, Somerville J. Self expanding stents in congenital heart disease. Br 129. Dickfeld T, Calkins H, Zviman M, et al. Anatomic stereotactic catheter ablation on
Heart J. 1994;72:378–383. three-dimensional magnetic resonance images in real time. Circulation. 2003;108:
114. van Dijk LC, van Holten J, van Dijk BP, et al. A precious metal alloy for construction of MR 2407–2413.
imaging-compatible balloon-expandable vascular stents. Radiology. 2001;219:284–287. 130. Lardo AC, McVeigh ER, Atalar E, et al. Magnetic resonance guided radiofrequency abla-
115. Buecker A, Spuentrup E, Ruebben A, et al. Artifact-free in-stent lumen visualization by tion: Visualization and temporal characterization of thermal lesions. Circulation. 1998;
standard magnetic resonance angiography using a new metallic magnetic resonance imag- 98:3385.
ing stent. Circulation. 2002;105:1772–1775. 131. Susil RC, Yeung CJ, Halperin HR, et al. Multifunctional interventional devices for MRI:
116. Raman VK, Karmarkar PV, Guttman MA, et al. Real-time magnetic resonance-guided A combined electrophysiology/MRI catheter. Magn Reson Med. 2002;47:594–600.
endovascular repair of experimental abdominal aortic aneurysm in swine. J Am Coll 132. Dukkipati SR, Mallozzi R, Schmidt EJ, et al. Electroanatomic mapping of the left ventricle
Cardiol. 2005;45:2069–2077. in a porcine model of chronic myocardial infarction with magnetic resonance-based cath-
117. Raval AN, Telep JD, Guttman MA, et al. Real-time magnetic resonance imaging-guided eter tracking. Circulation. 2008;118:853–862.
stenting of aortic coarctation with commercially available catheter devices in swine. 133. Lardo AC, McVeigh ER, Jumrussirikul P, et al. Visualization and temporal/spatial char-
Circulation. 2005;112:699–706. acterization of cardiac radiofrequency ablation lesions using magnetic resonance imaging.
118. Eggebrecht H, Heusch G, Erbel R, et al. Real-time vascular interventional magnetic reso- Circulation. 2000;102:698–705.
nance imaging: The future of aortic stent-graft placement? Basic Res Cardiol. 2007;102:1–8. 134. McGann CJ, Kholmovski EG, Oakes RS, et al. New magnetic resonance imaging-based
119. Nordbeck P, Bauer WR, Fidler F, et al. Feasibility of real-time MRI with a novel carbon cath- method for defining the extent of left atrial wall injury after the ablation of atrial fibrilla-
eter for interventional electrophysiology. Circ Arrhythm Electrophysiol. 2009;2:258–267. tion. J Am Coll Cardiol. 2008;52:1263–1271.
120. Mahnken AH, Chalabi K, Jalali F, et al. Magnetic resonance-guided placement of aor- 135. Krueger S, Lips O, David B, et al. Towards MR-guided EP interventions using an RF-safe
tic stents grafts: Feasibility with real-time magnetic resonance fluoroscopy. J Vasc Interv concept. J Cardiovasc Magn Reson. 2009;11:O84.
Radiol. 2004;15:189–195. 136. Bove EL, de Leval MR, Migliavacca F, et al. Toward optimal hemodynamics: Computer
121. Buecker A, Neuerburg JM, Adam GB, et al. Real-time MR fluoroscopy for MR-guided modeling of the Fontan circuit. Pediatr Cardiol. 2007;28:477–481.
iliac artery stent placement. J Magn Reson Imaging. 2000;12:616–622. 137. Troianowski G, Taylor CA, Feinstein JA, et al. Three-dimensional simulations in Glenn
122. Bonhoeffer P, Boudjemline Y, Saliba Z, et al. Percutaneous replacement of pulmonary patients: Clinically based boundary conditions, hemodynamic results and sensitivity to
valve in a right-ventricle to pulmonary-artery prosthetic conduit with valve dysfunction. input data. J Biomech Eng. 2011;133:111006.
Lancet. 2000;356:1403–1405. 138. LaDisa JF Jr, Alberto Figueroa C, Vignon-Clementel IE, et al. Computational simulations
123. Boudjemline Y, Bonhoeffer P. Steps toward percutaneous aortic valve replacement. for aortic coarctation: Representative results from a sampling of patients. J Biomech Eng.
Circulation. 2002;105:775–778. 2011;133:091008.
LWBK1209-ch38_p634-647.indd 647 16/05/13 9:17 PM

Chapter
39 Dara L. Kraitchman
Endovascular Delivery of Gene and

Stem Cell Therapy
■■ Endovascular Magnetic Resonance Imaging (b) the lack of ionizing radiation for the interventionalist
Delivery Devices and the patient; (c) the lack of iodinated contrast agents that
are often poorly tolerated by patients with cardiac or renal
■■ Interventional Magnetic Resonance Imaging compromise; and (d) the ability to better target the thera-
Techniques peutic whether it be to determine viable myocardium using
■■ Gene Transfer–Guided Magnetic Resonance delayed contrast-enhanced MRI for stem cell targeting or
Imaging vulnerable atherosclerotic plaques by MR vessel wall imag-
ing for gene therapy targeting. In addition, the development
■■ Contrast Agents for Gene Tracking and Cell of MR-visible labeling techniques offers a method to view
Labeling targeted delivery and track the fate of these therapeutics
■■ Imaging Gene Delivery over weeks to months.
Although computed tomography (CT) is emerging as a
■■ Imaging Gene Expression new technology for whole-body screening and rapid deter-
■■ Cellular Labeling Strategies mination of stenotic coronary vessels and global left ventric-
ular function, CT alone does not lend itself to endovascular
■■ Cellular Detection Limits delivery procedures. Nonetheless, the recent development
■■ Imaging Stem Cell Delivery of hybrid x-ray/MRI systems is an obvious extension that
enables reductions in radiation dose while deriving the ben-
■■ Summary
efits of MRI along with the high temporal resolution and
angiographic capabilities of x-ray imaging. As multislice CT
Most preclinical studies of the efficacy of gene and stem cell scanners permeate the market, fusion of these 3D CT images
therapies have relied on validation by postmortem tissue with conventional x-ray angiography are becoming more
analysis using techniques, such as immunohistochemistry routine for interventional procedures. The recent intro-
or polymerase chain reaction. These techniques can provide duction of flat-panel x-ray detectors in place of traditional
high spatial resolution information, which is microscopic, image intensifiers enables the acquisition of C-arm CTs,
and, when used in combination with serial animal sacrifice, which are intrinsically registered with live x-ray fluoroscopic
the distribution of stem cells or longevity of gene expression image (1), for 3D targeting and soft tissue detail. Fusion of
can also be obtained. However, noninvasive methodologies live x-ray fluoroscopy with previously acquired CT, MRIs,
that can be used to serially assess the presence of stem cells and C-arm CTs is an enabling technology for stem cell and
or gene expression to determine the safety and the efficacy of gene targeting on a platform that is familiar to the cardiac
these therapies would enhance clinical trials markedly. Thus, interventionalist using off-the-shelf devices.
the introduction of magnetic resonance imaging (MRI) scan-
ners with real-time imaging capabilities has expanded MRI
from a diagnostic imaging tool to interventional applications
directed at gene and cellular therapeutic delivery. However, Endovascular Magnetic
x-ray angiography remains the imaging modality of choice Resonance Imaging
because of the high temporal resolution for minimally inva- Delivery Devices
sive cardiac stem cell and gene delivery.
MR delivery of endovascular and transmyocardial thera- In concert with the development of MR scanner with real-
pies offers four major advantages over other imaging modal- time acquisition capabilities, another obvious hurdle is the
ities: (a) The ability to acquire full three-dimensional (3D) development of delivery devices that are MR-compatible.
images of the cardiovascular system with softtissue detail; In the simplest form, a conventional x-ray delivery catheter
648
LWBK1209-ch39_p648-666.indd 648 17/05/13 5:23 PM

Chapter 39 ■ Endovascular Delivery of Gene and Stem Cell Therapy 649
Figure 39.1. Transvenous T1-weighted MR coronal image (left) of a platinum stent (Omniflex,
Angiodynamics), placed under x-ray fluoroscopy in the iliac artery, in a normal pig obtained using an MR
imaging guidewire placed in the iliac vein. Notice that the stent is virtually invisible in the MRI using a double
inversion fast spin-echo sequence (612 milliseconds repetition time, 11 milliseconds echo time, 6 signal aver-
ages, 0.19 mm2 in plane resolution). Placement of the stent is demonstrated in pathology (right) (Images cour-
tesy of Lawrence V. Hofmann, MD.)
could be redesigned using nonmagnetic materials. However, triangulation of several of these small coils (13,22,23). The
often these MR-compatible devices either create large cost of this improved resolution and localization is the in-
magnetic susceptibility artifacts or signal voids that either creased complexity of both device manufacture and image
limit or prevent passive tracking of the device. At the other acquisition control. In addition, increases in device diameter
extreme, nonmetallic materials, such as nitinol guidewires to accommodate the wiring from multiple coils and the in-
or platinum stents, are essentially MR-invisible (Fig. 39.1), creased risk of localized heating of the device have limited
making passive tracking of these devices difficult (2). their use to preclinical animal studies. Several MR imaging
Several mechanisms for improving the passive visualiza- guidewires have been safety tested and used in human clini-
tion of these devices have been developed. A simple approach cal trials in transesophageal and transvenous approaches
is to visualize a catheter lumen during injection of MR con- (Fig. 39.2) (24,25).
trast agents similar to an x-ray angiogram (3–5). Another Two other approaches have emerged in the past 5 years
approach is to incorporate markers or coatings of para- to attempt to circumvent the concerns of heating of active
magnetic or superparamagnetic iron oxide (SPIO) contrast MR interventional devices and poor handling characteristics
agents in the devices (6–10) that can be tracked. Similarly, of MR-compatible guidewires. One approach is based on
current-controlled susceptibility artifacts have also been transformers used in transmission lines for electric current
used historically for tracking (11,12). where the receiver circuit is divided into short cables with
Active tracking approaches incorporate small MR receiver miniature transformers to avoid long wire lengths that are
coil(s) into the device to receive the MR signal (13–16). In prone to heating (26). The second approach is to create a
the simplest approach, a guidewire is used as a receiver coil polymer-based guidewire that is inherently nonferrous and
or loopless antenna for imaging (17–21). The advantage of has stiffness characteristics that mimic conventional guide-
the active tracking approaches is the ability to obtain high- wires. Krueger et al. (27) have developed a fiber-reinforced
resolution imaging from the miniaturized coils as well as the guidewire that is formed using micro-pultrusion methods.
ability to perform catheter or catheter-tip localization by With handling characteristics that were very similar to
LWBK1209-ch39_p648-666.indd 649 17/05/13 5:23 PM

Figure 39.2. X-ray angiogram (left) obtained in a 61-year-old patient with claudication demonstrates a
large plaque (arrows, arrowhead) in the common iliac artery. An axial MRI image (right) shows the imaging
guidewire (IVMR) in the common iliac vein (V) and the bright fibrous cap with the hypointense lipid calcium
in the iliac artery. L, lumen. (Images courtesy of Lawrence V. Hofmann, MD.)
metallic guidewires, interventional MRI clinical trials were catheters, such as coils for tip tracking (Fig. 39.4). However,
initiated (28). However, the lack of visualization of the mi- most of the device development has been performed at 1.5
cropultruded fiber-reinforced guidewire proved unsafe as T rather than 3 T, which has ramifications in spatial and
the guidewire became severed in a patient during the early temporal resolution due to the decrease signal-to-noise ratio
clinical trials. Another polymer guidewire made from poly- (SNR) at 1.5 T.
etheretherketone that has an iron oxide coating for passive
visualization is being tested preclinically for MRI-guided an-
gioplasty and stenting (29). Interventional Magnetic
Specialized devices for stem cell injections under MRI Resonance Imaging Techniques
guidance were developed by several device companies. A
modified injection catheter system (Boston Scientific, Inc.), In addition to developing MR-compatible devices, a suite of
developed for x-ray–guided transmyocardial delivery of gene imaging platforms are emerging to enable real-time inter-
and stem cells, consisting of a curved guide catheter and ventional MR imaging of gene and stem cell delivery, using
spring-loaded injection needle, was modified for active cath- active delivery devices. Thus, in-room imaging displays
eter tracking and delivery in relevant large animal models of and monitoring capabilities, which were originally devel-
stem cell delivery for cardiac regeneration (30,31). A proto- oped for biopsy and neurointerventional procedures, are
type steerable catheter with a nitinol needle for endovascular rapidly being adapted for cardiovascular applications. The
injections (Bioheart, Inc.) has been coated with gadolinium inclusion of track balls and keyboards to enable full scan-
oxide markers for visualization under a hybrid x-ray/MR ner operation within the scan room has now been imple-
scanner (9). Using this hybrid imaging system, cardiac cath- mented by all major MR vendors (Fig. 39.5). Feedback
eterization can be performed under x-ray guidance with ac- systems to automatically adapt the imaging planes based
curate targeting to infarcted myocardium using MRI. A cus- on catheter position or velocity are now available (33).
tom injection catheter has also been developed for stem cell Graphical user interfaces that allow rapid plane manipula-
therapeutic delivery with MR fluoroscopy (32) that enables tion and table motion are becoming more common (34).
access to all portions of the left ventricle because of a flex- Multiple approaches have been developed to fuse the sig-
ible steerable and bendable distal portion (Fig. 39.3). While nal from the catheter coils with body coil images to incor-
commercial interest in MR-compatible devices has waned, the porate higher spatial resolution data from roadmaps with
completion of a number of positive cardiac stem cell trials has images acquired at a high temporal resolution to track
rekindled interest in MRI delivery to avoid ionizing radiation. motion. Improvements in the graphical interfaces have also
New catheter designs are being developed incorporating fea- been driven by the development of MR-compatible abla-
tures developed in MR-compatible electrophysiology ablation tion systems.
LWBK1209-ch39_p648-666.indd 650 17/05/13 5:23 PM

Figure 39.3. The distal tip of a custom MR-compatible loopless antenna for transendocardial injection
of genetic and cellular therapeutics will appear as a bright signal at the distal tip (arrow) on MR images. The
mechanical properties for steering and flexibility to reach all portions of the left ventricular endocardium are
incorporated into a pull-wire. Approximately 25% (left) and 50% (right) deflection of the catheter tip are
demonstrated. (From: Karmarkar PV, Kraitchman DL, Izbudak I, et al. MR-trackable intramyocardial injec-
tion catheter. Magn Reson Med. 2004;51:1163–1172, with permission.)
However, many hurdles still remain to the widespread in a moving heart. In addition, the high noise level of the
acceptance of interventional MRI. The implementation of MR suite requires specialized systems for communications
methods to rapidly adjust imaging parameters “on the fly” within the interventionalist team.
is still limited. Although techniques, such as SENSE (35)
and SMASH (36) have enabled increases in the temporal
resolution of interventional MRI, even at the highest frame Gene Transfer–Guided Magnetic
rates, the temporal and spatial resolutions are much lower Resonance Imaging
than x-ray fluoroscopy, making precise targeting difficult
The explosion of potential uses of gene therapy in clinical
medicine, along with the initiation of the first human clinical
trials in the United States in the 1990s, has brought forth the
importance of not only interventional techniques to deliver
genetic materials but also imaging methods to monitor the
efficacy of gene transfection and expression. Endovascular
delivery of gene products offers a method to apply high
concentration of vectors in a localized region as compared
to systemic intravenous administration. Besides using gene
therapy to replace defective or missing genes, genes may
be introduced to (a) up-regulate or inhibit expression, as
in vascular endothelial growth factor (VEGF) production,
to enhance angiogenesis in a stenotic vessel (37) or reduce
infarct size in angina patients (38), or phospholamban inhi-
bition of SERCA2a to enhance contractility in heart failure
(39); (b) deliver genes that may produce a toxic product
or cause cell suicide, as in the use of herpes simplex virus
thymidine kinase/ganciclovir (HSV-tk/GCV) systems (40); or
A B (c) act as reporter gene either to monitor the life span of a
transfected cell or to image an event, such as cell differentia-
Figure 39.4. A digital image (A) of a modified MR-compatible tion that causes production of a new protein, enzyme prod-
radiofrequency ablation catheter, which contains multiple coils along
uct, or cell receptor.
the length for tracking the catheter position interleaved with image
acquisition, includes a needle for stem cell injections (shown protrud-
There are two primary methods to transfect cells with
ing from the catheter tip). A screen capture (B) of rendered tip of the genes. Nonviral transfection methods use either naked
catheter guided into the left ventricle of a pig for injection of stem cell plasmid DNA alone or in combination with substances
therapeutics. Planned injection sites are shown as colored dots in the called transfection agents (TAs), like cationic liposomes.
myocardium. Viral transfection methods include retroviral, adenoviral,
LWBK1209-ch39_p648-666.indd 651 17/05/13 5:23 PM

Figure 39.5. Screen capture of the Siemens prototype interactive front end (IFE) graphical interface that
enables real-time scan plane manipulation and serial acquisition of up to three imaging planes. The image is
reconstructed with the active injection catheter colored green for enhanced visibility. Representative pseudo
long- and short-axis images are shown acquired in real-time in vivo in a canine reperfused myocardial infarc-
tion. Bookmark images (small images at bottom) facilitate rapid return to previous scan plane position using
a simple drag-n-drop of the image plane into one of three image acquisition planes. (From: Kraitchman
DL, Gilson WD, Lorenz CH. Stem cell therapy: MRI guidance and monitoring. J Magn Reson Imaging.
2008;27(2):299–310, with permission.)
adeno-associated viral (AAV), lentiviral, and herpes (sim- Contrast Agents for Gene
plex) viral vectors to transfer genetic material to the cell. Tracking and Cell Labeling
The primary advantage of viral transfection methods is
their more efficient transfection rate than nonviral trans- Because of the low background signal, labeling techniques
fection techniques. The negative aspect of this higher with radionuclides for radionuclide imaging are exquisitely
transfection efficiency is the tendency for these techniques sensitive to an extremely low number of molecular targets.
to be immunogenic. The inability to cause stable, long- MR or CT methods would require 10- to 100-fold large
term transfection and expression can be a problem if the amplification of these targets for detection. In addition, cel-
gene product must be expressed for the life-span of the pa- lular labeling for x-ray imaging would require an iodinated
tient, but can be an asset if the gene is only expressed for contrast agent load that is incompatible with maintaining
a short time period, thereby limiting the immunogenicity. normal cellular function and viability. However, using tech-
In particular, AAV gene therapy has a natural tropism for niques largely adapted from monoclonal antibody (MoAb)
cardiovascular cells, for example, smooth muscle cells and techniques of radionuclide imaging, cellular, and genetic
cardiomyocytes; exhibits less of an inflammatory response labeling strategies for MR imaging have been developed
than adenoviral vectors; and appears to result in expres- using conventional MR contrast agents. The inherent advan-
sion of the gene product for several months. Since drug- tage of labeling with MR contrast agents is that the cells
eluting stents have lessened much of the concern about are not exposed to radioactive species that can become espe-
stent restenosis, gene therapy targets are now more con- cially cytotoxic when internalized.
centrated in preventing stent graft restenosis, angiogenesis MR contrast agents can be divided into two major classes
for peripheral arterial disease, and calcium transport in of agents: The paramagnetic and SPIO contrast agents.
heart failure (41). Gadolinium chelates, which form the most widely used
LWBK1209-ch39_p648-666.indd 652 17/05/13 5:23 PM

clinically approved MR contrast agents, are in the paramag- 6% gadolinium contrast agent was delivered intra-arterially
netic class of agents. The adoption of gadolinium-based MR and monitored in real time (Fig. 39.6). Radiofrequency (RF)
contrast agents for noncellular imaging is largely a result heating of imaging guidewires, which is a concern of active
of the ability of these agents to decrease the T1 relaxation catheter systems, has been exploited in this application by
time at low doses. This decrease in T1 results in an increased this group to enhance transfection rates (48). In a similar
signal intensity (i.e., hyperintense signal) of tissues exposed study to visualize gene delivery, Barbash et al. (49) used a
to the contrast agent using T1-weighted imaging sequences. percutaneous approach in a rat infarction model to deliver
However, because of the high toxicity of free gadolinium, an adenoviral LacZ reporter gene mixed with gadolinium
chelation of gadolinium is paramount. Concerns about via a 22G needle. In 2013, Barbash et al. (50) performed
dechelation of the gadolinium compounds in applications transendocardial delivery in dogs of an AAV mixed with an
for genetic and cellular labeling have largely limited the iodinated contrast agent and SPIO to monitor delivery suc-
development of these agents for these purposes (42). More cess during x-ray fluoroscopic imaging fused with previously
problematic is the reduced ability of paramagnetic com- acquire MRIs that were volume rendered.
pounds when internalized to affect extracellular water, and A more sophisticated method to monitor sustained release
hence, the ability to alter signal intensity is greatly reduced. of a gene would be to encapsulate the gene and paramagnetic
The SPIO particles were developed shortly after the contrast agent in a biodegradable sphere. Faranesh et al. (51)
gadolinium-based contrast agents (43,44). The large mag- have developed a poly(DL-lactic-co-glycolic acid) (PLGA)
netic moments, which are more than three orders of mag- microsphere containing VEGF and gadolinium diethylene-
nitude greater than paramagnetic-based contrast agents, of triamine pentaacetic acid (Gd-DTPA), which releases the
SPIO particles cause a greater effect on T2 relaxation and a gene and contrast agent over 6 weeks that can be nonin-
smaller effect on T1 relaxation. Thus, on T2*-weighted vasively imaged as hyperenhancement on T1-weighted MRI.
images, SPIO particles appear hypointense and create a much Microsphere systems, such as this system, enable a sustained
larger signal change or contrast per unit of metal particle release of the gene product over many weeks to assist with
than paramagnetic contrast agents. Thus, small quantities of higher or sustained transfection rates, but what is unlikely
SPIO particles can be used for gene or cellular labeling, yet is that the kinetics of Gd-DTPA and VEGF release from the
with a much larger amplification effect than paramagnetic microsphere are equivalent, such that absolute quantitation
compounds. Importantly, less agent must be internalized of gene release from MRI can be obtained using this delivery
to create image contrast, thereby limiting cellular toxicity. technique. Other cutting-edge techniques, which have been
Moreover, if the SPIO is degraded, the free iron that is re- developed to determine cellular uptake of the gene product
leased does not appreciably expand the native iron pool and, include binding plasmid DNA to polymers that will stabilize
thus, can be degraded along normal iron recycling pathways. SPIOs (52,53) or loading micelles or nanoparticles with iron
Most commercially available forms of SPIOs and ultrasmall oxides and DNA (54,55) such that the transfected cells ap-
SPIOs (USPIOS) have coatings to prevent particle aggre- pear dark on T2*-weighted MRI.
gation. One of the most common USPIO coatings is dex-
tran, which is a convenient surface for binding ligands and
other functional groups for labeling. For economic reasons, Imaging Gene Expression
an FDA-approved formulation of a dextran-coated SPIO-
ceased product (Feridex, Bayer Healthcare), which was used Another area where MR delivery may be useful is in com-
extensively in preclinical stem cell labeling studies, ceased bination with ultrasonography. High-frequency ultrasound
production. Recently, several groups have demonstrated the has been used to increase vascular permeability and also
use of ferumoxytol (an FDA-approved USPIO for the treat- rupture ultrasound microbubbles (56,57). MR paramag-
ment of iron deficiency in renal patients) in combination netic agents encapsulated in liposomes could be combined
with protamine sulfate (an FDA-approved TA for reversing with naked DNA to enable high-resolution localization of
heparin therapy) to label stem cells (45–47). the genes followed by enhanced transfection using high-
frequency ultrasound.
Several groups have been able to develop amplification sys-
Imaging Gene Delivery tems to visualize gene expression using MRI. In particular,
Weissleder et al. (58) and Ichikawa et al. (59) have transfected
As stated previously, MR imaging for gene detection is tumor cells with an engineered transferrin receptor (Tfr).
largely inferior to radionuclide imaging techniques because Monocrystalline iron oxides (MION) can be conjugated to
of the adherently lower sensitivity for detecting genetic tar- transferrin (Tf), the natural ligand of Tfr. After systemic in-
gets. Until amplification strategies to increase MR sensitiv- jection, the MION-Tf will be taken up preferentially by the
ity are more fully developed, the ability to harness the high tumor cells overexpressing Tfr, resulting in a hypointensity
spatial resolution and interventional capabilities of MRI for on T2*-weighted images of the tumors. In this model system,
cardiac gene therapies will not be recognized. amplification has been achieved in two ways: (a) The overex-
In an early attempt to apply these techniques for cardio- pression of the receptor and (b) the ability of SPIO particles
vascular gene delivery, Du et al. (48) developed a novel MR with high magnetic moments to cause large dephasing of local
gene delivery system that was validated in a swine model protons, resulting in a dramatic shortening of T2 far beyond
on a clinical MR scanner. A gene delivery catheter was what could be achieved by conventional gadolinium chelates.
combined with an MR imaging guidewire, and a lentiviral In a similar strategy, the up-regulation of the enzyme tyrosi-
vector carrying green fluorescent protein (GFP) mixed with nase in melanomas results in increased melanin production.
LWBK1209-ch39_p648-666.indd 653 17/05/13 5:23 PM

A B C
D E F
G H
Figure 39.6. High-resolution MR images of the gadolinium/GFP lentivirus transfer in the iliac artery of a
pig. A: Before gadolinium/GFP lentivirus infusion, the balloon is inflated with 3% Magnevist. The open arrow
indicates the artery. B–F: During gadolinium/GFP lentivirus infusion from minute 3 to minute 15 (at 3-minute
intervals), the arterial wall is enhanced by the gadolinium coming from the gene infusion channels (arrow-
heads, B) of the gene delivery catheter. At minute 15, the arterial wall is enhanced as a ring (arrow, F). G,H:
Corresponding immunohistochemistry in both control (G) and GFP-targeted (H) arteries. H: GFP is detected
as brown-colored precipitates through all layers of the intima (arrows) and media as well as the adventi-
tia. Original magnification, 200X. V, vein. Scale = 1 mm. (Reprinted from: Yang X, Atalar E, Li D, et al.
Magnetic resonance imaging permits in vivo monitoring of catheter-based vascular gene delivery. Circulation.
2001;104:1588–1590, with permission.)
Like Tfr, melanin has a high affinity for iron, and gene ex- bound to multiple targets. Perez et al. (62–65), Bogdanov
pression can thus be probed using exogenous administration et al. (66), and Josephson et al. (67) have developed several
of MION (60,61). Internalization of the MION results in a of these enzyme-sensing particles that cluster in the presence
bright signal in T1-weighted MRI. of peroxidases. In particular, Perez et al. (65) have developed
The potential exists for the development of smart MR con- a superparamagnetic nanoparticle in this class which, in the
trast agents which either act as substrates that are cleaved by presence of myeloperoxidase (MPO), assembles to cause de-
specific enzymatic activity to expose the magnetic nanopar- creases in spin–spin relaxation times that can be detected by
ticle or are assembled into larger polypeptides when specific MRI. Although gadolinium substrates lack the amplification
enzymes or proteases are present to amplify the magnetic sig- of iron oxide nanoparticles, this same group has designed a
nal. In essence, the magnetic relaxation of the particle can nanoparticle probe with gadolinium substrates that polymer-
be changed as a result of the increased size of the magnetic ize in the presence of MPO (68). Myeloperoxidase, which
particle resulting from cross-linking, as multiple ligands are is produced by activated macrophages, has been implicated
LWBK1209-ch39_p648-666.indd 654 17/05/13 5:23 PM

in inflammation and particularly atherosclerotic plaques. In Thus, changes in T1 relaxation can be achieved in vivo. In
particular, MPO activity has been detected in a rabbit athero- addition, therapeutics can be incorporated into the surface
sclerotic model as hyperenhancing regions on T1-weighted of the nanoparticle for drug delivery (75). By modifying the
images as early as 2 hours after administration (69), that is antibody receptor complex, targeting of this agent has been
distinct from enhancement by clinically approved Gd contrast performed to fibrin (i.e., thrombus) for cardiovascular ap-
agents. Thus, the vulnerable atherosclerotic plaque could be plications (73,76). To decrease the potential toxicity of the
detected or the response of the plaque to genetic modification gadolinium chelate, several paramagnetic formulations have
could be performed using this targeted MPO agent. Particles been explored, which have also increased the relaxivity of
in this class can also be used as MR switches to sense the assem- the agent resulting in enhanced signal enhancement after in-
bly and dispersal of the particles as occurring in the presence of jection (77). The most exciting cardiovascular application is
restriction endonucleases during DNA cleavage (70). the targeting to markers of angiogenesis (i.e., aνb3 integrins)
Another approach that may increase the MR sensitivity for using this agent (Fig. 39.8) (78,79). The fluorine signal of
clinical applications is to image a nonproton reporter probe, the nanoparticle provides an independent method for vali-
such as fluorine-19 (19F). Since there are almost undetectable dating the bright Gd signal detected in proton imaging. In
levels of fluorine in the body, there will be a high sensitivity addition, the introduction of higher field strength magnets
to detect the presence of the reporter probe using either MR for clinical imaging has overcome the signal-to-noise issues
spectroscopy or 19F MRI. Mizukami et al. (71) have devel- often associated with 19F spectroscopy and imaging.
oped a reporter probe containing gadolinium and a fluorine The potential to measure gene expression has also been
moiety linked by galactose; the close proximity of the gado- demonstrated with 19F spectroscopy in a tumor model (80).
linium to the fluorine quenches the fluorine resonance peak. In this model, genetically modified tumor cells expressing
When b-galactosidase is produced by transfected cells, the cytosine deaminase from yeast were implanted in mice. The
reporter probe will be cleaved and the fluorine peak can now concept of gene-directed proenzyme therapy is to introduce
be imaged resulting as a hot spot on the MRI, which only a nonmammalian enzyme into tumor cells that efficiently
appears where transfected cells are present. Using a different converts a nontoxic prodrug into a cytotoxic metabolite.
reporter probe containing 19F, Yu et al. (72) showed the ability In this case, yeast cytosine deaminase efficiently converts
to detect b-galactosidase activity based on a chemical shift in 5-fluorocytosine into 5-fluorouracil (5-FU), thereby limiting
the fluorine peak and the trapping of ferric ions that can be the dose of the widely used chemotherapeutic agent 5-FU to
detected on proton MRI. the solid tumor. From the MR spectra, the pharmacokinet-
Lanza et al. (73) and Yu et al. (74) have developed a ics of the 5-FC metabolites can be determined and, hence,
targeted nanoparticle containing a liquid perfluorocarbon the gene expression in the tumor. It can be envisioned that
core with chelated gadolinium (Gd) complexes incorporated gene-directed proenzyme therapy for atherosclerosis could
into the outer surface. They have overcome the amplifica- be monitored in a similar manner.
tion problem with gadolinium-based agents by two means: The overexpression of protein products that can be
(a) Numerous Gd molecules can be bound to the surface of used in concert with chemical exchange saturation transfer
the nanoparticle and (b) antibody receptor complexes on the (CEST) MRI (81) have re-energized the field of MR reporter
nanoparticle are used for targeting so that the particle is not gene imaging. Using engineered cells to overexpress lysine-
internalized which would reduce T1 relaxivity (Fig. 39.7). rich residues, Gilad et al. (82) have assessed cellular tracking
in the brain in mice. Since one can theoretically distinguish
different exchangeable protein residues, a palette of labels
could be used similar to differently colored optical labels.
Given that naturally occurring proteins can be overexpressed
for CEST reporter gene imaging, the biocompatibility of this
reporter gene may be enhanced. Since CEST imaging is in-
herently a subtraction technique comparing saturation of
various resonant frequencies, the technique may be better
suited for imaging stationary organs like the brain than the
19F heart.
Cellular Labeling Strategies

As with gene therapy, most of the stem cell labeling strate-
gies have been directed toward labeling with iron oxide com-
pounds. Unlike gene therapy applications, surface labeling
for stem cells is not preferred because the label may become
Lipid coat Gd3+ detached and transferred to other cells. Thus, most stem cell
Figure 39.7. A targeted nanoparticle containing a liquid perfluoro- labeling techniques have been developed to optimize inter-
carbon core with chelated gadolinium (Gd3+) complexes incorporated nalization of the label. A few individuals have been devel-
into the outer surface. Monoclonal antibodies or targeted therapeutics oping techniques with paramagnetic compounds, but the
can be placed on the surface of the nanoparticle. (Image courtesy of focus at present has been on neurologic applications or have
Samuel Wickline, MD, and Greg Lanza, MD, PhD.) not yet been tested in vivo for cardiovascular applications.
LWBK1209-ch39_p648-666.indd 655 17/05/13 5:23 PM

Chol + Figure 39.8. Percent enhance-

avb3 + ment maps created from MR axial
images from individual aortic seg-
ments at renal artery (renal), mid-
aorta (mid), and diaphragm (diaph)
2 hours after treatment with avb3-
targeted paramagnetic nanoparticles
in cholesterol-fed rabbit (top), with
avb3-targeted nonparamagnetic
Chol + nanoparticles in cholesterol-fed rab-
avb3 − bit (middle), and with avb3-targeted

paramagnetic nanoparticles in rab-
bit on a normal diet (bottom). The
induction of angiogenesis in the
vasa vasorum seen as increased
signal enhancement is evident in
the cholesterol-fed rabbit receiving
the targeted gadolinium nanopar-
Chol − ticle (top). (Adapted from: Winter
avb3 +
PM, Morawski AM, Caruthers
SD, et al. Molecular imaging of
angiogenesis in early-stage athero-
sclerosis with alpha(v)beta3-integrin-
targeted nanoparticles. Circulation.
2003;108:2270–2274, with
permission.)
Modo et al. (83) have developed a gadolinium-based com- of iron oxides, immunoglobulins could be covalently linked
pound linked to dextran and a fluorescent dye, rhodamine, using a periodate oxidation/borohydride reduction method
that can be taken up by stem cells in vitro for intracellular in which the lysine groups of the monoclonal antibodies are
labeling. These exogenously labeled stem cells have been joined to the alcohol groups of the dextrans (94). Using this
implanted in a rat stroke model, and the migration has been technique, MION has been conjugated to an antimyosin
tracked in vivo by MRI and validated by detection of the flu- MoAb for detection of infarcted myocardium in the rat (95).
orescent label by histology (84). Using a calcium phosphate Although there are many other alternate methods to attach
transfection technique, Rudelius et al. (85) performed intra- MoAbs to magnetic nanoparticles, the lack of species cross-
cellular labeling of neuronal and embryonic stem cells with reactivity for many MoAbs requires that separate formula-
Gd-DTPA. Using this labeling technique, Daldrup-Link et al. tions be developed for animal validation and human studies.
(86) were able to achieve intracellular labeling of hematopoi- Meldrum et al. (96) attempted to circumvent the species
etic progenitor cells that resulted in significant lengthening of specificity problem by developing an iron oxide encapsulated
R1 such that detection of 500,000 cells or greater should be with a ferritin protein shell called magnetoferritin. Ferritin is
possible. However, as stated previously, the internalization highly conserved across species and thus, this agent has the po-
of these Gd particles will greatly reduce the potential con- tential for molecular imaging and cellular labeling. However, in
trast enhancement imparted by these agents. Thus, because vivo studies with magnetoferritin showed a rapid clearance of
of amplification problems and toxicity concerns, iron oxide the agent from the vascular pool that was unrelated to ferritin
compounds remain the preferred agents for cellular labeling. receptor binding (97), suggesting a limited use for this agent.
Cardiovascular applications of MR imaging with iron oxide Since dendrimers offer an efficient method to transfect cells
compounds were first implemented because of the recogni- (98–100), the development of dendrimer-encapsulated SPIOs
tion of the rapid uptake of these compounds by macrophages or “magnetodendrimers” offered another nonspecies specific
after systemic administration. Monocyte recruitment by the agent that could be used for cellular labeling. One formula-
atherosclerotic plaque has been demonstrated using MRI and tion of magnetodendrimers, MD-100, consists of iron oxide
accumulation of SPIOs in the hyperlipidemic rabbit and apoE nanoparticles coated with a generation 4.5 carboxylated den-
knockout mice models (87–90). Subsequently, retrospective drimer (101). These highly charged polymers bind to the cell
and prospective patient studies with USPIOs to identify ath- membrane and induce membrane binding and endocytosis.
eromas vulnerable to rupture have been undertaken (91,92). Many different cell types from many different animal species
Similarly, SPIO uptake by macrophages during allogeneic car- cultured in vitro for 24 to 48 hours with MD-100 show a
diac transplant rejection has been demonstrated in the rat (93). consistent, low iron concentration (10 to 25 mg Fe/mL) with
As with MR labeling for gene therapy, the initial meth- the agent stably maintained within endosomes (101–103).
ods for cellular MR labeling with iron oxides were based on Importantly, cell viability and proliferation was not affected
radionuclide MoAb techniques. Using the dextran coating by labeling with MD-100.
LWBK1209-ch39_p648-666.indd 656 17/05/13 5:23 PM

Billotey et al. (104) and Wilhelm et al. (105) have devel-

oped an anionic magnetic nanoparticle for cellular labeling
that is also maintained within endosomes. However, one
major disadvantage of these particles, as well as magneto-
dendrimers, is that they do not use commercially available
agents and are costly to produce with limited availabil-
ity. Thus, the production of commercially available SPIO/
USPIOs (e.g., Feridex or Combidex) created a large opportu-
nity for translation of stem cell labeling to clinical trials. The
SPIOs/USPIOs were mixed with commonly used TAs (e.g.,
SuperFect, poly-L-lysine, Lipofectin, FuGENE, and so forth)
that result in complexes via electrostatic interactions of the
two agents (106,107). The concentration of both agents
had to be carefully titrated to ensure good cellular uptake
without the formation of precipitates of the complex (108).
Initially, poly-L-lysine (PLL) was the preferred TA, because
it was inexpensive and available through several manufac-
turers. Subsequently, protamine sulfate replaced PLL as the
preferred TA, since it was already FDA approved and could
be used in an off-label manner (109). The TA-(U)SPIO com-
plex was then incubated with the cells in culture media for
24 to 48 hours, resulting in a consistent endosomal iron up-
take of 10 to 20 pg Fe/cell in a wide variety of cell lines with-
out species specificity (Fig. 39.9) (103,110–116). Although
ferumoxide-PLL labeling of human mesenchymal stem cells
(MSCs) does not affect cell proliferation or viability, chon-
drogenic differentiation assays performed in vitro were in-
hibited, whereas osteogenic and adipogenic differentiations
were not impaired (117,118). Interestingly, the inhibition
of chondrogenesis was not mediated by the PLL but by the
ferumoxides, suggesting some interaction of iron with cel-
lular signaling. Arbab et al. (109) have reported, when using
protamine sulfate instead of PLL for ferumoxide labeling,
that collagen X was still formed. These studies emphasize the
need for more detailed examination of the possible biologic
effects of labeling schemes in stem cells before initiating clini-
cal trials. Nonetheless, protamine sulfate has the advantage
over PLL of being an approved drug by the U.S. Food and Figure 39.9. Magnetic labeling of mesenchymal stem cells (MSCs),
Drug Administration (FDA) and thereby has less barriers to using superparamagnetic iron oxide particles (Feridex) coated with a
FDA approval of this labeling method. As mentioned earlier, transfection agent (poly-L-lysine, PLL). Cells were labeled for 48 hours
(top, middle) or 24 hours (bottom), with 25 mg Fe/mL Feridex and
the withdrawal of ferumoxides from commercial production
375 ng/mL PLL. Prussian blue staining of labeled human (top), porcine
created a hurdle for clinic translation, but several groups (middle), and canine (bottom) MSCs shows an efficient intracellular
have shown excellent labeling properties using the USPIO uptake of particles into endosomes (blue) that is nonspecific across
ferumoxytol in combination with protamine sulfate (45–47). species. (Adapted from: Bulte JW, Kraitchman DL. Monitoring cell
Another approach for cellular labeling has been developed therapy using iron oxide MR contrast agents. Curr Pharm Biotechnol.
using large styrene/divinyl benzene-coated magnetic micro- 2004;5:567–584, with permission.)
spheres (Bang Laboratories) (31,119). However, the biocom-
patibility of these large spheres, because the degradation
pathway is unknown, will probably limit their clinical applica- to demonstrate in vivo 19F MR spectroscopy at 11.7 T after
bility. Nonetheless, these iron oxide particles are so large that systemic injection and in situ 19F-MRI at 1.5 T after intramus-
a single particle can be detected by MRI as a hypointense spot cular injection. In 2011, Barnett et al. (124) performed in vivo
(119–121). MRI of PFCE-labeled human islet cells after transplantation
As with imaging gene delivery and expression, cellular into the renal capsule of rabbits on a clinical 3-T MRI scanner
labeling with perfluorocarbons combined with 19F MRI pro- (Fig. 39.10).
vides a highly sensitive method for detecting a small number Heretofore, cellular labeling with radiopaque contrast
of cells due to the lack of a native fluorine signal. Ruiz-Cabello agents was too cytotoxic to enable cell tracking using x-ray
et al. (122) labeled neural stem cells with perfluoro-15-crown- imaging techniques. Barnett et al. (125) modified a tech-
5-ether (PFCE) nanoparticles and imaged them for 2 weeks nique used to encapsulate islet cells with alginate to pre-
after injection in mouse brains using a high field MRI system. vent immunorejection to incorporate PFCEs and PFOB in
Partlow et al. (123) labeled human umbilical cord blood cells the capsules and demonstrated the ability to detect the cap-
with PFCE and perfluorooctylbromide (PFOB) and were able sules with a clinical multidetector CT due to the radiopacity
LWBK1209-ch39_p648-666.indd 657 17/05/13 5:23 PM

easily attained with approximately 10 pg Fe/cell commonly seen

with USPIO cellular labeling (128). Indeed, imaging of single
cells has been performed using both USPIO (128) and Bang
particles (119–121) at both 1.5 T and higher field strengths.
However, with cell division, the label is diluted and detection
may not be possible beyond several cell divisions, depending
on the initial labeling concentration. Short-term studies of stem
cells injected intramyocardially in animal models of myocardial
infarction (MI) have demonstrated persistence beyond 3 weeks
of injection (31,113) and as long as 8 weeks postinjection
(129), perhaps reflecting the low division rate of these cells.
Another problem with iron oxide labeling is that any cell
or genetic material that is degraded can potentially release the
iron oxide, which may be incorporated by adjacent phagocytic
cells. Thus, hypointense artifacts in images at later time points
after injection may represent cells other than the original ex-
ogenously labeled cells or cells that do not contain the trans-
fected gene. Therefore, strict quantification of cell numbers
based on the change in the volume of the hypointense artifact
will be difficult to determine.
Imaging Stem Cell Delivery

Many of the in vivo studies of cell trafficking and migration
using SPIO-labeled stem cells were originally performed in
neurovascular and oncologic applications (101,130–133).
Cardiovascular applications have recently been explored fol-
Figure 39.10. In vivo 3-T 19F combined with proton MR anatomi- lowing advances in interventional MRI techniques, devices,
cal image of a rabbit transplanted with 10,000 PFCE-labeled islets and labeling methodologies. Because of the limited regenera-
under the renal capsule demonstrates clear visualization of cell trans- tive capacity of the heart, the exogenous administration of stem
plants. The signal corresponds to fourteen 153-mg PFCE. (Adapted cells holds promise as a method to repair or regenerate cardiac
from: Barnett BP, Ruiz-Cabello J, Hota P, et al. Fluorocapsules tissue after an ischemic event. Hematopoietic bone marrow-
for improved function, immunoprotection, and visualization of derived stem cells, endothelial progenitor cells, multipotent
cellular therapeutics with MR, US, and CT imaging. Radiology. adult progenitor cells (MAPCs), MSCs, cardiac-derived stem
2011;258:182–191 with permission.)
cells, skeletal myoblasts, and inducible pluripotent stem cells
are all thought to be potential candidates to provide stromal
imparted by the bromine in PFOB. Using a gadolinium–gold support, angiogenesis, and/or new cardiomyocytes. Many stem
nanoparticle incorporated in the microcapsules, Arifin et al. cell clinical trials for ischemic heart disease (134–144) have been
(126) were able to track intraperitoneal transplantation of performed or are ongoing. Overall, modest improvements in left
these microcapsules in mice using small animal MRI and CT ventricular function by MRI after stem cell therapy have been
systems. Since the gadolinium and gold were now confined demonstrated in several trials (135–138). But it is difficult to
to the microcapsule rather than intracellularly, toxicity by determine whether further improvements in left ventricular func-
either agent was markedly reduced. Recently, Kedziorek et al. tion could be gained by administering different doses or dosing
(127) have used a barium sulfate-impregnated microcapsule intervals of the stem cells because of the inability to determine
containing allogeneic bone marrow-derived stem cells that whether the stem cells persist in the myocardium and, if so, the
could be detected on digital radiographs and cone-beam extent of engraftment. Thus, MRI or x-ray methods to monitor
C-arm CT using clinical angiographic systems (Fig. 39.11); delivery and track the persistence of engraftment using labeled
in this study, microencapsulation of the stem cells prevented stem cells appear poised to answer some of these questions.
stem cell destruction and enhanced angiogenesis in a rabbit Numerous MR studies of magnetically labeled MSCs
hind limb ischemic model. These x-ray–visible microcapsule have been performed for cardiac applications in the last few
formulations are very attractive since well-accepted x-ray years. Lederman et al. (145) were the first to demonstrate
angiography systems can be used for stem cell delivery and the use of a real-time imaging platform and prototype de-
tracking using existing delivery devices. vice proof of principle using transendocardial injections of
contrast agent under MR fluoroscopy on a clinical 1.5-T
scanner. This study as well as subsequent studies with stem
Cellular Detection Limits cells used fast gradient-echo and steady-state free precession
techniques to monitor real-time delivery.
Estimates of the required amount of iron oxide per cell, to Subsequently, Kraitchman et al. (113) demonstrated the
detect a single cell at the spatial resolution of 100 × 100 × 200 ability to track the migration of SPIO-PLL–labeled allogeneic
mm, range from 1.4 to 3 pg Fe/cell. Fortunately, this range is MSCs as hypointense artifacts on T2*-weighted images in a
LWBK1209-ch39_p648-666.indd 658 17/05/13 5:23 PM

A C E
D F
B
Figure 39.11. In vivo appearance of x-ray–visible barium sulfate microcapsules (Xcaps). A: Planned injec-
tion sites on the left thigh (arrow) of a rabbit subjected to superficial femoral artery (SFA) occlusion are marked
in black. B: Unencapsulated mesenchymal stem cells (i.e., naked MSCs) injected intramuscularly in the medial
thigh of the rabbit cannot be appreciated on fluoroscopic images. An arrow points to platinum thrombogenic
coils in SFA. C–F: Xcaps with (C,E) and without MSCs (D,F) injected intramuscularly in a rabbit peripheral
arterial disease model are visible as distinct radiopacities on anterior–posterior fluoroscopic images and appear
similar whether the capsules contain cells or not. In addition, the radiopacity is relatively unchanged at 14 days
post-transplantation (E,F). (Adapted from: Kedziorek DA, Hofmann LV, Fu Y, et al. X-ray-visible microcapsules
containing mesenchymal stem cells improve hind limb perfusion in a rabbit model of peripheral arterial disease.
Stem Cells. 2012;30:1286–1296.)
clinically relevant swine model of acute MI. Follow-up MRI cells could be detected. In a dosing study in a canine model
at several weeks after MSC showed changes in signal inten- of MI, Bulte and Kraitchman (129) were able to demon-
sity and volume of the hypoenhancing artifact indicative of strate that 105 autologous SPIO-PLL–labeled MSCs injected
migration of the MSC within the MI (Fig. 39.12). Delivery under MR fluoroscopy were still visible by MRI at 2 months
of the magnetically labeled MSCs was performed under x-ray postinjection as well as precise targeting to the peri-infarction
fluoroscopy with only about 70% of the injections visualized region (Fig. 39.13). Thus, magnetically labeled stem cells
by MRI. Thus, the potential for enhanced injection success can be successfully delivered noninvasively using MR fluo-
rate could be envisioned using MR delivery of magnetically roscopy with noninvasive serial follow-up of the persistence,
labeled stem. In an effort to more precisely target cells to the and hence engraftment of the MSCs.
infarcted tissue, Garot et al. (114) used electromechanical The advantage of all of these large animal MI models, such
mapping of iron oxide labeled myogenic progenitor cells in an as pigs and dogs, is that the devices used for endocardial injec-
acute swine MI. Within 2 hours after injection, MRI was used tion of stem cells are similar to those that would be used in
to determine the location of the magnetically labeled cells. human trials. In addition, the large animal closed-chest mod-
Several clinical trials have used electromechanical mapping els of MI are minimally invasive and, thus, more comparable
to guide unlabeled stem cell injections in patients (146–148). to human MI pathology than open-chest rodent procedures.
Subsequently, Dick et al. (30), using the Bang parti- Moreover, the imaging protocols developed are typically per-
cle labeled MSCs in a swine MI model, were able to pre- formed on 1.5-T MR clinical scanners and, therefore, demon-
cisely target the magnetically labeled MSCs to the infarct strate the feasibility for future human trials. Furthermore, it
with a 100% success rate using MR fluoroscopic deliv- can be expected that many of these studies with magnetically
ery. Furthermore, Hill et al. (31) in this same magnetically labeled stem cells will soon present results of regional changes
labeled MSC swine MI model, showed that as few as 105 in infarction size, perfusion, global, and regional function
LWBK1209-ch39_p648-666.indd 659 17/05/13 5:23 PM

Figure 39.12. Long-axis MRIs

acquired with a high-resolution, breath-
hold ECG-gated fast gradient-echo
pulse (500 × 500 × 5,000 mm3 resolu-
tion) sequence show hypointense lesions
(arrows) caused by Feridex-PLL–labeled
mesenchymal stem cells acquired within
24 hours (top) and 1 week (bottom)
of injection in a pig with a myocardial
infarction. The inset (right) shows a
magnified view demonstrating expansion
of lesion over 1 week and a change in
the shape of the hypointense lesion from
an ovoid shape to a shape with more
irregular borders, suggesting migration of
the MSCs. (Adapted from: Kraitchman
DL, Heldman AW, Atalar E, et al. In vivo
magnetic resonance imaging of mesen-
chymal stem cells in myocardial infarc-
tion. Circulation. 2003;107:2290–2293,
with permission.)
using the inherent advantages of the high spatial resolution ized muscle cells in a rabbit model, the dynamic removal
and noninvasiveness of techniques. In a dog model of chronic of the cells can be followed by MRI (150). Similarly, Bos
MI, Rickers et al. (149) have used MRI in this manner to et al. (151) have also watched the distribution kinetics of
assess improvements in myocardial perfusion reserve and intra-arterial and intravenous injection of SPIO-PLL–labeled
regional myocardial function in animals that received unla- rat MSCs as long as 7 to 11 days postinjection. Thus, stem
beled stem cells, for example, MAPCs, when compared to cell therapies for patients with peripheral artery disease that
control animals that received sham intramyocardial injec- may be more amenable to intravenous delivery of stem cells
tions via a surgical approach. could be monitored dynamically by MRI, but methods to
Finally, the tracking of systemic delivery of magnetically detect small numbers of cells will still be needed. However,
labeled stem cells will be highly dependent on the sensitiv- recently, techniques based on the methods of Seppenwoolde
ity of the imaging technique. Cahill et al. (150) have shown et al. (152) for white marker tracking of passive devices have
repeatable T2 measurements in murine skeletal muscle that been adapted for imaging of the hypointense susceptibility
imply that extremely small numbers of SPIO-labeled cells artifacts caused by iron oxide-labeled stem cells (153–157).
can be detected. Furthermore, this group has shown that Rather than detecting hypointensities that can often be con-
after intra-arterial delivery of SPIO-PLL–labeled immortal- fused with other image artifacts such as from metallic stents
LWBK1209-ch39_p648-666.indd 660 17/05/13 5:23 PM

MI
LV
Liver
Figure 39.13. Injection of Feridex-PLL–labeled autologous canine mesenchymal stem cells (MSCs) in a dog
myocardial infarct model. Still-frame long-axis view from real-time MRI (top left) demonstrating Feridex-labeled
MSCs as hypoenhancing artifacts (arrows) after initial two injections at 3 days postinfarction (MI). Top: Lesion is
7 × 106 Feridex-labeled MSCs; bottom: Lesion is 3 × 106 labeled MSC with 4 × 106 unlabeled MSCs. At 8 weeks
after injection, initial two injections (top right, upper arrows) are still visible, as well as additional injections (top
right, lower arrows), with as low as 1 × 105 labeled MSCs at initial injection in fast gradient-echo images (FGRE).
Hypoenhancing artifacts change from round lesions to linear lesions by 8 weeks. Targeting of the labeled MSCs
(arrows, bottom left) to the peri-infarction region using MR fluoroscopic delivery can be guided by delayed contrast-
enhanced short-axis MRI, which shows infarcted myocardium (bottom right, MI). Stem cell placement was validated
on high-resolution FGRE short-axis images (bottom left). (Adapted from: Bulte JW, Kraitchman DL. Monitoring cell
therapy using iron oxide MR contrast agents. Curr Pharm Biotechnol. 2004;5:567–584, with permission.)
LWBK1209-ch39_p648-666.indd 661 17/05/13 5:23 PM

A B C
Figure 39.14. Adjacent double-oblique slices (A–C) from an in vivo 3-T fast spin-echo 3D positive
contrast (Inversion Recovery with On-resonant water suppression or IRON) acquisition obtained in an isch-
emic rabbit hind limb with two injection sites of SPIO-labeled stem cells (250,000, dotted white arrow and
125,000, solid white arrow). Note the excellent background suppression that leads to clear visualization of the
stem cell injection sites with positive contrast and a larger volume of hyperintense signal for the 250,000-cell
injection site. In the region of the dashed white arrow, imperfections in both on-resonant water suppression
and off-resonant fat suppression are observed. This residual signal occurred in an area that was outside of the
second-order shim volume. B0 is the main magnetic field. (Adapted from: Stuber M, Gilson WD, Schar M,
et al. Positive contrast visualization of iron oxide-labeled stem cells using inversion-recovery with ON-resonant
water suppression (IRON). Magn Reson Med. 2007;58:1072–1077.)
and tissue interfaces, stem cells appear hyperintense using (127). However, the large size of the microcapsules (300
these techniques often with distinctive patterns (Fig. 39.14) to 500 microns) makes these therapies unlikely for direct
that lend promise to the ability to image even smaller num- intramyocardial and intracoronary injection. An approach
bers of magnetically labeled stem cells (157,158). that appears to combine the advantages of MRI for viabil-
Recent work suggests that x-ray–visible stem cell mi- ity detection and excellent soft tissue detail and the high
croencapsulation techniques can resolve as few as 1 to 5 temporal resolution for interactivity of x-ray fluoroscopy is
microcapsules containing 50 to 200 stem cells per capsule to fuse MRI ventricular boundary and viability maps with
A B
Figure 39.15. X-ray fused with MRI for targeted myocardial delivery of cellular therapeutics. A: Digital
fluoroscopic image taken using a conventional flat-panel x-ray angiographic system (Axiom dFA, Siemens
AG) demonstrating the lack of ability to see the myocardial borders for transmyocardial delivery of stem cells.
B: Live fluoroscopic image overlaid on x-ray fused with MRI from a cone beam C-arm CT (syngo DynaCT,
Siemens AG) acquired with the same flat-panel angiographic system combined with segmented myocardial
borders (blue and pink) from a whole-heart MRI (Espree, Siemens AG) using vendor software (i-pilot). This
enhanced visualization of vessels and myocardial wall may enable more precise targeting of stem cell therapeu-
tics. (Adapted from: Fu Y, Azene N, Xu Y, et al. Tracking stem cells for cardiovascular applications in vivo:
Focus on imaging techniques. Imaging Med. 2011;3:473–486.)
LWBK1209-ch39_p648-666.indd 662 17/05/13 5:23 PM

real-time fluoroscopic imaging to guide transmyocardial de- 20. Qiu B, Karmarkar P, Brushett C, et al. Development of a 0.014-inch magnetic resonance
imaging guidewire. Magn Reson Med. 2005;53:986–990.
livery (159,160). The ability to perform 3D–3D registration 21. Kocaturk O, Kim AH, Saikus CE, et al. Active two-channel 0.035’’ guidewire for interven-
of cone beam C-arm CTs acquired in the angiographic suite tional cardiovascular MRI. J Magn Reson Imaging. 2009;30:461–465.
22. Elgort DR, Wong EY, Hillenbrand CM, et al. Real-time catheter tracking and adaptive
with MRI (Fig. 39.15) shows tremendous promise for rapid imaging. J Magn Reson Imaging. 2003;18:621–626.
clinical adoption of this technique. Similar to cardiac CT, 23. Omary RA, Green JD, Fang WS, et al. Use of internal coils for independent and direct
methodologies to reduce radiation dose and cardiac gating MR imaging-guided endovascular device tracking. J Vasc Interv Radiol. 2003;14:
247–254.
techniques to freeze motion will need to be developed for 24. Shunk KA, Atalar E, Lima JA. Possibilities of transesophageal MRI for assessment of
C-arm CT for cardiac interventions. aortic disease: A review. Int J Cardiovasc Imaging. 2001;17:179–185.
25. Hofmann LV, Liddell RP, Eng J, et al. Human peripheral arteries: Feasibility of transve-
nous intravascular MR imaging of the arterial wall. Radiology. 2005;235:617–622.
26. Weiss S, Wirtz D, David B, et al. In vivo evaluation and proof of radiofrequency safety of
Summary a novel diagnostic MR-electrophysiology catheter. Magn Reson Med. 2011;65:770–777.
27. Krueger S, Schmitz S, Weiss S, et al. An MR guidewire based on micropultruded fiber-
reinforced material. Magn Reson Med. 2008;60:1190–1196.
Gene and stem cell therapeutics for the heart have been 28. Tzifa A, Krombach GA, Kramer N, et al. Magnetic resonance-guided cardiac interven-
advancing rapidly in the last few years. In concert, MR tions using magnetic resonance-compatible devices: A preclinical study and first-in-man
congenital interventions. Circ Cardiovasc Interv. 2010;3:585–592.
imaging methods and devices have been rapidly evolving to 29. Kos S, Huegli R, Hofmann E, et al. Feasibility of real-time magnetic resonance-guided
keep pace. Although CT offers many benefits in terms of angioplasty and stenting of renal arteries in vitro and in swine, using a new poly-
cost and rapid image acquisition, MR techniques for track- etheretherketone-based magnetic resonance-compatible guidewire. Invest Radiol. 2009;
44:234–241.
ing gene expression and cellular labeling will provide non- 30. Dick AJ, Guttman MA, Raman VK, et al. Magnetic resonance fluoroscopy allows tar-
invasive means to determine the success of these therapies as geted delivery of mesenchymal stem cells to infarct borders in swine. Circulation.
2003;108:2899–2904.
we move into clinical trials. The introduction of flat-panel 31. Hill JM, Dick AJ, Raman VK, et al. Serial cardiac magnetic resonance imaging of injected
angiographic systems that can acquire CT-like images repre- mesenchymal stem cells. Circulation. 2003;108:1009–1014.
sents an enabling technology that can exploit the real-time 32. Karmarkar PV, Kraitchman DL, Izbudak I, et al. MR-trackable intramyocardial injection
catheter. Magn Reson Med. 2004;51:1163–1172.
nature of x-ray fluoroscopy in combination with the high 33. Wacker FK, Elgort D, Hillenbrand CM, et al. The catheter-driven MRI scanner: A new
spatial resolution and soft tissue detail of CT and MRI. approach to intravascular catheter tracking and imaging-parameter adjustment for inter-
ventional MRI. AJR Am J Roentgenol. 2004;183:391–395.
34. Guttman MA, Kellman P, Dick AJ, et al. Real-time accelerated interactive MRI with adap-
tive TSENSE and UNFOLD. Magn Reson Med. 2003;50:315–321.
References 35. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: Sensitivity encoding for fast
MRI. Magn Reson Med. 1999;42:952–962.
1. Meyer BC, Frericks BB, Albrecht T, et al. Contrast-enhanced abdominal angiographic CT 36. Sodickson DK, Manning WJ. Simultaneous acquisition of spatial harmonics (SMASH):
for intra-abdominal tumor embolization: A new tool for vessel and soft tissue visualiza- Fast imaging with radiofrequency coil arrays. Magn Reson Med. 1997;38:591–603.
tion. Cardiovasc Intervent Radiol. 2007;30:743–749. 37. Isner JM, Walsh K, Symes J, et al. Arterial gene therapy for therapeutic angiogenesis in
2. Spuentrup E, Ruebben A, Schaeffter T, et al. Magnetic resonance–guided coronary artery patients with peripheral artery disease. Circulation. 1995;91:2687–2692.
stent placement in a swine model. Circulation. 2002;105:874–879. 38. Grines CL, Watkins MW, Helmer G, et al. Angiogenic Gene Therapy (AGENT) trial in
3. Omary RA, Unal O, Koscielski DS, et al. Real-time MR imaging-guided passive catheter patients with stable angina pectoris. Circulation. 2002;105:1291–1297.
tracking with use of gadolinium-filled catheters. J Vasc Interv Radiol. 2000;11:1079–1085. 39. del Monte F, Harding SE, Dec GW, et al. Targeting phospholamban by gene transfer in
4. Strother CM, Unal O, Frayne R, et al. Endovascular treatment of experimental canine human heart failure. Circulation. 2002;105:904–907.
aneurysms: Feasibility with MR imaging guidance. Radiology. 2000;215:516–519. 40. Moolten FL. Tumor chemosensitivity conferred by inserted herpes thymidine kinase genes:
5. Bakker CJ, Bos C, Weinmann HJ. Passive tracking of catheters and guidewires by contrast- Paradigm for a prospective cancer control strategy. Cancer Res. 1986;46:5276–5281.
enhanced MR fluoroscopy. Magn Reson Med. 2001;45:17–23. 41. Jessup M, Greenberg B, Mancini D, et al. Calcium Upregulation by Percutaneous
6. Buecker A, Spuentrup E, Schmitz-Rode T, et al. Use of a nonmetallic guide wire for mag- Administration of Gene Therapy in Cardiac Disease (CUPID): A phase 2 trial of intra-
netic resonance-guided coronary artery catheterization. Invest Radiol. 2004;39:656–660. coronary gene therapy of sarcoplasmic reticulum Ca2+ATPase in patients with advanced
7. Wacker FK, Reither K, Ebert W, et al. MR image-guided endovascular procedures with heart failure. Circulation. 2011;124:304–313.
the ultrasmall superparamagnetic iron oxide SH U 555 C as an intravascular contrast 42. Runge VM, Ai T, Hao D, et al. The developmental history of the gadolinium chelates as
agent: Study in pigs. Radiology. 2003;226:459–464. intravenous contrast media for magnetic resonance. Invest Radiol. 2011;46:807–816.
8. Bakker CJ, Smits HF, Bos C, et al. MR-guided balloon angioplasty: In vitro demonstration 43. Mendonca Dias MH, Lauterbur PC. Ferromagnetic particles as contrast agents for mag-
of the potential of MRI for guiding, monitoring, and evaluating endovascular interven- netic resonance imaging of liver and spleen. Magn Reson Med. 1986;3:328–330.
tions. J Magn Reson Imaging. 1998;8:245–250. 44. Renshaw PF, Owen CS, McLaughlin AC, et al. Ferromagnetic contrast agents: A new
9. Saeed M, Lee R, Martin A, et al. Transendocardial delivery of extracellular myocardial approach. Magn Reson Med. 1986;3:217–225.
markers by using combination X-ray/MR fluoroscopic guidance: Feasibility study in dogs. 45. Thu MS, Bryant LH, Coppola T, et al. Self-assembling nanocomplexes by combining feru-
Radiology. 2004;231:689–696. moxytol, heparin and protamine for cell tracking by magnetic resonance imaging. Nat
10. Mekle R, Hofmann E, Scheffler K, et al. A polymer-based MR-compatible guidewire: A Med. 2012;18:463–467.
study to explore new prospects for interventional peripheral magnetic resonance angiog- 46. Chen S, Alcantara D, Josephson L. A magnetofluorescent nanoparticle for ex-vivo cell
raphy (ipMRA). J Magn Reson Imaging. 2006;23:145–155. labeling by covalently linking the drugs protamine and Feraheme. J Nanosci Nanotechnol.
11. Glowinski A, Adam G, Bucker A, et al. Catheter visualization using locally induced, 2011;11:3058–3064.
actively controlled field inhomogeneities. Magn Reson Med. 1997;38:253–258. 47. Castaneda RT, Khurana A, Khan R, et al. Labeling stem cells with ferumoxytol, an FDA-
12. Adam G, Glowinski A, Neuerburg J, et al. Visualization of MR-compatible catheters approved iron oxide nanoparticle. J Vis Exp. 2011:e3482.
by electrically induced local field inhomogeneities: Evaluation in vivo. J Magn Reson 48. Du X, Qiu B, Zhan X, et al. Radiofrequency-enhanced vascular gene transduction and
Imaging. 1998;8:209–213. expression for intravascular MR imaging-guided therapy: Feasibility study in pigs.
13. Dumoulin CL, Souza SP, Darrow RD. Real-time position monitoring of invasive devices Radiology. 2005;236:939–944.
using magnetic resonance. Magn Reson Med. 1993;29:411–415. 49. Barbash IM, Leor J, Feinberg MS, et al. Interventional magnetic resonance imaging for
14. Leung DA, Debatin JF, Wildermuth S, et al. Intravascular MR tracking catheter: guiding gene and cell transfer in the heart. Heart. 2004;90:87–91.
Preliminary experimental evaluation. AJR Am J Roentgenol. 1995;164:1265–1270. 50. Barbash IM, Cecchini S, Faranesh AZ, et al. MRI roadmap-guided transendocardial deliv-
15. Worthley SG, Helft G, Fuster V, et al. A novel nonobstructive intravascular MRI coil: In ery of exon-skipping recombinant adeno-associated virus restores dystrophin expression
vivo imaging of experimental atherosclerosis. Arterioscler Thromb Vasc Biol. 2003;23: in a canine model of Duchenne muscular dystrophy. Gene Ther. 2013;20:274–282.
346–350. 51. Faranesh AZ, Nastley MT, Perez de la Cruz C, et al. In vitro release of vascular endo-
16. Hillenbrand CM, Elgort DR, Wong EY, et al. Active device tracking and high-resolution thelial growth factor from gadolinium-doped biodegradable microspheres. Magn Reson
intravascular MRI using a novel catheter-based, opposed-solenoid phased array coil. Med. 2004;51:1265–1271.
Magn Reson Med. 2004;51:668–675. 52. Wang Z, Liu G, Sun J, et al. N-alkyl-polyethylenimine stabilized iron oxide nanoparticles
17. Ladd ME, Zimmermann GG, Quick HH, et al. Active MR visualization of a vascular as MRI visible transfection agents. J Nanosci Nanotechnol. 2012;12:879–886.
guidewire in vivo. J Magn Reson Imaging. 1998;8:220–225. 53. Mykhaylyk O, Sobisch T, Almstatter I, et al. Silica-iron oxide magnetic nanoparticles
18. Ladd ME, Zimmermann GG, McKinnon GC, et al. Visualization of vascular guidewires modified for gene delivery: A search for optimum and quantitative criteria. Pharm Res.
using MR tracking. J Magn Reson Imaging. 1998;8:251–253. 2012;29:1344–1365.
19. Ocali O, Atalar E. Intravascular magnetic resonance imaging using a loopless catheter 54. Wang C, Ravi S, Martinez GV, et al. Dual-purpose magnetic micelles for MRI and gene
antenna. Magn Reson Med. 1997;37:112–118. delivery. J Control Release. 2012;163:82–92.
LWBK1209-ch39_p648-666.indd 663 17/05/13 5:23 PM

55. Lee SJ, Lee HJ, Moon MJ, et al. Superparamagnetic iron oxide nanoparticles-loaded 88. Schmitz SA, Coupland SE, Gust R, et al. Superparamagnetic iron oxide-enhanced MRI
polymersome-mediated gene delivery guided by enhanced magnetic resonance signal. of atherosclerotic plaques in Watanabe hereditable hyperlipidemic rabbits. Invest Radiol.
J Nanosci Nanotechnol. 2011;11:7057–7060. 2000;35:460–471.
56. Lawrie A, Brisken AF, Francis SE, et al. Microbubble-enhanced ultrasound for vascular 89. Ruehm SG, Corot C, Vogt P, et al. Magnetic resonance imaging of atherosclerotic plaque
gene delivery. Gene Ther. 2000;7:2023–2027. with ultrasmall superparamagnetic particles of iron oxide in hyperlipidemic rabbits.
57. Unger EC, McCreery TP, Sweitzer RH. Ultrasound enhances gene expression of liposomal Circulation. 2001;103:415–422.
transfection. Invest Radiol. 1997;32:723–727. 90. Litovsky S, Madjid M, Zarrabi A, et al. Superparamagnetic iron oxide-based method
58. Weissleder R, Moore A, Mahmood U, et al. In vivo magnetic resonance imaging of trans- for quantifying recruitment of monocytes to mouse atherosclerotic lesions in vivo:
gene expression. Nat Med. 2000;6:351–355. Enhancement by tissue necrosis factor-alpha, interleukin-1beta, and interferon-gamma.
59. Ichikawa T, Hogemann D, Saeki Y, et al. MRI of transgene expression: Correlation to Circulation. 2003;107:1545–1549.
therapeutic gene expression. Neoplasia. 2002;4:523–530. 91. Schmitz SA, Taupitz M, Wagner S, et al. Magnetic resonance imaging of atherosclerotic
60. Enochs WS, Petherick P, Bogdanova A, et al. Paramagnetic metal scavenging by melanin: plaques using superparamagnetic iron oxide particles. J Magn Reson Imaging. 2001;14:
MR imaging. Radiology. 1997;204:417–423. 355–361.
61. Weissleder R, Cheng HC, Bogdanova A, et al. Magnetically labeled cells can be detected 92. Kooi ME, Cappendijk VC, Cleutjens KB, et al. Accumulation of ultrasmall superparamag-
by MR imaging. J Magn Reson Imaging. 1997;7:258–263. netic particles of iron oxide in human atherosclerotic plaques can be detected by in vivo
62. Perez JM, Josephson L, Weissleder R. Use of magnetic nanoparticles as nanosensors to magnetic resonance imaging. Circulation. 2003;107:2453–2458.
probe for molecular interactions. Chembiochem. 2004;5:261–264. 93. Kanno S, Wu YJ, Lee PC, et al. Macrophage accumulation associated with rat cardiac
63. Perez JM, Josephson L, O’Loughlin T, et al. Magnetic relaxation switches capable of sens- allograft rejection detected by magnetic resonance imaging with ultrasmall superparamag-
ing molecular interactions. Nat Biotechnol. 2002;20:816–820. netic iron oxide particles. Circulation. 2001;104:934–938.
64. Perez JM, Simeone FJ, Saeki Y, et al. Viral-induced self-assembly of magnetic nanoparti- 94. Sanderson CJ, Wilson DV. Methods for coupling protein or polysaccharide to red cells by
cles allows the detection of viral particles in biological media. J Am Chem Soc. 2003;125: periodate oxidation. Immunochemistry. 1971;8:163–168.
10192–10193. 95. Weissleder R, Lee AS, Khaw BA, et al. Antimyosin-labeled monocrystalline iron oxide allows
65. Perez JM, Simeone FJ, Tsourkas A, et al. Peroxidase substrate nanosensors for MR imag- detection of myocardial infarct: MR antibody imaging. Radiology. 1992;182:381–385.
ing. Nano Letters. 2004;4:119–122. 96. Meldrum FC, Heywood BR, Mann S. Magnetoferritin: In vitro synthesis of a novel mag-
66. Bogdanov A Jr, Matuszewski L, Bremer C, et al. Oligomerization of paramagnetic sub- netic protein. Science. 1992;257:522–523.
strates result in signal amplification and can be used for MR imaging of molecular targets. 97. Bulte JW, Douglas T, Mann S, et al. Initial assessment of magnetoferritin biokinetics and
Mol Imaging. 2002;1:16–23. proton relaxation enhancement in rats. Acad Radiol. 1995;2:871–878.
67. Josephson L, Tung CH, Moore A, et al. High-efficiency intracellular magnetic labeling with 98. Kukowska-Latallo JF, Bielinska AU, Johnson J, et al. Efficient transfer of genetic material
novel superparamagnetic-Tat peptide conjugates. Bioconjug Chem. 1999;10:186–191. into mammalian cells using Starburst polyamidoamine dendrimers. Proc Natl Acad Sci
68. Chen JW, Parm W, Weissleder R, et al. Human myeloperoxidase: A potential target for U S A. 1996;93:4897–4902.
molecular MR imaging in atherosclerosis. Magn Reson Med. 2004;52:1021–1028. 99. Tang MX, Redemann CT, Szoka FC Jr. In vitro gene delivery by degraded polyamido-
69. Ronald JA, Chen JW, Chen Y, et al. Enzyme-sensitive magnetic resonance imaging target- amine dendrimers. Bioconjug Chem. 1996;7:703–714.
ing myeloperoxidase identifies active inflammation in experimental rabbit atherosclerotic 100. Delong R, Stephenson K, Loftus T, et al. Characterization of complexes of oligonucle-
plaques. Circulation. 2009;120:592–599. otides with polyamidoamine starburst dendrimers and effects on intracellular delivery.
70. Perez JM, O’Loughin T, Simeone FJ, et al. DNA-based magnetic nanoparticle assembly J Pharm Sci. 1997;86:762–764.
acts as a magnetic relaxation nanoswitch allowing screening of DNA-cleaving agents. J 101. Bulte JW, Douglas T, Witwer B, et al. Magnetodendrimers allow endosomal magnetic
Am Chem Soc. 2002;124:2856–2857. labeling and in vivo tracking of stem cells. Nat Biotechnol. 2001;19:1141–1147.
71. Mizukami S, Matsushita H, Takikawa R, et al. 19F MRI detection of β-galactosidase 102. Hakumaki JM, Savitt JM, Gearhart JD, et al. MRI detection of labeled neural progenitor
activity for imaging of gene expression. Chem Sci. 2011;2:1151–1155. cells in a mouse model of Parkinson’s disease. Dev Brain Res. 2001;132:A43–A44.
72. Yu JX, Kodibagkar VD, Hallac RR, et al. Dual 19F/1H MR gene reporter molecules for 103. Walter GA, Cahill KS, Huard J, et al. Noninvasive monitoring of stem cell transfer for
in vivo detection of β-galactosidase. Bioconjug Chem. 2012;23:596–603. muscle disorders. Magn Reson Med. 2004;51:273–277.
73. Lanza GM, Lorenz CH, Fischer SE, et al. Enhanced detection of thrombi with a novel fibrin- 104. Billotey C, Wilhelm C, Devaud M, et al. Cell internalization of anionic maghemite nanopar-
targeted magnetic resonance imaging agent. Acad Radiol. 1998;5 Suppl 1:S173–S176; ticles: Quantitative effect on magnetic resonance imaging. Magn Reson Med. 2003;49:
discussion S183–S184. 646–654.
74. Yu X, Song SK, Chen J, et al. High-resolution MRI characterization of human thrombus 105. Wilhelm C, Billotey C, Roger J, et al. Intracellular uptake of anionic superparamag-
using a novel fibrin-targeted paramagnetic nanoparticle contrast agent. Magn Reson Med. netic nanoparticles as a function of their surface coating. Biomaterials. 2003;24:1001–
2000;44:867–872. 1011.
75. Winter PM, Morawski AM, Caruthers SD, et al. Molecular imaging of angiogenesis in 106. Frank JA, Zywicke H, Jordan EK, et al. Magnetic intracellular labeling of mammalian
early-stage atherosclerosis with alpha(v)beta3-integrin-targeted nanoparticles. Circulation. cells by combining (FDA-approved) superparamagnetic iron oxide MR contrast agents
2003;108:2270–2274. and commonly used transfection agents. Acad Radiol. 2002;9 Suppl 2:S484–S487.
76. Flacke S, Fischer S, Scott MJ, et al. Novel MRI contrast agent for molecular imaging 107. Frank JA, Miller BR, Arbab AS, et al. Clinically applicable labeling of mammalian
of fibrin: Implications for detecting vulnerable plaques. Circulation. 2001;104:1280– and stem cells by combining superparamagnetic iron oxides and transfection agents.
1285. Radiology. 2003;228:480–487.
77. Lanza GM, Yu X, Winter PM, et al. Targeted antiproliferative drug delivery to vascu- 108. Guidance for Industry: Regulation of Human Cells, Tissues, and Cellular and Tissue-Based
lar smooth muscle cells with a magnetic resonance imaging nanoparticle contrast agent: Products (HCT/Ps) - Small Entity Compliance Guide.
Implications for rational therapy of restenosis. Circulation. 2002;106:2842–2847. 109. Arbab AS, Yocum GT, Kalish H, et al. Efficient magnetic cell labeling with protamine
78. Winter PM, Caruthers SD, Kassner A, et al. Molecular imaging of angiogenesis in nascent sulfate complexed to ferumoxides for cellular MRI. Blood. 2004;104:1217–1223.
Vx-2 rabbit tumors using a novel alpha(nu)beta3-targeted nanoparticle and 1.5 tesla mag- 110. Arbab AS, Yocum GT, Wilson LB, et al. Comparison of transfection agents in forming
netic resonance imaging. Cancer Res. 2003;63:5838–5843. complexes with ferumoxides, cell labeling efficiency, and cellular viability. Mol Imaging.
79. Anderson SA, Rader RK, Westlin WF, et al. Magnetic resonance contrast enhance- 2004;3:24–32.
ment of neovasculature with alpha(v)beta(3)-targeted nanoparticles. Magn Reson Med. 111. Arbab AS, Bashaw LA, Miller BR, et al. Characterization of biophysical and metabolic
2000;44:433–439. properties of cells labeled with superparamagnetic iron oxide nanoparticles and transfec-
80. Stegman LD, Rehemtulla A, Beattie B, et al. Noninvasive quantitation of cytosine deami- tion agent for cellular MR imaging. Radiology. 2003;229:838–846.
nase transgene expression in human tumor xenografts with in vivo magnetic resonance 112. Arbab AS, Bashaw LA, Miller BR, et al. Intracytoplasmic tagging of cells with ferumox-
spectroscopy. Proc Natl Acad Sci U S A. 1999;96:9821–9826. ides and transfection agent for cellular magnetic resonance imaging after cell transplanta-
81. McMahon MT, Gilad AA, DeLiso MA, et al. New “multicolor” polypeptide diamagnetic tion: Methods and techniques. Transplantation. 2003;76:1123–1130.
chemical exchange saturation transfer (DIACEST) contrast agents for MRI. Magn Reson 113. Kraitchman DL, Heldman AW, Atalar E, et al. In vivo magnetic resonance imaging of
Med. 2008;60:803–812. mesenchymal stem cells in myocardial infarction. Circulation. 2003;107:2290–2293.
82. Gilad AA, McMahon MT, Walczak P, et al. Artificial reporter gene providing MRI con- 114. Garot J, Unterseeh T, Teiger E, et al. Magnetic resonance imaging of targeted catheter-
trast based on proton exchange. Nat Biotechnol. 2007;25:217–219. based implantation of myogenic precursor cells into infarcted left ventricular myocar-
83. Modo M, Cash D, Mellodew K, et al. Tracking transplanted stem cell migration using dium. J Am Coll Cardiol. 2003;41:1841–1846.
bifunctional, contrast agent-enhanced, magnetic resonance imaging. Neuroimage. 2002;17: 115. Anderson SA, Glod J, Arbab AS, et al. Noninvasive MR imaging of magnetically labeled
803–811. stem cells to directly identify neovasculature in a glioma model. Blood. 2005;105:420–425.
84. Modo M, Roberts TJ, Sandhu JK, et al. In vivo monitoring of cellular transplants by 116. Arbab AS, Jordan EK, Wilson LB, et al. In vivo trafficking and targeted delivery of mag-
magnetic resonance imaging and positron emission tomography. Expert Opin Biol Ther. netically labeled stem cells. Hum Gene Ther. 2004;15:351–360.
2004;4:145–155. 117. Kostura L, Kraitchman DL, Mackay AM, et al. Feridex labeling of mesenchymal stem
85. Rudelius M, Daldrup-Link HE, Heinzmann U, et al. Highly efficient paramagnetic label- cells inhibits chondrogenesis but not adipogenesis or osteogenesis. NMR Biomed.
ling of embryonic and neuronal stem cells. Eur J Nucl Med Mol Imaging. 2003;30:1038– 2004;17:513–517.
1044. 118. Bulte JW, Kraitchman DL, Mackay AM, et al. Chondrogenic differentiation of mesenchy-
86. Daldrup-Link HE, Rudelius M, Oostendorp RA, et al. Targeting of hematopoietic pro- mal stem cells is inhibited after magnetic labeling with ferumoxides. Blood. 2004;104:
genitor cells with MR contrast agents. Radiology. 2003;228:760–767. 3410–3412.
87. Schmitz SA, Taupitz M, Wagner S, et al. Iron-oxide-enhanced magnetic resonance imag- 119. Hinds KA, Hill JM, Shapiro EM, et al. Highly efficient endosomal labeling of progenitor
ing of atherosclerotic plaques: Postmortem analysis of accuracy, inter-observer agreement, and stem cells with large magnetic particles allows magnetic resonance imaging of single
and pitfalls. Invest Radiol. 2002;37:405–411. cells. Blood. 2003;102:867–872.
LWBK1209-ch39_p648-666.indd 664 17/05/13 5:23 PM

120. Dodd SJ, Williams M, Suhan JP, et al. Detection of single mammalian cells by high-reso- 141. Lunde K, Solheim S, Aakhus S, et al. Intracoronary injection of mononuclear bone mar-
lution magnetic resonance imaging. Biophys J. 1999;76:103–109. row cells in acute myocardial infarction. N Engl J Med. 2006;355:1199–1209.
121. Shapiro EM, Skrtic S, Sharer K, et al. MRI detection of single particles for cellular imag- 142. Hare JM, Traverse JH, Henry TD, et al. A randomized, double-blind, placebo-controlled,
ing. Proc Natl Acad Sci U S A. 2004;101:10901–10906. dose-escalation study of intravenous adult human mesenchymal stem cells (prochymal)
122. Ruiz-Cabello J, Walczak P, Kedziorek DA, et al. In vivo “hot spot” MR imaging of neural after acute myocardial infarction. J Am Coll Cardiol. 2009;54:2277–2286.
stem cells using fluorinated nanoparticles. Magn Reson Med. 2008;60:1506–1511. 143. Bolli R, Chugh AR, D’Amario D, et al. Cardiac stem cells in patients with ischaemic
123. Partlow KC, Chen J, Brant JA, et al. 19F magnetic resonance imaging for stem/progeni- cardiomyopathy (SCIPIO): Initial results of a randomised phase 1 trial. Lancet. 2011;378:
tor cell tracking with multiple unique perfluorocarbon nanobeacons. FASEB J. 2007; 1847–1857.
21:1647–1654. 144. Traverse JH, Henry TD, Ellis SG, et al. Effect of intracoronary delivery of autologous
124. Barnett BP, Ruiz-Cabello J, Hota P, et al. Use of perfluorocarbon nanoparticles for bone marrow mononuclear cells 2 to 3 weeks following acute myocardial infarction on
non-invasive multimodal cell tracking of human pancreatic islets. Contrast Media Mol left ventricular function: The Late TIME randomized trial. JAMA. 2011;306:2110–2119.
Imaging. 2011;6:251–259. 145. Lederman RJ, Guttman MA, Peters DC, et al. Catheter-based endomyocardial injection
125. Barnett BP, Ruiz-Cabello J, Hota P, et al. Fluorocapsules for improved function, immu- with real-time magnetic resonance imaging. Circulation. 2002;105:1282–1284.
noprotection, and visualization of cellular therapeutics with MR, US, and CT imaging. 146. Perin EC, Dohmann HF, Borojevic R, et al. Transendocardial, autologous bone marrow
Radiology. 2011;258:182–191. cell transplantation for severe, chronic ischemic heart failure. Circulation. 2003;107:
126. Arifin DR, Long CM, Gilad AA, et al. Trimodal gadolinium-gold microcapsules contain- 2294–2302.
ing pancreatic islet cells restore normoglycemia in diabetic mice and can be tracked by 147. Perin EC, Dohmann HF, Borojevic R, et al. Improved exercise capacity and ischemia 6
using US, CT, and positive-contrast MR imaging. Radiology. 2011;260:790–798. and 12 months after transendocardial injection of autologous bone marrow mononuclear
127. Kedziorek DA, Hofmann LV, Fu Y, et al. X-ray-visible microcapsules containing mesen- cells for ischemic cardiomyopathy. Circulation. 2004;110:II213–II218.
chymal stem cells improve hind limb perfusion in a rabbit model of peripheral arterial 148. Silva GV, Perin EC, Dohmann HF, et al. Catheter-based transendocardial delivery of
disease. Stem Cells. 2012;30:1286–1296. autologous bone-marrow-derived mononuclear cells in patients listed for heart transplan-
128. Foster-Gareau P, Heyn C, Alejski A, et al. Imaging single mammalian cells with a 1.5 T tation. Tex Heart Inst J. 2004;31:214–219.
clinical MRI scanner. Magn Reson Med. 2003;49:968–971. 149. Rickers C, Gallegos R, Seethamraju RT, et al. Applications of magnetic resonance imaging
129. Bulte JW, Kraitchman DL. Monitoring cell therapy using iron oxide MR contrast agents. for cardiac stem cell therapy. J Interv Cardiol. 2004;17:37–46.
Curr Pharm Biotechnol. 2004;5:567–584. 150. Cahill KS, Gaidosh G, Huard J, et al. Noninvasive monitoring and tracking of muscle
130. Bulte JW, Zhang S, van Gelderen P, et al. Neurotransplantation of magnetically labeled stem cell transplants. Transplantation. 2004;78:1626–1633.
oligodendrocyte progenitors: Magnetic resonance tracking of cell migration and myelina- 151. Bos C, Delmas Y, Desmouliere A, et al. In vivo MR imaging of intravascularly injected
tion. Proc Natl Acad Sci U S A. 1999;96:15256–15261. magnetically labeled mesenchymal stem cells in rat kidney and liver. Radiology. 2004;233:
131. Hoehn M, Kustermann E, Blunk J, et al. Monitoring of implanted stem cell migration in 781–789.
vivo: A highly resolved in vivo magnetic resonance imaging investigation of experimental 152. Seppenwoolde JH, Viergever MA, Bakker CJ. Passive tracking exploiting local signal con-
stroke in rat. Proc Natl Acad Sci U S A. 2002;99:16267–16272. servation: The white marker phenomenon. Magn Reson Med. 2003;50:784–790.
132. Jendelova P, Herynek V, DeCroos J, et al. Imaging the fate of implanted bone marrow 153. Cunningham CH, Arai T, Yang PC, et al. Positive contrast magnetic resonance imaging of
stromal cells labeled with superparamagnetic nanoparticles. Magn Reson Med. 2003;50: cells labeled with magnetic nanoparticles. Magn Reson Med. 2005;53:999–1005.
767–776. 154. Bowen CV, Zhang X, Saab G, et al. Application of the static dephasing regime theory to
133. Yeh TC, Zhang W, Ildstad ST, et al. In vivo dynamic MRI tracking of rat T-cells labeled superparamagnetic iron-oxide loaded cells. Magn Reson Med. 2002;48:52–61.
with superparamagnetic iron-oxide particles. Magn Reson Med. 1995;33:200–208. 155. Mani V, Briley-Saebo KC, Hyafil F, et al. Positive magnetic resonance signal enhancement
134. Strauer BE, Brehm M, Zeus T, et al. Repair of infarcted myocardium by autologous from ferritin using a GRASP (GRE acquisition for superparamagnetic particles) sequence:
intracoronary mononuclear bone marrow cell transplantation in humans. Circulation. Ex vivo and in vivo study. J Cardiovasc Magn Reson. 2006;8:49–50.
2002;106:1913–1918. 156. Mani V, Briley-Saebo KC, Itskovich VV, et al. Gradient echo acquisition for superpara-
135. Assmus B, Schachinger V, Teupe C, et al. Transplantation of Progenitor Cells and magnetic particles with positive contrast (GRASP): Sequence characterization in mem-
Regeneration Enhancement in Acute Myocardial Infarction (TOPCARE-AMI). Circulation. brane and glass superparamagnetic iron oxide phantoms at 1.5T and 3T. Magn. Reson.
2002;106:3009–3017. Med. 2006;55:126–135.
136. Schachinger V, Assmus B, Britten MB, et al. Transplantation of progenitor cells and 157. Stuber M, Gilson WD, Schar M, et al. Positive contrast visualization of iron oxide-labeled
regeneration enhancement in acute myocardial infarction: Final one-year results of the stem cells using inversion-recovery with ON-resonant water suppression (IRON). Magn
TOPCARE-AMI Trial. J Am Coll Cardiol. 2004;44:1690–1699. Reson Med. 2007;58:1072–1077.
137. Wollert KC, Meyer GP, Lotz J, et al. Intracoronary autologous bone-marrow cell trans- 158. Mani V, Adler E, Briley-Saebo KC, et al. Serial in vivo positive contrast MRI of iron
fer after myocardial infarction: The BOOST randomised controlled clinical trial. Lancet. oxide-labeled embryonic stem cell-derived cardiac precursor cells in a mouse model of
2004;364:141–148. myocardial infarction. Magn Reson Med. 2008;60:73–81.
138. Smits PC, van Geuns RJ, Poldermans D, et al. Catheter-based intramyocardial injection 159. de Silva R, Gutierrez LF, Raval AN, et al. X-Ray fused with magnetic resonance imaging
of autologous skeletal myoblasts as a primary treatment of ischemic heart failure: Clinical (XFM) to target endomyocardial injections: Validation in a swine model of myocardial
experience with six-month follow-up. J Am Coll Cardiol. 2003;42:2063–2069. infarction. Circulation. 2006;114:2342–2350.
139. Schachinger V, Aicher A, Dobert N, et al. Pilot trial on determinants of progenitor cell 160. Fu Y, Azene N, Xu Y, et al. Tracking stem cells for cardiovascular applications in vivo:
recruitment to the infarcted human myocardium. Circulation. 2008;118:1425–1432. Focus on imaging techniques. Imaging Med. 2011;3:473–486.
140. Schachinger V, Erbs S, Elsasser A, et al. Intracoronary bone marrow-derived progenitor
cells in acute myocardial infarction. N Engl J Med. 2006;355:1210–1221.
LWBK1209-ch39_p648-666.indd 665 17/05/13 5:23 PM

LWBK1209-ch39_p648-666.indd 666 17/05/13 5:23 PM
Last H1 Head 667
Index
Note: Page number followed by “f” and “t” indicates figure and table respectively.
A clinical considerations, 483 Atherosclerosis, 582, 589

AAV gene therapy, 652 imaging protocol, 484 complications of
Abdominal angina, 571 Aortic flow, automated quantification of, 104 lumen stenosis, 402
Abdominal aortic aneurysms (AAAs), 545, 559, Aorticopulmonary window, 440, 441f rupture and acute events, 402
559f, 561–563, 562f, 563f Aortic coarctation, 493 development of, 366
classification of, 562 Aortic plaque, in vivo MRI studies on, high-risk plaques, 403
endovascular treatment of, 563–568, 410–411, 411f in lower limbs, 575 (see also Peripheral
564f–567f Aortic PWV, 551 arteries)
evaluation of, 562–563 Aortic stenosis (AS), 66, 204 risk factors for, 366
Abdominal arteries, CTA of, 558 congenital, 448, 451f type of lesions in, 366, 366f, 402 (see also
clinical applications Aortic stent grafting, 602, 604f Atherosclerotic plaques, MRI for
abdominal aorta, 561–568 Aortic valve imaging of)
intra-abdominal arterial bleeding, 571–573 aortic regurgitation, 206–207, 207f, 208f vulnerable plaques, 366, 402–403
mesenteric ischemia, 570–571 aortic stenosis, 204–206, 205f Atherosclerosis Risk in Community Study
renal applications, 568–570 area, 204 (ARIC), 410
vascular liver imaging, 570 implantation, transarterial, 602–605 Atherosclerotic plaques, MRI for imaging of, 403
techniques Aortic valve disease, MRI blood flow imaging applications
contrast administration, 558–559 in, 546 plaque composition, determination of, 408
postprocessing, 560–561 Aquilion 64, scan protocol for, 559 plaque size, measurement of, 408
scanning technique, 559–560 Aquilion 320, vascular protocols for, 560 in vivo monitoring of therapy, 413–414
ABI. See Ankle bracelet index Arch anomalies, 444, 447f in vivo MRI studies on aortic plaque,
ACC/AHA. See American College of Cardiology Arrhythmias and cardiovascular computed 410–411, 411f
and American Heart Association tomography, 52, 53f in vivo MRI studies on carotid artery
ACE. See Angiotensin-converting enzyme Arrhythmogenic right ventricular cardiomyopathy plaque, 408–410, 409f, 410t
Acquired heart disease (ARVC), 142, 142f, 149 in vivo MRI studies on coronary artery
cardiac failure, 85–86, 85f cardiac magnetic resonance techniques plaque, 411–412
coronary disease, 86–87 black blood imaging, 150 in vivo MRI studies on peripheral arteries,
valvular heart disease, 86, 86f bright blood imaging, 150–151 412–413, 412f
Active appearance models (AAM), 103 diastolic dysfunction of right ventricle, 154 and future challenges, 414, 415f
Active catheter-visualization, 608f intramyocardial fat, 152–153 imaging techniques, 403
Active shape models (ASM), 101 myocardial fibrosis, 154 dynamic contrast-enhanced MRI, 405, 406f
Acute hemopericardium, 160 right ventricular outflow tract, 153–154 magnetic resonance angiography
Acute pericarditis, 163, 164f right ventricular volumes, global and approach, 403
etiologies of, 163t regional dysfunction, 151, 151f plaque biomechanics, imaging of, 407–408
Acylcholesterol acyltransferase (ACAT) right ventricular wall hypertrophy, 153 plaque characterization, 404–405, 404f
inhibitors, 413 right ventricular wall thinning, 153 plaque morphometrics, 403–404
AFD. See Anderson–Fabry disease trabecular disarray of right ventricle, 153 targeted molecular contrast agents
Agatston calcium score, 366, 367, 422 characteristics of, 142, 149 imaging, 405–407, 407f
AHA. See American Heart Association clinical background, 149–150 limitations of, 414
American College of Cardiology and American diagnosis of, 149 Atria, 457, 457t, 459
Heart Association (ACC/AHA), 252 features of, 142 Atrial septal defect (ASD), 436–438, 494, 634
American Heart Association (AHA), 108, 109f, triangle of, 149 inferior type of sinus venous, 437f
243 Arrhythmogenic right ventricular large secundum, 437f
on CT imaging of CAC, 373 cardiomyopathy/dysplasia primum ASD, 437f, 442f
Amyloidosis, 143, 143f (ARVC/D), 137 secundum, 436f
Anderson–Fabry disease (AFD), 141, 288 Arrhythmogenic right ventricular dysplasia size, 438f
Angiosarcoma, 187–189 (ARVD), 289, 289f superior type of sinus venosus, 436f
Angiotensin-converting enzyme (ACE), 207 Arterial bleeding, intra-abdominal, 571–573, VEC cine MRI images, 439f
Ankle bracelet index (ABI), 421 572f Atrioventricular canal defect, 440, 442, 442f
Anomalous coronary artery identification, ARVC. See Arrhythmogenic right ventricular Atrioventricular connections, abnormalities of
325–326, 326f, 326t cardiomyopathy atresia of atrioventricular valve, 460–461,
Aorta, 545 ARVC/D. See Arrhythmogenic right ventricular 461f, 462f
blood flow imaging, with MRI, 545–546 (see cardiomyopathy/dysplasia discordant connections, 459–460, 460f
also Thoracic aorta disease, MRI in) ARVD. See Arrhythmogenic right ventricular double-inlet ventricle, 460
Aortic aneurysmal disease, 545 dysplasia Atrioventricular septal defect, 440. See also
Aortic coarctation ASD. See Atrial septal defect Endocardial cushion defects
blood flow imaging in, with MRI, 546–547 ASM. See Active shape models Atrioventricular septum, 440, 442f
postoperative MRI ATHEROMA study, 413 AVA. See Aortic valve area
667
LWBK1209-Ind_p667-680.indd 667 16/05/13 9:15 PM

668 Index
B valves, 59–60 hypertrophic cardiomyopathy, 140–141

Balanced steady-state free precession (b-SSFP), ventricles, 60, 60f metabolism, 138–139
197–198 Cardiac event risk, estimation of, 420–421 morphology and function, 138
Benign primary cardiac tumors, 179–181, 181t Cardiac failure myocarditis, 140, 140f
fibroma, 185–186, 185f cardiac output, 85–86, 85f restrictive cardiomyopathy, 142
hemangioma, 186, 187f ventricular function, 85 Takotsubo cardiomyopathy, 141–142
lipoma and lipomatous hypertrophy, 183, Cardiac function, evaluation of tissue, 138
184f deformation, 68–69, 68f characteristics of, 138
myxoma, 181–182, 181f–183f dimensions, 68 infiltrative
papillary fibroelastoma, 183–184 flow, 69 Aderson–Fabry disease, 288
pheochromocytoma, 186, 186f Cardiac hemangioma, 186 cardiac amyloidosis, 286–287, 286f, 287f
rhabdomyoma, 184–185 Cardiac magnetic resonance (CMR), 108, cardiac sarcoidosis, 287–288, 287f, 288f
Bernoulli equation, 80, 203 125, 149 Danon disease, 288
Black blood coronary MRA, 319–320, 320f advantage of, 150 endomyocardial fibrosis, 288
Black blood imaging, 150. See also Cardiac anatomic features, 112–114, 112f–114f systemic sclerosis, 288
magnetic resonance (CMR) and arrhythmogenic right ventricular infiltrative secondary cardiomyopathies and
Black blood MRI, for atherosclerotic plaques, 404 cardiomyopathy endomyocardial diseases, 142
Black blood turbo spin-echo technique, 8, 8f black blood imaging, 150 amyloidosis, 143, 143f
Blood flow assessment, by MRI, 545–547. See bright blood imaging, 150–151 diabetic cardiomyopathy, 144
also Thoracic aorta disease, MRI in diastolic dysfunction of right ventricle, 154 endomyocardial diseases, 144
Blood flow measurements intramyocardial fat, 152–153 hemochromatosis, 144
background myocardial fibrosis, 154 sarcoidosis, 143, 143f
flow measurements with magnetic right ventricular outflow tract, 153–154, transplant cardiomyopathy, 143–144
resonance imaging, 73–74 154f noncompaction, 144, 145f
physiology, 87–88, 87f right ventricular volumes, global and Cardiovascular computed tomography (CVCT),
techniques, 74f regional dysfunction, 151, 151f 39–41, 40f, 391, 392f, 400f
modulus-based techniques, 74–75 right ventricular wall hypertrophy, 153 applications, 391, 392f
phase-sensitive techniques, 75–81, 75f, right ventricular wall thinning, 153 coronary artery calcification score, 49
76f, 80f trabecular disarray of right ventricle, 153 heart, pediatric computed tomography of,
validation body axes imaging planes, 111–112, 111f, 112f 49, 50f
in vitro, 81, 82f bright-blood and black-blood imaging, 109, myocardial imaging, 49
in vivo, 81–83 109f pulmonary veins, assessment of, 49, 50f
Blood flow velocity, CMR and, 351, 353, cardiac axis imaging planes, 109–111, 110f, ventricular function, assessment of, 47–49,
354f–355f 111f 48f
Blood oxygen level dependent (BOLD), 223 in infants and children artifacts
Blood pool gadolinium-based MR contrast anatomic imaging, 129–131 arrhythmias, 52, 53f
agents, 595–596 anatomy, 125 cardiac motion, 51, 52f
BOLD. See Blood oxygen level dependent cine CMR, 130f, 131, 131f high-contrast, 50, 51f
Bolus tracking, 378 coronary imaging, 128 respiratory motion, 51–52, 52f
Breath-holding techniques, for coronary MRA, delayed enhancement, 132–133 technical errors, 52, 53f
307–308, 307f gadolinium-based techniques, 129, 129f, challenge of, 42
Bright blood coronary MRA, 315–318, 315f–319f 130f for diastolic heart function, 396, 397f
Bright blood imaging, 150–151. See also gadolinium imaging, 131–132, 131f future perspectives, 54–55, 54t, 55f
Cardiac magnetic resonance (CMR) inability to breath-hold, 128–129 myocardial perfusion, evaluation of, 391–396,
Brugada syndrome, 153, 154 limitations of, 126–127 393f, 394f (see also Stress CT MPI)
b-SSFP. See Balanced steady-state free precession physiology/function, 125, 126f, 127f for myocardial scar tissue, 398–400
spatial resolution, 127–128 patient preparation, 47
C technical challenges, 126 phase selection and temporal resolution,
CABG. See Coronary artery bypass grafting temporal resolution, 128 42–45, 43f, 44f
CAC. See Coronary artery calcification velocity mapping, 132, 132f for systolic heart function, 396, 398, 398f
CAC software scoring packages, 367–368, 368f issues in, 108 Cardiovascular disease (CVD), 365
CAD. See Coronary artery disease ultra high-field imaging, 114–115 Cardiovascular magnetic resonance (CVMR), 100,
Calcium homeostasis, 63 Cardiac masses 137, 139–140, 139f, 239, 292, 330
Calcium scores, 366–367, 367–368, 367f, location of, 179, 180f in cardiomyopathies, 138
368f. See also Coronary artery from normal anatomic variants, of coronary artery bypass grafts
calcification (CAC) diferentiation of, 194 anatomy assessment, 350–351, 351t,
Carcinoid heart disease, 210 techniques 352f, 353f
Cardiac amyloidosis, 286–287, 286f, 287f computed tomography, 177 functional assessment, 351–357,
characteristics of, 286 magnetic resonance imaging, 177–179 352f–355f, 356t
Cardiac anatomy, 57–58, 58f tomographic imaging gadolinium-enhanced, 144
contraction and relaxation, 62–63 goals and indications, 176–177 left ventricular systolic function and, 292–293
cardiac cycle, 64, 64f Cardiac motion, 51, 52f. See also Cardiovascular myocardial ischemia by, assessment of,
heart sounds, 64–65 computed tomography 293–296, 294f, 295f, 296t
myocytes, 63–64, 63f compensation, 4–5, 4f–5f myocardial tissue.characterization by
functional anatomy Cardiac receiver coils, 312 microvascular obstruction and myocardial
atria, 59 Cardiac resynchronization therapy (CRT), 293 hemorrhage, 298–299, 300t
circulation, 59 Cardiac sarcoidosis, 287–288, 287f, 288f myocardial edema, 301–302, 302f
collagen matrix, 61 Cardiac tamponade, 163 myocardial necrosis and scar, 296–297,
conduction system, 62 Cardiac transplant rejection arteriolopathy, 289 297f, 298t
defining regions and surfaces, 59 Cardiac tumors, sites of origin of, 181t peri-infarct zone, 299–301, 300f, 301f, 301t
endocardium, 62 Cardiomyopathies, 137–138 to patient with cardiomyopathy
general anatomy, 58–59 cardiovascular magnetic resonance in, 138 arrhythmogenic right ventricular
interventricular dependence, 61–62 arrhythmogenic right ventricular cardiomyopathy, 142, 142f
myofiber structure, 60–61, 61f cardiomyopathy, 142 dilated cardiomyopathy, 139
pericardium, 61 dilated cardiomyopathy, 139 hypertrophic cardiomyopathy, 140–141
LWBK1209-Ind_p667-680.indd 668 16/05/13 9:15 PM

Index 669
metabolism, 138–139 Chylous effusion, 161 in adults, 469–470

morphology and function, 138 Cine imaging, 150 aorta, coarctation of, 83, 84f
myocarditis, 140, 140f Cine MRI, 434, 434f. See also Congenital heart aortic coarctation and
restrictive cardiomyopathy, 142 disease (CHD), MRI in clinical considerations, 483
Takotsubo cardiomyopathy, 141–142 CK-MB. See CK-myocardial band imaging protocol, 484
tissue, 138 CK-myocardial band (CK-MB), 265 atrioventricular connections, abnormalities
right ventricular systolic function and, 293, CMR. See Cardiac magnetic resonance of
293f CNR. See Contrast-to-noise ratio atresia of atrioventricular valve, 460–461,
SPECT myocardial perfusion imaging, 335t Coarctation of aorta, 442, 443f, 444f, 445f, 461f, 462f
techniques 446f discordant connections, 459–460, 460f
anatomy, 7–8, 8f Coils, advantage of, 3 double-inlet ventricle, 460
calculations on K-space, 19, 19f Computational fluid dynamics (CFD), 81, 407 conduits, 85
cardiac motion compensation, 4–5, 4f–5f Computed tomographic angiography (CTA), cyanotic lesions and
coils, 3, 4f 558, 618 Ebstein malformation, 461–462, 464f
coronary MRA, 14–15 of peripheral arteries, 579–581, 582f–588f hypoplastic left heart syndrome, 462–463,
delayed enhancement, 11 plaque imaging by, 423–425 464f
flow, 11–14, 11f–15f Computed tomography (CT), 39, 108 pulmonary atresia, 452–453, 453f, 454f
function, 8–10, 9f, 10f for acquired cardiovascular diseases pulmonary atresia with intact ventricular
image processing, 15–16 assessment, 39 septum, 453–454, 455f
perfusion, 10–11 acquisition parameters for, 421t situs abnormalities, 465, 465f
plan scan, 7 for cardiac masses determination, 176, 177, tetralogy of Fallot, 452, 453f
reduced data acquisition methods, 16–18, 177f, 178f total anomalous pulmonary venous
17f, 18f for congenital cardiovascular diseases connection, 464–465
respiratory motion compensation, 5–7, 6f assessment, 39 truncus arteriosus, 455–457, 459f
survey, 7, 7f limitations of, 39 ventriculoarterial connections,
vessel wall imaging, 15 multislice abnormalities of, 454–455, 454t,
usage of, 139 imaging requirements, 45–46, 45f, 46t 456f, 457f, 458f
Carney’s syndrome, 181 patient dose in, 52–54, 54t Fontan circulation and
Carotid artery plaque, in vivo MRI studies on, technical principles of, 41–42, 41f, 42f clinical considerations, 484–485,
408–410, 409f, 410t for paracardiac masses determination, 176, 484f–486f
Carotid–femoral PWV, 551 177, 177f, 178f imaging protocol, 485
Cartesian scanning, 620 of pericardial diseases future applications of, 485–487
Cartesian trajectory, 548 acute pericarditis, 163, 164f indications for, 434
Catheter(s) cardiac tamponade, 163 postoperative, 290, 290f, 465–467, 465f
future perspectives for, 615–616 constrictive pericarditis, 163–168, total correction of TOF, 467, 467f, 468f
visualization of 166f–171f pulmonary vessels, 83–84
active instrument 605–606, 607f, 623–626 goals and indications for, 156 segmental approach, 457
passive instrument, 601–605, 603f, 622 normal pericardium and pericardial great artery relationships, 459
MR-guided cardiovascular interventions, recesses, 157–158, 157f–160f ventricular loop, 459, 459t
606f pericardial cysts and diverticula, 171–172, visceroatrial situs, 457, 459
MRI and, 566–567, 568 172f, 173f shunts, quantification of, 84–85, 84f
active stent visualization in, 612–614 pericardial effusion, 159–163, 160f–162f techniques, 434, 471
vascular stents in, 612 pericardial hematomas, 172–173 anatomy, assessment of, 434
PEEK-reinforced guidewire, 602, 603f pericardial tumors, 173, 173f, 174f black blood imaging, 471
radiofrequency pericardium, absence of, 169–171, 171f blood flow, quantification of, 435
markers, 607–609 techniques, 156–157 contrast-enhanced MRI, 473–475, 475f
profiling, 606, 607f plaque analysis, 425 delayed enhancement imaging, 474–475,
safety of, 614–615 technique, 421–422 476f
wireless active, 609, 610f, 614f Computed tomography angiography (CTA), flow mapping, 472, 473f, 474f
x-ray fluoroscopy and , 610–611 505 gradient-echo balanced fast-field echo
CE-MRA. See Contrast-enhanced MR Computed Tomography Dose Index technique, 471–472, 472f
angiography (CTDI), 52 magnetic resonance angiography, 473–474,
CFD. See Computational fluid dynamics Conduit obstruction, 495–496 475f, 476f
CFR. See Coronary flow reserve Congenital heart disease (CHD), 125, 433–434, SENSE technology, 472
Chagas disease, 289–290 471 stress testing, 475–476
CHD. See Congenital heart disease acyanotic lesions and T1-mapping, 475
Children, cardiac magnetic resonance in aorticopulmonary window, 440, 441f ventricular function, evaluation of, 434
anatomy, 125 arch anomalies, 444, 447f Tetralogy of Fallot, 85
limitations of, 126–127 atrial septal defect, 436–438, 436f, 437f, tetralogy of Fallot and
physiology/function, 125, 126f, 127f 438f, 439f clinical considerations, 476–479,
protocol coarctation of the aorta, 442–444, 477f–479f
anatomic imaging, 129–131 443f–446f imaging protocol, 479
cine CMR, 130f, 131, 131f congenital aortic stenosis, 448, 451f transposition of the great arteries and
delayed enhancement, 132–133 coronary artery anomalies, 449, 451f, clinical considerations, 479–482,
gadolinium imaging, 131–132, 131f 452f 480f–483f
velocity mapping, 132, 132f endocardial cushion defects, 440, 442, imaging protocol, 482–483
technical aspects 442f valvular obstruction and regurgitation, 85
coronary imaging, 128 left-to-right shunts, 436–440 Congestive cardiomyopathy, 284
gadolinium-based techniques, 129, 129f, partial anomalous pulmonary venous Constrictive pericarditis, 163–168, 166f–171f
130f connection, 438, 440f acute inflammatory, 165
inability to breath-hold, 128–129 pulmonary arterial anomalies, 444, CT and MRI features of, 165t
spatial resolution, 127–128 447f–450f, 448 effusive, 165
temporal resolution, 128 pulmonary valvular stenosis, 448, 448f, etiologies of chronic, 165t
technical challenges, 126 450f functional features of, 169t
Churg–strauss syndrome, 288–289 ventricular septal defect, 438, 441f Contrast administration, in CTA, 558–559
LWBK1209-Ind_p667-680.indd 669 16/05/13 9:15 PM

670 Index
Contrast-enhanced magnetic resonance left ventricular systolic dysfunction in, 292 future perspective, 320
angiography (CE-MRA), 22, 31, multidetector computed tomography and, 305 high field strength coronary MRA, 321
31f, 412 pathophysiology of interventional coronary MRA, 320–321
acquisition timing, 32 coronary artery lesions, 222 stents, 320
dynamic subtraction contrast-enhanced, risk stratification of patients with, 292 (see sequences, 314–315
33–34, 34f also Cardiovascular magnetic technical considerations, 306
signal initiated acquisition, 33 resonance imaging (CMR)) cardiac-specific coils, 312
test bolus, 32–33 x-ray coronary angiography and, 305 contrast enhancement, 310–312, 311f
advantages of, 35, 35f Coronary artery imaging, 133 2D verses 3D data acquisition, 312–313
imaging parameters, 35 clinical applications of, 324–339 magnetic field strength, 314
limitations of, 35–36 Coronary artery lesions, development of, 222 motion artifacts, suppression of, 306–310
phase encoding considerations, 32, 33f Coronary artery plaques, imaging of, 411–412 (see also Extrinsic myocardial
T1 shortening, 32 Coronary artery stenosis motion; Intrinsic myocardial
Contrast enhancement, in coronary MRA, 310 assessment of, imaging modalities in, 370 motion)
endogenous, 310–311, 311f coronary artery calcification and, 369–370 parallel imaging techniques, 313–314,
exogenous, 311–312 zero score, value of, 370 313f
Contrast media (CM), 226, 226f Coronary atherosclerosis, 70 scout scanning and imaging volume
hyperpolarized, 227 development of, 420, 421f definition, 312, 313f
intravascular, 227 Coronary blood flow spatial resolution, 312
T1-enhancing extravascular, 226–227 measurements, 342–344 visualization, 314, 314f
Contrast-to-noise ratio (CNR), 330 in coronary artery bypass graft, 345–347, Coronary magnetic resonance imaging
Coronary arteries, magnetic resonance imaging 346f advantages of, 325
of, 324–325, 325t in coronary sinus, 347–348, 347f for coronary artery bypass graft patency
Coronary artery aneurysms, 326, 327f, 328f of coronary flow reserve, 344–345, 344f assessment, 337t
Coronary artery anomalies, 449, 452f and myocardial perfusion, 70–71, 70f diagnostic accuracy of, 338t
anomalous origin of left coronary artery, Coronary calcium free-breathing 3D k-space segmented
451f, 452f clinical value of, 425 gradient-echo, 327t
ectopic origin of right coronary artery, 452f screening, 422–423, 423f, 424f free-breathing, electrocardiogram-triggered
Coronary artery bypass graft (CABG), 251, 350 Coronary computed tomography angiography 1.5 T “whole-heart” 3D SSFP, 331t
assessment, magnetic resonance imaging for, (CTA), 116, 376 free-breathing targeted 3D k-space gradient-
335–337, 336f–339f clinical application of, 376, 377f, 377t echo navigator, 329t
CMR of, 350–357 arteriovenous coronary artery fistula, for identification of native vessel coronary
anatomy assessment, 350–351, 351t, 378f stenoses, 326–332, 329f, 329t, 330f,
352f, 353f coronary artery disease, 376, 377f 331f, 332f, 333f
angiographically controlled studies, 355 malignant-type coronary artery anomaly, comparative studies versus coronary CTA,
arterial grafts and, 354 378f 332–335, 334f, 334t
bypass graft flow at 3-T MR, 356 coronary artery stenosis by, detection of, 376 limitation of, 337
bypass grafts and, 355–356 coronary in-stent restenosis and, 383, 389f sensitivity, specificity, and accuracy of, 337t
flow velocity, 351–356 dataset for reconstructions, 116 Coronary plaque imaging, 55
other functional assessment, 356–357 future perspective on, 385, 388 Creatine kinase (CK), 253
vein grafts and, 353–354 and image artifacts, 382–383 CTA. See Computed tomographic angiography
CT of, 357–361 issuses related to, 383, 385, 389f CTDI. See Computed Tomography Dose Index
assessment of graft disease, 357–358, nitroglycerin, use of, 377 CTDIvol. See Volume CTDI
358t, 359f–362f performing and interpreting, 377–378, Curved multiplanar reconstruction, 117–120,
patency of bypass grafts, 357 381–383, 381f–388f 119f–122f
preoperative assessment, 358–361 preparation for, 377 application for, 120
multimodality imaging for, 361 reconstruction of images for, 381–382, 383f Curved planar reformation (CPR), 560, 560f
Coronary artery calcification (CAC), 365–366 technical implementation, 376–377, CVMR. See Cardiovascular magnetic resonance
atherosclerotic plaque development and, 379f–381f
366, 366f Coronary disease, coronary flow measurements, D
calcium score reporting and, 367–368, 368f 86–87 Dalcetrapib, 414
and coronary artery stenosis, 369–370 Coronary flow reserve (CFR), 222, 351 dal-PLAQUE multicenter study, 414
elderly patients and, 371 Coronary lesions, detection of, 222 Danon disease, 288
in ER patients with chest pain, 372 Coronary magnetic resonance angiography, DCM. See Dilated cardiomyopathy
imaging of 14–15, 305–306 Delayed contrast-enhanced magnetic resonance
history of, 367 black blood, 319–320 imaging, 104–106, 105f, 106f
importance of, 367 2D and 3D fast spin-echo coronary MRA, Delayed enhancement (DE), 129
multidetector CT scanners for, 367 319–320, 320f imaging, 11, 474–475, 476f (see also
as screening tool prior to angiography, 372 2D spin-echo coronary MRA, 319 Cardiovascular magnetic resonance
summary and recommendations for, bright blood, 315–318 (CVMR))
373–374 contrast-enhanced MRA, 316–318, 319f Delayed enhancement magnetic resonance
as predictor of cardiac risk, 370–371 coronary MRA spin-labeling methods, imaging (DE-MRI), 253
progression of, 371–372 318, 319f assumptions and clinical applications
radiation dosage and, 368–369 2D and 3D spiral coronary MRA, 316, clinical interpretation, 279
and reproducibility issues, 368 317f functional recovery, 278
score, 49 (see also Cardiovascular computed 3D radial coronary MRA, 316 viability, intermediate levels of, 276–278,
tomography) 3D segmented k-space echo-planar 278f
standardization of CT scanners and, 368 coronary MRA, 316, 318f background, 254
statin therapy, and progression of, 372 2D segmented k-space gradient-echo current technique, 254–255, 255t
Coronary artery disease (CAD), 221, 239, 252, coronary MRA, 315 imaging parameters, 256–258, 257f, 257t
292, 305, 324, 365, 420. See also 3D segmented k-space gradient-echo imaging time after contrast administration,
Coronary artery calcification (CAC) coronary MRA, 315, 315f 259–260, 260f
coronary magnetic resonance angiography 3D steady-state free-precession coronary inversion time, 258–259, 258f, 259f
for, 305–306 MRA, 315–316, 316f, 317f pathophysiologic validation
LWBK1209-Ind_p667-680.indd 670 16/05/13 9:15 PM

Index 671
animal studies, 260–265, 261f–266f Endoleak, 565–566, 565f, 567f Fontan circulation
background, 260 Endomyocardial biopsy, 140 postoperative MRI
human studies, 265–271, 266f–270f Endomyocardial diseases, 144 clinical considerations, 484–485,
procedure, 255–256, 256f Endomyocardial fibrosis, 288, 289f 484f–486f
pulse sequence timing, 256, 256f Endotension, 566–567, 568f imaging protocol, 485
technique, 271–276, 272f–277f Endovascular aneurysm repair (EVAR), Fontan procedure, 133
Delayed gadolinium enhancement (DGE), 283 563–568, 564f–567f Four-dimensional (4D) flow, 547, 547f. See also
and nonischemic myocardial disease, Endovascular applications, of interventional MRI Thoracic aorta disease, MRI in
283–284 MR-guided cardiac interventions, 630–632 mapping, 483, 484
DE-MRI. See Delayed enhancement magnetic MR-guided coil embolization, 630 FOV. See Field of view
resonance imaging MR-guided percutaneous dilatation, 628 Fractional flow reserve (FFR), 222
DGE. See Delayed gadolinium enhancement MR-guided placement of vena cava filters, 628 Framingham Heart Study (FHS), 411
Diabetes mellitus type 2 (DM2), 144 MR-guided radiofrequency ablation of heart, Framingham risk
Diabetic cardiomyopathy, 144 628 assessment, 425
characteristics of, 144 MR-guided stent placement, 628 score, 366
Diastolic heart function, CT assessment of, MR-guided tips procedure, 628 Frank–Starling law, 57, 67
396, 397f Endovascular interventional magnetic resonance Free-breathing navigator methods, for 3D
Diffuse myocardial fibrosis, 290 imaging, 618 coronary MRA, 308–310, 308f,
Dilated cardiomyopathy (DCM), 137, 284, 285f safety aspects, 627, 627f 310f
characteristics of, 139 techniques, 618–619 French Aortic Plaque Study (FAPS), 403
DIR-FSE. See Double-inversion recovery fast active visualization, 626–627, 626f FSE. See Fast spin-echo
spin-echo field inhomogeneity concept, passive
DLP. See Dose Length Product visualization with, 623–626, 625f G
Dobutamine stress echocardiography (DSE), 252 hardware considerations, 619 Gadobenate dimeglumine, 332, 596
Dobutamine stress magnetic resonance instrument visualization, 621–622 Gadobutrol 1M, 595
(DSMR), 241 passive visualization, 622–623, 624f Gadofluorine M, 406
diagnostic accuracy of, 245t real-time MR imaging, 619 Gadofosveset sodium, usage of, 331
imaging prognostic value of, 247 sequence design, 619–621, 619f–623f Gadolinium (Gd), 138
viability imaging with, 248–249, 248f Endovascular interventions Gadolinium diethylenetriamine pentaacetic acid
Dobutamine stress testing applications (Gd-DTPA), 178, 601
ergometric stress, 240 MRI-guided cardiac catheterization, Gadolinium-enhanced magnetic resonance
image acquisition, 241, 242f 642–643, 642f angiography, 473, 475f
image interpretation, 243–244, 244f MRI-guided catheter-based intervention, Gadolinium-enhanced 3D MRA, 491
perfusion imaging during, 244–247, 246f 643–645 Gd-DTPA. See Gadolinium diethylenetriamine
pharmacologic stress, 240–241 techniques pentaacetic acid
safety, 241–243 endovascular catheters and guidewires, Gene therapy
Dobutamine termination criteria, 241t MRI tracking of, 639–641 devices, 648–650
Dose Length Product (DLP), 52 imaging techniques, 635–636, 637f, 638f endovascular delivery of, 648
Double aortic arch, 447f interventional MRI unit, 634–635, 635f, cellular detection limits during, 658
Double-inlet ventricle, 460 636f cellular labeling strategies during,
Double inversion methods, defined, 28 medical implants, MRI characteristics of, 655–658
Double inversion recovery (DIR) techniques, 404 636–639 gene delivery, 653
Double-inversion recovery fast spin-echo End-systolic volume (ESV), 47 gene expression, 653–654
(DIR-FSE), 150 EOA. See Efficient orifice area Glenn shunt, 463f
Double-outlet right ventricle (DORV), 455, 458f EPI. See Echo planar imaging Gradient-echo balanced fast-field echo
DSE. See Dobutamine stress echocardiography EPXMA. See Electron probe x-ray technique, 471–472, 472f
DSMR. See Dobutamine stress magnetic microanalysis Gradient-echo images, 507
resonance Ergometric stress, 240. See also Dobutamine Graft stenosis, evaluation of, 351
Duplex Doppler ultrasonography (US), 589 stress testing GRAPPA (GeneRalized Autocalibrating
Dynamic contrast enhancement (DCE) MRI, ESV. See End-systolic volume Partially Parallel Acquisitions), 548
405, 406f Eustachian valve (Ev), 113
Dyskinesia, 151 EVAR. See Endovascular aneurysm repair H
Exercise stress EGG testing, 370 Half Fourier single-shot turbo spin-echo
E Extrinsic myocardial motion, 306 (HASTE), 150
Early gadolinium enhancement ratio (EGEr), suppression of, 307–310 Harmonic phase (HARP), 247
140 breath-holding techniques, 307–308, 307f HARP. See Harmonic phase
EBCT. See Electron beam CT free-breathing navigator methods, HASTE. See Half Fourier single-shot turbo
Ebstein malformation, 461–462, 464f 308–310, 308f, 310f spin-echo
Ebstein’s anomaly, 210, 214f, 496 respiratory belts, 308 HCM. See Hypertrophic cardiomyopathy
ECG. See Electrocardiogram signal averaging, 308 Heart
ECG signal, 506 atria, 59
Echo planar imaging (EPI), 619 F blood circulation, 59
EDV. See End-diastolic volume Fast spin-echo (FSE), 30 collagen matrix, 61
Efficient orifice area (EOA), 86 Ferritin, 656 conduction system of, 62
EGEr. See Early gadolinium enhancement ratio FFR. See Fractional flow reserve defining regions and surfaces, 59
Ejection fraction (EF), 47, 252 Fibroma, 185–186, 185f endocardium, 62
Electrocardiogram (ECG), 4, 42, 253 Fick’s principle, 642 general anatomy of, 58–59
triggering methods, 5 Field of view (FOV), 32, 257 interventricular dependence, 61–62
Electron-beam Computed Tomography (EBCT), Flow compensation and MRA, 22–23. See also myofiber structure, 60–61, 61f
367, 421 Magnetic resonance angiography pediatric computed tomography of, 49, 50f
Electron probe x-ray microanalysis (EPXMA), (MRA) pericardium, 61
264 Flow eccentricity, 554–555, 555f sounds, 64–65
End-diastolic volume (EDV), 47 Flow mapping, 472, 473f, 474f valves, 59–60
Endocardial cushion defects, 440, 442, 442f Flow velocity acquisitions, 490–491 ventricles, 60, 60f
LWBK1209-Ind_p667-680.indd 671 16/05/13 9:15 PM

672 Index
Heart disease, congenital International Task Force, 149 Living-related kidney donors, CTA for
adult, 489–490 Interventional coronary MRA, 320–321 evaluation of, 568–569, 569f
applications, 492–501 Intima-media thickness (IMT), 421 Löffler endocarditis, 144
aortic coarctation, 493–494 Intracardiac thrombus, 192–193, 192f–194f Low-contrast resolution, 46, 47f. See also
atrial septal defects, 494–495 Intracoronary stents, 335, 335f, 336f Multislice computed tomography
hypertension, pulmonary, 495 Intramyocardial fat, 152–153. See also Cardiac (MSCT)
patent ductus arteriosus, 494 magnetic resonance (CMR) Low-dose dobutamine stress magnetic
Ross operation, 492–493 Intravascular ultrasound (IVUS), 425 resonance (LDDSMR), 248
ventricular outflow tract, 492 Intravoxel phase dispersion, 22 LVEF. See Left ventricular ejection fraction
ventricular septal defects, 494–495 Intrinsic myocardial motion, 306 LVOT. See Left ventricular outflow tract
summary, 501 suppression of, 306–307, 306f, 307f Lyme disease, 290
techniques, 490–492 IRA. See Infarct-related artery Lymphoma, 190, 190f
flow velocity acquisitions, 490–491 Ischemic heart disease, 239
multislice scout imaging, 490 pathophysiology of, 239–240 M
right ventricular, 492 stress-induced wall motion abnormalities in, Magnetic resonance angiography (MRA), 6, 618
three-dimensional acquisitions, 244 additional noncontrast-enhanced MRA
491–492 wall motion, quantitative methods for methods
Helical pitch factor, defined, 42 evaluating, 247–248 3D fast spin-echo ECG-gated MRA
Hemangioma, 186, 187f Isolated noncompaction cardiomyopathy, 144 methods, 30–31
Hemochromatosis, 144 IVC. See Inferior vena cava slab nulling preparation and inflow MRA,
characteristics of, 144 IVUS. See Intravascular ultrasound 31, 31f
occurrence of, 144 for atherosclerotic plaques, 403 (see also
Hemorrhagic versus nonhemorrhagic J Atherosclerotic plaques, MRI for
pericardial effusion, 163t Jatene procedure (arterial switch) for TGA, imaging of)
Hepatic hydroxymethylglutaryl coenzyme A 466, 468f contrast-enhanced, 31, 31f
(HMG CoA) reductase inhibitors, 413 acquisition timing, 32
Hibernating myocardium, 292 K advantages of, 35, 35f
High-contrast artifacts, 50, 51f. See also Kaposi sarcoma, 188 dynamic subtraction contrast-enhanced,
Cardiovascular computed Kawasaki disease, 326, 327f, 328f 33–34
tomography k-space segmentation, 548 imaging parameters, 35
HLA. See Horizontal long-axis K-t BLAST technique, 13 limitations of, 35–36
Horizontal long-axis (HLA), 109 defined, 6 (see also Cardiovascular magnetic phase encoding considerations, 32
Hounsfield units (HU), 41 resonance (CVMR)) signal initiated acquisition, 33
range of, 41f and K-t SENSE, 16–18 test bolus, 32–33
Hypertension, pulmonary, 495 k-t broad-use linear acquisition speed-up T1 shortening, 32
Hypertrophic cardiomyopathy (HCM), 137, technique (k-t BLAST), 472 flow compensation, 22–23
140–141, 141f, 284, 285, 285f, 286f flow dynamics, 22
Hypoplastic left heart syndrome, 462–463, 464f L parameter choices, 24–25
LacZ reporter gene, 653 of peripheral arteries, 576–579, 577f–582f
I LAD. See Left anterior descending phase difference methods, 29–30
IGT. See Impaired glucose tolerance Laplace’s law, 66 postprocessing and display
Impaired glucose tolerance (IGT), 144 Larmor equation, 75 maximum intensity projection algorithm,
IMT. See Intima-media thickness Late gadolinium enhancement inversion 36
Infants, cardiac magnetic resonance in recovery gradient recalled echo reformatting, 36–37
anatomy, 125 (LGE-IR-GRE), 150 presaturation, 24, 24f
limitations of, 126–127 Late gadolinium enhancement (LGE), 138, 140, primary methods, 22
physiology/function, 125, 126f, 127f 204, 248 repetition time, 25–26, 25f
protocol LBBB. See Left bundle branch block flip angle, 26–27, 26f
anatomic imaging, 129–131 LCA. See Left main coronary artery techniques, 21–22
cine CMR, 130f, 131, 131f LCX. See Left circumflex coronary artery time-of-flight methods, 23
delayed enhancement, 132–133 LDDSMR. See Low-dose dobutamine stress three-dimensional, 24, 24f
gadolinium imaging, 131–132, 131f magnetic resonance two-dimensional, 23–24, 23f
velocity mapping, 132, 132f Left anterior descending (LAD), 51, 87, 265, 324 TOF considerations and strategies
technical aspects Left atrium (LA), 59 black blood, 28, 29f
coronary imaging, 128 Left bundle branch block (LBBB), 152 disordered blood flow, 27
gadolinium-based techniques, 129, 129f, Left circumflex coronary artery (LCX), 324 flow independent magnetic resonance
130f Left internal mammary (LIMA) graft, 117f angiography, 27
inability to breath-hold, 128–129 multiplanar reconstruction of, 117f radiofrequency excitation pulses, 28–29,
spatial resolution, 127–128 Left main coronary artery (LCA), 117f 30f
temporal resolution, 128 Left ventricle (LV), 108, 138 sequential three-dimensional methods,
technical challenges, 126 Left ventricular ejection fraction (LVEF), 247, 27–28, 28f
Infarct-related artery (IRA), 265 252, 292–293 short echo sequences, 27
Inferior vena cava (IVC), 24, 59 Left ventricular outflow tract (LVOT), 140, 204 spiral scan, 28
Infiltrative secondary cardiomyopathies and Left ventricular partial noncompaction, 284 Magnetic resonance fluoroscopy, 33
endomyocardial diseases, 142 LGE. See Late gadolinium enhancement Magnetic resonance imaging (MRI), 73
amyloidosis, 143, 143f LGE-IR-GRE. See Late gadolinium automated contour detection in short-axis
diabetic cardiomyopathy, 144 enhancement inversion recovery multislice cine, 100–104, 102t, 103f
endomyocardial diseases, 144 gradient recalled echo for cardiac masses determination, 176–179,
hemochromatosis, 144 LIMA graft. See Left internal mammary graft 178f–180f
sarcoidosis, 143, 143f Lipoma, 183, 184f characteristic of, 642
transplant cardiomyopathy, 143–144 Lipomatous hypertrophy, 183, 184f characteristics of medical implants, 636–639
Internal cardiac defibrillator (ICD), 293 Liver transplantation, CT for, 570 atrial and ventricular septal occluders,
Internal mammary arteries (IMAs) grafts, 354 Living-donor liver transplantation, CTA for 639, 640f
International Consensus Group, 140 evaluation of, 570, 571f embolization coils and duct occluders, 638
LWBK1209-Ind_p667-680.indd 672 16/05/13 9:15 PM

Index 673
endovascular stents, 637–638, 639f, 640f noncontrast medium-based approaches, MultiHance (Bracco), 596
mechanical heart valve prostheses, 639 223, 223f Multiplanar reconstruction (MPR), 51,
for coronary artery bypass graft assessment, Magnetic resonance pulse sequences, automated 116–117, 118f, 129, 560, 560f
335–337, 336f–339f contour detection optimization for, applications of, 116–117
delayed contrast-enhanced, 104–106, 105f, 102–103, 102f for assessment of valvular function, 117
106f Magnetic resonance spectroscopy (MRS), 138 for visualization of cardiac anatomy and
endovascular interventional, 618 Magnetic resonance stress tests, function, 117
active visualization, 626–627, 626f contraindications to, 243t Multiple overlapping thin slab acquisition
field inhomogeneity concept, passive Magnetization transfer saturation, 29 (MOTSA), 27
visualization with, 623–626, 625f Magnetoferritin, 656 Multislice computed tomography (MSCT),
hardware considerations, 619 Main pulmonary artery (MPA), 117f, 39, 93
instrument visualization, 621–622 121, 123f automated segmentation of, 101, 101f
passive visualization, 622–623, 624f Major Adverse Cardiac Events (MACE), 371 imaging requirements for
real-time MR imaging, 619 Malignant primary cardiac tumors, 187t, 188f, low-contrast resolution, 46, 47f
safety aspects, 627, 627f 189f required spatial resolution, 45–46
sequence design, 619–621, 619f–623f angiosarcoma, 187–189 required temporal resolution, 46
techniques, 618–627 features of, 187t scan time, 46
flow quantification, 104 fibrosarcoma, 190 patient dose in, 52–54
future perspectives of, 645 leiomyosarcoma, 190 technical principles of, 41–42, 41f, 42f
guided cardiac catheterization, 642–643, 642f liposarcoma, 190 Multislice scout imaging, 490
guided catheter-based intervention, 643–645 lymphoma, 190 Myocardial edema, and CMR imaging,
guided percutaneous balloon angioplasty, 644 osteosarcoma, 190 301–302, 302f
of normal coronary arteries, 324–325, 325t rhabdomyosarcoma, 189 Myocardial fibrosis, detection of, 154
for paracardiac tumor determination, Mass score, calcium, 366–367 Myocardial hibernation, 251
176–179, 178f–180f Maximum intensity projection (MIP), 36, 120, Myocardial imaging, 49. See also
of pericardial diseases 122f, 314, 561, 561f, 581, 618 Cardiovascular computed
acute pericarditis, 163, 164f algorithm, 36 tomography
cardiac tamponade, 163 limitations of, 36 Myocardial infarction (MI), 252, 420
constrictive pericarditis, 163–168, McGoon ratio, 485 Myocardial involvement, chronic infections with
166f–171f MDCT. See Multidetector computed Chagas disease, 289–290
goals and indications for, 156 tomography Lyme disease, 290
normal pericardium and pericardial Medical implants, MRI characteristics of, Myocardial ischemia
recesses, 157–158, 157f–160f 636–639 techniques for evaluation of, 240t
pericardial cysts and diverticula, 171–172, Mesenteric ischemia, 570–571, 572f testing by CMR, 293–296, 294f, 295f,
172f, 173f Metastasis, 191, 191f 296t
pericardial effusion, 159–163, 160f–162f Microvascular obstruction Myocardial mass, 95–96
pericardial hematomas, 172–173 and CMR imaging techniques, 298–299, left ventricular mass, 96–97
pericardial tumors, 173, 173f, 174f 298f, 299f right ventricular mass, 97
pericardium, absence of, 169–171, 171f definition of, 298 Myocardial necrosis and scar imaging,
techniques, 156–157 and intramyocardial hemorrhage, 299, 300t 296–297, 297f, 298t
regional function analysis using, 99–100 MIP. See Maximum intensity projection Myocardial perfusion, 10
tagging, 99–100 Misregistration artifact, 28 Myocardial perfusion imaging (MPI), 293, 328
tracking of endovascular catheters and Mitral regurgitation, 207–209, 210f, 211f Myocardial perfusion scintigraphy (MPS), 370
guidewires, 639–641 Mitral stenosis, 207 Myocardial stress CT perfusion, 376, 385, 388
active tracking, 640f, 641 Mitral valve Myocardial stunning, 251, 292
guidewires, tracking of, 641 mitral regurgitation, 207–209, 210f, 211f Myocardial vasculitides
hybrid tracking, 641 mitral stenosis, 207 churg–strauss syndrome, 288–289
passive tracking, 641, 641f prolapse, 209, 212f Myocardial viability, 251
for valvular heart disease evaluation, regurgitation, 141 background of, 251
196–197 Mixed valvular heart disease, 214 clinical importance of, 251–253
assessment of non, 204 Moment center time, defined, 80 definition of, 253–254, 253f, 254f, 254t
associated findings in, 204 Motion compensation gradients, 22 Myocarditis, 140, 140f
valvular morphology, 197–198, 197f, 198f MOTSA. See Multiple overlapping thin slab causes of, 140
valvular regurgitation, 198–201, acquisition CVMR in, 140
199f–201f MPA. See Main pulmonary artery features of, 140
valvular stenosis, 201–203, 202f–203f MPI. See Myocardial perfusion imaging Myocytes, 63–64, 63f
velocity-encoded, 99–100 MPR. See Multiplanar reconstruction Myxoma, 181–182, 181f–183f
Magnetic resonance (MR), 240, 283 MR contrast gene, 652
advantage of, 632 MRA. See Magnetic resonance angiography N
in cardiovascular diagnosis, 11 MRA technique, 505 National Institutes of Health study (MESA
guided cardiac interventions, 630–632, 631f MRI. See Magnetic resonance imaging study), on CT imaging of CAC, 373
guided coil embolization, 630 MRS. See Magnetic resonance spectroscopy Navigators, defined, 77
guided percutaneous dilatation, 628 MSCT. See Multislice computed tomography Negative predictive value (NPV), 248
guided placement of vena cava filters, 628, Mucocutaneous lymph node syndrome, 326 Nephrogenic systemic fibrosis (NSF), 575
630f Multidetector computed tomography (MDCT), Nitroglycerin, 377
guided radiofrequency ablation of heart, 628 305, 328, 558 NMR. See Nuclear magnetic resonance
guided stent placement, 628, 629f in assessment of graft disease, 357–358, Noninvasive vascular system, 505
guided tips procedure, 628 358t, 359f–362f Nonischemic myocardial disease, 283
perfusion imaging, performance of for coronary artery disease, 305 DGE mechanism in, 283–284
animal studies, 231, 232t coronary calcification and, 365–367 (see also goals and indications for, 283
human studies, 231–235, 233t, 234f–236f Coronary artery calcification (CAC)) Norwood procedure, for hypoplastic left heart
technical aspects of perfusion imaging in preoperative assessment, 358–361 syndrome, 463
magnetic resonance pulse sequences, Multi-Ethnic Study of Atherosclerosis (MESA), NPV. See Negative predictive value
223–226, 224f 368, 410 Nuclear magnetic resonance (NMR), 73
LWBK1209-Ind_p667-680.indd 673 16/05/13 9:15 PM

674 Index
O new developments in, 596 Pulse wave velocity (PWV), 80, 551, 551f
Oscillatory shear index (OSI), 80, 553 patient with diabetes, 587f, 588f Purkinje system, 62
OSI. See Oscillatory shear index technique for, 579–581 PWV. See Pulse wave velocity
Oxidized low-density lipoprotein (oxLDL), 406 visualization technique for, 581, 582f–588f
volume rendering projection and MIP R
P image, 582f, 583f Radiofrequency excitation pulses, 28–29, 30f
Papillary fibroelastoma, 183–184 CTA and MRA, comparison between, Radio-frequency (RF), 22, 114
Paracardiac masses 594–595 Rapid extended coverage (REX) method, 404
location of, 179, 180f evaluation of diseases of, 575–576 Rastelli procedure, 457, 466, 468f
techniques magnetic resonance angiography of RCA. See Right coronary artery
computed tomography, 177 enhanced vs. unenhanced MRA, 576 RCM. See Restrictive cardiomyopathy
magnetic resonance imaging, 177–179 new developments in, 595–596 Real-time interactive cine imaging, 133
tomographic imaging peripheral contrast-enhanced MRA, Receiver–operator curve (ROC), 228
goals and indications, 176–177 576–579, 577f–580f Regional hypokinesia, 151
Parallel imaging techniques, 313–314, 313f pitfalls of, 579, 581f, 582f Region of interest (ROI), 77
Partial anomalous pulmonary venous Peripheral circulation, 69–70, 69f, 70f Renal applications, of CTA, 568–570, 569f, 570f
connection (PAPVC), 438, 440f Peripheral contrast-enhanced MRA, techniques Renal artery stenosis, CTA for evaluation of,
Patent ductus arteriosus, 494 for, 576–579, 577f–580f 569, 570f
PC images, Phase contrast images basic principle, 576–577, 577f Renal tumors, CT imaging for, 570
PC-MRI. See Phase contrast MRI fluoroscopic triggering, 577 Reperfusion injury, 299
PEAR. See Phase encoding artifact reduction image presentation, 579, 580f Repetition time (TR), 23, 25–26, 25f
Pectinate muscles, 59 imaging strategies, 578 flip angle, 26–27, 26f
Pediatric computed tomography, of heart, 49, moving-table technique, 578–579, 579f Resonance circuits (RC), 641
50f multistation/multi-injection technique, 578 Respiratory belts, use of, in coronary MRA, 308
PEEK-reinforced guidewire, 602, 603f test bolus, 577 Respiratory motion, 51–52, 52f. See also
Percutaneous transluminal angioplasties (PTA), time-resolved imaging, 578 Cardiovascular computed
628 timing of image acquisition, 577 tomography
Perfusion data, analysis of, 227, 228f Peroxisomal proliferator-activated receptor- compensation, 5–7, 6f
absolute tissue perfusion, 228–230, 230f gamma (PPAR-gamma), 413 Respiratory-ordered phase encoding (ROPE),
tissue perfusion, parameters linked to, 228 PET. See Positron emission tomography 5, 548
visual assessment, 228 Pharmacologic stress, 240–241. See also Restrictive cardiomyopathy (RCM), 137, 142,
Pericardial cysts and diverticula, 171–172, Dobutamine stress testing 286
172f, 173f Phase contrast MRI (PC-MRI), 75 causes of, 286
Pericardial diseases strategies for, 76–77, 76f, 77f Retrospective ECG gating, defined, 422
CT and MRI of technical extensions of, 77–78 Reynolds number, 79
acute pericarditis, 163, 164f Phase contrast (PC) images, 22 Rhabdomyoma, 184–185
cardiac tamponade, 163 Phase encoding artifact reduction (PEAR), 5 Rhabdomyosarcoma, 189
constrictive pericarditis, 163–168, Phase image, defined, 74 Right atrium (RA), 59
166f–170f Pheochromocytoma, 186, 186f Right coronary artery (RCA), 51, 87, 265, 324
goals and indications for, 156 Physics Task Group of the International Right ventricle (RV), 97, 108, 139
normal pericardium and pericardial Consortium on Standardization in diastolic dysfunction of, 154
recesses, 157–158, 157f–160f Cardiac CT, 367 trabecular disarray of, 153 (see also Cardiac
pericardial cysts and diverticula, 171–172 Plaque, atherosclerotic, in coronary artery, 366, magnetic resonance (CMR))
pericardial effusion, 159–163, 160f–162f 366f Right ventricular outflow tract, enlargement of,
pericardial hematomas, 172–173 Plaque imaging, in symptomatic patients, 426, 153–154, 154f
pericardial tumors, 173 427f Right ventricular wall. See also Cardiac
pericardium, absence of, 169–171, 171f Popliteal aneurysms, 592, 593f magnetic resonance (CMR)
techniques, 156–157 Positive predictive value (PPV), 248 hypertrophy, 153
echocardiography, limitation of, 156 Positron emission tomography (PET), 108, thinning, 153
Pericardial effusions, 159–163, 160f–162f 239, 253 ROC. See Receiver–operator curve
etiologies of, 160t Postinfarction remodeling, 292 ROI. See Region of interest
Pericardial hematomas, 172–173 PPV. See Positive predictive value ROPE. See Respiratory ordered phase encoding
Pericardial tumors, 173, 173f, 174f Pressure mapping, 552–553 Ross operation, 492
Pericardium, absence of, 169–171, 171f Prosthetic valves, 214–215, 215f, 216f Rotation time, defined, 42
Peri-infarct zone, quantitative evaluation of, by Pseudo-normalization, defined, 66 Rotterdam Coronary Calcification Study, 371
CMR, 299–301, 300f, 301f, 301t PTA. See Percutaneous transluminal The Rule Out Myocardial Infarction using
Peripartum-dilated cardiomyopathy, 285 angioplasties Computer-Assisted Tomography
Peripheral arteries, 575 Pulmonary arterial anomalies, 444, 447f–450f, (ROMICAT) trial, 396
clinical applications of imaging techniques 448
for absent pulmonary artery, 448f S
aneurysms, 589, 592 branch pulmonary artery stenoses, 447f, SAR. See Specific absorption rate
atherosclerosis, 582, 589 448f Sarcoidosis, 143, 143f
diabetes, 592–593, 593f pulmonary artery sling, 449f Saturation, defined, 23
peripheral bypass grafts, surveillance of, pulmonary valvular stenosis, 448f Scan projection radiograph (SPR), 47
593–594, 594f, 595f Pulmonary atresia, 452–453 Scan time, defined, 46. See also Multislice
computed tomography angiography of, 579 with intact ventricular septum, 453–454, computed tomography (MSCT)
atherosclerosis after stent placement, 583f 455f Secondary cardiac tumors, 190t
atherosclerotic lesions in distal aorta and with ventricular septal defect, 453f, 454f direct extension from adjacent tumors,
iliac arteries, 584f Pulmonary valve, 209. See also Valvular heart 190–191, 191f
calcifications in distal aorta and iliac disease (VHD) metastasis, 191, 191f
arteries, 586f Pulmonary valvular stenosis, 448, 448f, 450f transvenous extension into the heart, 192,
calcified plaques in aorta, 585f Pulmonary veins (PV), 111 192f
factors determining quality of images in, assessment of, 49, 50f (see also Cardiovascular Seg IR-FGE. See Segmented inversion recovery
579–580 computed tomography) fast gradient-echo
LWBK1209-Ind_p667-680.indd 674 16/05/13 9:15 PM

Index 675
Segmented inversion recovery fast gradient-echo Systemic sclerosis, 288 SDCT and MDCT scan protocols, 509t
(Seg IR-FGE), 256 Systolic and diastolic function spin-echo MRI, 506, 506f
SE imaging. See Spin-echo imaging loading and contractility of strategy of MR study of, 507t
SENC. See Strain-encoding contractility, 67–68, 68f thoracic and abdominal aortae, 511, 512f
SENSE (SENSitivity Encoding) technology, contraction, velocity of, 66–67 intramural hematoma, 516–518
472, 548 intrinsic deformation, 65, 65f CT and, 519
SENSE technique and cardiovascular magnetic loading, 65–66, 66f, 67f MRI and, 518–519, 519f
resonance, 3 Systolic heart function, CT assessment of, 396, Thoracic aorta disease, MRI in, 545
Sequential two-dimensional time-of-flight, 23 398, 398f aortic imaging and, 545–546, 546f
Short tau inversion recovery (STIR), 138 clinical flow imaging applications, 546
Signal averaging, for coronary MRA, 308 T aortic coarctation, 546–547
Signal-to-noise ratio (SNR), 22, 76, 114, 330, Takotsubo cardiomyopathy (TTC), 141–142 aortic valve disease, 546
636 TAVI catheter, 605, 606f emerging, 547, 547f
Simpson’s method, 94, 117 Taylor expansion, 79 multidimensional flow imaging, 547–548, 547f
Single photon emission computed tomography TCPC. See Total cavopulmonary connection data quality, 550–551
(SPECT), 108, 239, 252, 328 TEE. See Transesophageal echocardiography scan time and acceleration, 548–549,
Sinogram space, 42, 42f TEI. See Transmural extent of infarction 548t, 549t
180-degree segmented, 44 Tetralogy of Fallot (TOF), 85, 452, 453f visualization, 549–550, 550f
horizontal axis in, 42 after transannular patch, 469f quantitative hemodynamic biomarkers and,
vertical axis of, 42 postoperative MRI 551
Situs abnormalities, 465 clinical considerations, 476–479, flow eccentricity, 554–555, 555f
left-sided isomerism, 465f 477f–479f pressure mapping, 552–553
polysplenia with interrupted inferior vena imaging protocol, 479 pulse-wave velocity, 551, 551f
cava, 466f with pulmonary regurgitation, 469f turbulence, 551–552, 552f
right-sided isomerism, 465f with repair in past, 478f wall shear stress, 553–554, 553f, 554f
Slice tracking, defined, 626 with right ventricular outflow tract Three-dimensional cine phase-contrast MRI,
SNR. See Signal-to-noise ratio aneurysm, 469f 547, 547f. See also Thoracic aorta
Society of Cardiovascular Magnetic Resonance, TGA. See Transposition of great arteries disease, MRI in
241, 243 Thoracic aneurysms, 545 Three-dimensional coronary MRA, 312–313.
SPAMM. See Spatial modulation of Thoracic aorta See also Coronary magnetic
magnetization acquired diseases resonance angiography (MRA)
Spatial modulation of magnetization (SPAMM), aortic dissection, 510–511, 511f Time-of-flight (TOF), 22
68 CT and, 514–516, 516f–517f methods, 23
Spatial resolution, 45–46, 45f, 46t. See also MRI and, 511–514, 512f three-dimensional, 24, 24f
Multislice computed tomography aortic aneurysms two-dimensional, 23–24, 23f
(MSCT) abdominal, 529 MRA considerations and strategies
Spatial resolution, for coronary MRA, 312, 312f CT and, 526–528 black blood, 28, 29f
Specific absorption rate (SAR), 114 MRI and, 524–526 disordered blood flow, 27
SPECT. See Single photon emission computed thoracic, 524 flow independent magnetic resonance
tomography aortic dissection angiography, 27
Spin-echo (SE) imaging, 150 CT, 523, 526f radiofrequency excitation pulses, 28–29,
Spin phase, defined, 11 MRI, 522, 524f 30f
SPIO contrast gene, 652 postoperative evaluation, 521–522, 523f sequential three-dimensional methods,
Spiral scan method, 28. See also Time-of-flight aortic trauma 27–28, 28f
(TOF) CT and, 531 short echo sequences, 27
SPR. See Scan projection radiograph imaging and, 530, 530f spiral scan, 28
SSFP. See Steady-state free precession MRI and, 530 TIPS. See Transjugular intrahepatic
Statin therapy, and CAC progression, 372 aortic ulcers portosystemic shunts
Steady state free precession (SSFP), 8, 77, 95, CT and, 521, 523f TKE. See Turbulent kinetic energy
109, 138, 179, 241, 324, 636 MRI and, 520–521, 520f T1 mapping, 475
Stents, 320 aortitis importance of, 138
STIR. See Short tau inversion recovery MRI, 535–536 TOF. See Tetralogy of Fallot; Time-of-flight
Strain-encoding (SENC), 247 CT, 536–537 Total anomalous pulmonary venous connection
Stress CT MPI, 392 congenital diseases (TAPVC), 464–465
accuracy of, for detecting ischemia, 395 aneurysms of Valsalva Sinus as, 540 Total cavopulmonary connection (TCPC), 84
advantages of, 392 aortic arch anomalies as TR. See Repetition time
blood flow, calculation of, 394–395 MRI, 538, 538f Training data, defined, 16
dynamic, 394, 394f aortic coarctation Transesophageal echocardiography (TEE), 137,
limitations of, 395–396 MRI, 538–539 634
patient preparation for, 393 postoperative findings, 539–540 Transit time methods, 551
patterns, 394 aortic pseudocoarctation, 540 Transjugular intrahepatic portosystemic shunts
static, 393, 393f MRI, 541 (TIPS), 628
Stress-induced myocardial wall motion imaging, endovascular aortic repair Transmural DGE. See Subendocardial DGE
293 indication, 532–533 Transmural extent of infarction (TEI), 258
Stress magnetic resonance imaging, monitoring preoperative evaluation, 533 Transmural perfusion ratio (TPR), 394
requirements for, 243t postprocedure, 533–534 Transplant cardiomyopathy, 143–144
Stress myocardial perfusion imaging (MPI), 391 imaging techniques Transposition of great arteries (TGA), 454,
CT protocols for, 392 (see also Stress CT contrast injection, 509–510 455, 457f, 458f
MPI) CT techniques, 509–510 after arterial switch repair, 481f, 482f, 483f
MRI for, 391–392 flow mapping, 507 after Mustard repair, 476f, 480f, 481f
SPECT for, 391–392 gadolinium-enhanced MRA, 507–509, postoperative MRI
Stress protocols, 241t 508f clinical considerations, 479–482, 480f–483f
Stroke volume (SV), 47 gradient-echo techniques, 506–507 imaging protocol, 482–483
Subendocardial DGE, 284 MRI techniques, 506 and Senning repair in past, 480f
LWBK1209-Ind_p667-680.indd 675 16/05/13 9:15 PM

676 Index
Transthoracic echocardiography, 137 valvular morphology, 197–198, 197f Ventricular function, assessment of, 47–49, 48f.
Transverse sinus, 157 valvular regurgitation, 198–201, See also Cardiovascular computed
Tricuspid atresia, 461, 462f 199f–201f tomography
Tricuspid valve, 209–214, 213f, 214f. See also valvular stenosis, 201–203, 202f–203f Ventricular measurements, 492
Valvular heart disease (VHD) prosthetic valves, 86, 214–215, 215f, 216f Ventricular outflow tract, 495–496
Triphenyl tetrazolium chloride (TTC), 253 pulmonary valve, 209 Ventricular septal defect (VSD), 438, 441f
TrueFISP, defined, 619 regurgitation, 86, 86f supracristal, 441f
Truncus arteriosus, 455–457, 459f stenosis, 86 Ventricular stroke volume measurements, 104
TTC. See Takotsubo cardiomyopathy; tricuspid valve, 209–214, 213f, 214f Ventricular volumes and global function,
Triphenyl tetrazolium chloride VCG. See Vector ECG quantification of, 97–98, 97f, 98f
Tumor and blood clot, differentiation between, VEC-MRI. See Velocity-encoded cine MRI Ventriculoarterial connections, abnormalities
193–194, 194t Vector ECG (VCG), 5 of, 454–455, 454t, 456f, 457f, 458f
Turbulence, 551–552, 552f Velocity-encoded cine MRI (VEC-MRI), 93 Ventriculoarterial coupling, 71, 71f
Turbulent kinetic energy (TKE), 80 application of, 104 Vertical long-axis (VLA), 109
in congenital heart disease evaluation, 104 Vessel wall imaging, 121, 123f
V in global ventricular function evaluation, 104 VHD. See Valvular heart disease
Valsalva maneuver, 66 in left and right ventricular diastolic function Viral myocarditis, 284, 284f
Valve tracking, 472, 473f evaluation, 104 Vitrea workstation, 560–561
Valvular function Velocity-encoded (VEC) cine MRI, 435, 435f, VLA. See Vertical long-axis
insufficiency, 72 436. See also Congenital heart VNR. See Velocity-to-noise ratio
stenosis, 72 disease (CHD), MRI in Volume CTDI (CTDIvol), 52
Valvular heart disease (VHD), 196 Velocity encoding (VENC), 77, 546 Volume rendering (VR), 121
aortic valve Fourier, 77 application of, 121
aortic regurgitation, 206–207 Velocity map, 76 for detection and quantification of coronary
aortic stenosis, 204–206, 205f defined, 11 stenosis, 121
future directions, 215–217 Velocity time integral (VTI), 204 presentations, 561, 561f, 581
mitral valve Velocity-to-noise ratio (VNR), 76 Volume score, calcium, 366
mitral regurgitation, 207–209, 210f, 211f VENC. See Velocity encoding VTI. See Velocity time integral
mitral stenosis, 207 Ventricles, morphologic features of, 459, 459t
prolapse, 209 Ventricular dimensions and global function, W
mixed, 214 quantification of Wall shear stress (WSS), 74, 553–554, 553f,
MRI strategies to evaluate, 196–197 global function assessment, 95, 96f 554f
assessment of non, 204 multislice short-axis acquisitions, 94 Windkessel effect, 64, 69
associated findings in, 204 slice thickness and slice gap, 94–95, 94f, 95f WSS. See Wall shear stress
LWBK1209-Ind_p667-680.indd 676 16/05/13 9:15 PM

MRI and CT of The Cardiovascular System

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MRI and CT of The Cardiovascular System

Uploaded by

Copyright:

Available Formats

MRI and CT of the

LWBK1209-FM_pi-xvi.indd 1 5/18/13 1:22 PM

LWBK1209-FM_pi-xvi.indd 3 5/18/13 1:22 PM

Two Commerce Square

Library of Congress Cataloging-in-Publication Data

MRI and CT of the cardiovascular system / [edited by] Charles B. Higgins,

LWBK1209-FM_pi-xvi.indd 4 5/18/13 1:22 PM

LWBK1209-FM_pi-xvi.indd 5 5/18/13 1:22 PM

Evan Appelbaum, MD Jens Bremerich, MD, PhD

LWBK1209-FM_pi-xvi.indd 7 5/18/13 1:22 PM

Zahi A. Fayad, PhD J. Wouter Jukema, MD, PhD

Mark A. Fogel, MD, FACC, FAHA, FAAP Han W. Kim, MD

LWBK1209-FM_pi-xvi.indd 8 5/18/13 1:22 PM

Venkatesh Mani, PhD Arno A.W. Roest, MD, PhD

LWBK1209-FM_pi-xvi.indd 9 5/18/13 1:22 PM

William Stanford, MD Martin N. Wasser, MD, PhD

Andrew M. Taylor, MD, FRCP, FRCR Oliver M. Weber, PhD

Rutger W. van der Meer, MD, PhD Michael Zellweger, FESC

Harrie C.M. van den Bosch, MD

LWBK1209-FM_pi-xvi.indd 10 5/18/13 1:22 PM

LWBK1209-FM_pi-xvi.indd 11 5/18/13 1:22 PM

contributing authors vii

PART 1 Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Magnetic Resonance Angiography Techniques . . . . . . . . . . . . . . . . . . . . . . . . 21

3 Techniques for Cardiovascular Computed Tomography . . . . . . . . . . . . . . . . . 39

4 Cardiac Anatomy and Physiolog y: Imaging Aspects . . . . . . . . . . . . . . . . . . . . 57

5 Blood Flow Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

6 Quantification in Cardiac Magnetic Resonance Imaging and

7 CVMR: Imaging Planes and Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

8 Coronary CTA: Display Techniques and Planes . . . . . . . . . . . . . . . . . . . . . . 116

9 Cardiovascular MR in Infants and Children . . . . . . . . . . . . . . . . . . . . . . . . . 125

PART 2 Acquired Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

LWBK1209-FM_pi-xvi.indd 13 5/18/13 1:22 PM

11 Arrhythmogenic Right Ventricular Cardiomyopathy/Dysplasia . . . . . . . . . . 149

12 MRI and CT of Pericardial Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

13 Cardiac and Paracardiac Masses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

14 Valvular Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

PART 3 Ischemic Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

16 Assessment of Ischemic Heart Disease Utilizing Dobutamine . . . . . . . . . . . . 239

17 Assessment of Myocardial Viability by Contrast Enhancement . . . . . . . . . . . 251

18 Delayed Enhancement in Nonischemic Myocardial Disease . . . . . . . . . . . . . . 283

19 Magnetic Resonance Imaging Risk Assessment in Ischemic Heart Disease . . 292

20 Coronary Magnetic Resonance Angiography: Technical Approaches . . . . . . . 305

21 Coronary Artery Imaging—Clinical Applications . . . . . . . . . . . . . . . . . . . . . 324

22 Coronary Blood Flow Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342

23 Cardiovascular Magnetic Resonance and Computed Tomography

24 CT Imaging of Coronary Artery Calcification . . . . . . . . . . . . . . . . . . . . . . . 365

25 Coronary Computed Tomography Angiography . . . . . . . . . . . . . . . . . . . . . 376

LWBK1209-FM_pi-xvi.indd 14 5/18/13 1:22 PM

26 CT Evaluation of Myocardial Perfusion, Function,

27 Atherosclerotic Plaque Imaging: MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402

28 Plaque Characterization—Computed Tomography . . . . . . . . . . . . . . . . . . . . 420

PART 4 Congenital Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . 431

30 Magnetic Resonance Imaging of Function and Flow in Postoperative

31 Adult Congenital Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489

PART 5 Vascular Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503

33 Thoracic Aorta Disease: Flow Evaluation by MR . . . . . . . . . . . . . . . . . . . . . 545

34 Computed Tomographic Angiography of the Abdominal Arteries . . . . . . . . . 558

35 Magnetic Resonance and Computed Tomography

PART 6 Interventional MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599

LWBK1209-FM_pi-xvi.indd 15 5/18/13 1:22 PM

37 Endovascular Interventional Magnetic Resonance Imaging . . . . . . . . . . . . . 618

38 Endovascular Interventions—Congenital Heart Disease . . . . . . . . . . . . . . . . 634

resented showing full ventricular coverage in real time

f requency encoding gradients. The phase encoding direction