Professional Documents
Culture Documents
Quality Control of Nuclear Medicine Instrumentation and Protocol
Quality Control of Nuclear Medicine Instrumentation and Protocol
OF NUCLEAR
MEDICINE
INSTRUMENTATION
AND PROTOCOL
STANDARDISATION
EANM TECHNOLOGISTS GUIDE
29
43
53
67
acquisition and reconstruction
Maximilien Vermandel and Hélène Lahousse
Chapter 7 Optimisation of PET/CT — acquisition and reconstruction* 85
Leesa Ross and Dusty York
Chapter 8 Accreditation for clinical trials: the EANM EARL Project 103
117
131
141
153
Søren Holm
*Articles were written with the kind support of and in cooperation with the
Foreword
In the ever-evolving field of nuclear medicine (NM),
technologists are at the intersection between the clinical,
research and academic domains, embodying the bridge
towards patients. Quality has become an unavoidable word
in NM practice. In the last two decades, NM has earned an
established place in various clinical areas on the basis of the
advances in respect of evidence-based practice and high-level
research. Naturally, nuclear medicine technologists (NMTs) have
been and will be involved in the clinical, research and academic
domains and therefore require the necessary tools to carry out
their tasks in compliance with best practice. This is the main
motivation for choosing Quality Control of Nuclear Medicine
Instrumentation and Protocol Standardisation as the topic for
our annual Technologist’s Guide.
The Technologist’s Guide is an annual pub- work to achieve the same outcome in
lication envisioned and edited by mem- the domain of NM: maintenance of the
bers of the EANM Technologist Committee best practice standards to ensure opti-
(EANM-TC). It is one of the many EANM mal implementation of patient-focussed
educational initiatives and has the aim of diagnostic and therapeutic procedures.
completing the training of NMTs and en- I am extremely grateful to all authors for
couraging scientific exchange within the sharing their expertise, which has been
NM community. The Technologist’s Guide fundamental to the successful completion
was and continues to be a reference for of this Technologist’s Guide. I would like to
educational standards inside and outside thank the EANM Physics Committee, the
Europe. SNMMI-TS (Society of Nuclear Medicine
and Molecular Imaging Technologist Sec-
As the topic of this guide encompasses a tion) and the International Atomic Energy
very broad range of applications, we de- Agency (IAEA) for their help in ensuring
cided to divide the book into three parts. the outstanding quality of this book. I am
The first part focusses on the principles very much indebted to the EANM-TC edi-
of quality and standardisation, an under- torial and language revision group for their
standing of which is needed in order to dedication in reviewing and editing this
completely grasp the more practically guide. Finally, thanks are due to the EANM
oriented parts 2 and 3, which are devoted Board, the EANM Technologist Committee
to imaging procedures and non-imaging and all of those involved in the Technolo-
instrumentation, respectively. gist’s Guide project.
Introduction
Technologists are members of the team required for
implementation of diagnostic imaging in nuclear medicine
(NM). In many hospitals, the technologists are responsible
for the quality assurance (QA) duties. The development of
hybrid imaging has increased further the need for strict
implementation of quality control (QC) and also rendered
QC more demanding. These new guidelines from the EANM
Technologist Committee address the tasks necessary for the
smooth implementation of QC in NM departments.
A SSU R A N C E
PRINC I PL E S
CON T RO L
by Goran Vuleta
CHAPTER 1
Activities Set of activities for ensuring quality in processes Set of activities for ensuring quality in products
Aim To prevent defects To identify (and correct) defects
Process Proactive Reactive
Means of Planned and systematic activities, including Activities or techniques to achieve and maintain
achievement documentation the product quality, process and service
Responsible Everyone involved in the process Specific team that
persons tests the product
Tool QA is a managerial tool QC is a corrective tool
Figure 1
Comparison of results
Performance check (status test) with reference values
QA and QC cycle for a medical imaging device (based on information in [3] and [4])
Figure 2
At the same time, reference tests should
A ND Q UA L I T Y A S S URA NC E
PRINC I PL E S O F Q UA L I T Y CO NTR O L
tics and performance. The recommended »» Availability of space (sufficient space for
A ND Q UA L I T Y A S S URA NC E
PRINC I PL E S O F Q UA L I T Y CO NTR O L
frequency of QC tests depends on the sta- the instrument, for the clinical practice
bility of the equipment. The routine test- and for QC and maintenance procedures)
ing can be performed on a daily, monthly, »» Electrical power supplies (which must
quarterly or annual basis. follow instrument specifications regard-
The siting of an instrument in the de- ing voltage and frequency)
partment is largely determined by its ex- »» Temperature, humidity and air pollution
pected use. The selection of a location for (stable temperature and slow tempera-
mounting the instrument can affect the ture gradients, low humidity and clear air)
performance of the instrument (e.g. crystals »» Background radiation levels (location of
in scintillation detectors), and therefore the the hot cell, the storage and movement
QC. Further parameters that can affect the of radioactive materials, and the move-
proper functioning of the instrument are: ment of patients)
Figure 3
MAINTENANCE
PREVENTIVE CORRECTIVE
MAINTENANCE MAINTENANCE
Types of maintenance
All of the instruments used in nuclear results of QC, there is a written description
A ND Q UA L I T Y A S S URA NC E
PRINC I PL E S O F Q UA L I T Y CO NTR O L
medicine are complex systems built from of all QC procedures, acceptable levels of
mechanical, electrical and electronic parts. tolerance and corrective measures in the
Any of these components can fail at some event that results are not within the level
point in time. For this reason, the mainte- of tolerance.
nance of instruments is necessary. Records must be maintained to provide
The goal of maintenance is to avoid the evidence of conformity to requirements
consequences of equipment failure. Main- and of the effective operation of the qual-
tenance of instruments is divided into two ity management system[6]. All records
categories (Fig. 3): must remain legible, readily identifiable
»» Preventive maintenance: a fundamen- and retrievable. They must be permanent
tal, routine and planned maintenance and non-erasable, as must changes to a
activity to keep equipment in operating record[6]. Record keeping is a main com-
status and to avoid unplanned mainte- ponent of an internal QC programme. Re-
nance activity. cords showing frequent malfunction and
»» Corrective maintenance: a set of activi- degradation of equipment performance
ties to detect and rectify a defect so that provide evidence of the need for com-
the equipment is returned to its normal plete instrument repair or replacement.
state.
Record keeping may include (depending
Preventive maintenance in nuclear med- on the equipment):
icine means the maintenance of equip- »» Instrument condition (physical,
ment in a given functional state through mechanical and electronic)
continued overviews, QC and detection »» All calibration records
and elimination of possible failures. Cor- »» QC results
rective maintenance in nuclear medicine »» Instrument maintenance records
means the restoration of equipment to a
functional status by means of repairs. The record keeping can be used to:
As already noted, all obtained re- »» Monitor compliance with QC procedures
sults, from the installation of equipment »» Educate employees
through to acceptance tests and then »» Help prevent instrument breakdowns
QC procedures, need to be recorded and »» Evaluate service personnel
stored. It is necessary that, apart from the »» Help ensure reliable patient results
Recognition by the head of department If the results of certain tests show a de-
A ND Q UA L I T Y A S S URA NC E
PRINC I PL E S O F Q UA L I T Y CO NTR O L
and the management of the institution viation from the allowed tolerance and ac-
of the need for QC is essential to its sat- ceptability it is necessary to decide wheth-
isfactory implementation and adequate er the instrument is fit to be used or needs
funding. to be put out of operation. This decision
It is necessary to clearly define who is must be the responsibility of people with
responsible for each aspect of QC as well clearly defined responsibilities.
as who will supervise the entire QC plan. The choice of tests and the frequency
That person must know all the technical of their performance must be specified
details and should be involved in the eval- for each device in nuclear medicine in
uation of the results. It is important that order to take account of their condition
tests for certain instruments are carried and their status. Protocols should be
out by people familiar with their use, and adapted to suit individual instruments. It
responsibility for daily and operational is essential that these protocols are strictly
tests should rest with the operators who followed.
regularly use these devices.
REFERENCES
1. International Atomic Energy Agency. Quality control surance for PET and PET/CT systems. IAEA Human
of nuclear medicine instruments 1991. IAEA-TEC- Health Series No. 1. Vienna: IAEA; 2009.
DOC-602. Vienna: IAEA; 1991. 5. Busemann Sokole E, Plachcinska A, Britten A,
2. International Atomic Energy Agency. Quality assur- Georgosopoulou ML, Tindale W, Klett R. Routine
ance for SPECT systems. IAEA Human Health Series quality control recommendations for nuclear med-
No. 6. Vienna: IAEA; 2009. icine instrumentation. Eur J Nucl Med Mol Imaging
3. International Electrotechnical Commission. Evalua- 2010;37:662–671.
tion and routine testing in medical imaging depart- 6. ISO 13485. Medical devices – Quality management
ments - Part 1: General aspects. IEC 61223-1. Geneva: systems – Requirements for regulatory purposes, 3rd
IEC; 1993. ed. Switzerland; 2016.
4. International Atomic Energy Agency. Quality as-
by Maurizio Dondi,
Thomas Pascual
and Diana Paez
CHAPTER 2
INTRODUCTION
INTE RNATI O NA L ATO M I C E NE R GY AGE N C Y
MA NAGE M E NT S YS TE M S: THE A PPR OAC H O F THE
INTE RNA L A ND E X TE RNA L AUD I TI NG O F Q UA L I T Y
This requires the identification of quality quality measurement systems and reward
policies and objectives and the produc- mechanisms.
tion of a documentation system with
clearly defined processes, procedures
and responsibilities. Such a system is usu- NUCLEAR MEDICINE AND
ally referred to as a quality management QUALITY MANAGEMENT
system (QMS) and its purpose is to help Nuclear medicine services (NMS) are mul-
coordinate and direct activities in order tidisciplinary by nature as several different
to meet customer and regulatory require- professional competencies are involved,
ments and to improve effectiveness and each with its own regulations, processes
efficiency on a continuous basis. In health and outputs. However, they all contribute
care, effective quality management is fo- to the success of the discipline. For this rea-
cussed on the needs of patients because son, a comprehensive QMS is not limited
they are the ones who judge the effec- to quality assurance/quality control (QA/
tiveness of treatments and the appropri- QC) but has to involve all aspects of NMS,
ateness of the service. including but not limited to: clinical appli-
Quality management in health care cations, including machinery handling and
requires the close cooperation of people their QA/QC; radiopharmaceutical prepa-
with diverse expertise and is essentially rations and, again, their QA/QC; radiation
about delivering consistent quality, which, protection of both patients and staff and
in turn, is dependent upon reliable pro- of the environment; and the ability of final
cesses. Reliability requires the existence reports to satisfy clinical questions.
of performance goals, risk reduction pro- Several factors may influence the struc-
cedures, quality improvement policies, ture of a QMS in nuclear medicine, includ-
ing the size and structure of the NMS and »» Written standard operating procedures
ensures the availability of necessary re- middle-income countries with a tool that
sources and information to support the would help them identify areas of weak-
operation and for monitoring of process- ness in their practices, raise awareness of
es. The management also ensures the international standards and eventually en-
effectiveness of the QMS through self-as- courage the implementation of an annual
sessments, data analysis, verification of ac- systematic audit process for the nuclear
tivities and management reviews. medicine practice as a whole. Following
various interactions among meeting par-
ticipants, a manual entitled Quality Man-
THE QUANUM PROJECT agement Audits in Nuclear Medicine Practic-
OF THE IAEA es (QUANUM) was published in 2008.
The IAEA (www.iaea.org) is one of the or- For the first time, a unique and holistic
ganisations of the United Nations (UN) programme covering all the disciplines
system and has the mission of supporting involved in the delivery of nuclear medi-
peaceful applications of nuclear technolo- cine was made available to practitioners
gies in the UN member states, which num- worldwide, and in the subsequent years
ber more than 170. Therefore, in addition to the QUANUM tool was successfully ap-
activities in many other fields, it has a long plied across the world. Lessons were
history of providing assistance in the spe- learned from the first IAEA expert mis-
cialty of nuclear medicine. In many of these sions, and throughout this period the spe-
countries, NMS are rare and often isolated cialty of nuclear medicine continued to
and practitioners have serious difficulties develop rapidly. Consequently, the IAEA
in exchanging experiences with their peers recognised that there was a need to up-
and even in accessing scientific journals. date the manual so that it would reflect
On the basis of its awareness of the very current and best practices in NMS. The
different levels of quality of practice and new edition was published in 2015 under
the need to raise them to internationally the title “Quality Management Audits in
recognised standard levels, the Nuclear Nuclear Medicine Practices, Second Edi-
Medicine Section of the IAEA planned tion. IAEA Human Health Series No. 33”. It
the preparation and the implementation is accessible at http://www-pub.iaea.org/
of a project called Quality Management books/IAEABooks/10714/Quality-Manage-
Audits in Nuclear Medicine (QUANUM). ment-Audits-in-Nuclear-Medicine-Practic-
The aim of the project was to provide es-Second-Edition.
Figure 1
INTE RNATI O NA L ATO M I C E NE R GY AGE N C Y
MA NAGE M E NT S YS TE M S: THE A PPR OAC H O F THE
INTE RNA L A ND E X TE RNA L AUD I TI NG O F Q UA L I T Y
Regular
Auditing Circle
Programmed
Internal audit Entrance Assessments
team is formed Briefing
External
audit
team is
Managerial review formed
Radiation safety
Clinical review
QA/QC euipment External
Radiopharmacy Organisations
Routine
YES
Nuclear Standard
Medicine met?
Activities
NO
Preventive/
Corrective Action NO
Need external
assessment? YES
Follow-up
YES Standard met? NO
The internal and external auditing cycle (adapted from: Quality Management Audits in Nuclear
Medicine Practices, Second Edition. IAEA Human Health Series No. 33. Vienna, 2014)
medicine administrative and nursing staff auditors should assure the staff that
Table 1: Use of the scoring system for evaluation of the level of conformance: the example of
documentation of clinical procedures
Figure 2
1. Strategies
INTE RNATI O NA L ATO M I C E NE R GY AGE N C Y
MA NAGE M E NT S YS TE M S: THE A PPR OAC H O F THE
INTE RNA L A ND E X TE RNA L AUD I TI NG O F Q UA L I T Y
100
17. HT Markers 2. Adm&Man
75
25
15. RP Lev 2 4. Radiat Reg
8. IT Syst
plemented and requires that countries in practices assessed and audited by col-
the European Union formally establish, leagues from other countries. Particularly
as a policy, that clinical audits are per- appreciated is the exchange of informa-
formed under published and regulated tion and advice received during the visits.
national guidelines. In almost all cases, the visit from inter-
Following this directive, the QUANUM national experts for an external audit and
programme, in slightly modified form, has the work done in preparing for this audit,
become the reference standard in some including filling of the checklist, has trig-
European countries, such as Belgium. gered the implementation of a quality
system; this is in itself represents an ex-
cellent outcome. Indeed, the QUANUM
CONCLUSIONS programme was conceived exactly for this
The QUANUM programme has been very purpose.
well received by counterparts, who wel- In the majority of the externally audit-
come the opportunity to have their daily ed practices, the audit has shown that
SELECTED BIBLIOGRAPHY
» Applying radiation safety standards in nuclear med- » IAEA Human Health Series No. 1. Quality Assurance
icine. Jointly sponsored by the International Atomic for PET AND PET/CT systems. International Atomic
Energy Agency, International Labour Office, Interna- Energy Agency, Vienna, 2009.
tional Organization for Medical Physics, Pan American » IAEA Human Health Series no. 6. Quality assurance for
Health Organization, World Federation of Nuclear SPECT systems. International Atomic Energy Agency,
Medicine and Biology and World Health Organization. Vienna, 2009.
International Atomic Energy Agency, Vienna, 2005. » IAEA Safety Standards Series No. GSR Part 3. Radiation
» Dondi M, Kashyap R, Pascual T, Paez D, Nunez-Miller protection and safety of radiation sources: Interna-
R. Quality management in nuclear medicine for bet- tional basic safety standards. General safety require-
ter patient care: the IAEA program. Semin Nucl Med ments. Jointly sponsored by: EC; FAO; UN; IAEA; ILO;
2013;43:167–171. OECD; PAHO; UNEP; WHO; IAEA. International Atomic
» D ondi M, Andreo P. Developing nuclear medicine in Energy Agency, Vienna, 2011.
developing countries: IAEA’s possible mission. Eur J » Operational guidance on hospital radiopharmacy: a
Nucl Med Mol Imaging 2006;33:514–515. safe and effective approach. International Atomic En-
» http://www.eanm.org/publications/guidelines/in- ergy Agency, Vienna, 2008.
dex.php?navId=37 » Pascual TN. Examining quality management audits
» https://humanhealth.iaea.org/HHW/NuclearMedicine/ in nuclear medicine practice as a lifelong learning
QUANUM_2.0_Excel_Tool_and_QNUMED/index.html process: opportunities and challenges to the nuclear
» http://nucmedicine.iaea.org/ medicine professional and beyond. Nucl Med Com-
» https://www.iaea.org/services/key-programmes/hu- mun 2016;37:785–791.
man-health-programme » Radiation protection and safety of radiation sources:
» http://www-pub.iaea.org/books/IAEABooks/10714/ international basic safety standards. Interim edition.
Quality-Management-Audits-in-Nuclear-Medi- General safety requirements. International Atomic
cine-Practices-Second-Edition. Energy Agency, Vienna, 2011.
» http://www.snmmi.org/ClinicalPractice/content.as- » Strategies for Clinical Implementation and Quality
px?ItemNumber=6414&navItemNumber=10790 Management of PET Tracers. International Atomic
» h ttp://www.who.int/management/quality/assur- Energy Agency, Vienna, 2009.
ance/QualityCare_B.Def.pdf?ua=1 » www.iaea.org
CON T RO L
Q UAL I T Y
C AME R A
PL AN A R
SYSTE M
by Mario Medvedec
CHAPTER 3
INTRODUCTION
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
Planar and SPECT scintillation cameras quarterly, half-yearly and annually: colli-
should also undergo different evaluation mator hole angulation, tilt-angle check,
steps during their life-cycle: factory tomographic spatial resolution, SPECT/
testing before shipment, acceptance CT alignment;
and reference testing after their on-site »» describe the Jaszczak phantom and un-
installation and before their clinical use, derstand and explain the procedure for
and routine periodic quality control (QC) testing of overall SPECT system perfor-
testing thereafter. mance;
After reading this chapter, readers should »» describe the most common artefacts and
be able to: explain how to proceed in such cases.
»» define acceptance testing on planar and
SPECT scintillation cameras, and under-
stand the reasons for performing such ACCEPTANCE TESTING
testing; Acceptance testing is a set of standard
»» define, understand and explain the procedures intended to verify that the im-
need for and the procedures and peri- aging equipment performs in accordance
odicity of recommended QC tests on with the manufacturer’s specifications and
planar and SPECT scintillation cameras: intended clinical use without any deficien-
daily: visual and physical inspections, cies or defects. The standard procedures
collimator touch pads and emergency and performance measurements usual-
stop buttons, energy window settings, ly employed in acceptance testing are
uniformity and sensitivity; those published by the National Electrical
weekly and monthly: centre of rotation, Manufacturers Association (NEMA), In-
bar phantom spatial resolution and lin- ternational Electrotechnical Commission
earity, high-count flood uniformity/sen- (IEC) or other international authorities. In
sitivity correction map; addition, some patient studies should be
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
procedure. In this way, acceptance and For many years, QC testing was perceived
reference testing provides baseline per- to be the responsibility of the individual,
formance data to be referred to in future to be performed at the individual’s dis-
QC tests, and supports the final user’s de- cretion. In the 1990s, however, aspects of
cision to accept or reject a particular piece QC testing were incorporated within the
of equipment for safe routine clinical use. European legal framework by European
The warranty period for the imaging sys- Council Directives, with subsequent merg-
tem should begin only when the system ing and updating in 2013. Now, as stated
has passed acceptance testing by achiev- above, the requirement for QC is to be fully
ing at least the minimum acceptable re- implemented in European Union Mem-
sults under clinical conditions. It should be ber States through the introduction, by
noted not only that good clinical practice 6 February 2018, of laws, regulations and
for medical devices entails compliance administrative provisions necessary in or-
with professionally agreed and widely ac- der to comply with European Council Di-
cepted technical standards, but also that, rective 2013/59. In this context, QC means
at the time of writing, acceptance testing the set of operations (programming, co-
is becoming a legal requirement within ordinating, implementing) intended to
the European Union: Member States are maintain or to improve quality. It includes
to bring into force European Council Di- monitoring, evaluation and maintenance
rective 2013/59-compatible laws, regula- at the required levels of all characteristics
tions and administrative provisions by 6 of performance of equipment that can be
February 2018 and ensure that acceptance defined, measured and controlled. Mem-
testing is carried out before the first use of ber States shall ensure that performance
equipment for clinical purposes[1–15]. testing is carried out on a regular basis, and
after any maintenance procedure liable to
QUALITY CONTROL TESTING affect performance. Furthermore, Member
A basic requirement for the successful es- States shall ensure that the competent
tablishment of quality management sys- authority takes steps to ensure that the
tems, quality assurance programmes and necessary measures are taken by the un-
QC procedures is that the leadership of dertaking to improve inadequate or defec-
healthcare institutions, including nuclear tive performance of medical radiological
medicine departments, recognises and equipment in use, and also adopt specific
understands quality-related principles and criteria for the acceptability of equipment
and shut down all motor-driven system are then due to the patient only and not
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
movements when pressed [1–3, 6, 7, 11–13, 15]. the scintillation camera itself. QC testing
of intrinsic or extrinsic uniformity and sen-
Energy window sitivity of the imaging equipment should
Daily operational checks of energy win- be performed daily in order to check the
dow settings should be performed to system’s response to spatially uniform flux
confirm that all preset pulse height anal- of 99mTc or 57Co photons. Such a flood field
yser energy windows are properly centred uniformity may be tested qualitatively
around the energy photopeaks of the ra- by visual inspection or quantitatively by
dionuclides to be used with the scintilla- calculation of the integral and differen-
tion camera for clinical imaging purposes, tial image uniformity within the camera’s
thus suggesting correct energy calibration central field of view and useful field of
of the system [1–7, 11–13, 15]. view. If a daily intrinsic low-count unifor-
mity test is selected, then each collimator
Background should be checked weekly or monthly by
Operational check of the background count an extrinsic high-count uniformity test.
rates with or without collimators and with- Overall sensitivity of the detection system
in one or more energy windows should be is calculated as count rate per unit activ-
performed daily to detect radiation caused ity (cps/MBq) of the imaged radioactive
by possible radioactive contamination of source [1–9, 11–13, 15].
the scintillation camera, floor or walls, ra-
diation from some neighbouring unshield- CT checkup and quality
ed source or an excess of electronic noise. Daily X-ray CT QC testing of the SPECT/CT
Under constant measuring conditions the system should be performed according to
background count rates should be approx- the manufacturer’s recommended proce-
imately constant in all detector directions dures and medical physics expert advice.
used for clinical imaging [1–4, 11, 13]. For instance, it may be recommended to
perform daily CT checkup and CT quality
Uniformity and sensitivity procedures which automatically execute a
One of the basic assumptions in nuclear set of CT tube warm-up acquisitions, au-
medicine imaging is that the response of tomatic function checks, and different air
the imaging system to a uniform irradia- and water calibration steps for all available
tion is uniform within defined limits. Ob- voltage settings, in order to guarantee op-
served differences in activity distribution timum image quality [2, 11, 12, 15].
of the measured peak location from the mination of extrinsic spatial resolution and
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
best-fit peak location. spatial linearity, the bar phantom is placed
The bar phantom is a rectangular or directly on the collimated detector. The de-
circular sheet of plastic material in which tector is irradiated either by the flood source
a number of lead bars are embedded in a placed directly on top of the bar phantom
pattern of parallel stripes, usually arranged or by the point source placed several metres
into four quadrants of parallel bars. Lead away from the bar phantom. After collect-
bars of a given thickness are supposed ing the required number of counts with an
to stop radiation, whereas plastic stripes appropriately set-up camera for the imaged
are supposed to be transparent to radi- radionuclide, spatial resolution is expressed
ation. The width of the lead bars and the in terms of the quadrant pattern, with the
distance between two bars is equal with- narrowest stripes still resolvable on the ac-
in one quadrant, but different for each of quired images. When used for determina-
the four quadrants (e.g. 2, 2.5, 3 and 3.5 tion of intrinsic spatial resolution and spatial
mm or 3.2, 4.0, 4.8 and 6.4 mm). The incre- linearity, the bar phantom is placed directly
ment of bar separations and widths from on the uncollimated detector and irradiated
one to another quadrant should be small by the point source placed away from the
enough to provide reasonably accurate fi- bar phantom at a distance which is at least
nal semi-quantitative estimation of spatial five times the largest dimension of the de-
resolution, for instance FWHM ≤8 mm. tector. After collecting the required number
The bar phantom should match the spatial of counts with an appropriately set-up cam-
resolution of the scintillation camera in such era for the imaged radionuclide, the intrinsic
a way that at least one quadrant of stripes spatial resolution can be approximated as
cannot be fully resolved in the acquired bar FWHM = 1.75·B, where B is the width of the
phantom image. Bar phantoms can be used narrowest bars that the scintillation camera
weekly, biweekly or more infrequently in can still resolve. For the purpose of thor-
routine QC testing for visual determination ough evaluation of spatial resolution and
of extrinsic or intrinsic spatial resolution of linearity, the bar phantom can furthermore
the scintillation camera, whichever appears be rotated and inverted in such a way that
more convenient. The same high-count im- the quadrant of the bar phantom with the
ages of bar phantoms can also serve for the narrowest bars is imaged in each quadrant
evaluation of the scintillation camera spatial of the detector in each direction, i.e. with
linearity, which is normally measured by us- the imaged stripes parallel to the x- and
ing the slit phantom. When used for deter- the y-axis of the detector. Moreover, the bar
phantom can be imaged in air at a certain rection of its detection system after it has
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
distance from the detector equipped with been properly tuned and adjusted. This cor-
different parallel-hole collimators or in tis- rection is basically applied by multiplying
sue-equivalent material added between each particular pixel in acquired images by
the bar phantom and the collimator. a factor calculated as the ratio of the aver-
The purpose of checking spatial reso- age counts in the high-count flood image
lution and linearity is to detect gradual to the counts in the corresponding pixel in
long-term deterioration of spatial resolu- the high-count flood image. Intrinsic uni-
tion, and to display imaged linear objects formity correction corrects for non-unifor-
as exactly linear as possible, as compared mities in the detector only, whereas extrin-
with acceptance and reference measure- sic uniformity correction corrects for both
ments. Bar phantom image acquisition detector and collimator non-uniformities.
may or may not be required by imaging The total number of counts to be collected
equipment manufacturers, but is done at in high-count flood images depends upon
the discretion of the user [1–9, 11–13, 15]. the particular procedure and equipment
but is typically in the range of tens to one
High-count flood and uniformity or a few hundred million. Modern scintil-
A flood source is typically a rectangular lation cameras include on-line corrections
source of uniformly distributed radioactiv- for detection system variations in energy
ity, in the form of either a sealed 57Co sheet response, spatial linearity and spatial unifor-
source or a plastic phantom fillable with mity across the field of view of the scintilla-
a solution of the selected radionuclide. tion camera, but these should be periodi-
Detector irradiation can also be consid- cally verified and re-created if necessary to
ered uniform if a point radioactive source assure acceptable integral and differential
is placed away from the detector face at uniformity [1–9, 11–13, 15].
a distance five or more times greater than
the largest linear dimension of the detector.
Ideally, a uniform irradiation of a scintillation QUARTERLY AND
camera detector should produce an image ANNUAL QC TESTS
of homogeneously distributed radioactivi-
ty. The purpose of acquiring a high-count Collimator hole angulation
flood image is to verify uniformity within Collimator hole angulation is the geomet-
the field of view of the scintillation camera ric relationship of the actual collimator
and to provide a uniformity/sensitivity cor- holes and septa to the crystal face of the
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
the axis of rotation of the SPECT scintilla- The angle of tilt of the SPECT scintillation
tion camera. In the case of an ideally col- camera detector is the angle between
limated and rotating scintillation camera the detector plane and the axis of rota-
detector, all edges of the holes and septa in tion, measured along the axis of rota-
a parallel hole collimator should be parallel tion. Assuming that the axis of rotation
to each other and exactly perpendicular to is horizontal, the parallel hole collimators
the crystal and the axis of rotation. Any dif- should also be levelled exactly horizon-
ferences in these angles are referred to as tally. This adjustment is usually done by
collimator angulation error. Non-orthogo- careful use of a spirit level or an angle
nality in the x-direction (perpendicular to gauge. Head tilt should normally be 0°
the axis of rotation) actually represents a at the beginning of tomographic acqui-
centre of rotation offset, whereas non-or- sition of the correctly set up system, and
thogonality in the y-direction (parallel to should remain 0° for all angles of rotation.
the axis of rotation) represents scintillation The angle of tilt can be determined from
camera head tilt. Both non-orthogonalities summed projection images over 360° of
can deteriorate the quality of reconstruct- a radioactive point source placed off the
ed images. axis of rotation.
Quality control testing of the collimator If there is no head tilt, the amplitude of
hole angulation checks the septal align- the sinusoidal motion of such a radioac-
ment and angulation for all parallel hole tive point source in the y-direction will be
collimators used. It is performed by using equal to zero, showing a constant rather
a point radioactive source placed a few than a sinusoidal pattern in the y-direction
metres from the face of the collimator, in of the projection images and a flat rather
the centre of each parallel hole collima- than ellipsoidal shape (i.e. a short ellipse
tor and in four or more other positions axis equals zero, while a long ellipse axis
approximately halfway to the edge of the equals a distance 2r off the axis of rota-
field of view. Acquired images should be tion) when all projection data in the x- and
visually inspected and checked for any y-directions are taken together. If there
asymmetries, streaks and distortions. If the is a head tilt, it can be determined from
collimator holes and septa do not appear the length of the short ellipse axis and
appropriately aligned and angulated, the the known radius of the radioactive point
manufacturer should provide a new colli- source [1–6, 11, 13, 14].
mator [1–4, 6, 11, 13, 14].
how the imaging system performs in such SPECT data to be reconstructed with fil-
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
situations. Cylindrical phantoms, used in tered backprojection and a ramp filter. Dif-
QC testing of system performance, are ferent total performance phantoms and
plastic cylindrical tanks that have differ- studies are used in acceptance, reference
ent shapes, dimensions, inner structures, or QC testing at less frequent intervals to
inserts and other physical characteristics check for possible slow degradation in the
and are fillable with solutions containing performance characteristics of different
different radionuclides. Image quality pa- low- to ultra-high resolution SPECT sys-
rameters which may be evaluated during tems. Total performance phantom studies
system performance tests include tomo- are also useful to assess the performance
graphic uniformity, contrast, resolution, characteristics of SPECT system hardware
attenuation, noise, linearity and lesion and software after significant preventive
detectability. Two examples of commer- or corrective maintenance and upgrades,
cially available cylindrical tomographic or when conducting research activities [1–4,
phantoms are the Jaszczak phantom and 6, 11, 13, 15]
.
the Carlson phantom. A Data Spectrum
ECT (emission computed tomography) Artefacts
phantom, usually known as the Jaszczak Artefacts in biomedical imaging are
phantom, consists of a main plastic cir- misperceptions and misrepresentations of
cular or elliptical tank which contains few the imaged objects caused by the imaging
parts: a segment of homogeneous radio- equipment or the employed image acqui-
activity, a segment of non-radioactive sol- sition and processing techniques. Various
id spheres of different sizes and a segment sudden or gradually developing problems
of (non-)radioactive rods of different sizes. may become evident at any time, but
These segments are used after image re- taking appropriate preventive measures
construction in order to detect possible – planning, preparing and organising the
ring artefacts and distorted spheres and nuclear medicine facility, conditioning the
rods, to evaluate the contrast and spatial electrical power supply, performing ac-
resolution of objects of a known size and ceptance, reference and QC testing, carry-
to calculate the linear attenuation coeffi- ing out regular maintenance, conducting
cient for attenuation correction if related overall clinical practice competently, etc.
software is available. The phantoms are – decreases the likelihood of artefacts.
typically filled with 99mTc and are imaged The most common imaging artefacts
for tens of minutes to acquire high-count are: full or partial ring and bull’s eye arte-
facts; different forms of distortion; blur- requirement before proceeding with rou-
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
rings; variations in intensity; lines and tine clinical practice. If full correction of
stripes or other discontinuities; local spots detected problems is possible locally and
of higher or lower intensity; artefacts due in a timely way, this should be implement-
to patient movements or metallic objects; ed, with subsequent successful repetition
artefacts caused by energy spectrum dis- of QC procedures before continuation
tortions and energy resolution degrada- with daily clinical practice. If correction is
tion; and artefacts due to decreases in de- not possible, limited use of the imaging
tector sensitivity, poor spatial uniformity, system should be considered. For this pur-
poor contrast and spatial resolution, inad- pose a call for service is usually made. De-
equate image acquisition and processing pending on the outcome of this call, lim-
etc. All of these artefacts are extensively ited clinical practice may be continued. If
illustrated elsewhere. immediate service is needed, it should be
According to the general troubleshoot- followed by successful QC testing before
ing flowchart (including but not limited again proceeding with full routine clinical
to artefacts), the acquisition of QC results practice [1, 3, 4, 6, 13].
within acceptable limits is a necessary
SYSTE M Q UA L I T Y CO NTR O L
PL A NA R C AM E RA A ND S PE C T
REFERENCES
1. International Atomic Energy Agency. Nuclear medi- Guide. Vienna: EANM Technologist Committee and
cine resources manual. Vienna: International Atomic Technologist Education Subcommittee; 2007.
Energy Agency; 2006. 10. Council Directive 2013/59/Euroatom of 5 December
2. EANM Physics Committee, Busemann Sokole E, 2013 laying down basic safety standards for protec-
Płachcínska A, Britten A. Acceptance testing for nu- tion against the dangers arising from exposure to
clear medicine instrumentation. Eur J Nucl Med Mol ionising radiation, and repealing Directives 89/618/
Imaging 2010;37:672–681. Euratom, 90/641/Euratom, 96/29/Euratom, 97/43/
3. International Atomic Energy Agency. Quality assur- Euratom and 2003/122/Euratom. Official Journal of
ance for SPECT systems. Human Health Series No. 6. the European Communities 17.1.2014.; No. L-13: 1-73.
Vienna: International Atomic Energy Agency; 2009. 11. EANM Physics Committee, Busemann Sokole E, Płach-
4. Bailey DL, Humm JL, Todd-Pokropek A, van Aswegen cínska A, Britten A, EANM Working Group on Nuclear
A. Nuclear medicine physics - a handbook for teach- Medicine Instrumentation Quality Control, Lyra Geor-
ers and students. Vienna: International Atomic Ener- gosopoulou M, et al. Routine quality control recom-
gy Agency; 2014. mendations for nuclear medicine instrumentation.
5. Cherry SR, Sorensen JA, Phelps ME. Physics in nuclear Eur J Nucl Med Mol Imaging 2010;37:662–671.
medicine. 4th ed. Philadelphia: Saunders/Elsevier; 2012. 12. Siemens Medical Solutions. Operating Instructions
6. International Atomic Energy Agency. Quality control Symbia System T16/T6/T2/T Series. Erlangen: Sie-
of nuclear medicine instruments 1991. Technical mens Medical Solutions; 2009.
document 602. Vienna: International Atomic Energy 13. International Atomic Energy Agency. Quality con-
Agency; 1991. trol atlas for scintillation camera systems. Vienna:
7. Murphy PH. Acceptance testing and quality control International Atomic Energy Agency; 2003.
of gamma cameras, including SPECT. J Nucl Med 14. Busemann-Sokole E. Measurement of collimator
1987;28:1221–1227. hole angulation and camera head tilt for slant and
8. National Electrical Manufacturers Association. NEMA parallel hole collimators used in SPECT. J Nucl Med
NU 1. Performance measurement of scintillation 1987;28:1592–1598.
cameras. Rosslyn, VA: National Electrical Manufactur- 15. Jones DW, Hogg P, Seeram E. Practical SPECT/CT in
ers Association; 2001, 2007, 2013. nuclear medicine. London: Springer; 2013.
9. European Association of Nuclear Medicine. Best
Practice in Nuclear Medicine. Part 2. A Technologist’s
by Claudiu Peștean
CHAPTER 4
INTRODUCTION
Q UA L I T Y CO NTR O L
PE T SYS TE M
This trend is due not only to the contribu- stand the basic principles of acquisition
tion that PET/CT has made to medicine and reconstruction. Like all diagnostic
by virtue of its combined morphological procedures in nuclear medicine, PET
and functional perspective, but also to its technology uses the ability of scintilla-
wide application in various pathologies, tor detectors to record signals produced
including above all oncologic diseases, by the interaction of gamma photons
against a background of increasing prev- with the detector crystals. PET is a tech-
alence. In relation to the PET part of the nique based on the administration of a
PET/CT technology, it is necessary to un- radiotracer containing positron-emitting
derline the importance of quantification isotopes. With respect to QA/QC, it is
of radiotracer uptake in pathological tis- not of great importance and is beyond
sues, commonly achieved by calculation the purpose of this chapter to detail as-
of the standardised uptake value (SUV) pects relating to the radiopharmaceutical
for fluorine-18 fluorodeoxyglucose (FDG), properties of the radiotracers used in PET.
which reflects the metabolic rate of the This information may be found in differ-
tissue and, implicitly, the aggressivity of ent papers, including the technologist’s
the tumour. Importantly, PET/CT has also guide published by EANM Technologist
started to be used for radiation therapy Committee on the subject: Principles and
planning. As PET/CT has developed into Practice of PET/CT. Part 2 – A technologist’s
an indispensable imaging procedure in di- guide[2]. The QC and QA for PET are direct-
agnostic and therapeutic strategies, it has ly related to the nuclear physics exploited
become ever more important that reliabil- in this technique and to the technology
ity of the produced images is ensured. This used to create PET/CT images. After ad-
is achieved by the quality control (QC) and ministration of the radiopharmaceutical
quality assurance (QA) procedures em- and following allowance for the uptake
ployed to ensure correct scanner set-up time, the patient is placed in the scan-
and operation[1]. ner and the data are acquired. First, the
To understand the concepts of QC and CT data are recorded based on measure-
QA tests for PET, it is essential to under- ment of the attenuation in the tissues of
the X-rays emitted by an X-ray tube. For data are recorded after the CT acquisition.
Q UA L I T Y CO NTR O L
PE T SYS TE M
this purpose, the patient is placed in the The table moves the patient inside the PET
gantry and X-rays are emitted from dif- detector. The PET detector (consisting in
ferent angles, traversing the patient and a multitude of scintillation crystals organ-
being recorded on the opposite side by ised in blocks and modules distributed on
a solid state detector. The CT image is a ring with several rows) detects, very close
produced on the basis of the recorded at- to the point of release, the annihilation
tenuation values from all the projections, photons that are produced when the posi-
and using specific reconstruction and fil- trons released by the PET isotopes interact
tration methods. with electrons in the matter. The two pho-
tons produced by each annihilation trav-
The most important aspect of the CT el through matter in opposite directions
image in PET/CT, besides the anatomical and interact with the detector crystals.
information that it provides, is its utility in If these photons are recorded by the de-
correcting the attenuation within the pa- tector within the same time window, they
tient’s tissues of the gamma photons used are called coincident. If two photons are
to produce the PET image. This process, recorded by two crystals, the system will
termed CT attenuation correction (CTAC), assume that the annihilation process has
is necessary to minimise introduction of been produced somewhere on the line
bias into the PET image, which should between these crystals, the so-called line
precisely reflect the biodistribution of the of response (LOR). Faster detectors may
radiotracer. Based on the attenuation val- calculate the time interval between the ar-
ues calculated with the CT technique, also rivals of the two annihilation photons and
referred to as CT numbers or Hounsfield can more accurately estimate the origin of
units, the tissue densities are calculated; the annihilation process and the presence
once the density map of the investigated of the radiotracer; as a consequence, this
tissues is known, the attenuation of gam- so-called time-of-flight (TOF) technique
ma photons detected by PET is calculat- offers superior spatial resolution. LOR or
ed. An attenuation map is elaborated and TOF information is recorded on sinograms
applied to the PET non-corrected image and, using specific reconstruction meth-
based on the acquired data and the CTAC ods, the PET image is obtained[3]. In order
PET image is produced. This attenuation to gain a fuller understanding of the basic
correction process is performed after the principles of PET and PET/CT, the reader
PET data have been acquired. These PET can access the EANM website to consult
the technologist’s guide produced for this ing any relevant differences in relation to
Q UA L I T Y CO NTR O L
PE T SYS TE M
purpose: Principles and Practice of PET/CT. the various PET scanners on the market.
Part 1 – A technologist’s guide{4]. While we shall describe the principles of
Before any PET image is analysed, it the procedures, we shall not detail the
is mandatory to guarantee that quality practical aspects.
standards for the equipment are met. This
is the aim of the QC and QA procedures
which we shall describe in this chapter. ACCEPTANCE TESTING FOR PET
We shall try to define the concepts of QC Since PET and PET/CT techniques were
and QA, starting with acceptance test- first introduced into clinical medical im-
ing procedures and then describing the aging, the need for standardisation has
basic QC tests that are required for PET grown tremendously as the sensitivi-
equipment. There are various vendors on ty and the applicability of the method
the market who provide several types of have increased. In response to this trend,
PET/CT scanner, designed according to American and European associations
the same principles but with slight differ- and bodies, such as the Society of Nu-
ences in both the technologies used for clear Medicine, the National Electrical
detection or reconstruction and the QC Manufacturers Association (NEMA) and
and QA procedures. In modern PET scan- the International Electrotechnical Com-
ners, the QC and QA procedures are in- mission, started to establish and develop
creasingly automatic or semi-automatic, sets of standards for positron emission
which excludes potential human errors tomographs. The benefit is that now all
to the greatest possible extent. Never- the manufacturers can specify the per-
theless, it is important to understand the formance of their equipment using the
principles of the technique and to be same set of parameters and standards,
able to recognise the artefacts than can which can be measured and verified af-
be produced if QC or QA fails. It is also ter installation. NEMA standards have
necessary to know the measures to be been imposed over time and now all
taken in this circumstance in order to the vendors relate the performance of
eliminate errors from the diagnostic in- their equipment to these standards. Fol-
formation and thereby benefit patients. lowing the elaboration of the first NEMA
Here we aim to review the QC and QA NU2-1994 standard, the need for further
procedures that are of the greatest im- standardisation gradually grew, and sev-
portance in the practice of PET, identify- eral updates have since been produced:
NEMA NU2-2001, NEMA NU2-2007 and analysed by calculating the count rate for
Q UA L I T Y CO NTR O L
PE T SYS TE M
NEMA NU2-2012[5]. each slice and for each sleeve using a de-
After installation of the equipment, ac- cay correction formula and an attenuation
ceptance testing procedures need to be correction for each sleeve[6].
performed in order to ensure that the lev-
els of performance specified by the man- Spatial resolution
ufacturer are met. The acceptance tests The spatial resolution should be assessed
and the parameters verified usually refer during the acceptance testing, at the end
to the NEMA standards. If the acceptance of the warranty and whenever there is sus-
tests are failed, the system will need to be picion that the performance of the equip-
corrected. It is also important to perform ment has suffered importance changes.
a new set of acceptance tests prior to the The principle of the test consists in image
end of the warranty. acquisition using three point sources (with
Depending on the type of scanner, var- an activity of approx. 1 MBq) that have a di-
ious tests will form part of the acceptance ameter less than 1 mm and are suspended
testing procedures and these are dis- in the air to avoid any effect of scattered
cussed below. radiation; these point sources are placed 1
cm vertically from the centre of the FOV,
Sensitivity 10 cm vertically from the centre of the FOV
The sensitivity is determined during the and 10 cm horizontally from the centre of
acceptance tests, at the end of the war- the FOV. Two acquisitions are performed:
ranty and when the system suffers import- one in the centre of the axial FOV and one
ant changes in its performance. For this at a quarter of the axial FOV. The spatial
test, a 7-cm line source of 5 MBq is used, resolution is derived on the basis of the
placed in a phantom with five aluminium full-width at half-maximum (FWHM) of
sleeves with a wall thickness of 1.25 mm. the response function calculated for each
Acquisitions are performed starting with source[6].
the smallest sleeve and with the source in
the centre of the field of view (FOV) (or at Energy resolution
10 cm offset from the centre if variations A test relevant for scanners that use only
in sensitivity within the FOV are being singles-based attenuation correction and
checked); for subsequent acquisitions the calibration is performed by acquiring more
attenuation is increased by adding the than 10 kcts after placing a point source
other sleeves one by one. Sensitivity is of 18F with an activity of approx. 1 MBq in
the centre of the FOV. Usually, the energy of clinical imaging conditions. This can
Q UA L I T Y CO NTR O L
PE T SYS TE M
resolution is analysed using the vendor’s be achieved by acquiring the data using
procedure for energy testing[6]. two types of phantom. The first is a body
phantom with spheres, a so-called image
Scatter fraction, count losses and quality phantom, which consists in three
random measurements compartments: a body compartment, six
This test is very important since the scat- hollow spheres with diameters of 1, 1.3,
ter and random photons may affect im- 1.7, 2.2, 2.8 and 3.7 cm and a cylindrical
age quality. The test reflects the count rate central insert filled with material that sim-
performance of the scanner, expressed as ulates lung tissue’s attenuation properties.
noise equivalent count rate. To test this, This body compartment of this phantom
a polyethylene cylindrical phantom of 70 is filled with 18F solution. A sphere-to-
cm length and 200 mm diameter is used; background ratio of 4:1 is used for the
the cylinder is traversed from end to end first study for the four smallest spheres
by a hole that is parallel to the axis and at (the other two are filled with cold water)
a radial offset of 45 mm, to contain the line and a second study is performed using a
source. From the acquisition data, prompt ratio of 8:1. The second phantom is used
and random sinograms are obtained and, to mimic the radioactivity present outside
based on Poisson statistics, the scatter of the FOV in clinical situations. This is a cy-
fraction is calculated using the vendor’s lindrical phantom similar to the phantom
specifications and applying complex sta- described for the determination of scatter
tistical calculations[6]. fraction, with a concentration activity in
the line source equal to the concentration
Image quality and accuracy of of the background compartment of the
attenuation and scatter correction and image quality phantom.
quantitation One transverse slice centered on the
It is very difficult to simulate the distri- spheres is used for image analysis. ROIs are
bution of the radiopharmaceutical in drawn on each sphere and in addition 12
the patient’s body and to assess the im- ROIs are drawn in the background com-
age quality using a phantom. The image partment at a distance of 15 mm from the
quality, the attenuation and scatter cor- edge. The per cent contrast is calculated
rection and even the quantitation on the for each sphere (hot and cold) in relation
PET image can be evaluated reproducing to the background average counts mea-
as closely as possible the particularities sured on the previously drawn ROIs.
The accuracy of the attenuation and the tolerance values established by the
Q UA L I T Y CO NTR O L
PE T SYS TE M
scatter correction is estimated using ROIs manufacturer.
drawn on the lung compartment. The
counts measured in the lung insert of PET normalisation
the phantom are related to the counts Even when non-uniformities are within
measured in the ROIs placed in the back- the tolerance level, it is important that the
ground compartment as described above. reconstructed image is normalised to en-
Using the option provided by the soft- sure optimal uniformity. For this purpose,
ware, the concentration of the radioactivi- a rotating 68Ge or a uniform 68Ge cylindrical
ty is displayed and compared with the true phantom may be used. Calibration data
radioactivity concentration, known from are obtained which are used to normalise
the time of preparation of the 18F solution the acquired data in the clinical mode.
when filling the phantom[6].
Radioactivity concentration calibration
Coincidence timing resolution For this test, a cylindrical fillable 18F phan-
Coincidence timing resolution is estimat- tom is used, performing an acquisition
ed by histogramming the differences in with consistent data statistics. It is also pos-
the arrival times of annihilation photons sible to evaluate the SUV accuracy by scan-
after acquiring data using a point source ning a phantom with a known activity us-
placed in the centre of the FOV within a ing a multibed protocol. If the information
scattering material[6]. relating to the tracer and phantom (e.g.
activity, calibration time, phantom weight,
Uniformity of the reconstructed image phantom volume) is recorded as part of
This test is a measure of the system re- the patient data, then the measured SUV
sponse to a homogeneous radioactivity should be 1. Practically, this calibration test
distribution. A cylindrical hollow phan- evaluates the ability of the system to cor-
tom filled with 18F solution or, alterna- rectly measure the SUV.
tively, a cylindrical 68Ge/68Ga cylindrical
phantom is used. On the reconstructed The above represent a minimum set of
slices, the non-uniformities are assessed tests that need to be performed in relation
by determining the counts on square to the PET system during the acceptance
ROIs of 1×1 cm. The evaluation is per- testing for a PET/CT scanner. It is recom-
formed on each slice and the maximum mended that the tests are repeated before
value of non-uniformity should be within the end of the warranty to ensure that the
system will conform to the original accep- daily QC procedures for the main types of
Q UA L I T Y CO NTR O L
PE T SYS TE M
Figure 1
Q UA L I T Y CO NTR O L
PE T SYS TE M
Calibration system based on a 68Ge rod source placed in the FOV of a GE PET/CT scanner
Discovery series, Optima 560 (courtesy of CT-Clinic Positron Emission Tomography)
recommendations and the national reg- for acquisition of accurate diagnostic in-
ulations established for nuclear medicine formation and also for reproducibility of
practice. These recommendations and scanning conditions, which in turn is nec-
regulations should be strictly respect- essary to allow correct comparison and
ed to ensure proper functionality of the evaluation of images during the diagnos-
equipment; the latter is a prerequisite tic sequences.
REFERENCES
1. International Atomic Energy Agency. Standard Op- tice of PET/CT. Part 1. A technologist’s guide. Vienna:
erating Procedures for PET/CT: A Practical Approach European Association of Nuclear Medicine; 2010.
for Use in Adult Oncology. Vienna: International 5. International Atomic Energy Agency. PET/CT atlas
Atomic Energy Agency; 2013. on quality control and image artefacts. Vienna: Inter-
2. Testanera G, van den Broek W, eds. Principles and national Atomic Energy Agency; 2014.
practice of PET/CT. Part 2. A technologist’s guide. 6. International Atomic Energy Agency. Quality assur-
European Association of Nuclear Medicine; 2011. ance for PET and PET/CT systems. Vienna: Interna-
3. Saha GB. Basics of PET imaging - physics, chemistry tional Atomic Energy Agency; 2009.
and regulations [internet]. New York: Springer Sci- 7. Busemann Sokole E, Płachcínska A, Britten A, Lyra
ence+Business Media Inc.; 2005. Available from: http:// Georgosopoulou M, Tindale W, Klett R. Routine quality
www.springerlink.com/index/10.1007/b138655 control recommendations for nuclear medicine instru-
4. Hogg P, Testanera G, eds. EANM principles and prac- mentation. Eur J Nucl Med Mol Imaging 2010;37:662–71.
INTRODUCTION
Q UA L I T Y CO NTR O L
C T SYS TE M
CT scanners are used not only in diag- by the dose to the patient – a higher dose
nostic radiology but also in radiotherapy usually means better image quality, but
and nuclear medicine. In radiotherapy sole attention to the latter can cause doses
departments, they are used for treatment to become inappropriately high. An effec-
planning and in nuclear medicine depart- tive quality assurance (QA) system should
ments mostly as a part of SPECT/CT or therefore be implemented to assure ad-
PET/CT systems. In spite of rapid techno- equate performance and optimised use
logical development, a drawback of CT is of the CT system. QA begins already with
still the relatively high patient radiation the specifications of the equipment being
exposure. Patient doses per procedure are purchased. When equipment is installed,
rarely below 1 mSv and can easily reach an acceptance test is performed; this is
several tens of mSv. In some procedures best done on a collaborative basis by the
(e.g. perfusion imaging), doses can even service engineer, medical physicist and
reach the threshold for deterministic ef- radiographer who will be responsible for
fects [1]. An exception to this can be nu- the scanner. At acceptance testing, the CT
clear medicine, if CT images are used for scanner and its operation are tested to es-
attenuation correction only and are there- tablish whether the scanner functions as
fore not of diagnostic quality. designed, whether it complies with regu-
latory requirements and whether it meets
The performance of the CT scanner and the requirements set in the purchase
its components has a significant influence specification. Acceptance testing is also
on the outcome (image) and the patient useful in establishing baselines for future
exposure. CT image quality is influenced QC testing.
Quality control (QC) is an important teria. However, one should not forget
Q UA L I T Y CO NTR O L
C T SYS TE M
element of the QA system. It is intended that proper functioning of equipment is
to verify that CT scanner performance is merely a necessary condition for the per-
consistent with the predetermined re- formance of optimised diagnostic pro-
quirements. The QC programme should cedures. Appropriate use (protocols) of
therefore include a list of tested parame- procedures is also of vast importance and
ters, a description of testing methods and should therefore also be monitored. This
the tolerances used to check that quality can be done using different techniques,
requirements are met. It should also in- for example clinical audits and dose mon-
clude anticipated actions to adjust or cor- itoring (tracking).
rect performance if the tolerance limits are
exceeded.
TESTS TO BE CONDUCTED BY
There are many books and guidelines MEDICAL PHYSICISTS
dealing with QA and QC in CT (e.g. [2–5]). Tests described in this section are divided
QC tests are usually divided into basic into three types: (a) geometry, positioning
and optional or according to the recom- and movement tests, (b) image quality
mended frequency. Some QC tests need tests and (c) CT dosimetry tests. The basic
to be performed frequently (daily, weekly equipment needed to perform these tests
or monthly), so it is recommended that is an image quality phantom and a CT
these tests be performed by radiographers dosimetry set (dosimetry phantom, ioni-
working on a CT scanner. More compre- sation chamber and electrometer). As test
hensive testing is done semi-annually or procedures are at least partially depen-
annually and is usually performed by med- dent on the available equipment, the user
ical physicists. should read the instructions provided with
the equipment. We have tried to describe
The focus of this guide is technical the test procedures listed below as gener-
QC, which refers to procedures designed ally as possible, so that the information is
to show that the CT equipment and not specific to particular equipment.
its performance conform to the set cri-
Geometry, positioning
and movement tests
Q UA L I T Y CO NTR O L
C T SYS TE M
Q UA L I T Y CO NTR O L
C T SYS TE M
Accuracy of indicated table movements
Purpose. The test is performed to check that the table movements are consis-
tent with the planned movements.
Description. The movement test should be performed with a weighted table (e.g.
with a person sitting on it):
» Attach a tape measure to the table in such a way that the central laser
crosses it at a certain number.
» Set the table position to 0.
» Move the table for a certain distance (e.g. 30 cm) and see whether the
laser crosses the tape measure at the expected distance.
» Return the table to the previous position and check it with the laser
line on the tape measure.
Description. The test can be performed together with the test of alignment of lasers
with the scan plane (see above). Proceed as follows:
» Position the film strip on the table or on the phantom and centre it
vertically and horizontally using lasers.
» Using the central laser, position the film so that the laser crosses the
film at 0 mm.
» Scan the film strip using the axial head protocol and a certain thickness.
» Move the table for a certain distance (e.g. 30 mm) and scan with a
different thickness.
» Observe the image (see Fig. 1 for example).
» If strip film is scanned, radiation profiles can be deduced and the slice
width accurately measured (Fig. 2).
Geometry, positioning
and movement tests
Q UA L I T Y CO NTR O L
C T SYS TE M
Q UA L I T Y CO NTR O L
C T SYS TE M
Figure 1
Figure 2
Figure 3
Artefact evaluation
Purpose. To evaluate the presence/absence of artefacts on images.
Q UA L I T Y CO NTR O L
C T SYS TE M
CT number accuracy
Purpose. To check that the CT numbers of different materials are within the
range of expected values.
Equipment. Image quality phantom with different materials with known CT numbers
(for example: CATPHAN, PRO-CT phantom, Mini CT QC Phantom, ACR CT
accreditation phantom).
CT dosimetry tests
Q UA L I T Y CO NTR O L
C T SYS TE M
Accuracy of indicated dose parameters
Purpose. To check the accuracy of CT dose parameters (CTDIvol and/or DLP) displayed on the
CT scanner control console.
Equipment. Calibrated dosimeter with 100-mm pencil ionisation chamber (CTDI probe) and
PMMA CTDI phantom (16 cm and 32 cm PMMA).
Description. Proceed as follows:
» Place the CTDI head phantom on the head support and centre it vertically and
horizontally using lasers.
» Place the CTDI probe in the centre hole of the phantom.
» Prepare the CTDI probe for the measurements as described in the manufacturer’s
manual.
» Place the CTDI probe in the CTDI phantom in such a way that the centre of the
CTDI probe will be in the centre of the CTDI phantom.
» Centre the table at this position (put the position to zero or note the actual position).
» Make a scout image of the CTDI phantom. Set the start and end of the scan in
such a way that the centre of the CTDI phantom (CTDI probe) is in the middle of
the scanning area.
» Set the table movement to 0. During this test, the table should stay in a fixed position.
» Select the axial scan using exposure parameters close to the most frequently
used head scan.
»M ake a scan (1 rotation) and record the measured dose (PKL). Repeat to obtain at
least three measurements. The ionisation chamber measures the product of air-ker-
ma and length (PKL), which is usually called the dose–length product (DLP). To calcu-
late the CTDI, use the following formula:
PKL
CTDI = where N•T is the nominal beam collimation (N slices of thickness T).
N•T
» The dose parameter, which is usually displayed on the CT scanner control
console, is the weighted CT dose index CTDIw, which is a combination of mea-
surements in the centre of the phantom and at the periphery (CTDIc and CTDIp,
respectively):
1 2
CTDIW = CTDIC + CTDIP
3 3
» R epeat the measurements by placing the CTDI probe in all four periphery holes (po-
sitions: 12 h, 3 h, 6 h, 9 h). All unused holes of the phantom must be filled with plugs.
» Calculate the deviation of measured CTDI value to CTDI displayed on the control
console.
Tolerances. Deviation of measured dose from indicated dose <20%.
Notes. • Measurements should be done for all tube voltages that are used in practice.
• The same procedure is followed using a body phantom (32 cm PMMA) and a standard body protocol.
• Measurements of PKL in the centre can also be used to test some other properties of the CT scan-
ner, such as linearity with tube load (mAs), repeatability and dependence on beam width.
• Instead of measurements within the phantom, measurements free in air can be performed. In this
case, an ionisation chamber is attached to the centre of the CT scanner gantry.
Q UA L I T Y CO NTR O L
C T SYS TE M
»» Panel switches/lights/meters working portunity to expand learning outcomes
»» Door interlocks functioning for educational institutions and improve
»» Warning labels present the competencies of radiographers who
»» Intercom system functioning already perform CT procedures. In order
»» Postings present to ensure the provision of an effective, safe
»» Service records present and efficient service, radiographers’ train-
ing should incorporate instruction on QA
Most modern CT scanners have their own and QC practices, to include: legislation,
QC procedure in which the test phantom regulations and guidelines, test equip-
provided with the scanner is used. In CT ment and methodologies, programme de-
software QC packages, tests are automat- sign and implementation and reporting[7].
ed and measurement data are stored in With effective analysis and collaboration
CT software and can be accessed and an- with medical physicists and radiologists,
alysed. Such tests are not time consuming QC tests can also become an important
and can therefore be performed relatively training tool regarding the ways in which
quickly. As such they can be easily incor- parameters monitored during these tests
porated into a QA programme. affect image quality and dose. Although
QC phantoms cannot fully simulate pa-
tients, they can still provide radiographers
CONCLUSION with useful knowledge on the perfor-
Radiographers must be a part of the QC mance of their CT scanner and can be
team in order to secure, maintain or im- used as a valuable tool in the optimisation
prove the health and well-being of the process.
patient. Radiographers have a specific re-
sponsibility to actively participate in QA
and therefore should have the compe-
tences required to perform QC tests for
CT scanners and to evaluate the results of
routine QA and QC tests.
The document “Guidelines on Radiation
Protection Education and Training of Med-
ical Professionals in the European Union”[7]
is useful for the evaluation of existing ed-
REFERENCES
1. Bauhs JA, Vrieze TJ, Primak AN, Bruesewitz MR, Mc- 5. American College of Radiology. ACR–AAPM techni-
Collough CH. CT dosimetry: comparison of mea- cal standard for diagnostic medical physics perfor-
surement techniques and devices. Radiographics mance monitoring of computed tomography (CT)
2008;28:245–253. equipment. ACR Technical Standard 2012 Resolu-
2. IAEA HHS-19. Quality Assurance Programme for tion No. 34.
Computed Tomography: Diagnostic and Therapy 6. American College of Radiology. 2012 Computed
Applications. International Atomic Energy Agency, Tomography. Quality Control Manual. 2012. https://
Human Health Series No. 19, 2012. www.acr.org/~/media/ACR%20No%20Index/Docu-
3. International Atomic Energy Agency. IAEA Human ments/QC%20Manual/2012CTQCManual1a
Health Series No. 19. Quality assurance programme 7. European Commission. Radiation Protection No.
for computed tomography: diagnostic and ther- 175. Guidelines on radiation protection education
apy applications. 2012. http://www-pub.iaea.org/ and training of medical professionals in the Euro-
MTCD/Publications/PDF/Pub1557_web.pdf pean Union. 2014. https://ec.europa.eu/energy/en/
4. ImPACT CT Scanner Evaluation Group. ImPACT Infor- radiation-protection-publications
mation Leaflet 1. CT scanner acceptance testing. 2001.
http://www.impactscan.org/download/acceptan-
cetesting.pdf
6
O PTIM I SAT I O N
O F SP E C T A N D
ACQ UI SI T I O N A N D
RECO N ST RU C T ION
by Maximilien Vermandel
and Hélène Lahousse
I N S TRUME N TATION AND PROTO CO L S TA NDA RDIS AT IO N
QUA L I T Y CON TROL OF NUC LE A R ME DIC INE
CHAPTER 6
INTRODUCTION
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
The range of medical indications for SPECT remains of interest. Indeed, for nephrology
is wide and includes bone resorption, indications, quantification is expected to
studied using a bisphosphonate labelled estimate renal function or cortical transit
with 99mTc, and renal function, investigated time. In cardiology, cardiac function can
using 99mTc-labelled mercaptoacetyltrigly- also be analysed using 99mTc-pertechne-
cine (MAG3). 99mTc is also used with sesta- tate-labelled red blood cells and calculat-
mibi to study myocardial perfusion. In neu- ing the portion of isotope ejected during
rology, brain perfusion analysis is achieved each heartbeat to determine the ejection
using 99mTc coupled with a neutral and li- fraction. In particular, radiotracer uptake is
pophilic complex whose molecular mass is measured at the end of systole and dias-
sufficiently low to allow passage across the tole to estimate the ejection fraction, fol-
blood-brain barrier. However, some other lowed by evaluation of the kinetics in 3D.
isotopes are also used, such as iodine-123 Moreover, prior to therapy, the avidity of
in thyroid scintigraphy for hyperthyroidism the thyroid for iodine-123 may be assessed
or congenital hypothyroidism, indium-111 through measurement of the uptake rate,
with a somatostatin analogue for neuro- which serves as the basis for estimation of
endocrine tumours, krypton-81m for lung the iodine-131 activity to be prescribed.
ventilation, thallium-201 for evaluation
of cardiac viability and iodine-131 for the Since the range of applications of SPECT
treatment of thyroid cancer. imaging is wide, acquisition parameters
must be chosen carefully, in accordance
For some of these applications, the vi- with the indication being investigated.
sual inspection of SPECT images provides Indeed, SPECT studies depend on the set-
sufficient information (e.g. bone fracture, tings and methods used during acquisi-
arthritis, metastasis). For other applications, tion, reconstruction and post-processing.
however, quantification on SPECT images These parameters may drastically affect
the image quality and have to be carefully SPECT AND SPECT/CT IMAGING
considered both when performing the ac- INSTRUMENTATION
quisition itself and when interpreting the
images. Gamma camera basics
The first section of this chapter presents A gamma camera is a set of planar detec-
important background information on the tors, usually two (dual-head), mounted on a
principles of SPECT instrumentation. The gantry that enables simultaneous detector
following section is devoted to the factors acquisition of parts of a patient’s body from
that affect image reconstruction quality different orientations. Detection of the gam-
and the parameters that must be consid- ma photons emitted by the decay of the ra-
ered in order to achieve optimal acquisi- dioactive isotopes relies on the scintillation
tion. Finally, some new advancements in principle[1]. A basic sketch of a detector is
SPECT are discussed. shown in Fig. 1.
Figure 1
Photomultiplier tubes
Light guide
Scintillator
Collimators
SPECT
When used in static or whole-body mode,
gamma camera detectors only allow for
planar imaging of the radiotracer distribu-
tion. However, in SPECT imaging, acquisi-
SS acquisition requires just a single cycle, pose of attenuation and scatter correction
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
which basically entails rotation of each de- during tomographic reconstruction.
tector through 180° around the subject in The use of hybrid technologies may
order to cover the full 360° range (for de- entail significantly increased radiation ex-
tector configuration in 180° or “H” mode). posure for the patient. In order to reduce
Step duration or speed and number of cy- the exposure while maintaining good im-
cles in continuous mode depend primarily age quality, manufacturers have released
on the intensity of the radiotracer uptake systems that automatically modulate the
and the patient’s condition. tube current (mAs) according to the den-
Finally, image reconstruction is achieved sity of tissues being explored. Basically, the
using specific algorithms to render the system automatically adapts, in real time,
spatial distribution of the radiotracer into the tube current to the thickness of tissue
matrices of 64×64, 128×128 or 256×256 crossed by the X-ray beam as deduced
pixels according to the location and size of from the topogram acquisition (Fig. 3). For
the organ studied. example, the X-ray beam is more strongly
attenuated by the shoulders than by the
SPECT/CT abdomen owing to the lower density of
Image reconstruction with conventional the latter; thus decreasing the tube current
SPECT acquisitions cannot be corrected for for the abdominal exploration allows for a
attenuation or scatter when considering substantial dose reduction. However, in
non-heterogeneous areas, an exception specific cases the dose reduction system
being the brain, for which Chang model- may not be activated. For instance, when
ling gives satisfactory results. For the pur- the CT scan is only acquired for the pur-
pose of such correction, the SPECT gantry pose of attenuation correction, minimal
can be coupled to a computed tomograph dose exposure is expected and thus high
(CT) so that a CT scan can be acquired in voltage and tube current are carefully bal-
conjunction with SPECT imaging. This hy- anced to achieve this goal.
brid imaging system has two major advan-
tages: First, the quasi-simultaneous acquisi- Collimator
tion of SPECT and CT allows for the fusion In the chain of acquisition, the collimator
of functional and anatomical data with type is the first element to be considered
minimal concerns regarding registration. for optimisation of SPECT image acquisition
Second, the CT images, converted to a map since it closely depends on (1) the energy of
of electronic densities, are used for the pur- the emitted photons to be measured and
Figure 3
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
(a)
(b)
(2) the location being investigated. As de- limit the study duration and thus restrict
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
scribed previously, the collimator is an array the patient’s movements as far as possible.
of holes or honeycomb structure. Thus, the
main factors influencing the acquisition are Depending on the energy, the collima-
hole length, septa width and hole diameter tor is classified as low energy (99mTc, 123I,
(Fig. 4). For instance, increasing the length 201
Tl), medium energy (111In, 67Ga) or high
of the hole improves the collimation and energy (131I).
spatial resolution but leads to a lower sen-
sitivity. Septa thickness is related to emitted Collimator types also differ according
photon energy and is thicker for higher en- to the geometry of the holes. Most colli-
ergy photons (e.g. 0.4 mm for 140-keV pho- mators used in SPECT imaging are of two
tons versus 1.9 mm for >300-keV photons). types: parallel-hole collimators and con-
Figure 4 Figure 5
3
1 – Hole diameter | 2 – Hole length | 3 – Septa width (a) (b)
Hole diameter (or hole size) influences the verging-hole (fanbeam) collimators (Fig. 5).
spatial resolution: a smaller hole diameter The collimator type to be used for a given
increases the spatial resolution but decreas- imaging application depends on the ratio
es the sensitivity. In general, the choice of between the size of the field of view (FOV),
collimator for SPECT imaging should pri- the size of the detector and the expected
oritise sensitivity over spatial resolution to spatial resolution and sensitivity[2].
The parallel-hole collimator is the histori- improve photon interactions and the sen-
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
cal design by Anger and this is still the stan- sitivity. Indeed, a thick crystal increases
dard collimator employed in clinical prac- the probability of total absorption of the
tice. For a parallel-hole collimator, the point incident photons when the energy is high.
source sensitivity is the same over the whole However, the resultant gain in sensitivity
FOV. When the organ or location under in- comes at the cost of a reduction in spatial
vestigation is smaller than the FOV, the use resolution.
of fanbeam collimators results in improved The thickness of the crystal/scintillator
performance in terms of both sensitivity is specific to the gamma camera and de-
and spatial resolution[3]. This latter configu- pends on the type of radiotracer used. For
ration may be preferred for brain imaging, a low-energy tracer, a thin crystal should
for instance. be preferred (3/8 inch) while a thicker
Finally, for cardiology imaging, specific crystal (5/8 inch) is to be preferred for
collimators allowing for faster cardiac SPECT higher energy tracers.
acquisition have recently been released[4]. Finally, the choice of crystal thickness
depends on the type of daily use. For in-
Crystal stance, SPECT with 99mTc will be acquired
Interaction of emitted photons with the using a thin crystal, whereas if most uses
scintillator crystal leads to the emission of involve 131I, a thick crystal will be pre-
fluorescence photons (light) in all direc- ferred.
tions. Only a limited fraction of light trav-
els to the PMT. Thus, the scintillator is sur- Photomultiplier tubes
rounded by a reflector material on all sur- Light emitted inside the crystal is collect-
faces, except the surface in contact with ed by the PMTs in order to convert the
the PMT, in order to maximise light col- light signal into an electrical signal (Fig.
lection at the PMT. Additionally, the crys- 7). PMTs comprise an evacuated glass
tal is coupled to the PMT array through a package with a photocathode, to convert
transparent medium of the same index of light to electrons, followed by a string of
refraction as the crystal so that reflection electron-multiplying dynodes. The ampli-
is minimised at the interface between the fication gain is high (>106). High voltage
PMT and the scintillator. for the whole PMT array and gain for each
The thickness of the crystal impacts PMT have to be carefully fine-tuned in or-
on both sensitivity and spatial resolution. der to achieve a similar response on all the
As illustrated in Fig. 6, a thick crystal will FOV for both the energy of incident pho-
Figure 6
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
Illustration of the effect of the crystal thickness on light collection at the PMT array.
In (a), the crystal is thicker, improving the sensitivity but limiting the spatial resolution (the cone
of light collected at the PMT array is wider). In (b), not all of the photons interact with the crystal
but the spatial resolution is higher (i.e. the cone of light collected at the PMT is narrower).
(a) (b)
Figure 7
tons and the sensitivity. Furthermore, the the gantry. Variation in the PMT response
response of each detector has to be ho- according to angle may have an impact on
mogeneous independently of the angle of the reconstructed image.
Figure 8
10 16 8 10 16 8
9 X X X 9 4 2 3
14 X X X 14 2 10 2
11 X X X 11 4 4 3
(a) (b)
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
when reconstructed using FBP methods. ly, the number of subsets and the number
Iterative methods address this issue and, of iterations should be carefully balanced.
thanks to the performance of the most Indeed, if the number of subsets and it-
recent generations of computers, they are erations is too small, the algorithm does
now commonly used for SPECT image re- not converge and the result is a poorly
construction. The general concept of itera- contrasted and blurred image, while if the
tive methods is to solve p=A×f where p is number of iterations is too large in respect
a vector of the projection data (i.e. projec- to the number of subsets, the reconstruct-
tions acquired), f is a vector representing ed image displays an increased noise level
the three-dimensional distribution to be (Figs. 10 and 11).
reconstructed (i.e. tomographic images)
and A is a function of projection (i.e. geom- Other benefits of iterative reconstruc-
etry of acquisition). The principle of itera- tion are that the projection function may
tive algorithms is to find a solution (i.e. f) include different acquisition parameters in
by successive estimates. Ordered subsets order to optimise image rendering. Indeed,
expectation maximisation (OSEM) is the spatial resolution (both intrinsic and ex-
most commonly applied algorithm and trinsic, depending on the collimator used),
is based on the maximum likelihood ex- attenuation by the patient, Compton scat-
pectation maximisation (MLEM). The aim ter (in the patient, collimator and/or crys-
of MLEM algorithms is to find a solution Figure 9
as the best estimate of f than can produce
projections p with the highest likelihood.
OSEM allows acceleration of computation
by dividing the set of projections into sub-
sets (Fig. 9). Thus, independently from the
acquisition duration and the numbers of
projections, the main factors influencing
the image quality are the number of sub-
sets and the numbers of iterations. (a) (b)
Finally, the higher the number of itera-
tions, the higher is the spatial resolution, Examples of ordered subsets:
but with an increase in the noise level, a) set of projections divided into four subsets;
and the higher the number of subsets, b) set of projections divided into two subsets
Figure 10
Example of the poor contrast and blurring that occur if the number of subsets and iterations
is too small. Tomographic reconstruction of a bone study (99mTc), no attenuation reconstruction,
5-mm Gaussian post-filter, 64 angular positions (128 projections). Reconstruction was achieved
with OSEM, with 4 iterations and either 4 (a) or 16 subsets (b)
Figure 11
Example of the increase in noise level that occurs with a higher number of iterations. Tomographic
reconstruction of a bone study (99mTc), no attenuation reconstruction, 5-mm Gaussian post-filter,
64 angular positions (128 projections). Reconstruction was achieved with OSEM, 4 subsets and
an increasing number of iterations: 4 (a), 8 (b), 16 (c), and 64 (d)
tal) and collimator septal penetration can subject to higher attenuation on their
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
be modelled to optimise the reconstruc- paths. In other words, a photon emitted
tion. These reconstruction options with from the surface of the patient is less atten-
the possibility of applying post-processing uated than a photon emitted from a deep
filters must be implemented within acqui- location. As a consequence, the emission
sition and/or post-processing software to of photons from deeper locations are un-
facilitate the optimisation of image quality. derestimated on the projections (Fig. 12).
Figure 12
Example of photons emitted from different depths, with more attenuation occurring for photons
emitted from deeper locations
Figure 13
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
(a) (b)
(c) (d)
(e) (f )
Examples of images reconstructed without and with the use of attenuation correction based on
CT acquisition. a,b) A phantom filled with 99mTc solution and images reconstructed without (a)
and with (b) attenuation correction. c,d) A plain phantom with two sources of 177Lu in the centre
and images reconstructed without (c) and with (d) attenuation correction. e,f) Patient images
reconstructed from a 99mTc study without (e) and with (f) attenuation correction. The images clearly
show that without attenuation correction, events at depth are underestimated.
as the brain. For heterogeneous areas, an 256×256 up to 512×512. The size of the ac-
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
electronic density map of tissues acquired quisition matrix will define the size of the
from CT is preferred for attenuation cor- tomographic images to be reconstructed.
rection. Figure 13 shows examples of re- Thus, a higher matrix size will result in re-
construction without and with the use of cording of finer details within the limits of
attenuation correction based on CT acqui- the intrinsic spatial resolution. Matrix size
sition. selection directly affects the signal to noise
ratio (SNR) of acquired data. Indeed, larger
Matrix size and zoom matrix sizes lead to events being spread
Events detected on a projection are distrib- over a greater number of pixels; according-
uted on a grid, a so-called matrix. Each pic- ly, to maintain the same SNR with a larger
ture element (pixel) of the matrix basically matrix, the acquisition duration must be
covers the entire FOV and contains the longer so as to increase the number of
total number of events detected during events. Thus, acquisition duration and ma-
the acquisition. A matrix is described by trix size must be carefully balanced.
the number of rows (m) and columns (n) Zoom is used for the magnification of
defining its size: m × n. For SPECT imag- the object (e.g. specific body location,
ing, standard matrix sizes are 128×128 or Fig. 14) while keeping the matrix size un-
Figure 14
Illustration of a 4x4 acquisition matrix for the study of a kidney. a) The matrix covers the entire
FOV. b) A zoom of 2 is applied during the acquisition to achieve higher spatial resolution.
(a) (b)
age with a 128x128 matrix and a pixel size Optimisation of SPECT/CT acquisition re-
of 5 mm will allow a pixel size of 2.5 mm lies on multiple factors. Obviously, the pro-
with the use of a zoom of 2. Finally, the tocol acquisition may be optimised taking
zoom allows for acquisition of the same into account both the technical aspects
matrix size on a limited FOV area but the detailed above and the ALARA principle
duration of frame acquisition may need (whereby the radiation exposure is kept
to be altered in order to maintain the SNR. “as low as reasonably achievable”). Howev-
er, patient status must be taken into con-
sideration. Indeed, the patient’s age and
NEW ADVANCEMENTS IN SPECT comorbidities may influence the design
Today, the newest CT scan reconstruction of the acquisition protocol, which thus
technologies, including iterative recon- may vary depending on the context. For
struction, are contributing in reducing instance, the exploration of pelvic pain by
patients’ radiation exposure. Additionally, means of bone scintigraphy in an elderly
new detection technologies will also help patient largely justifies a CT acquisition
to minimise this exposure. Indeed, the with SPECT because (1) risk of fracture of
new detectors based on semiconductor the pelvis is high and the bladder uptake
technology (CZT: cadmium-zinc-telluride) may prevent study of the sacrum and (2)
directly convert emitted photons into an the CT dose exposure remains acceptable
electrical pulse without the use of a scintil- since SPECT alone contributes the major-
lator or PMT. Thus, CZT detectors increase ity of the patient’s overall exposure[8]. Fur-
spatial resolution, energy resolution and thermore, the duration of the SPECT/CT
sensitivity. Finally, the improved sensitiv- acquisition may be reduced if uptake on
ity of this technology, initially developed planar scintigraphy is high. On the other
for cardiology applications[6], means that hand, in a child with gait abnormalities,
less radiotracer activity is required for an acquisition of a SPECT/CT after planar
equivalent image quality and allows re- study may be considered unnecessary if a
duction of the study duration, thereby complementary examination is scheduled
avoiding movement artefacts and reduc- (e.g. MRI), thereby avoiding needless dose
ing overall patient exposure[7]. exposure.
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F S PE C T A ND S PE C T/C T –
REFERENCES
1. Peterson TE, Furenlid LR. SPECT detectors: The Anger 6. Mouden M, Timmer JR, Ottervanger JP, Reiffers S,
camera and beyond. Phys Med Biol 2011;56:R145– Oostdijk AH, Knollema S, et al. Impact of a new ul-
R182. trafast CZT SPECT camera for myocardial perfusion
2. Van Audenhaege K, Van Holen R, Vandenberghe S, imaging: Fewer equivocal results and lower radiation
Vanhove C, Metzler SD, Moore SC. Review of SPECT dose. Eur J Nucl Med Mol Imaging 2012;39:1048–1055.
collimator selection, optimization, and fabrica- 7. Agostini D, Marie PY, Ben-Haim S, Rouzet F, Songy
tion for clinical and preclinical imaging. Med Phys B, Giordano A, et al. Performance of cardiac cadmi-
2015;42:4796–4813. um-zinc-telluride gamma camera imaging in coro-
3. Tsui BM, Gullberg GT, Edgerton ER, Gilland DR, Perry nary artery disease: A review from the cardiovascular
JR, McCartney WH. Design and clinical utility of a fan committee of the european association of nucle-
beam collimator for SPECT imaging of the head. J ar medicine (EANM). Eur J Nucl Med Mol Imaging
Nucl Med 1986;27:810–819. 2016;43:2423–2432.
4. Gremillet E, Agostini D. How to use cardiac IQ•SPECT 8. Rausch I, Fuchsel FG, Kuderer C, Hentschel M, Beyer T.
routinely? An overview of tips and tricks from prac- Radiation exposure levels of routine SPECT/CT imag-
tical experience to the literature. Eur J Nucl Med Mol ing protocols. Eur J Radiol 2016;85:1627–1636.
Imaging 2016;43:707–710.
5. Bruyant PP. Analytic and iterative reconstruction al-
gorithms in SPECT. J Nucl Med 2002;43:1343–1358.
INTRODUCTION
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
Figure 1
A: A PET/CT image consisting of coronal whole-body CT image. B: PET image with CT attenuation
correction. C: Fused PET/CT image (C). Courtesy of the Journal of Nuclear Medicine Technology
(JNMT)[1]
Patient preparation for a whole-body and by the sides for head and neck im-
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
PET scan with 18F-FDG includes no con- aging. Elevating the arms when imaging
sumption of food, simple carbohydrates the torso will reduce beam-hardening ar-
or liquids other than plain (unflavoured) tefacts, as well as artefacts caused by trun-
water for 4 h prior to 18F-FDG injection for cation of the measured field of view (FOV).
the non-diabetic patient. It is also recom- If imaging the head and neck area, as well
mended that the patient should not con- as the torso, it is beneficial to acquire two
sume food after midnight the night before separate acquisitions — one with the arms
the study. A low carbohydrate diet is pre- elevated for torso acquisition and one
ferred for one day prior to the study[2]. with the arms by the sides for head and
The patient’s blood glucose level should neck imaging. In radiation therapy treat-
be evaluated prior to injection of 18F-FDG. ment planning, the position of the patient
For clinical studies involving the whole should be the same as for radiotherapy
body, the blood glucose level should be treatment set-up.
lower than 11 mmol/L (~200 mg/dL). Pa- Skull base to proximal thigh imaging is
tients with glucose levels out of this range recommended for most tumour imaging,
may be excluded from the study[3]. The typically from the external auditory me-
patient should be injected in a quiet area atus to the mid-thigh region. For tumours
and remain in that location with the lights with an increased potential for scalp, skull,
dimmed for 60–120 min post injection. or brain involvement or for lower extrem-
The patient should be kept warm and re- ity involvement, whole-body imaging is
quested to sit still during the waiting pe- performed. Limited-area imaging can be
riod. For the purposes of radiation safety performed when there is a specific area of
and compliance with ALARA, explanation interest.
of the procedure should be completed pri- PET acquisition and reconstruction
or to injection, thereby reducing radiation protocols should always begin with the
exposure to the technologist. factory-recommended settings. Any mod-
Before the patient is positioned on the ifications in protocol should be evaluated
table, he or she should be instructed to carefully to avoid affecting patient images
empty the bladder to enhance comfort in a negative manner. Over time, software
and decrease the opportunity for artefacts. upgrades and changes in professional rec-
The patient is then placed in the supine ommendations may occur that warrant
position, with arms elevated and support- modification of the factory-recommended
ed above the head for torso acquisition settings.
for use by the nuclear medicine commu- The acquisition time varies from 2 to 5
nity. Selection is based on patient popula- min or longer per bed position for body
tion, department needs, costs and facility imaging. Continuous bed movement
vendor preference. may also be utilised to image the pa-
Various professional organisations and tient, if the scanner has that capability.
societies have developed performance Acquisition time varies according to the
guidelines, including the American Col- administered activity, the patient body
lege of Radiology (ACR). The ACR-rec- weight, 2D versus 3D and the count rate
ommended scanner specifications are as capability and sensitivity of the PET scan-
follows: ner. The average imaging time for skull
to mid-thigh acquisition is 15–45 min. A
For the CT scanner: variation in time per bed position can be
a. Spiral scan time: <5 s (<2 s is preferable) used if the system is capable of that tech-
b. S lice thickness and collimation: nique. This enables a faster scan time by
<5 mm (<2 mm is preferable) permitting reduction in the time per bed
c. Limiting spatial resolution: >8 lp/cm position in areas outside of the torso be-
for >32-cm display field of view (DFOV) cause those areas have less attenuation.
and >10 lp/cm for <24-cm DFOV Administration of a higher activity can
also reduce the image acquisition time,
For the PET scanner: but compliance with ALARA principles
a. In-plane spatial resolution: <6.5 mm requires a lower dose and longer acqui-
b. A xial resolution: <6.5 mm sition. Formulas are available to calculate
c. S ensitivity (3D): >4.0 cps/kBq dose taking into account bed overlap and
d. S ensitivity (2D): >1.0 cps/kBq patient weight, and whether maximum or
e. U niformity: <5% minimum activity is used.
When using systems with a high count
For the combined PET/CT scanner: rate capability, 18F-FDG activity and scan
a. M
aximum co-scan range (CT and PET): duration for each bed position must
>160 cm be adjusted so that the product of the
b. M aximum patient weight: 18
F-FDG activity and scan duration ±10% is
>159 kg equal to or greater than the specifications
c. P
atient port diameter: >59 cm set out below[3]:
The figures for systems with the scanner. Acquisition systems with sep-
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
bed overlap of <25% are: ta are 2D scanners. Recent scanners devel-
»P
roduct of <MBq/kg × min/bed > oped in the early 2000s use 3D technology
27.5 for 2D scans and do not use septa[5]. Some scanners are
»P
roduct of MBq/kg × min/bed > capable of 2D and 3D acquisition[6].
13.8 for 3D scans 3D acquisition increases sensitivity four
to ten times due to an increase in lines of
The dosage is then calculated as follows: response (LORs) resulting from the detec-
»» FDG activity in MBq for 2D scans = tion of cross-plane events[7]. Removal of
27.5 × weight/(min/bed) septa in 3D acquisitions provides a higher
»» FDG activity in MBq for 3D scans = detection probability. Increased sensitivity
13.8 × weight/(min/bed) also results in increased random and scat-
ter contributions. To compensate, these
And for systems with scanners are also equipped with new and
a bed overlap of 50%: fast detectors using new scintillation crys-
»» Product of MBq/kg × min/bed > tals. Faster crystals provide better count
6.9 (3D only) rate performance. A short coincidence
»» FDG activity in MBq = 6.9 × weight/ timing window can be applied, which re-
(min/bed) sults in a reduction of random and scatter
coincidences. Due to low sensitivity at the
No matter what technique is used to deter- edge of the detector ring in 3D acquisition,
mine the administered activity, the activity bed overlap from one bed position to the
within the FOV should not exceed the peak next may be warranted. In continuous
count rate capability of the system. motion acquisition there is no need for
overlap, as there is uniform detector sen-
sitivity over the patient[6].
2D VERSUS 3D ACQUISITION
PET images can be acquired in 2D or 3D
mode. Earlier scanners used septa made DYNAMIC IMAGES
of lead or tungsten that were positioned Dynamic images may also be obtained
within the FOV. The septa provided a way when a specific area is to be imaged as a
to limit photons emitted or scattered out- flow study. A bed position of 15–26 cm
side the transverse or transaxial plane to is capable of imaging smaller regions. Dy-
the detector. This limited the sensitivity of namic PET images are similar to dynamic
aging acquires complete volume images Reported data should include the type of
within a set time frame. Typical times for isotope utilised, the patient’s height and
evaluating arterial flow are 2 s per image body weight, FDG activity administered,
followed by a longer imaging time to assay time and time of injection. This
demonstrate further uptake of the tracer aids in standardised uptake value (SUV)
over time[7]. reporting. SUV is used to quantitate ac-
tivity in body organs, abnormalities or
the total body and supplements visual
LIST MODE interpretation of an area. It is derived as
If the system is capable of it, images may follows[8]:
be acquired in list mode. Events, timing
markers and physiological gating markers (Region of interest (kBq/mL)
are stored as lists in list mode; the data SUVBW =
Injected activity with
may be rebinned (recreated) during use of decay-corrected residual
cardiac or respiratory gating. These data subtracted (MBq/kg)/patient
can then be formatted into sinograms for weight (kg) where BW is
reconstruction. Time of flight scanners are body weight.
capable of acquiring in list mode; there-
fore, list mode may become the standard
method for acquiring data[7].
ACQUISITION PARAMETERS
Further instructions on obtaining the
DUAL TIME POINT emission image include that the online
For certain disease states, dual time point randoms correction should be based on
imaging may be beneficial. This involves the “delayed coincidence time window”
an initial scan, a waiting period and then technique or randoms correction using
a delayed scan. In the case of tumour im- a model based on (block) single count
aging there will be an increase in activity rates. This has an effect on image con-
over time, whereas in infections there will trast. Decay correction must be ‘on’, as
be a decrease over time[6]. Thus delayed well. Shorter image times result because
imaging can increase the specificity of of the increase in contrast that occurs
whether activity is due to either tumour or from the high-quality image obtained
inflammation. with time of flight[3].
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
The typical PET/CT scan involves the ac- of less than 1.0 results in overlap, while
quisition of a topogram followed by single a pitch of 2.0, or higher, results in spaces
or multiple helical CT scans. CT can be between slices, equal to the thickness of
performed for attenuation correction, ana- the slice. Pitch is inversely proportional to
tomical alignment or acquisition of an op- patient dose. Increasing the table speed
timised diagnostic scan. When obtaining or pitch reduces the radiation dose in pro-
attenuation-corrected/anatomical align- portion to the increase in pitch (unless the
ment images, a low milliampere-seconds scanner automatically increases mA with
setting is used to reduce the radiation increase in pitch). However, image quality
dose. In an optimised diagnostic CT scan, will be reduced[9].
standard CT milliampere-seconds settings Slice thickness and increment define
are used. This results in better spatial res- the positional relationship of image slices
olution of the CT scan. Modulation of the and the amount of separation between
tube current may also be used to reduce the slices or potential overlap. If the slice
radiation dose in patients without metal thickness is equal to the increment, there
implants[3]. This can be accomplished with is no separation. If the slice thickness is
dose reduction software developed by the less than the increment there is overlap.
manufacturer of the scanner, when avail- It is beneficial to have some overlap when
able, and enables optimal low mA and kV attempting to detect smaller lesions[9].
settings to be used. It is always the case Regarding rotation time, at slower times
that selection of CT acquisition param- a compensatory increase in mA is required
eters, such as pitch, rotation time, slice to add more X-rays to the image. The
thickness, increment, voltage and tube slower the rotation time, the more blurred
current, depends on the purpose of the the image. A faster rotation time is ben-
CT scan. If the CT image is obtained solely eficial when body motion is a concern.
for the purpose of attenuation correction, Increasing the rotation time will result in
voltage and current can be reduced to better resolution because each image is
near the lowest the scanner is capable of generated from more projections[9].
producing and still be acceptable. Other CT acquisition parameters are
Pitch is the ratio of table movement voltage and tube current. Denser areas re-
through the gantry during one 360° rota- quire a higher voltage. This voltage is mea-
tion, relative to beam collimation. The low- sured in kVp. The higher the mA setting, or
er the pitch, the more the overlap. A pitch tube current, the higher the quality of the
image acquired; however, radiation dose no longer available. Since 2006, scanners
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
Pitch 0.8 s
a
Check vendor specifications for scanner utilised. For
acquisition and reconstruction protocols involving
Reconstruction for AC B30s, medium smooth
the brain and heart, please refer to the Tech Guide
FOV 780 mm
publications by the EANM on Brain Imaging and
Reconstruction for B30f medium smooth Myocardial Perfusion Imaging
imaging 5×5 mm, 500 mm FOV
a
Check vendor specifications for scanner utilised. For
acquisition and reconstruction protocols involving
the brain and heart, please refer to the Tech Guide
publications by the EANM on Brain Imaging and
Myocardial Perfusion Imaging
Figure 2
USE OF CONTRAST
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
Contrast may be used when performing the
CT portion of the scan to enhance demon-
stration of the anatomy. Contrast can be
administered orally or intravenously. Oral
contrast is administered when the area of in-
terest is in the gastrointestinal tract. Intrave-
nous (IV) contrast is administered when vas-
cular structures need to be differentiated[3].
The patient’s creatinine level and/or glo-
merular filtration rate should be tested pri-
or to administration of IV contrast in order
to reduce the possibility of damage to the
kidneys. Metformin, an oral hypoglycae-
mic agent, should be discontinued for 48
h post IV contrast injection[6].
BREATHING PROTOCOL
The CT for attenuation correction should be A curvilinear cold artefact (arrow) is commonly
acquired during the resting phase or shallow seen at the dome of the diaphragm/liver or at
breathing. If the CT scanner has six or fewer the lung base because of a respiration mismatch
detector rings, use of a breath-hold protocol on PET images with CT attenuation correction.
Courtesy of the Journal of Nuclear Medicine
in normal expiration should be considered
Technology (JNMT)[1]
for the duration of scanning of the thorax
and upper abdomen. Deep breath-hold for
chest CT acquisition can cause misregistra- facturer and age of the equipment. There is
tion artefacts on the emission image (Fig. 2). typically a proprietary file format, unique to
each vendor and model, which contains the
raw data. Raw data are first formatted into
IMAGE FORMATION AND sinograms. The sinograms represent the
RECONSTRUCTION activity across the detector at each projec-
The method of reconstruction of PET imag- tion angle. The sinograms are reconstructed
es can vary greatly according to the manu- into axial image slices and corrections are
applied. Other projection planes are then the process of back-projecting data across
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
created from the PET and the CT, to include the imaging matrix. It is a quick and sim-
coronal and sagittal. The CT scan data are ple method. However, the produced im-
used to create an attenuation correction ages suffer from extra noise and streaking
map for use in image reconstruction. artefacts which result in an overall reduc-
tion of contrast. FBP has consequently
been replaced by iterative methods.
IMAGE RECONSTRUCTION Iterative reconstruction is the process
METHODS of creating estimates of actual radiophar-
PET/CT image reconstruction is similar to maceutical distribution in an object. Each
SPECT. Image reconstruction is a process projection estimate is compared with the
of calculating the 3D activity distribution measured projection data, and the differ-
from the sinograms. During reconstruc- ence between them is then used to create
tion the data are normalised, with appli- a new estimate[10]. This process continues
cation of corrections for random coin- until the difference between the esti-
cidences, scatter radiation, attenuation, mate and the measured values reaches a
dead time and decay. Reconstruction specified value. A point is reached in the
algorithms and order of application vary number of iterations where there will be
between equipment manufacturers and no further improvement in image quality.
may be proprietary. Once this point has been reached, further
Reconstruction algorithms are general- iterations can actually begin to degrade
ly classified into two methods: analytical the image quality. It is important to deter-
and iterative. Image reconstruction differs mine the number of iterations which will
for 2D and 3D imaging. One sinogram is yield optimal image quality. The number
created for each 2D slice during 2D recon- of iterations is preset by the user. While
struction and one sinogram is created for iterative reconstruction produces images
each projection plane during 3D mode with less noise and artefacts than FBP, it is
image reconstruction[9]. The majority of computationally intense and takes signifi-
current scanners are 3D; therefore 3D im- cantly longer than FBP.
age reconstruction will predominantly be Maximum likelihood expectation max-
discussed in this chapter. imisation (MLEM) was one of the first
In the past, the most commonly used iterative algorithms. MLEM requires the
analytical image reconstruction method use of all the image data. The process
was filtered back-projection (FBP). FBP is proved to be time consuming and oth-
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
faster and more efficient. Ordered subset by completion of the process of iterative
expectation maximisation (OSEM) is one reconstruction. Methods available to com-
of the most common iterative algorithms. plete this conversion of data from 3D to 2D
OSEM does not require the use of all the include single slice rebinning (SSRB) and
image data; instead only a subset of the Fourier rebinning (FORE). SSRB rebins 3D
projection data is used, resulting in a dras- data into 2D projection sinograms. SSRB
tic reduction in reconstruction time. Two results in a blurring of image data and is not
approaches are used: (a) a higher number commonly used today since there are bet-
of subsets (8–24) and small number of it- ter approaches. FORE incorporates a more
erations (2–8) and (b) fewer subsets (2–8) accurate process for converting data from
but more iterations (8–24). Both methods 3D to 2D projection data and is therefore
maintain the image quality despite the the most widely used technique. Rebin-
reduction in reconstruction time. In clin- ning results in reduced resolution as the
ical practice, use of fewer iterations and distance from the axial centre of the FOV
increased subsets appears to be the most increases. Full 3D reconstruction does not
widely used approach. Other iterative require the conversion of data and is the
methods have been developed by propri- optimal method to avoid the resolution
etary vendors as well[7]. loss associated with rebinning. Fully 3D
Use of 3D image reconstruction requires reconstruction is only available on mod-
additional steps beyond the requirements ern scanners but is the standard on new
of 2D image reconstruction. Reconstruction equipment. There is a significant increase
options for 3D image data are determined in data in the reconstruction process due
by the age of the scanner and software lim- to the increase in the number of LORs con-
its. The reconstruction of 3D data requires tained in each sinogram, which means that
that the raw data, 3D sinograms, be either additional computer components must be
rebinned into 2D information or recon- devoted to this step[11]. The process of full
structed completely as a full 3D volume[10]. 3D reconstruction is very complicated and
Significantly more data are acquired in 3D is beyond the scope of this chapter.
mode due to the increase in sensitivity. Along with reconstruction algorithms, the
Reconstructing 3D data in full 3D mode is data must also undergo filtering. The filter
computationally intense and currently this most commonly used in PET whole-body
option is only available on newer scanners. imaging, for the purpose of smoothing the
Traditional methods require 3D data to be image, is the Gaussian filter. This filter is not
like most filters used in nuclear medicine; Time of flight (ToF) PET systems are not a
rather, it is defined by its full-width half-max- new concept. ToF PET coincidence imag-
imum (FWHM) in pixel space, which is more ing was used during the 1980s on exper-
commonly known as the spatial domain. imental systems[12]. The first PET scanners
The width of the filter is measured in milli- used bismuth germinate (BGO) crystals,
metres. Increasing the FWHM on the Gauss- which have a low stopping power and
ian filter will result in a smoother image. A limited spatial resolution and cannot sup-
high-resolution image can be produced by port ToF PET. ToF systems were limited to
applying a FWHM value of 5 mm. research until the early 2000s. Current PET
systems provide scintillators with a higher
density, shorter decay time and modern
SCATTER CORRECTION electronics which allow for faster com-
Scatter correction is essential to achieve puting power. Lutetium oxyorthosilicate
accurate PET data. In 3D acquisition (LSO) and lutetium-yttrium oxyorthosili-
modes, scatter coincidences are even cate (LYSO) both have good timing reso-
greater than in 2D mode. Therefore, scat- lution and increased stopping power and
ter correction methods are even more energy resolution. ToF PET is common on
important. There are two scatter correc- today’s scanners. ToF PET allows for more
tion methods: energy window-based accurate determination of where an event
methods and calculation-based methods. has taken place. If an event occurs any-
Accurate scatter correction is difficult to where other than the midline of an LOR,
determine. Energy window-based meth- the crystal closer to the annihilation event
ods require acquisition of a second en- will detect its photon first. The time differ-
ergy window below the original energy ence between the arrival of the two pho-
window. Calculation-based methods are tons is used to determine where along
preferred. Such methods are modelled the LOR the event took place. As a result,
rather than measured, the amount of scat- ToF PET increases image resolution. The
ter correction depending on the model greatest benefit is found in larger patients,
applied. Single scatter simulation is a who suffer most from poor image quality.
currently used calculation-based method. ToF PET requires additional considerations
Scatter correction is applied during image during reconstruction. The addition of ToF
reconstruction and should take place prior data prompts a shift from sinogram to list
to attenuation correction. mode reconstruction[12].
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
CORRECTION plants, respiratory motion, contrast medi-
Point spread function (PSF) correction is a um and truncation.
software option that corrects for the blur-
ring arising from mispositioning of events Partial volume effect
due to variation in depth of detection The partial volume effect is the apparent
within the patient. Some vendors refer to loss of density or concentration of a small
such correction as high-definition imag- object which occurs on PET and CT when
ing. PSF corrections are applied during re- lesions are small compared with the reso-
construction as part of the reconstruction lution of the scanner. The size of an object
algorithm. Detections along the edges of or lesion needs to be approximately three
the axial FOV can be slightly misplaced, times greater than the image resolution for
resulting in blurring of small outer lesions. the activity to be accurately represented[7].
PSF corrects for the blurring and enhances
visualisation of these lesions. PSF correc- Metallic implants
tions are available on new scanners. Streaking artefacts are generated when me-
tallic objects exist within the patient, such
as dental fillings, hip prosthetics or chemo-
PITFALLS AND therapy ports. The metallic object leads to
ARTEFACTS an overestimation of PET activity in the re-
Due to issues beyond the control of the gion of the object when the CT attenuation
technologist, common artefacts can be correction map is applied. This will lead to
demonstrated on the PET/CT scan due a false positive PET finding. However, not
A) Example of the way in which high-density metallic implants generate streaking artefacts and high CT
numbers (arrow) on CT images. B) High CT numbers will then be mapped to high PET attenuation
coefficients, leading to overestimation of activity concentration. C) PET image without attenuation
correction helps rule out metal-induced artefacts. Courtesy of the Journal of Nuclear Medicine
Technology (JNMT)[1]
all metallic objects produce false positives. use of the CT image to define the location
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
For example, in the case of hip prosthetics, of the lesion will help in interpretation. To
the implants produce high CT numbers minimise respiratory motion artefacts, the
because of their high photon absorption. breathing procedure should be explained
These implants will also attenuate the PET to the patient thoroughly before begin-
511-keV photons. This results in no emis- ning the study (Fig. 4).
sion data and demonstration of a cold area.
When CT attenuation correction is per- Figure 4
formed, the PET images show diminished
uptake in that area. In these cases, the
non-attenuated PET images can be useful
for interpretation of the study (Fig. 3).
Respiratory motion
Due to the nature of the human body, re-
spiratory motion is the most common arte-
fact demonstrated on a PET/CT image. The
artefact is due to the difference between
the chest position on the PET image and
the chest position on the CT image. The A 58-year-old man with colon cancer.
image resulting from a PET scan is an aver- A) Lesion at the dome of the liver is mislocalised
age of many breathing cycles while the CT to the right lung (arrow) because of respiratory
image is obtained at a single stage of one motion. B) Image without attenuation
breathing cycle. To partially compensate correction shows that all lesions are confined to
for this possible artefact, the CT image can the liver. Courtesy of the Journal of Nuclear
Medicine Technology (JNMT)[1]
be obtained at mid-expiration, at mid-in-
spiration or during shallow breathing. The
image produced has a curvilinear cold
area at the lung–diaphragm interface. The Contrast media
most critical impact of this artefact is on liv- As previously mentioned, contrast agents
er lesions. Because of this motion, a liver can be administered in PET/CT studies.
lesion may appear to be in the lung, simu- Contrast agents such as barium sulphate
lating a lung nodule. This type of artefact is and iodine that are within the patient from
a misregistration of lesions. In these cases, a previous study can give rise to a false pos-
itive finding on the PET image because they mon in larger patients or patients imaged
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
mimic metallic object artefacts. The higher with their arms down by their sides. When
the concentration of the contrast, the high- a patient extends beyond the CT FOV, the
er the CT number, resulting in higher PET at- extended anatomy is truncated and not
tenuation coefficients. This causes an over- represented in the reconstructed CT im-
estimation of the PET tracer uptake, which age. This results in no attenuation correc-
A 61-year-old patient with lung cancer who ingested barium for an oesophagogram one day
before PET/CT scan. A) Concentration of contrast medium in colon (arrow) is increased because
of significant water reabsorption, shown on the CT image. B) High CT numbers of residual barium
overcorrect attenuation of PET emission data and mimic increased 18F-FDG uptake on the PET image
with CT attenuation correction. C) No increase in 18F-FDG uptake is seen on the image without
attenuation correction. Courtesy of the Journal of Nuclear Medicine Technology (JNMT)[1]
is the reason for the false positive study. A tion values for the corresponding region.
technologist should be aware of whether The consequence is a bias on the PET at-
the patient has had a previous study involv- tenuation-corrected images, which under-
ing use of contrast. In the case of prior con- estimate the SUV in the region in question.
trast use and presence of the resulting arte- Truncation also produces streaking arte-
fact, the non-attenuated PET images can be facts at the edge of the CT image, resulting
used for interpretation (Fig. 5). in an overestimation of the attenuation
coefficients to be used to correct the PET
Truncation data. To avoid this artefact, the patient
Truncation artefacts occur because of the should be positioned in the centre of the
difference in size of the FOV between the FOV and imaged with the arms above his
CT and PET images. They are most com- or her head, when possible (Fig. 6).
A 54-year-old man with a history of metastatic melanoma (arrow). The CT image (A) appears truncated
at the sides and biases the PET attenuation-corrected image (B). Courtesy of the Journal of Nuclear
Medicine Technology (JNMT)[1]
ACQ UI S I TI O N A ND RE CO NS TRUC TI O N
O PTIM I S ATI O N O F PE T/C T —
REFERENCES
1. Sureshbabu W, Mawlawi O. PET/CT imaging artifacts. 7. Christian PE, Swanston NM. PET study guide. 2nd ed.
J Nucl Med Technol 2005;33:156–161. Reston, VA: SNMMI; 2015.
2. Boellaard R, O’Doherty MJ, Weber WA, Mottaghy 8.
Thie JA. Understanding the standardized uptake
FM, Lonsdale MN, Stroobants SG, et al. FDG PET and value, its methods, and implications for usage. J Nucl
PET/CT: EANM procedure guidelines for tumour PET Med 2004;45:1431–1434.
imaging: Version 1.0. Eur J Nucl Med Mol Imaging 9. Seeram E. Computed tomography: Physical princi-
2009;37:181–200. ples, clinical applications, and quality control. 3rd ed.
3. Boellaard R, Delgado-Bolton R, Oyen WJG, Giam- United States: Elsevier Health Sciences; 2008.
marile F, Tatsch K, Eschner W, et al. FDG PET/CT: EANM 10. Prekeges J. Nuclear medicine instrumentation. 2nd
procedure guidelines for tumour imaging: Version ed. Sudbury, MA: Jones & Bartlett; 2012.
2.0. Eur J Nucl Med Mol Imaging 2014;42:328–354. 11. Lartizien C, Kinahan P, Swensson R, Comtat C, Lin
4. The American College of Radiology. ACR–SPR M, Villemagne V, Trébossen R. Evaluating image
practice parameter for performing FDG-PET/CT reconstruction methods for tumor detection in
in oncology. 2016. https://www.acr.org/~/me- 3-dimensional whole-body PET oncology imaging.
dia/71B746780F934F6D8A1BA5CCA5167EDB.pdf. J Nucl Med 2003;44:276–290.
Accessed January 12, 2017. 12. Vandenberghe S, Mikhaylova E, D’Hoe E, Mollet P,
5. Townsend DW. Combined PET/CT: the historical per- Karp JS. Recent developments in time-of-flight PET.
spective. Semin Ultrasound CT MR 2008;29:232–235. EJNMMI Phys 2016;3:3.
6. Hogg P, Testanera G, eds. Principles and practice
of PET/CT. Part 1. A technologist’s guide. EANM;
2012. http://www.eanm.org/content-eanm/up-
loads/2016/11/gl_Principles_and_Practice_of_PET-
CT_Part_1.pdf. Accessed January 11, 2017.
by Giorgio Testanera and Michel de Groot I N S TRUME N TATION AND PROTO CO L S TA NDA RDIS AT IO N
QUA L I T Y CON TROL OF NUC LE A R ME DIC INE
CHAPTER 8
INTRODUCTION
Multiple research studies have shown the benefit of combined positron
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
Both quantification and the development greatest tools for establishing the position
of new and promising PET radiopharma- of PET/CT as a state of the art modality in
ceuticals have become more important oncology patient care within the compet-
in clinical decision making and improved itive environment of diagnostic imaging.
quality of care for the individual patient[1, 2]. Owing to its widespread use in tumour
The use of FDG-PET as a surrogate tool response assessment studies, the SUV, de-
for monitoring of therapy response offers fined as the activity concentration ratio in
better patient care by allowing individu- tissues, can be considered the measure to
alisation of treatment and avoidance of be used in order to standardise PET scans.
ineffective treatments: surgery, radiation Absolute SUV is used for definition of pa-
therapy or chemotherapy. Imaging bio- tient eligibility, patient stratification and
markers are of paramount importance, lesion selection, while relative SUV (per-
not only for patient evaluation but also for centage changes from pre-treatment val-
drug development. ues) can also be used to evaluate response
to therapy. However, SUV measurements
present uncertainties that may result in
QUANTITATIVE PET/CT resistance to their use in clinical trials of
Quantification in PET/CT can be defined cancer therapies. Possible causes of un-
as the use of a combination of methods certainties are as follows[3]:
to enhance image resolution, reduce im-
age noise and perform data analysis, and Technical factors
use of the maximum standardised uptake »»
Relative calibration between the PET
value (SUVmax) can be considered as the scanner and dose calibrator (10%)
standard quantitative method in practice. »» Residual activity in the syringe (5%)
Assessment of cellular uptake is one of the »» Incorrect synchronisation of clocks (10%)
»» Injection vs. calibration time (10%) within the scope of EANM Research Lim-
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
»» Quality of administration (50%) ited (EARL) activities, was originally based
on FDG-PET and PET/CT: EANM procedure
Physics-related factors guidelines for tumour PET imaging: version
»» Scan acquisition parameters (15%) 1.0, published in the European Journal of
»» Image reconstruction parameters (30%) Nuclear Medicine and Molecular Imaging
»» Use of contrast agents (15%) (EJNMMI). This widely accepted guideline,
»» ROI (50%) revised to version 2.0 in 2014, aims at har-
monising quantification in multicentre
Biological factors studies and providing a minimum stan-
»» Uptake period (15%) dard for the acquisition and interpretation
»» Patient motion and breathing (30%) of PET and PET/CT scans obtained with
»» Blood glucose levels (15%) FDG. The guideline specifically addresses
patient preparation, dosage of FDG as a
Furthermore, the variability in methodolo- function of scanner model, patient weight
gy across centres hinders the exchange of and scan duration, data acquisition, image
SUV measurements. Therefore, standardi- reconstruction, data analysis and QC pro-
sation of FDG-PET whole-body procedures cedures[2, 4].
is essential in multicentre trials. Accord-
ingly, the nuclear medicine community is The FDG-PET/CT accreditation ensures
working to achieve standardisation of PET/ harmonised quantitative performance
CT imaging and to improve the quality of of PET/CT systems within a multicentre
the images generated, both in clinical rou- setting through the standardisation of ac-
tine and in clinical trials [1, 2]. quisition and processing of PET/CT scans.
This rigorous harmonisation of imaging
systems enables PET/CT sites to compare,
EARL exchange and combine FDG-PET/CT find-
In 2010, the European Association of Nu- ings as data are collected and processed.
clear Medicine (EANM) initiated a pro- SUVs can also be reliably used owing to
gramme for the accreditation of PET/CT the resultant reduction in inter- and in-
scanners using FDG in order to support tra-institute variability. The standardisation
compliance with requirements regarding achieved by the accreditation programme
quality control (QC) and quality assurance relates to imaging procedures and meth-
of PET/CT systems. The programme, run odology, including patient preparation,
scan acquisition and image processing gramme. The Italian Lymphoma Founda-
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
and analysis, which is of the utmost im- tion (FIL), Lymphoma Study Association
portance for quality assurance in daily (LYSA) and GELA (Groupe d’Etude des
clinical practice as well as in multicentre Lymphomes de l’Adulte, i.e. Adult Lym-
trials[2]. phoma Study Group) are further groups
that have taken the EANM imaging guide-
line as the basis for PET/CT QC within their
COLLABORATION WITH OTHER studies. The Society of Nuclear Medicine,
INSTITUTIONS India (SNMI) and the Brazilian Society of
The EARL FDG-PET/CT accreditation pro- Nuclear Medicine (Sociedade Brasileira
gramme is strongly supported by the de Medicina Nuclear/SBMN) have also
European Organisation for Research and expressed great interest in the EARL FDG-
Treatment of Cancer Imaging Group (EO- PET/CT accreditation programme.
RTC-IG), which was founded in 2009 to en- The EANM imaging guideline is wide-
sure standardisation of image acquisition ly accepted and is also well represented
and quality assurance for EORTC clinical within the draft of the Uniform Protocols
trials with regard to CT, PET, magnetic res- for Imaging in Clinical Trials (UPICT), es-
onance imaging (MRI) and other imaging tablished by the Quantitative Imaging
modalities as they become available[5]. Biomarkers Alliance (QIBA).
EANM/EARL and the International Atom-
ic Energy Agency (IAEA) are also working
together with respect to quality stan- MULTICENTRE CLINICAL TRIALS
dards in nuclear medicine. EANM/EARL is The goal of the EARL FDG-PET/CT ac-
aware of the different harmonisation and creditation programme is to enhance the
standardisation efforts regarding PET/ quality standard of PET/CT investigations
CT worldwide and is eager to contribute for both daily use and multicentre studies.
to a possible alignment of accreditation FDG-PET/CT accreditation ensures similar
programmes. To this end, discussions are performance of PET/CT systems within a
being held with the Society of Nuclear multicentre setting through the harmoni-
Medicine and Molecular Imaging (SNMMI) sation of acquisition and processing of PET/
Clinical Trials Network (CTN). CT scans. Centres can compare, exchange
Furthermore, the Japanese Society of and combine FDG-PET/CT findings, includ-
Nuclear Medicine (JSNM) has adopted ing SUVs, since data are collected and pro-
the basis of the EARL accreditation pro- cessed in a standardised manner[2].
In order to ensure that data can be ex- sation. Centres seeking accreditation also
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
changed between centres participating in benefit from:
a multicentre trial organised in collabora- »» EARL’s knowledge, which is based on the
tion with EARL, it is essential that all PET/ contributions of worldwide-recognised
CT systems used in the trial are accredited imaging experts, with a global leading
by EARL. Furthermore, all centres should role in PET/CT research, who provide ad-
be required to apply the EARL-approved vice within the programme and monitor
parameter settings (used for processing its development.
of the QC phantom images) for the recon- »» Imaging results that can be compared,
struction of imaging during the course of exchanged and combined, since the
the trial. The sites can apply additional re- prerequisite for evaluation of imaging re-
constructions for other purposes, local use sults within preclinical and clinical (mul-
or diagnostic interpretations. The EANM ticentre) trials is comparable scanner
provides recommendations for acquisition performance across multiple sites (re-
times per bed position in combination duction of inter- and intra-institute vari-
with dosage per kg patient weight for var- ability in SUV results; provision of lower/
ious types of PET/CT system in the EANM upper limits of recovery coefficients; cal-
imaging guideline[2]. ibration factor within ±10%).
»» Accurate, reproducible and quantitative
An up-to-date list of accredited centres assessment enabled through standard-
participating in the EARL FDG-PET/CT ac- isation of methodology, including pa-
creditation programme, also known as the tient preparation, scan acquisition, im-
Centres of Excellence (CoE), can be found age processing and analysis.
via the EARL website. »» Reliable and quantitative imaging biomark-
er results generated within multicentre
clinical trials, leading to an enhanced
ACCREDITATION BENEFITS outcome (e.g. information on biological/
The current established and scientifical- pathological processes and response to
ly validated FDG-PET/CT accreditation therapeutic intervention) and thus ac-
programme provides independent qual- celerating compound development and
ity control/assurance within multicentre approval by the regulatory agencies and
clinical trials and specifically addresses lowering costs in the long run[4].
the needs of the pharmaceutical industry »» An exploratory further optimisation,
regarding harmonisation and standardi- presently being evaluated by EARL. This
the administered FDG activity for PET/ requested at fixed quarterly intervals to
CT systems with higher sensitivity or meet the standard requirements, as de-
improved performance using new en- scribed in the most recent versions of the
hanced technology (e.g. better time- EANM imaging guideline and manual, in
of-flight performance, continuous bed order to retain their accreditation. Failure
motion or extended axial field of view, to adhere to the deadlines for submission
i.e. length of bed position). may lead to EARL putting a centre’s ac-
creditation on hold. The documentation in
the initial accreditation procedure needs
PROCESSES INVOLVED IN to be submitted only once, whereas the
ACCREDITATION QC measurements and documents need
Sites which are seeking EARL FDG-PET/CT to be regularly performed and submitted
accreditation for the first time need to pass to retain the accreditation. An update of
through the initial procedure. This proce- the online questionnaire in the first quar-
dure includes submission of the online ter of each year is a prerequisite for sub-
questionnaire and signing of the statement mission of the QC documents[2, 6].
and signet policy. Additionally, sites are
asked to perform and submit QC data with-
in the subsequent 3 weeks. For this pur- SPECIFIC QUALITY CONTROLS
pose, sites have to perform calibration QC
measurements using a cylindrical calibra- Calibration QC
tion phantom and image quality QC mea- Calibration QC measurements should be
surements using a NEMA NU2-2001/2007 performed for initial accreditation and
image quality phantom and to submit on- have to be repeated every 3 months. The
line the image data in DICOM format and aim is to verify that the average activi-
the results to EARL. A detailed step-by-step ty concentration and/or SUV within the
procedure is published in the FDG-PET/CT phantom is within 10% of the expected
accreditation manual to assist sites in ade- value. For this purpose a cylindrical cali-
quately performing these measurements. bration phantom of any dimension but
After submission and independent review with a precisely known volume is required
of the results by EARL, the site is granted (it is preferable to obtain the phantom via
EARL FDG-PET/CT accreditation, assuming the manufacturer, ensuring that it is spe-
the results meet EARL requirements [2, 6]. cific for the used PET/CT model; if this is
not possible, the phantom should be 20 ly 70 MBq (65–75 MBq) FDG. The FDG ac-
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
cm in diameter and 20–30 cm long) (Fig. tivity is dispensed into the phantom when
11). The process of cross-calibration de- it is completely full with water, ensuring
termines the correct and direct (cross- or that all the activity is in the phantom by
relative) calibration of the PET/CT system flushing the syringe a few times. The phan-
with the institution’s own dose calibrator tom must then be vigorously shaken in
or against another one which is used to order to homogenise the distribution of
determine patient-specific FDG activities[4]. the activity.
Figure 1
Calibration QC setup (© Giorgio Testanera and Fabio Provenzano, Humanitas Research Hospital)
All relevant information on the standard After positioning the phantom in the
operating procedure (SOP) needs to be centre of the gantry, a PET/CT scan consist-
recorded accurately, including the exact ing of at least two PET bed positions needs
phantom volume, the scanner hardware to be acquired. PET/CT scan acquisition
and software, the exact doses and the time and reconstruction should be performed
of performance of every step. identically to patient studies as prescribed
It is first necessary to prepare a 5- to 10- in the clinical protocol. However, for statis-
ml syringe with an activity of approximate- tical reasons somewhat longer emission
times (e.g. 5–10 min per bed position) » to determine/check the correctness of
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
Figure 2
recovery coefficients using the NEMA mixed with the FDG contained in one
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
NU2-2007 image quality phantom (Fig. of the syringes (20 MBq), flushing the
3). The recovery coefficient is determined syringe several times and homogenis-
as a function of the sphere size, using the ing the solution (Fig. 4). After this step,
maximum pixel value and the A50 vol- the different diameter spheres inside the
ume of interest (VOI). As for calibration phantom should be filled, leaving no air
QC, all relevant information on the SOP inside of the spheres.
must be recorded accurately, including
the exact volume of the phantom, the The next step is to remove 30 ml wa-
scanner hardware and software, the ex- ter from the background compartment
Figure 3 Figure 4
act doses and the time of performance of of the phantom and to add 20 MBq of
every step [4, 6]. FDG, flushing several times to ensure that
The first step is to prepare two syring- all activity has been dispensed from the
es, each containing 20 MBq of FDG at syringe into the phantom. At this point,
the expected phantom acquisition time. the background compartment should be
To prepare the solution for use in the filled entirely with water and the solution
spheres, a bottle should be filled with homogenised by shaking the phantom
exactly 1,000 ml of water, which is then vigorously (Fig. 5).
Figure 5 Figure 6
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
Figure 7
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
Typical image quality phantom images
Figure 8
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
THE E A NM E A RL PR OJ E C T
ACC RE D I TATI O N FO R C L I NI C A L TRI A L S:
cross-calibration factor between the PET/ technologists to develop this new mindset
CT system and the dose calibrator. This fac- and to raise operating standards. There is
tor needs to be within ±10% from 1. The a great ongoing debate in Europe regard-
image quality QC phantom measurements ing the basic competencies and advanced
are analysed in order to determine the practice of nuclear medicine technolo-
background calibration factor by using sev- gists. A consensus document[7] has distin-
eral VOIs placed in the uniform background guished two levels, as follows:
compartment. The maximum allowable
calibration deviation is again ±10% from 1. Entry level
Additionally, the SUV recovery coefficients »» A competence and skill set that is con-
are calculated for each sphere (Fig. 8). A sidered necessary to ensure that nuclear
cold spot recovery using a central insert to medicine procedures are conducted to
verify the accuracy of scatter correction is an appropriate level
also performed. The software automatically »» This competence and skill set would be
compares the results of both QC phantom acquired during basic training/formative
analyses with the EARL specifications [2]. professional education
Advanced practice
TECHNOLOGIST’S ROLE »» A competence and skill set that is ac-
AND DAILY PRACTICE quired after basic training
In PET centres, technologists have the ma- »» The competence and skill set would be
jor role of acquiring images with PET and at a higher cognitive and clinical level
PET/CT scanners and ensuring that the ac- than basic training/formative profession-
quired images meet the quality standards al education
established by departmental criteria. The »» The competence and skill set would
first step in assuring that quantitative im- seek to improve patient care and man-
aging is accurate is compliance with all QC agement
procedures stipulated for the PET scanner »» The competence and skill set would
by national regulations, the vendor and seek to offer clinical career progression
local regulations. Working at an accred- opportunities
ited site for clinical trials implies not only
greater accuracy in standard daily practice It is true that the technologist’s role in EARL
but also a different mindset. Involvement accreditation can be considered a basic
REFERENCES
1. Boellaard R, Oyen WJG, Hoekstra CJ, Hoekstra OS, 4. Boellaard R, Delgado-Bolton R, Oyen WJG, Giam-
Visser EP, Willemsen AT, et al. The Netherlands pro- marile F, Tatsch K, Eschner W. FDG PET/CT: EANM
tocol for standardisation and quantification of FDG procedure guidelines for tumour imaging: version
whole body PET studies in multi-centre trials. Eur J 2.0. Eur J Nucl Med Mol Imaging 2015;42:328–354.
Nucl Med Mol Imaging 2008;35:2320–2333. 5. DeSouza N, Hoekstra OS, Nestle U, Stroobants S,
2. Ettinger S, Testanera G, Sera T, Boellaard R, Verzijlber- Boellaard R, Schaefer-Prokop C, et al. EORTC Imaging
gen F, Chiti A. Quality visits: The EANM/EARL FDG-PET/ Group. Eur J Cancer Suppl 2012;10:82–7.
CT accreditation programme. In: Glaudemans AWJM, 6. Manual for EARL FDG-PET/CT Accreditation, version
Medema J, van Zanten AK, Dierckx RAJO, Ahaus CTBK, 2.1 (April 2013). EARL, EANM Research Ltd.
eds. Quality in nuclear medicine. Heidelberg Berlin New 7. Waterstram-Rich K, Hogg P, Testanera G, Medvedec
York: Springer International Publishing; 2017:415–427. H, Dennan SE, Knapp W, et al. Euro-American discus-
3. Boellard R. Standards for PET image acquisition and sion document on entry-level and advance practice.
quantitative data analysis. J Nucl Med 2009;50 Suppl Nucl Med Technol 2011;39:240–248.
1:11S–20S.
» Flyer. EARL FDG-PET/CT Accreditation (March 2013). » Tatsch K. Standardisation and harmonisation boost
EARL, EANM Research Ltd. the credibility of nuclear medicine procedures. Eur J
Nucl Med Mol Imaging 2012;39:186–7.
117
CHAPTER 8
by Aljaz Socan
CHAPTER 9
INTRODUCTION
When undertaking any practices involving the human administration of
RA DIO NUC L I D E C A L I BRATO R
Figure 5
The electrons produced by ionisation cal current) increases. When the voltage
are attracted to the positive electrode becomes sufficient to cause complete
and the ionised atoms to the negative collection of all of the charges produced,
electrode, causing a momentary flow of a the saturation region begins. The voltage
small amount of electrical current. at which the saturation region begins is
For maximum efficiency of operation, called the saturation voltage (Vs, typically
the voltage between the electrodes must Vs ≈ 50–300 V). Ionisation chambers are
be sufficient to ensure complete collec- operated at voltages in the saturation
tion of ions and electrons produced by ra- region (Fig. 2). This ensures a maximum
diation within the chamber. If the voltage response to radiation and also that the
is too low, some of the ions and electrons response will be relatively insensitive
simply recombine with one another with- to instabilities in the voltage applied to
out contributing to electrical current flow the electrodes. The amount of electrical
(recombination region)[1]. charge current released in an ionisation
As the voltage increases there is less chamber by a single ionising radiation
recombination and the response (electri- event is very small[1].
Because of the small amount of electri- and measured. Small amounts of electrical
RA DIO NUC L I D E C A L I BRATO R
cal charge or current involved, ionisation current are measured using sensitive cur-
chambers generally are not used to record rent-measuring devices called electrome-
or count individual radiation events. In- ters[1].
stead, the total amount of current passing
through the chamber caused by a beam
of radiation is measured. Alternatively, the RADIONUCLIDE CALIBRATOR
electrical charge released in the chamber A radionuclide calibrator is a device con-
by the radiation beam may be collected sisting of an ionisation chamber and an
Figure 2
Voltage response curve (charge collected vs. voltage applied to the electrodes) for a typical ionisation
chamber. In usual operation, applied voltage exceeds saturation voltage Vs to ensure complete collection
of liberated charge. (Adopted from [1])
Figure 3
electrometer (Fig. 3). Unlike other types large for assay with NaI(Tl) detector sys-
of ionisation chamber, radionuclide cal- tems.
ibrators employ sealed and pressurised
chambers filled with argon gas. This elim- Because ionisation chambers have no
inates the effect of changing barometric inherent ability for energy discrimination,
pressure on output readings. Ionisation they cannot be used to select different
chamber radionuclide calibrators assay gamma ray energies for measurement, as
the total amount of activity present by is possible with detectors having pulse-
measuring the total amount of ionisation height analysis capabilities. An approach
produced by the sample. Radionuclide that is used to distinguish low-energy
calibrators typically are calibrated to read from high-energy gamma ray emitters
directly in units of activity (Bq or Ci), with (e.g. 99mTc vs. 99Mo) is to measure the sam-
switches to set the display for different ple with and without a few millimetres of
radionuclides. They are used for assaying lead shielding around the source (Fig. 4).
relatively large quantities (i.e. MBq range) Effectively, only the activity of the high-en-
of gamma ray-emitting radioactivity too ergy emitter is recorded with the shielding
in place, whereas the total activity of both counter depends on sample positioning[1].
RA DIO NUC L I D E C A L I BRATO R
emitters is recorded with the shielding ab- If a small volume of radioactive solution of
sent. This technique can be used to detect constant activity in a test tube is diluted
tens of kBq quantities of 99Mo in the pres- progressively by adding water to it, the
ence of tens or even hundreds of MBq of counting rate recorded from the sample
99m
Tc [1]. in a standard well detector progressively
decreases, even though the total activity
Figure 4 in the sample remains constant. In es-
sence, the geometric efficiency for the
sample decreases as portions of the ac-
tivity are displaced to the top of the well.
If the volume of a sample is increased by
adding radioactive solution at a constant
concentration, the counting rate first in-
creases linearly with sample volume (or
activity) but the proportionality is lost as
the volume approaches and then exceeds
the top of the well. Eventually there is little
change with increasing sample volume,
although the total activity is increasing.
For example, an increase in sample vol-
99
Mo breakthrough test lead shield
ume in a standard test tube from 7 to 8
(“radionuclide purity testing”) mL, i.e. a 14% increase in volume, increases
the counting rate by only about 1%[1]. Thus
the sample volume has significant effects
As with the NaI(Tl) well counter, radio- on the counting rate with well counters.
nuclide calibrators are subject to sample For the stated reasons, sample volumes
volume effects (the fraction of gamma should be the same when comparing two
rays escaping through the hole at the samples. A technique applicable when
end of the well depends on the position adequate sample volumes are available is
of the source in the well: it is about 7% if to use identical test tubes for all samples
the source is near the bottom of the well and to fill them such that the volume of
but increases to 50% if it is near the top) activity inside the well itself does not differ
and the geometric efficiency of a well between samples.
fection, the radionuclide calibrator must tine tests[3]. The most thorough assess-
RA DIO NUC L I D E C A L I BRATO R
inspection source holders and other accessible for physical inspection, but
accessories for damage the loose accessories should be checked
High voltage To check the constancy and Daily/as Essential for an accurate activity
correct operating voltage recommended measurement
by manufacturer
Clock To check that the calibrator Daily Essential for calibrating radioactivity
accuracy clock is the same as the time to a specific time of day; clock time
of day throughout the department must be
the same (i.e. all wall clocks and internal
computer clocks)
Zero To check that the display is at Daily Record the zero setting (before any
adjustment zero when no radioactivity is adjustment); a drift in the “zero” reading
present may indicate that the instrument needs
repair
Background To check background response Daily Perform the test with the source holder/
counts under operational conditions liner in place in the chamber; remove
appropriate for a particular nearby radioactive sources that might
radionuclide; to detect cause an incorrect background reading
contamination
Constancy To check the stability and Daily Measure a long half-life radionuclide, e.g.
reproducibility of the ionisation 137Cs, with its own calibration factor;
chamber, electrometer and also, obtain relative measurements for
calibrator nuclide settings each nuclide setting to be used that day
Stability To check the short-term Yearly Counting precision is a measure of the
counting precision stability of the whole system, and is
measured by repeated measurements
and application of the chi-square test
Accuracy To check the accuracy of the Yearly This requires readings of sources of
activity reading known activity; refer to the supplier and
national measurement standards for
guidance
Linearity To confirm that the calibration Six monthly/ The change in response when the
setting for a particular yearly measurement range is changed should
radionuclide indicates the be minimal; the range of use should be
correct activity over the entire chosen between the maximum activity
range of use to be measured (e.g. in the GBq range
for a 99mTc eluate) and the lowest
activity to be measured (e.g. 1 MBq) for a
particular radionuclide
Table 1: Routine QC tests for a radionuclide calibrator. Equipment type: gas ionisation
chamber. (Adapted from Eur J Nucl Med Mol Imaging 2010;37:662–671 [6])
by radioactive sources that might cause results kept for the life of the chamber. The
stability. Ideally, the test should use stan- ed source method involves manipula-
dards of radionuclides that are employed tion and accurate measurement of stock
by the radionuclide calibrator manufac- solution aliquots. The decaying source
turer to set the system when transporting method is recommended for secondary
calibration settings to production models standard radionuclide calibrators and ref-
(e.g. 57Co and 137Cs or 57Co and 60Co). As- erence radionuclide calibrators. For “field”
certaining which radionuclides were used instruments, the decaying source meth-
by the manufacturer should be part of the od should be used at acceptance testing
purchase process. The use of one source and following repair. The shield method
(e.g. 137Cs) in combination with the routine should be sufficient for annual testing;
“relative response tests” should be suffi- alternatively, the facility can employ the
cient for most medical facilities. For more decaying source method. Measurements
complex programmes, instrument stabili- are taken following the daily tests. The
ty should also be checked annually with decaying source measurements are tak-
at least two traceable reference sources en with the source in the source holder
and the radionuclides used should vary in the measurement position. The graded
from year to year[2]. Measurements of the shielding measurements are taken in ac-
long-lived standards and the two trace- cordance with the shield manufacturer’s
able reference sources should be within instructions. At acceptance testing and
±5% of the decay-corrected initial values. following repair, measurements should be
Secondary standard radionuclide calibra- taken using the decaying source method
tors and reference radionuclide calibrators from the highest activity (highest current)
should be within ±2%[3]. The measure- measured down to approximately 1 MBq.
ments should be recorded and available Annually, measurements should be taken
for regulatory review. between the maximum activity admin-
istered and 1 MBq over the range of use.
System linearity Measurements should be within ±5%
A calibrator is considered linear if the ra- of the expected values. For secondary
tio of the measured response to the pre- standard radionuclide calibrators and ref-
dicted response remains constant over erence radionuclide calibrators, linearity
the range of current inputs. The decaying testing should be performed quarterly us-
source method, the shield method (Fig. ing the decaying source method and the
5) and the graded source method may measurements should be within ±2%[3].
REFERENCES
1. Cherry SR, Sorenson JA, Phelps ME. Physics in nu- ing 2010;37:672–681.
clear medicine. 4th ed. Philadelphia: Elsevier; 2012. 5. EANM Physics Committee, Busemann Sokole
2. American Association of Physicists in Medicine. The E, Płachcínska A, Britten A; EANM Working Group
selection, use, calibration, and quality assurance of on Nuclear Medicine Instrumentation Quality
radionuclide calibrators used in nuclear medicine. Control, Lyra Georgosopoulou M, et al. Routine
Report of AAPM Task Group 181, June 2012. quality control recommendations for nuclear med-
3. International Atomic Energy Agency. Quality as- icine instrumentation. Eur J Nucl Med Mol Imaging
surance for radioactivity measurements in nuclear 2010;37:662–671.
medicine. Technical Report Series No. 454. Vienna: 6. Measurement Good Practice Guide No.93; Protocol
IAEA; 2006. for Establishing and Maintaining the Calibration of
4. Busemann Sokole E, Płachcínska A, Britten A; EANM Medical Radionuclide Calibrators and their Quality
Physics Committee. Acceptance testing for nuclear Control; National Physical Laboratory, May 2006,
medicine instrumentation. Eur J Nucl Med Mol Imag- ISSN: 1368-6550.
UPTA KE PRO B E S
by Kristof Baete
CHAPTER 10
PR O BE S A ND O R GA N UPTA KE PR O BE S
GAMMA PR O BE S: I NTRAO PE RATI V E
much more extensive. tectors are usually fabricated out of crys-
talline materials, such as CdTe, CdZnTe (or
Two types of radiation detector are used CZT) and HgI2.
for the manufacture of gamma probes: (a)
scintillation detectors, or scintillators, and Both detector types enable energy dis-
(b) semiconductors. crimination of the radiation and allow for
the selection of upper and lower energy
A scintillator is mostly a crystal that is level discriminators. Events that are re-
able to absorb incident gamma rays. The corded within the energy range can be ac-
energy of the charged particles, created cepted for further processing, while events
by the ionisation of the incident gamma outside of that range can be rejected or
rays, is re-emitted as scintillation light. used for the purpose of scatter correction.
This happens such that the light output In uptake probes, the entire energy spec-
of the crystal is proportional to the energy trum can be recorded.
of the incident gamma ray. Subsequently,
a light detector converts the scintillation The higher density and effective atom-
light to a measurable signal that is again ic number of inorganic scintillators allow
proportional to the energy of the incident for a better detector efficiency at higher
gamma ray. Scintillator crystals for gam- energies. However, scintillators in combi-
ma probes are usually inorganic materials, nation with a PMT are sensitive to environ-
such as NaI(Tl) or CsI(Tl). For higher ener- mental changes and therefore have a less
gy applications, BGO, LSO or GSO may be stable energy calibration. Scintillator-based
used owing to their better stopping pow- probes with a photodiode have shown an
er. Photomultiplier tubes (PMTs) or photo- improved ruggedness compared with PMT-
diodes are used for the detection of the based systems. Nevertheless, scintillators
scintillation photons. have a poorer energy resolution than semi-
conductors; this means that a semiconduc-
In a semiconductor, the incident radia- tor has a better scatter rejection capability.
tion is measured directly by the amount
of electron-hole pairs that are created in
proportion to the energy of the gamma QUALITY ASSURANCE
ray. The charge carriers are collected using The international Basic Safety Standards,
electrodes which then produce a measur- laid down in European Council Directive
analysis encompasses more than those The power supply of the device is crit-
PR O BE S A ND O R GA N UPTA KE PR O BE S
GAMMA PR O BE S: I NTRAO PE RATI V E
aspects that need to be verified once the ical and should be verified. Also, the bat-
equipment is in routine use. Performance tery-supplied intraoperative probe unit
tests for gamma probes have been pro- should be checked. The energy response
posed in the literature[7, 11, 13]. The most of an uptake probe depends on the stabil-
common acceptance and reference tests, ity of the high voltage and detector gain,
in general, are: but also on the environmental conditions,
such as temperature. Therefore, it is ex-
»» Physical inspection, shielding tremely important to assess the detector
»» Power supply, high voltage, energy calibration or peaking. After recep-
detector gain, battery tion, the energy settings should be verified
»» Energy peaking or calibration, for all used radionuclides, and if necessary
energy settings the instrument should be (re)calibrated.
»» Background count rate For an uptake probe, direct spectral analy-
»» Energy resolution sis of the energy peak of a reference source
»» Sensitivity (e.g. caesium-137) is usually obvious. How-
»» Counting precision ever, visualisation of the energy spectrum
»» Linearity of the response of an intraoperative probe by the end-user
is not always possible.
Some device-specific acceptance tests
should be added, such as the source ge- The background count rate of a gamma
ometry influence for uptake probes and probe is an essential property. Each mea-
spatial resolution for intraoperative probes. surement should in fact be preceded or
followed by a reading of the background
First of all, physical inspection of the count rate. This system parameter de-
probe should be carried out immediately af- pends on the energy settings. For any of
ter its reception. Any problem or deficiency these settings, a typical background count
observed during the warranty period should rate should be analysed and reference val-
be reported and repaired at once. Physical ues should be determined. It is important
impact to the detector, or its shielding, can to perform a follow-up of background
lead to malfunctioning or changes in effi- measurements for various detector loca-
ciency. For intraoperative probes, the signal tions and positions. In addition to the am-
from a nearby injection site, caused by a leak bient conditions, the electronics and con-
in the shielding, may conceal an SLN. necting cables can have an impact.
probe indicates how well it is able to dis- random process that obeys Poisson statis-
tinguish between closely spaced energy tics. A chi-square test can be used to verify
peaks of detected gamma rays in the en- that the short-term detector sensitivity is
ergy spectrum. A good energy resolution in accordance with the Poisson distribu-
enables the gamma probe to reject scat- tion [1]. With a series of measurements, a
tered radiation. Hence, measuring energy chi-square value can be computed and
resolution can identify the potential for a converted to a probability. A high prob-
good system performance. The energy ability (>0.99) means that the variations
resolution of the probe is quantified by in the measurements are smaller than
the full-width at half-maximum (FWHM) statistically expected. This may indicate
of the gamma peak at a certain energy. In that, for example, periodic noise is being
uptake probes, FWHM can be reported by measured instead of a radioactive source.
the console as part of the energy calibra- A low probability (<0.01) indicates that
tion or peaking procedure. the measurements are too irregular to
follow the Poisson nature of the source,
An important parameter is the sensitiv- and there may be something wrong with
ity of the gamma probe. In a standardised the device. A probability of 0.5 indicates a
setup, with a reproducible source-to-de- perfect detector behaviour according to
tector distance and a reference source Poisson statistics.
activity, the count rate response can be
determined for the available energy set- The linearity of the detector response,
tings. The results of these sensitivity mea- or the count rate capability, verifies the
surements are the reference values for the sensitivity at different levels of source ac-
subsequent QC tests. These QC measure- tivity. Radiation detectors are influenced
ments are therefore usually referred to as by a number of effects at increased count
a test of the constancy or the long-term rates. Some of these effects can lead to a
stability of the sensitivity. For a more re- shift of the measured energy spectrum.
alistic analysis, sensitivity may also be as- Even though this effect is small for most
sessed taking into account the influence applications, it should be recognised and
of scatter [11]. measured at least during acceptance test-
ing. The most important effect, however,
The short-term stability of the sensitivity is dead time or the loss of counts. With a
is often referred to as the count rate preci- decaying source, for example, one can
compare the observed detector count rate such that the QC results of the device are
PR O BE S A ND O R GA N UPTA KE PR O BE S
GAMMA PR O BE S: I NTRAO PE RATI V E
with the true count rate of the source. brought back in line with the reference
values. Moreover, trend analysis of the QC
For the uptake probe, a specific accep- results is a powerful tool for observation of
tance test should be added. Since the in- the performance of the instrument over
strument is used to quantify an amount of time.
activity, it is important to know the influ-
ence of different source geometries on the QC recommendations and test fre-
response of the detector. The result of this quency proposals have been described
assessment reveals the robustness of the in a number of publications [8–10]. The most
measurement setup and source (i.e. organ) important routine QC tests for gamma
positioning. probes, in general, are:
especially for those mobile units that use property that should be observed prior
PR O BE S A ND O R GA N UPTA KE PR O BE S
GAMMA PR O BE S: I NTRAO PE RATI V E
PR O BE S A ND O R GA N UPTA KE PR O BE S
GAMMA PR O BE S: I NTRAO PE RATI V E
REFERENCES
1. Cherry SR, Sorenson JA, Phelps ME, eds. Physics in quality control recommendations for nuclear med-
nuclear medicine. 4th ed. Philadelphia: Saunders; icine instrumentation. Eur J Nucl Med Mol Imaging
2012. 2010;37:662–671.
2. Bailey DL, Humm JL, Todd-Pokropek A, van Aswegen 9.
Quality control of nuclear medicine instruments.
A. Nuclear medicine physics. Vienna: International IAEA TECDOC 602. Vienna: International Atomic En-
Atomic Energy Agency; 2015. Available from: http:// ergy Agency; 1991.
www-pub.iaea.org/books/IAEABooks/10368/Nucle- 10. International Electrotechnical Commission. Nuclear
ar-Medicine-Physics medicine instrumentation – Routine tests – Part 1:
3. Heller S, Zanzonico P. Nuclear probes and intra- Radiation counting systems. 2001. IEC/TR 61948-1.
operative gamma cameras. Semin Nucl Med 2011; http://www.iec.ch.
41:166–181. 11. NEMA NU-3. Performance measurements and qual-
4. Povoski SP, Neff RL, Mojzisik CM, O’Malley DM, Hin- ity control guidelines for non-imaging intraopera-
kle GH, Hall NC, et al. A comprehensive overview of tive gamma probes. Rosslyn, VA: National Electrical
radioguided surgery using gamma detection probe Manufacturers Association; 2004.
technology. World J Surg Oncol 2009;7:11. 12. Wengenmair W, Kopp J, Sciuk J. Quality criteria of
5. Hänscheid H, Canzi C, Eschner W, Flux G, Luster gamma probes: Requirements and future develop-
M, Strigari L, Lassmann M. EANM Dosimetry Com- ments. In: Shauer AJ, Becker W, Reiser MF, Possinger
mittee Series on Standard Operational Procedures K, eds. The sentinel lymph node concept. Berlin Hei-
for Pre-Therapeutic Dosimetry – II. Dosimetry prior to delberg New York: Springer; 2005.
radioiodine therapy of benign thyroid diseases. Eur J 13. Chiesa C, Toscano F, Mariani M, Bomnardieri E.
Nucl Med Mol Imaging 2013;40:1126–1134. Physical performance parameters of intraoperative
6. International Electrotechnical Commission. Evalua- probes. In: Mariani G, Giuliano AE, Strauss HW, eds.
tion and routine testing in medical imaging depart- Radioguided surgery: A comprehensive team ap-
ments – Part 1: General aspects. 1993; IEC/TR 61223- proach. New York: Springer; 2008.
1. http://www.iec.ch. 14. Matheoud R, Giorgione R, Valzano S, Sacchetti
7. Busemann Sokole E, Płachcínska A, Britten A; EANM G, Colombo E, Brambilla M. Minimum acceptable
Physics Committee. Acceptance testing for nuclear sensitivity of intraoperative gamma probes used
medicine instrumentation. Eur J Nucl Med Mol Imag- for sentinel lymph node detection in melanoma
ing 2010;37:672–681. patients. Phys Med 2014;30:822–826.
8. EANM Physics Committee, Busemann Sokole 15. European Commission. Radiation Protection 162.
E, Płachcínska A, Britten A; EANM Working Group Criteria for acceptability of medical radiological
on Nuclear Medicine Instrumentation Quality equipment used in diagnostic radiology, nuclear
Control, Lyra Georgosopoulou M, et al. Routine medicine and radiotherapy. 2012.
Figure 1
W E L L CO UNTE RS
Voltage-response curve for gas-filled detectors
Figure 2
W E L L CO UNTE RS
Left panel: Schematic diagram of well counter.
Right panel: Most automated multiple sample systems use a so-called through-hole detector
est stopping powers and (c) impurities studies of glomerular filtration rate, cere-
are present even in relatively pure crystal brospinal fluid leak, red cell mass, plasma
and capture electrons released in ionisa- volume and radioimmunoassay tests with
tion, leading to lower detection efficiency. 125
I-radiolabelled probes[7,8]. Preclinically,
Compound semiconductors, such as cad- well counters are used in in vitro cell stud-
mium zinc telluride (CZT), can overcome ies, biodistribution studies, determination
these drawbacks[2,5]. CZT detectors have of radionuclidic impurities (e.g. determi-
been applied to various medical imaging nation of 68Ge in the 68Ge/68Ga generator
modalities in CT, PET and SPECT[6]. eluate) etc. Well counters usually comprise
a well-type detector, which surrounds the
sample and therefore provides high detec-
NaI(Tl) WELL COUNTERS tion efficiency (Fig. 2). A photomultiplier
Well counters are an important instrument tube (PMT) with associated electronics
in nuclear medicine since they play a key backs the detector. Very high detection ef-
role in preclinical and clinical practice. The ficiency of the NaI(Tl) well counters limits
most important clinical applications are the amount of activity that can be count-
ed (~37 kBq). With higher levels of activity, by self-absorption, even with large crys-
W E L L CO UNTE RS
Figure 3
W E L L CO UNTE RS
Left panel: Example of an automated multisample well type counter.
Right panel: Example of computerised routine QC tests
neous electron emission and thermal pendently from the sample volume. Sys-
noise contribute to an increased count tems with an automated sample arm can
rate[1, 9]. save a lot of time, since the sample vials
The overall advantages of NaI(Tl) make can be loaded into a rack and counted
these detectors the first choice for almost automatically. One disadvantage of auto-
all routine applications in nuclear med- mated systems is that they are not as well
icine involving the detection of gamma shielded as manual well counters, since
rays in the 50–250 keV energy range. there is no lead shielding directly on the
In daily practice, rather than manual top of the sample being counted. This
well counters, automated multiple sam- can increase background counting rates,
ple systems are used (Fig. 3). Most such particularly from the other samples in the
systems use a so-called through-hole rack, which can be problematic, especial-
detector, where the sample hole pass- ly when low-activity samples are counted
es throughout the NaI(Tl) crystal and together with high-activity samples in the
the PMT is connected to the side of the rack.
scintillator (Fig. 2). Samples in such a de- For very high throughput one can use
tector can be automatically positioned multidetector systems which can contain
at the centre of the NaI(Tl) crystal, inde- multiple NaI(Tl) scintillation detectors. This
simultaneously. The individual detectors to verify that the instrument meets spec-
are separated and shielded from each ifications and performs according to its
other to prevent crosstalk, which can give clinical purpose. Acceptance testing re-
rise to substantial counting error when sults are the reference data for future QC
counting high-energy gamma rays. Back- tests, and some of them may be repeated
ground measurements in one detector periodically (every 6 or 12 months) or after
while counting a sample in an adjacent a major service or component change[14].
one can be used to estimate the extent of At the time of installation of a well count-
the crosstalk[2]. er, physical inspection, energy window
Because of the high-energy photons, calibration and checking of the energy
in the case of PET applications the NaI(Tl) resolution, sensitivity and linearity are per-
crystal thickness needs to be optimised formed as acceptance testing[15].
and the shielding around the detector Since the electronic components and
has to be sufficient to prevent crosstalk detectors of well counters can fail or de-
from either the samples in adjacent rack teriorate over time, a good quality assur-
positions or the adjacent detector in the ance programme should be employed
case of multidetector systems[10,11]. Re- to ensure consistently accurate results.
cently, a high-sensitivity well counter with The routine QC tests for well counters in-
a bismuth germanate (BGO) detector clude energy window calibration (energy
was developed for preclinical animal PET peaking), measurement of background,
investigations in which quantitative mea- energy resolution, constancy, efficiency,
surements of small blood volumes and reproducibility (chi-square test), linearity
activities are needed[12]. and minimal detectable activity[13,14]. Table
2 provides an overview of basic QC tests,
their purpose and their frequency. Some
QUALITY CONTROL of the tests are described in detail below:
Quality control (QC) is defined as an es- »» When a well counter is equipped with
tablished set of ongoing measurements a multi-channel analyser, the energy
and analyses designed to ensure that the spectrum should be checked to verify
performance of a procedure or instrument that the photopeak of the radionuclide
is within a predefined acceptable range[13]. coincides with the preset photopeak
Every nuclear medicine instrument, in- energy window. Usually, energy peaking
cluding well counters, must undergo is done with a 20% photopeak energy
W E L L CO UNTE RS
Background Daily To detect contamination and to determine background
radiation
Constancy Daily Day-to-day counting of long-lived reference sources
Energy window Daily / quarterly To verify that the photopeak of the radionuclide coincides
calibration with the preset photopeak energy window
Energy resolution Quarterly To evaluate the sharpness of a photopeak of reference
source radionuclide
Reproducibility Quarterly / To check variations in the set of measurements of reference
half-yearly source (chi-square test)
Efficiency Annually To determine the sensitivity (ε: efficiency) in cpm/Bq for each
radionuclide
a
The type of test and the frequency depend on the manufacturer’s instructions, national legislation and
different guidelines.
brator. On the other hand, a surrogate or by aliquoting the initial solution with
W E L L CO UNTE RS
W E L L CO UNTE RS
REFERENCES
by Søren Holm
CHAPTER 12
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
of 10 mm in tissue (equivalent) material, were already present in previous directives,
and personal dose equivalent Hp(d) (unit but may be interpreted differently in the
sievert) defined as the equivalent dose at new implementation, are: categorisation
a depth d in soft tissue below a specified of (exposed) workers and classification of
point in the body. Relevant values of d are workplaces. Workplaces must be divided
d=10 mm (used for estimating effective into controlled areas with more restricted
dose), d=0.07 mm (used for skin dose) access and surveillance and supervised
and d=0.3 mm (for dose to the eye lens). areas with a lower risk and therefore more
These entities are listed here because ready access. Workers are divided into two
radiation protection equipment will nor- groups: group A, who are “liable to receive
mally be specified (calibrated) in terms of an effective dose greater than 6 mSv per
H* or Hp(d). year or an equivalent dose greater than
15 mSv per year for the lens of the eye or
greater than 150 mSv per year for skin and
RULES AND REGULATIONS IN extremities”, and group B, who are not. It
RADIATION PROTECTION is possible to design nuclear medicine de-
The fundamental concepts are laid out by partments and to plan the work so that it
ICRP in their recommendations, the most is highly unlikely that anyone (technolo-
recent general one being from 2007 (ICRP gists, physicians or medical physicists) will
103[1]). Based on this, international (IAEA) exceed the dose limits listed above. How-
and regional (EU) Basic Safety Standards ever, even then it will still be necessary to
(BSS) have been issued[3,4]. The EU BSS-di- monitor (most of ) the staff to document
rective (Euratom 2013/59[4]) must be im- this. The rest of this chapter shows the
plemented in the national legislations types of equipment that can be used for
of all the EU member states by February monitoring of persons and workplaces.
2018. Despite the European cooperation
between authorities and attempts to har-
monise, there will be some differences DOSIMETERS FOR PERSONAL
between countries. Further, local institu- MONITORING
tions may have their own (stronger) de- Monitoring of effective dose and finger
mands, and therefore it is recommended dose has been routine for decades. The
always to look (also) at what local rules recently lowered dose limit to the eye
may apply. lens[5] has created some uncertainty as to
whether special measures are necessary (rates) and also provide an alarm function
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
for documenting compliance with this for “active” protection. These two kinds of
limit. Currently there is no consensus on dosimeter form a useful supplement to
the issue, but in nuclear medicine (unlike each other: the passive one for legal use
interventional radiology) it seems likely and the active one as a reminder in daily
that the measure of effective dose will be processes. Dosimeters for personal moni-
sufficient to cover also the dose to the eye toring are normally calibrated to provide
lens. Radiochemistry may be an exception Hp(d) and results are reported in mSv.
to this. The legal, integrating dosimeters
Two kinds of dosimeter are used: pas- should be worn at all times in the depart-
sive, integrating devices that require a ment by all staff for whom this is relevant
Figure 1
a) Personal film dosimeter. b) Open (empty) film holder showing the metal filters and openings
allowing the distinction of different radiation qualities
special process to provide a reading and (as defined by the medical physics expert
active instruments (battery included!) or a radiation protection officer). They
that can immediately show actual dose should normally be worn at the front of
the body and at belt height. When work- terial can be stored and will retain the in-
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
ing in front of a shielded workplace, the formation for potential later re-evaluation.
dosimeter may be placed at chest level in Against the film material speaks the rather
order to avoid underestimation of the ex- high uncertainty at low exposure, which
posure here. Pregnant staff members (to over the years has become more import-
the extent that they are allowed to work at ant as the dose limits have been reduced
all in the department) should always wear and the price of the silver in the emulsion
the dosimeter at or near belt level to give has increased.
a better representation of the dose to the An alternative that is becoming increas-
foetus. ingly common is the use of thermolumi-
Electronic dosimeters can be considered nescence dosimetry (TLD). Many materials
optional. They are most important for new have the property, when irradiated, that
staff or when introducing new procedures. some of the absorbed energy will get
They may also be given to short-term visi- “trapped”. Electrons are excited out of their
tors for whom the monthly (or 3-monthly) stable ground state and end up in a high-
dosimeters do not make sense. er energy state that is also quite stable at
The classical personal dose meter is film normal (room) temperatures. By heating
based (Fig. 1a); the interaction of ionising (thermo = T) the material to 180–260°C,
radiation with a photographic film was the electrons are “lifted” out of the traps
what in 1896 led Becquerel to the detec- and will fall back emitting light (lumines-
tion of radioactivity. As with other film ma- cence = L). The amount of light can be
terials, processing (developing, fixing) is re- collected and will be proportional to the
quired to show the exposure, and a special energy absorbed in the first place (dose
device is needed to read and convert the = D). For dosimetry, the most common
gray scale into a dose value. The placement basis material is lithium-fluoride (LiF) with
of the film in a specially designed film certain impurities added to determine its
holder (Fig. 1b) with different thickness- properties in detail. The material is “tissue
es of plastic material and metal filters (Al, equivalent” and can be formed into tab-
Sn, In, Cd, Pb) to some extent allows de- lets (Fig. 2); in addition to being used for
termination of the radiation quality (soft/ whole-body monitoring (Fig. 3), these tab-
hard gamma, soft/hard beta, neutron). lets are small enough that they not only
Further, contamination by small drops of can be mounted in finger rings or wrist do-
radioactivity can easily be identified. Film simeters (Fig. 4) but can even be attached
has the advantage that the developed ma- to the fingertips (and covered by gloves).
As soon as the reading (heating) has been (unexpected) higher dose levels. The set-
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
performed, the information vanishes by ting of the alarm level requires some con-
“annealing” and the tablet can be reused. sideration when working with patients
The minimum detection limit for TLD is who may be unnecessarily scared if they
lower than for film, but TLD provides little observe that they are triggering the signal.
information on radiation quality, contam- Film and TLD require no quality control
ination is more difficult to observe and from the user except for a simple check
re-examination after the first reading is that the film holder or TLD package is
not possible. not damaged. Most often the handling
Electronic dosimeters (Fig. 5) are usual- of film or the TLD is performed by cen-
ly based on silicon solid-state detectors, tralised institutions that must be accred-
Figure 2 Figure 3
The size of a TLD tablet is only a few The author’s current TLD
millimetres
where ionising radiation produces pairs ited to perform that task and adhere to
of electrons/holes that can be collected certain standards, e.g. ISO/IEC 17025, ISO
and converted to a dose reading in µSv/h. 14146. Electronic dosimeters will typically
The advantage of such an instrument is loudly signify that they are running low
straightforward. The user can keep an eye on battery; if they do not do so, then it
on the dose rate during different work pro- is the responsibility of the user to check
cesses and can even be warned against and detect this. The function of alarm
levels can be checked (again, by the user) tering material in such a measurement
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
simply by moving the device gradually should be minimised.
towards any (sufficiently strong) gamma Most personal electronic dosimeters
source, e.g. a 57Co flood source, a 137Cs or have an energy response (efficiency) that
68
Ge calibration source or a vial contain- rapidly falls off towards zero for photon
ing99mTc or 18F (in this last case, the device energies below 30–50 keV, dependent on
must be brought just above the vial as the type. This may result in underestima-
it is shielded at the sides!). The absolute tion of doses, in particular if a major part of
calibration should be certified at delivery, the radiation consists of scattered radiation
and constancy of the reading can then be at low energy. However, a comparison in a
checked using, for example, a long-lived PET department has shown that results ob-
Figure 4 Figure 5
TLD tablets can be used for finger and wrist Two examples of personal
dosimetry and even attached to the fingertips electronic dosimeters
calibration source in a fixed geometry. tained using dosimeter film, a TLD and an
Proving the correct absolute calibration electronic dosimeter are reasonably consis-
from first principles is not easy even with tent[6]. In such a comparison, one key issue
a known, certified calibration source. Re- is the handling of (natural) background
member (if trying this) that the distance radiation, which is corrected for in the in-
square law is valid only for the direct radi- tegrating film or TLD systems by subtrac-
ation and therefore the presence of scat- tion of the measured value from an “unex-
posed” sample, whereas for the electronic ment. There are many manufacturers and
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
Contamination monitors exist in many detector for particles (alpha, beta) must
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
forms, but typically they are either large-ar- have a thin entry window to facilitate
ea (100–200 cm2), rectangular detectors their access, and a detector for photons
(Fig. 7a,b) or circular (cylindrical) detectors (X-ray, gamma) must have a high densi-
that detect at an “end-window” of diame- ty and high atomic number to provide
ter 2–6 cm (Fig. 7c). Obviously the large-ar- stopping power. Figure 8 shows a gam-
ea detectors in general have a higher ef- ma source (137Cs) in the upper half and
Figure 7
a) Contamination monitor with a large detection area. b) For particle measurements, the bottom plate
must be slid off. Note the mesh protecting the thin cover foil. c) Small circular contamination monitor
ficiency which is useful when surveying a a beta source (36Cl) in the lower half. The
large surface. The smaller ones, on the oth- instrument to the right has an end-win-
er hand, are easier to apply in the search dow GM tube which is sensitive to beta
for localized spills (single drops), where the particles and will even detect some alpha
large-area monitors yield a rather constant particles. It has a rather low sensitivity
signal when moving over the spot. for gamma photons because it is a gas
Some instruments look quite similar, detector. To the left is a scintillation (NaI)
and yet their probes have very different detector which is highly sensitive to gam-
properties. It is easily understandable that ma rays (and X-rays). The alumina cover
a detector can function only if the radi- of the crystal, however, does not allow
ation of concern has a high probability beta particles (at least below 1 MeV) to
of entering the active detector volume enter. It can therefore be very important
and a high probability of being stopped to choose the right instrument in a given
(interacting) in that volume. Therefore a situation. Some instruments (like the one
in Fig. 7a) use a plastic scintillator covered stancy with a relevant source, preferably
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
with a thin foil (lightproof ) and allow for long lived, in a fixed geometry. For the
detection of both particles and photons end-window detectors (and cylindrical
with reasonable efficiency. NaI detectors) a point source will be a nat-
Quality control of contamination ural choice. For large-area monitors, a foil
monitors should include battery check, of, for example, 100 cm2 with a weak gam-
background check and check for con- ma or beta activity is convenient (Fig. 9).
Figure 8
(a)
(b) (c)
A demonstration of the difference in detection properties between a scintillation counter (left) and
a GM tube (right). In (a), with the detectors pointing to a gamma source, the scintillation counter
yields 100 cps while the GM tube only shows 5–6. With the beta source, the scintillation counter (b)
is down to 5 cps (mostly background) while the GM tube in (c) shows almost 50 cps
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
trate the cover foil and trigger the detec-
tor, which may use energy discrimination
to distinguish alpha from beta particles
(and noise). A ~3-μm cover foil may cor-
respond to 0.5 cm of air, so to be sure of
detection the distance should not exceed
2 cm (Fig. 10). Then it is obvious that there
is a high probability of contaminating the
foil by touching potential activity directly.
It is possible to cover the instrument by a
single layer of standard household plastic
foil (8–10 µm) and still have a signal. In this
way a potential contamination can easily
∫
be removed by replacing the plastic foil.
Alpha monitors should be checked for
constancy with an appropriate source,
which could be a point source of 241Am,
Example of calibration/test sources which has a half-life of 433 years.
for large-area detectors
Stationary instruments
Not all contamination monitors can A hand-foot monitor (Fig. 11) is a conve-
detect alpha particles. A particularly thin nient piece of equipment for a fast check
window is required for the particles to en- of personnel contamination. One of the
ter. When measuring alpha radiation on a hand detectors can be removed from its
surface, it is also important to remember position to scan clothes. In older instru-
that the range even in air is quite limited ments, the panels used to be proportion-
and that the energy is gradually lost along al counters (gas-based) but today plastic
the path. For protection, it is a reasonably scintillator systems are common. Quality
easy rule of thumb that the range in air is control of this kind of monitor is the same
always below 10 cm. For a 5-MeV alpha as for portable contamination monitors,
particle, however, the range is really only i.e. checking at regular intervals for back-
4.5 cm, and after passing the air gap from ground and constancy, e.g. with a long-
the source to the detector, the particle lived beta-active sheet (14C or 90Sr).
Figure 10
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
When measuring for alpha-particles, a close distance to the (potential) source is mandatory
Figure 11
ME A S URE M E NT E Q UI PM E NT
RA DIATI O N PR OTE C TI O N
REFERENCES
1. ICRP. The 2007 Recommendations of the Interna- gers arising from exposure to ionising radiation and
tional Commission on Radiological Protection. ICRP repealing Directives 89/618/Euratom, 90/641/Eura-
publication 103. Ann ICRP 2007;37(2–4):1–332. tom, 96/29/Euratom, 97/43/Euratom and 2003/122/
2. International Atomic Energy Agency. Nuclear Euratom. OJ of the EU. L13; 57: 1–73 (2014).
medicine physics. Vienna: IAEA, Vienna; 2015. Web: 5. ICRP statement on tissue reactions and early and
http://www-pub.iaea.org/books/IAEABooks/10368/ late effects of radiation in normal tissues and organs
Nuclear-Medicine-Physics-A-Handbook-for-Teach- – threshold doses for tissue reactions in a radiation
ers-and-Students. Accessed 1 March, 2017. protection context. ICRP publication 118. Ann ICRP
3. International Basic Safety Standards for Protection 2012;41(1–2):1–322.
Against Ionizing Radiation and for the Safety of 6. Dall B, Svalling S, Holm S. A comparison of three dif-
Radiation Sources. IAEA Safety Series No. 115. ISBN: ferent methods for measuring whole body radiation
92-0-104295-7. exposure. Eur J Nucl Med Mol Imaging 2008;35 (Suppl
4. European Council Directive 2013/59/Euratom on ba- 2):S214.
sic safety standards for protection against the dan-
ISBN: 978-3-902785-13-8
Publisher:
European Association of Nuclear Medicine
Schmalzhofgasse 26, 1060 Vienna, Austria
Phone: +43-(0)1-890 27 44 | Fax: +43-(0)1-890 44 27-9
Email: office@eanm.org | URL: www.eanm.org
Main-Editor:
Sebastijan Rep
Co-Editors:
Marie Claire Attard
Luca Camoni
Andrea Santos
Project Management:
Magdalena Schwartz
Content:
No responsibility is taken for the correctness of this information.
Information as per date of printing September 2017.
Printing
“agensketterl” Druckerei GmbH
Druckhausstraße 1, A-2540 Bad Vöslau, Austria
Email: info@diedrucker.biz | URL: www.agensketterl.at