Professional Documents
Culture Documents
Diagnostic Imaging: Medical Radiology
Diagnostic Imaging: Medical Radiology
Diagnostic Imaging
Softcover Edition
Editors:
A. 1. Baert, Leuven
K. Sartor, Heidelberg
Springer-Verlag Berlin Heidelberg GmbH
A.M. Davies, Y.N. Cassar-Pullicino (Eds.)
Imaging
ofthe Knee
Techniques and Applications
With Contributions by
Foreword by
A.L.Baert
Springer
A. M. DAVIES. MD
Consultant Radiologist
MRI Centre
Royal Orthopaedic Hospital
Bristol Road
Birmingham, B31 2AP
UK
V. CASSAR-PULLICINO, MD
Consultant Radiologist
Department of Radiology
Robert Jones and Agnes Hunt Orthopaedic Hospital
Oswestry
Shropshire, SYIO 7AG
UK
ISBN 978-3-540-00250-5
Library of Congress Cataloging-in-Publicalion Dala
http/lwww.springer.de
CI Springu-Verlag Berlin Heidelberg 2003
Originally published by Springer-Verlag Berlin Heidelbcrg New York in 2003
The use of general descriptive namn, trademarks, etc. in this publication does not imply,even in the absente of a sp«ific
Slatement, that such names art exe mpl from the rtlevant proteetive law$ a nd egulalions
r and therefort frec for general
,~.
Produci liabilily: The publisherş cannot guarantee the accuracy of any informalion about do.age and applitalion contained
in this book. ln every case the user must cheek such information by consulting th e relevant literature.
Gover-Design and 1'ypesetting: Verlags.service Teichmann, 69256 Mauer
2113150 - S 4 3 2[O - Printed on acid-free paper
Foreword
Developments in imaging continue apace, not least in musculoskeletal imaging, with the
knee arguably the most commonly affected area. As our understanding of disease pro-
cesses and biomechanics improves there is a need to continuously update the knowledge
of radiologists, orthopaedic surgeons and others working in this field. Whereas many
previous texts on the knee have concentrated on a single imaging technique or a single
topic such as trauma, this book takes a dual approach to the subject. The first section
acquaints the reader with the full range of techniques available for imaging knee pathol-
ogy, emphasising indications and contraindications. The five chapters include contribu-
tions on radiography, computed tomography (CT) and CT arthrography, magnetic reso-
nance (MR) imaging and MR arthrography and ultrasound. The remaining 13 chapters
discuss the optimal application of these techniques to specific pathologies, highlighting
practical solutions to everyday clinical problems.
The editors are grateful to the international panel of authors for their contributions
to this book, which aims to provide a comprehensive overview of current imaging of the
knee.
Birmingham A.M.DAVIES
Oswestry V. N. CASSAR-PULLICINO
Contents
Imaging Techniques 1
1 Radiography
KJELL JONSSON and TORSTEN BOEGÄRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Arthrography
A. MARK DAVIES and VICTOR N. CASSAR-PULLICINO .. . . . . . . . . . . . . . . . . . . . . . . .. 19
5 Ultrasound
STEFANO BIANCHI, CARLO MARTINOLl, and JEAN GARCIA . . . . . . . . . . . . . . . . . . . . .. 65
7 Non-accidental Injury
STEPHEN CHAPMAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 97
8 Bone Trauma
PRUDENCIA TYRRELL and VICTOR N. CASSAR-PULLICINO ...................... 109
9 The Menisci
JAVIER BELTRAN and STEVEN SHANKMAN .................................... 129
14 Infection
MOHAMED E. ABD EL BAGI, MONA S. AL SHAHED, and BASSAM M. SAMMAK ....... 249
15 Arthritis
CHARLES WAKELEY and IAIN WATT .......................................... 269
reducing the dose burden to the patient due to faulty allows some image processing (windowing, zoom,
exposure factors. The radiation dose to the patient can image rotation and so on) depending on the software
be decreased without affecting the diagnostic accura- in the workstation. The free choice of data processing
cy, especially with measurements of lines and angles means that from a single exposure, it is possible to
(J6NSSON et al. 1996; SANFRIDSSON et al. 1998). With obtain the image that is best suited for excluding and
the analogue technique it is possible to overexpose a displaying the pathology looked for in each individu-
film, rendering the pathology less obvious. The broad al case. In knee examinations this is advantageous for
exposure range with CR may, however, be a disad- the diagnosis of suprapatellar pouch distension with
vantage. A heavily overexposed picture does not show blood (Fig.l.l) and ßake fragments (Fig.1.2) and for
on the image obtained, and appears as a normally the disclosure of small intra-articular calcifications.
exposed film. This is a threat, because the patient may Image output on a film makes it possible to show
receive a heavy radiation dose, which is not easily two images on one film, where the images are pro-
detected. This can be checked by looking at the sensi- cessed independently of each other. One image is
tivity value (S-value) that is available for every image. normally processed to give the impression of a con-
The S-value can be used for selection of the correct ventional image, and the other image is usually pro-
exposure settings on the X-ray generator. A high S-val- cessed with a higher degree of edge enhancement to
ue corresponds to a low dose and a low S-value indi- increase the contrast for the desired spatial frequen-
cates a high exposure. The S-value is related to the cy. Computed radiography has a lower spatial resolu-
speed of a screen-film system and it indicates which tion than conventional radiography, but this has not
speed a screen-film system should have to produce the proven to be a limiting factor with modern systems
correct density on the film. Images with a low S-value (J6NSSON et al. 1995; SCOTT et al. 1993).
have low noise while those with a high S-value are A ßat panel detector is used for digital radiogra-
grainier due to high noise. phy. The size of the detector is approximately that of
The image is either transferred to a laser image a large cassette, up to 43x43 cm. The detector con-
printer for film output or to a PACS system (picture verts the X-ray image into a digital signal, which is
archiving and communication system), where the transferred to a computer for image processing, dis-
image is reviewed on a workstation and stored in a play and transfer to a PACS system or a laser printer.
digital archive. It is possible to perform a complete The detector is usually mounted on a stand and some
reprocessing of the images if the IP reader transfers stands allow the detector to be moved for different
the image raw data to a dedicated workstation. The X-ray projections. The conversion of X-ray photons
reprocessing capabilities can also be incorporated in into an electrical signal can be done with two differ-
the IP reader in some models. A PACS workstation ent processes, either an indirect or a direct process.
a b
Flat panels using an indirect process convert the sibility of obtaining a high resolution because the
X-ray photons into light, with a scintillator, often a X-ray photons are direcdy converted into electrical
layer of cesium iodide (CsI), covering the entrance charge without the conversion to light, which might
side of the detector. The light is detected with an get scattered, thereby reducing resolution. The char-
amorphous silicon photodetector (photodiode) array acteristies of flat panel detectors will probably change
placed against the scintillator, converted into an elec- during the next few years owing to the rapid develop-
trical signal, read out with a thin film transistor array ment of this new technology.
and converted into a digital signal. Each photodiode It must be stressed that radiography, analogue or
represents a pixel. digital, is excellent for the evaluation of bone detail,
Flat panels using a direct process convert the X-ray whereas the soft tissues are poorly visualised. A large
photons into an electrical charge in an amorphous fluid effusion of the knee is seen as a distension of the
selenium layer. A bias voltage is applied between a suprapatellar pouch (Fig. 1.1) but the radiographie
surface electrode on the detector and an electrode appearances are the same whether the fluid is blood
array below the selenium layer in order to collect the in haemarthrosis or effusion in active arthritis or pus
electrical charge. The charge on the electrode array in septic arthritis.
is read out with a thin film transistor array and con- Conventional tomography may be used to increase
verted into a digital signal. Each electrode in the elec- the information. This is applicable for both analogue
trode array represents a pixel. and digital radiography. The conventional X-ray tube
The pixel size for flat panel detectors is around and the film-screen combination (or image plate) are
0.15 mm. The advantage of the indirect process is used to define a predetermined plane in the body,
the possibility of achieving a high detective quantum while the structures above and below this plane are
efficiency because of the high X-ray absorption of eliminated or blurred. This is achieved by moving
Cs!. The advantage of the direct process is the pos- the X-ray tube and the film-screen combination in a
6 K. Jonsson and T. Boegärd.
defined mann er in relation to each other, while the choiee depends on the available equipment. The min-
examined part of the body remains stationary. The imum radiographie projections are AP (anteropos-
motion of the X-ray tube and the film may be either terior) and lateral views. The knee joint should be
unidirectional (lineartomography) or pluridirection- straight in the AP view and slightly flexed in the lat-
al (circular, elliptieal, spiral or hypocycloidal tomog- eral projection. These radiographs are taken with the
raphy). The more complex the movement, the better patient in the supine position.
the quality of the image with less longitudinal streak- The tunnel view ("einblick") is taken with knee
ing, which may be seen in unidirectional tomogra- flexion of 40-50° and with the X-ray beam angulated
phy. The disadvantage with tomography is the long to the same degree to visualise the intercondylar
examination time and the high radiation exposure to notch and tibial spine; ideally the beam will be
the patient. Computed tomography (CT) has super- tangential to the tibial plateau (Fig. 1.4). Other meth-
seded tomography with multiplanar reconstruction. ods employed to obtain the tunnel view have been
If CT is not available, conventional tomography can described (BALLINGER 1991; BONTRAGER 1997).
be used to demonstrate the degree of depression of In the evaluation of the joint space, the radio-
a tibial plateau fracture (Fig. 1.3), to assess prema- graphs should be taken with the patient weight bear-
ture growth plate fusion, to evaluate healing of bony
pseudarthrosis and to identify sequestra in chronie
osteomyelitis.
1.3
Radiographie Projections
Fig.l.3. a AP view of the knee after trauma. There appears to be compression of the lateral tibial condyle (arrow). b Conventional
hypocycloidal tomography in the AP projection. There is moderate depression of a fragment from the lateral tibial condyle. c
Conventional hypocycloidal tomography in the lateral projection. The fragment is located posteriorly. With tomography it is
possible to outline the size and degree of depression of a fragment
Radiography 7
ing and with the knee in slight flexion. At the Univer- Other protocols for reproducible examination of
sity Hospital, Lund, the weight load is placed almost the weight-bearing knee have been reported.
entirely on the leg being examined. The lateral view BUCKLAND-WRIGHT (1995) designed a protocol
is obtained after fluoroscopic monitoring, so that the where the patient flexes the knee until the tibial pla-
dorsal aspects of the femoral condyles are superim- teau is horizontal and parallel to the central X-ray
posed (Figs. 1.5, 1.12b). Using fluoroscopic controt, an beam. This requires aknee flexion of 11-20°,depend-
AP view is obtained with the X-ray beam centred over ing on the inclination of the tibial plateau. The posi-
the joint space and tangential to the tibia plateau, tion is checked with fluoroscopy. The foot is rotated
with the tibial spines centred in the femoral notch. internally or externally until the tibial spines appear
BOEGARD et al. (1997) described a similar technique centrally placed relative to the femoral notch. The
with the PA (postero-anterior) projection, where the position of the foot is drawn and recorded on a piece
patella and the big toe touch the upright examination of paper placed in a defined position on the floor.
table. To obtain reproducible projections, the medial Reproducible views at follow-up examinations are
border of the foot is parallel with the X-ray beam. then possible using the same position of the foot.
Radiographs obtained with weight bearing on both In another protocol for flexed PA knee examina-
legs have, however, proved to be advantageous and tion, the patient is positioned with the patella and the
are increasingly used. hip in contact with the surface of the upright exam-
ination table, the feet pointing straight ahead verti-
cally relative to the knee and with knee flexion of
around 30°, the "schuss view" (PIPERNO et al. 1998).
Using fluoroscopy the X-ray beam is adjusted to be
tangential with the tibial plateau.
One group has developed a technique for AP knee
examination in extension. (RAVAUD et al. 1996) The
patient stands distributing his weight on both feet.
The posterior aspect of the knee is placed as elose
to the X-ray film as possible. During fluoroscopy
the inclination of the X-ray beam (approximately 5°
downward) is checked to be tangential to the medial
tibia plateau. The foot is rotated until the tibial spines
are centred beneath the femoral notch.
In our practice, the axial view of the patella is
taken with weight bearing and 30-60° knee flexion
with a verticalX-raybeam (Fig.1.6) to ensure that the
a ~----- middle portion of the articular surface of the patella
is in contact with the articular surface of the femur.
The central X-ray beam is tangential with the articu-
lar surface of the patella. The same technique is used
by BUCKLAND-WRIGHT (1995). One has to be aware
that with this technique, patellar pathology seen with
flexion below 30° is missed. In trauma patients, an
axial view of the patella can be obtained by various
techniques (Fig. 1.7).
1.3.1
Standard Series
b
The minimum projections are AP, lateral and tunnel
Fig.1.5. a Position of the patient for lateral weight-bearing views. The last-mentioned projection is often very
view of the knee. b Position of the patient for AP weight-
bearing view of the knee. Almost a11 the body load is on the
helpful when looking for loose bodies in the joint.
examined leg; there is knee flexion of 20-30°. The projection Only radio-opaque loose bodies are seen, while carti-
is checked with fluoroscopy lage loose bodies are not visualised. The tunnel view
8 K. Jonsson and T. Boegärd.
a b
different settings for window width and level. If
this is not done, subtle changes such as small cal-
m----------_/I ~D--------- /
~~~ cified loose bodies in osteochondromatosis may be
missed. In tumours and osteomyelitis the first find-
c d
ings are often subtle, with a faint periosteal reaction
such as Codman's tri angle that may be missed. Mar-
ginal osteophytes are important indicators of OA.
lt has been shown that when small marginal osteo-
Il'
phytes are present on the femoral or tibial condyles,
there are always local degenerative changes of the
cartilage, seen with magnetic resonance imaging
e f (MRI) (BOEGARD et al. 1998a) or arthroscopy. How-
Fig. 1.7a-f. Different radiographie techniques used to obtain
ever, in some instances of degenerative changes
axial non-weight-bearing views of the patella. [From MER- there are no accompanying osteophytes. Osteo-
CHANT et al. 1994. Reproduced with permission from the Jour- phytes at the tibial spine and at the intercondylar
nal 0/ Bone and Joint Surgery (Ameriean edition)] fossa are unreliable signs of knee OA (BOEGARD et
al. 1998a).
In children with pain over the tibial tub ero sity,
is of value to delineate osteochondritis dissecans, Osgood-Schlatter disease is clinically suspected. A
which in most cases is located on the inner aspect of coned down view with soft tissue exposure should
the medial femoral condyle (Fig.1.8). be obtained to reveal bone fragmentation (Fig. 1.9).
In standard series it is important not to over- Fragments mayaiso be seen with a skyline view of the
or underexposed analogue films, and to check the patella, where the fragments overlie the patellofemo-
computed radiography picture on the monitor with ral joint. A similar type of osteochondrosis is Sind-
Radiography 9
a b
Most fractures of the knee are apparent on the stan- nosed by the presence of marginal osteophytes, while
dard trauma series. Sometimes a finding may be subtle the degree of the degeneration is indicated by the
or equivocal, particularly in the case of tibial condylar severity of joint space narrowing. AHLBÄCK (l968) and
fractures. There may be a subtle and localised depres- LEACH et al. (l970) found that weight-bearing exam-
sion of the joint surface, but the degree and extension ination was superior to that obtained in the supine
are not clear on plain films. In such cases the stan- position for demonstration of joint space narrowing
dard examination should be completed with conven- in tibiofemoral OA. With general degeneration of the
tional tomography or CT with reconstruction in mul- joint cartilage, joint space narrowing is easily found
tiple planes for evaluation (Fig.1.3). with examination under weight bearing, but with local
Stress views of the knee may be valuable for indi- chondrallesions this may be difficult. It is important
reet demonstration of ligamentous disruption. Stress to obtain tangential views of the joint cartilage in
may be exerted manually during fluoroscopy or by the optimal position (Fig. 1.l1a). A number of factors
using a special device to produce standardised stress influence the possibility of achieving reproducible and
(STEDTFELD and STROBEL 1983). If stress radio- optimal views of the joint space. One factor is that the
graphs are done in the acute stage after trauma, the slope of the tibia plateau differs between the medial
test should be done under local or general anaesthe- and lateral condyles. The shape of the tibial condyles
sia to avoid muscle spasm. Stress radiographs seem differs in that the medial condyle is slightly concave
to be of limited value and should, according to STRO- and a tangential view shows the superimposition of
BEL and STEDTFELD (l990), be limited to anterior
and posterior drawer tests in acute and chronic cap-
suloligamentous injury.
If there is any clinical suspicion of a patellar frac-
ture, a sunrise view of the patella should be included
in the examination. Several techniques may be used
~~~-'-------ID
to obtain this view in a trauma patient (Fig. l.7).After
patellar dislocation, a small fragment is often avulsed a d
from the medial border of the patella (Fig.l.lO),
which is a helpful clue to the mechanism of trauma
in instances of post-traumatic pain.
I____. -~r-=----
t-_ \
b
L_((
r----~
~
1) .._
~
c
Table 1.1. Classification of knee GA. Comparison between the Ahlbäck and the Kellgren and Law-
rence grading systems
Ahlbäck Ahlbäck Kellgren and Kellgren and
grade definition Lawrence grade Lawrence definition
It has to be pointed out that this technique is not nations, differences in position between the bodies
generally available, because it requires special equip- can be revealed.
ment and computer pro grams for evaluation. The The tantalum beads can be easily introduced per-
method is used at some centres for routine evaluation cutaneously under local anaesthesia. The knee is
ofknee alignment (COOK et al. 1999). ideal for RSA studies. The method may be used for
analysis of the knee kinematies and prosthesis fixa-
tion (Fig.1.l5). The accuracy of measurements from
1.4.3 conventional radiography is in the order of 2-3 mm,
Roentgen Stereophotogrammetric Analysis but with RSA the accuracy is 0.2-0.3 mm. Also, rota-
tion can be estimated with the same degree of accu-
Stereophotogrammetry (RSA) was described by racy (RYD 1992; FR IDEN et al.1992).
SELVIK (1989). The idea behind the method is that There is no other method available that can com-
mobility between two structures can be defined and pete with RSA in terms of accuracy of measurements,
analysed by a stereophotogrammetrie technique. but the method requires considerable time and effort.
The structures are defined by small metallic beads The markers must be implanted with great care, and
of known size and density. The beads are placed the radiographs must be taken with patients cor-
in two structures that may move in relationship to rectly positioned. Digitisation and analysis are time
each other, for instance the tibial component of a consuming and demand thorough knowledge of the
prosthesis in relation to the proximal tibia. Abso- technique. RSA is best suited for smaller well-defined
lutely spherieal 0.8-mm beads made of tantalum are investigations that address kinematie problems. So
implanted into the bone and/or the plastie coating far it is not a method for general use.
of a prosthesis. At least three beads are required for
each structure to define a "rigid body". Two radio-
graphs are taken simultaneously with two angled 1.4.4
X-ray tubes in stereoscopie convergent-ray mode or Femoral Condyle Configuration
at 90° to each other. Before these radiographs are
taken, special calibration procedures must be per- The configuration of the femoral condyles may influ-
formed. The films are mounted in a special analy- ence knee stability. In a study by FRIDEN et al. (1993),
ser that, due to the stereoscopie radiographie tech- 100 consecutive patients with anterior cruciate liga-
nique, can locate the bodies in space and define the ment (ACL) rupture were studied prospectively for
position of each body in relation to the other by 5 years. During this time 16 patients developed dis-
using a special software program. At repeat exami- ability, which required reconstructive surgery. The
a b
Fig. l.lSa, b. Total knee prosthesis examined with weight bearing. a Postoperative examination 8 weeks after surgery. b Repeat
examination 1 year after surgery. The first examination (a) is not performed correctly: the X-ray beam is not tangential to the
tibial plateau. By contrast, the second examination (b) is correct. Already at the first examination there was clinical suspicion of
infection. The zone under the tibial plateau cannot be evaluated in a, while in b the zone is weH seen. Note the tantalum beads
for stereophotogrammetry in the tibia and in the plastic coating of the tibial component. In b, but not in a, both the anterior
and the posterior tantalum implants are seen. This is a further check on correct X-ray beam direction
16 K. Jonsson and T. Boegard.
Boegard T, Rudling 0, Petersson lF et al (1998a) Correlation levels in injured knees: do they always represent lipohem-
between radiographieally diagnosed osteophytes and mag- arthrosis? Radiology 198:499-502
netie resonance detected cartilage defects in the tibiofemo- Mazzuca ST, Brandt KD (1999) Plain radiography as an out-
ral joint.Ann Rheum Dis 57:401-407 come measure in clinieal trials involving patients with knee
Boegard T, Rudling 0, Petersson lF et al (1998b) Joint-space osteoarthritis. Rheum Dis Clin North Am 25:467-480
width in the weight-bearing radiogram of the tibio-femo- Merchant AC, Mercer RL,Jacobsen RH et al (1974) Roentgeno-
ral joint. Should the patient stand on one leg or two? Acta graphie analysis of patellofemoral congruence. J Bone Joint
RadioI39:32-35 Surg [Am] 56A:1391-1396
Boegard T, Rudling 0, Petersson lF et al (1998c) Joint -space Odenbring S, Berggren AM, Peil L (1993) Roentgenographie
width in the axial view of the patello-femoral joint. Def- assessment of the hip-knee-ankle axis in medial gonarthro-
initions and comparison with MR imaging. Acta Radiol sis. A study of reproducibility. Clin Orthop 289: 195-196
39:24-31 Pavlov H, Burke M, Giesa M, Seager KR, White E (1999) Ortho-
Boegard T, Rudling 0, Petersson lF et al (1998d) Correlation paedist's guide to plain film imaging. Thieme, New York
between radiographieally diagnosed osteophytes and mag- Piperno M, Conrozier T, Bochu M, Carillon Y, Fantino 0,
netie resonance detected cartilage defects in the patello- Vignon E (1998) Re-evaluation of conventional radiogra-
femoral joint. Ann Rheum Dis 57:395-400 phy in arthritis (in French). Rev Prat 48 SuppI17:S5-S8
Bontrager KL (1997) Textbook of radiographie positioning Ravaud P,Aule1ey GR, Chastang C, Rousselin B Paolozzi L,Arnor
and related anatomy, 4th edn. Mosby, St Louis B, Dougados M (1996) Knee joint space width measurement:
Brossmann J, Freyschmidt J, Czerny C (2001) Grenzen des an experimental study of the influence of radiographie pro-
Normalen und Anfänge des Pathologischen in der Radi- cedure and joint positioning. Br J RheumatoI35:761-766
ologie des kindlichen und erwachsenen Skeletts, 14th edn. Resniek D, Vint V (1980) The "tunnel" view in assessment of carti-
Thieme, Stuttgart lage loss in osteoarthritis of the knee. Radiology 137:547-548
Buckland-Wright C (1995) Protocols for precise radio-ana- Ryd L (1992) The role of roentgen stereophotogrammetrie
tomieal positioning of the tibiofemoral and patellofemoral analysis (RSA) in knee surgery. Am J Knee Surg 5:44-54
compartments of the knee. Osteoarthritis Cartilage 3 Sanfridsson J, Ryd L, Eklund K et al (1996) Angular configu-
[Suppl A]:71-80 ration of the knee. Comparison of conventional measure-
Cook TDV, Kelly BP, Harrison L, Mohamed G, Khan B (1999) ments and the QUESTOR Precision Radiography system.
Radiographie grading for knee osteoarthritis. A revised Acta RadioI37:633-638
scheme that relates to alignment and deformity. J Rheuma- Sanfridsson J, Svahn G, Jonsson K et al (1997) Computed radi-
toI16:641-644 ography for characterisation of the weight-bearing knee.
De Pablos J, Azcarate J, Barrios C (1995) Progressive open- Acta RadioI38:514-519
ing-wedge osteotomy for angular long-bone deformities in Sanfridsson J, Svahn G, Ryd L et al (1998) Assessment ofimage
adolescents. J Bone Joint Surg [Br] 77:387-391 post-processing and of measuring assistance tools in com-
Egund N, Lundin A, Wallengren NO (1988) The vertieal posi- puted radiography. 39:642-648
tion of the patella. A new radiographie method for routine Sanfridsson J, Holje G, Svahn G et al (2000) Radiation dose and
use. Acta RadioI29:555-558 image information in computed radiography. A phantom
Edholm P, Lindahl 0, Lindholm B et al (1976) Knee instability. study of angle measurements in the weight-bearing knee.
An orthoradiographie study. Acta Orthop Scand Acta RadioI41:310-316
47:658-663 Sanfridsson J, Ryd L, Svahn G, Friden T, Jonsson K (2001) Radio-
Friden T, Ryd L, Lindstrand A (1992) Laxity and graft fixation graphie measurement of femorotibial rotation in weight
after reconstruction of the anterior cruciate ligament. A bearing. The influence of flexion and extension in the knee
roentgen stereophotogrammetrie analysis of 11 patients. on the extensor mechanism and angles of the lower extrem-
Acta Orthop Scand 63:80-84 ity in a healthy populaton. Acta Radio142:207 - 217
Friden T, Jonsson A, Erlandsson T et al (1993) Effect offemoral Scott WW Jr, Rosenbaum JE, Ackerman SJ et al (1993) Subtle
condy1e configuration on disability after an anterior cruci- orthopaedie fractures: teleradiology workstation versus
ate ligament rupture. 100 patients followed for 5 years. Acta film interpretation. Radiology 187:811-815
Orthop Scand 64:571-574 Seil R, Müller B, Georg T et al (2000) Reliability and interob-
lnsall J, Shoji H, Mayer V (1974) High tibial osteotomy. J Bone server variability in radiologie al patellar height ratios. Knee
Joint Surg [A]56:1397-1405 Surg Sports Traumatol Arthrosc 8:231-236
J6nsson A, Hannesson P, Herrlin K et al (1995) Computed Se1vik GA (1989) A roentgen stereophotogrammetrie system
vs film-screen magnification radiography of fingers for the study of the kinematies of the skeletal systems. Uni-
in hyperparathyroidism. An ROC analysis. Acta Radio1 versity of Lund, Sweden, Thesis 1974 (reprint). Acta Orthop
36:71-75 Scand 60 [SuppI232]:1-51
J6nsson A, Herrlin K, Jonsson K et al (1996) Radiation dose Siu D, Cook DV, Broekhoven LD et al (1991) A standardized
reduction in computed skeletal radiography: effect on technique for lower limb radiography. luvest Radio126:71
image quality. Acta RadioI37:128-133 Spector TD, Co oper C, Cushnaghan Jet al (1992) A radiograph-
Keats TE (1996) Atlas of normal roentgen variants that may ie atlas of knee osteoarthritis. Springer, Berlin Heide1berg
simulate disease, 6th edn. Mosby-Year Book, st. Louis NewYork
Kellgren JH, Lawrence JS (1957) Radiologic assessment of Stedtfeld HW, Strobel M (1983) A new holding deviee for the
osteoarthritis. Ann Rheum Dis 16:494-501 preparation of stress roentgenograms of the knee joint.
Leach RE, Gregg T, Silber FJ (1970) Weight bearing radiogra- Unfallheilkunde 86:230-235
phy in osteoarthritis of the knee. Radiology 97:265-268 Strobel M, Stedtfeld HW (1990) Diagnostie evaluation of the
Lugo-Olivieri CH, Scott WW Jr, Zerhouni EA (1996) Fluid-fluid knee. Springer, Berlin Heidelberg New York
2 Arthrography
A. MARK DAVIES and VICTOR CASSAR-PULLICINO
CONTENTS 2.2
2.1 Introduction 19
Indications
2.2 Indications 19
2.3 Contraindications 19 Historically the main indication for knee arthrog-
2.4 Technique 19 raphy was soft tissue injury, be it meniscal and/or
2.5 Complications 20 ligamentous. Although in the developed world knee
2.6 Interpretation 20
2.7 Accuracy 21
arthrography has been superseded by other tech-
2.8 Conclusion 21 niques, it may be of value if the cost of the procedure
References 21 is critical or the access to MR imaging is limited. In
developing countries, however, it is debatable wheth-
er knee problems are a health priority.
2.1
Introduction
2.5
Complications
CONTENTS upon the condition being imaged, the model and age
of scanners used, the scanning protocol and the expe-
3.1 Introduction 23
3.2 Developments in CT 23 rience and ability of the scanner operators and the
3.2.1 Slip Rings 24 radiologist. Whilst CT is "losing ground"to MR imag-
3.2.2 X-ray Tubes 24 ing for imaging of the musculoskeletal system, it
3.2.3 X-ray Detectors 24 should not be forgotten that the methods are comple-
3.2.4 Helical CT (Spiral or Volume Scanning) 24
mentary. In addition, CT "came first", is more widely
3.2.5 CT "Fluoroscopy" 25
3.2.6 Data Manipulation 25 available, cheaper and quicker, and can be easier to
3.2.7 Reformatted Images 26 perform weH and interpret. CT remains more suit-
3.3 Scan Image Quality 27 able than MR imaging in the assessment of acute
3.3.1 Internal Metalwork from Fixation Devices 27 trauma (e.g. intra-articular fractures). The addition
3.3.2 CT Number, Hounsfield Units,
Window Width and Levels 28
of arthrography further increases the specificity and
3.4 CT of the Knee 29 sensitivity of both MR imaging and CT for articular
3.4.1 Anatomy 30 and meniscallesions. There is a law of diminishing
3.4.2 Immobilisation 31 returns as these sensitivities and specificities creep
3.5 Indications 31 ever doser to 100%, which also reduces the real
3.5.1 Trauma 31
3.5.2 Knee Morphology and Surgery 32
difference between the two techniques. CT remains
3.5.3 Patellofemoral Joint 33 essential in the assessment of patients in whom MR
3.5.4 Articular Cartilage 34 imaging is contra -indicated (e.g. owing to intracrani-
3.5.5 Soft Tissues 35 al aneurysm dips or cardiac pacemakers). CT there-
3.6 Arthrography 36 fore continues to have a role in the diagnosis and
3.6.1 Role and Indications 36
3.6.2 Technique 37
management of many pathologies of the knee. Having
3.7 CT-Guided Interventions 37 decided that CT is an appropriate investigation for an
3.8 Conclusion 38 individual, the precise format of the examination will
References 39 depend upon the suspected pathology and the equip-
ment available. Whilst this chapter starts by describ-
ing the recent developments in CT scanners and the
value of these to knee imaging, the main aim is to
3.1 outline those considerations that should optimise the
Introduction images obtained, whatever CT scanner is used.
re cent developments have been made include helical same but the time in which the slice was acquired was
scanning, multislice acquisition and real-time CT reduced. In addition, for helical scanning continu-
"fluoroscopy". These developments have been made ous X-ray output for up to 60 s may be required. The
on the back of improving technology which includes dis advantage of these X-ray tubes is the increased
slip rings for power and data transmission to and ease with which high radiation doses can be given to
from the gantry, higher heat loading X-ray tubes, patients during CT investigations. This disadvantage
high-efficiency solid state X-ray detectors, and faster may be mitigated by the developments of solid state
data transmission and processing abilities of the elec- and multislice detectors.
tronics.
3.2.3
3.2.1 X-ray Detectors
Slip Rings
Xenon gas detectors, used in CT scanners for many
In order to acquire X-ray transmission data in all years, have a conversion efficiency (X-rays to signal
directions across a slice of the patient, the X-ray tube strength) of around 60%, which can diminish further
has to travel around the entire circumference of a if the detectors are not maintained. Solid state crystal
circle around the slice. If the tube is supplied with detectors may have conversion efficiencies of nearly
power by cables, then these have to wrap around 100%, resulting in a 40% reduction in patient radia-
the circle as the tube moves. In order to unwrap the tion dose for the equivalent scan appearances. The
cables, the next slice is performed by rotating the tendency for solid state detectors to continue emit-
tube in the opposite direction. This design requires ting light after the X-rays had terminated (afterglow)
more than 360° of tube rotation as initial acceleration and other technical problems with respect to the size
and final deceleration distances are also required. of the front face of the individual detectors and the
Powerful motors and brakes are required to cope interspace material between adjacent detectors have
with the inertia of this system (which may include been largely overcome.
the X-ray detectors and counterweights to balance Although one of the earliest EMI CT head scan-
the gantry) and a significant time delay is necessary ners acquired two slices at the same time with a
between each slice acquisition to allow for these pencil beam of X-rays passing to two adjacent detec-
acceleration and deceleration phases. Replacing the tors, the ease with which solid state detectors can
cables with slip rings (large-circumference electrical- be stacked in parallel adjacent channels has facilitat-
ly conducting rings) which encircle the X-ray tube ed the re-development of multislice scanners. These
path, and transferring power from the rings to the scanners can acquire several sections simultaneously,
X-ray tube via conducting brushes on the X-ray tube which can be separately proeessed to give large num-
gantry, allows the gantry to be continuously rotated bers of thin seetions, or recombined to give fewer
in one direction. This has several advantages - rapid thicker seetions with lower noise.
acceleration and deceleration of the gantry are no
longer required yet a faster rotation speed can be
achieved, giving shorter scan acquisition tim es. The 3.2.4
time delay between slices need be no longer than that Helical eT (Spiral or Volume Scanning)
required for table movement in conventional acquisi-
tion mode and the potential for acquiring continu- The requirement for a break in the X-ray emission
ously updated X-ray transmission data paves the way whilst the table is moved to the next slice position
for both helical scanning and CT fluoroscopy. was overeome by the development of helical sean-
ning. Helical scanning is performed by moving the
table continuously during the exposure, from the first
3.2.2 slice loeation to the last. Thus a helix of X-ray trans-
X-rayTubes mission data through the sean volume is acquired.
To generate a CT image the data from adjaeent turns
The development of slip rings resulted in a require- of the helix are interpolated to produee transmission
ment for X-ray tubes to have both a higher heat data which are effeetively from a single slice loeation.
capacity and a higher maximum tube current, as the This process ean be performed at any loeation within
mAs required for a single slice remained much the the helix (except the first and last 1800 's, where there
Computed Tomography (CT) and CT Arthrography 25
other image plane will have the same image quality 3.2.7
as the acquisition images, potentially requiring even Reformatted Images
further hard copies.
Fast workstations, allowing rapid reformatting As spiral multislice scanning produces overlapping
and display of examinations in the most appropri- sections and thinner slice collimation (less than
ate plane for the pathology being demonstrated, 1 mm), in-plane and reformatted plane spatial resolu-
are therefore necessary, with hard copy restricted tions are now potentially similar, even for CT images
to representative images. Other image reconstruc- from small fields of view (Table 3.1). Volume acqui-
tion methods [curved planes, surface-rendered 3D sitions obtained in any plane can therefore be refor-
images, minimum or maximum intensity projec- matted into other planes without marked loss of
tions (Fig.3.2)] can produce a bewildering array of image quality. For scanners not capable of such fine
visually stunning images, though demonstration of collimation, CT in the most appropriate plane for the
the clinical utility of these methods is currently lim- expected pathology is still preferable, if achievable.
ited. As the best effective z-axis resolution for a spiral
The current state of the art device has a mul- scan acquisition is approximately half the X-ray
tislice helical scanner with solid state detectors, sub- beam collimation thickness, isometric voxels can be
second scan acquisition and image reconstruction achieved by selecting a combination of scan param-
times, CT fluoroscopy capability and a link to a eters such as those identified in Table 3.1. For non-
powerful workstation with real-time image manip- helical scanners, overlapping transverse sections will
ulation software. provide better z-axis resolution (e.g. 25-cm field of
a b
t ....... ....
view, I-mm slice collimation, D.S-mm table increment
between sections) but at twice the radiation dose to
the patient.
- 10
.... ----- ..
O+-----~----~----~--~-----,
More intrusive streak artefact is seen when the CT of all materials change with X-ray beam energy, there
plane is through locking screws in intramedullary are consequently only two fixed points on the Houn-
rods, bone surface plates or fixation screws. Care in sfield scale. These are -1,000, which is the HU value
patient positioning (including decubitus positions for no X-ray attenuation (i.e. a vacuum), and zero,
where necessary), combined with gantry angulation which corresponds to the HU value for water (at the
in order to align the scan plane with the long axis calibration pressure and temperature for the scan-
of any screws present, will reduce the number of sec- ner). The HU scale is, in fact, open ended, with high
tions degraded by streak artefact from the screws to a atomic number, high -density materials having values
minimum. In scanners with operator-selectable kVp, way in excess of the upper end of the usual scale (even
the use of the highest kVp setting will reduce streak on "extended sc ale" scanners) (Table 3.2).
artefact, as will the selection of a higher mAs (though
the combination of increased kVp and mAs results Table 3.2. Theoretical RU values for a variety of materials at
65keV
in considerably greater tube loading and patient irra-
diation). Streak artefact also may appear visually less Material RUvalue
intrusive on volume-rendered (3D) images (PRETO-
Adipose tissue -80
RIUS and FISHMAN 1999). Water o
Collagen 250
Dense cortical bone 1,600
3.3.2 Aluminium 2,300
Cl Number, Hounsfield Units, Window Width lron 34,000
Iodine 141,300
and Levels Lead 205,000
3.4
Cl of the Knee
For lower atomic number materials such as are
present in soft tissues, the X-ray attenuation and con- Most of the research literature on applying CT imag-
sequent CT number is predominantly influenced by ing to the knee was published in the 1980s, as the
the electron density of the material, which is, in turn, technique became widely available and equipment
closely related to the physical density of the material. improved. Descriptions of patient positioning, immo-
Even the CT number of water is influenced by differ- bilisation methods and reformatting image data sets
ences in temperature, and differences in density exist (including 3D reconstructions) were all described,
between water at room and at body temperature. The and much of what was said is still relevant to modern
presence of protein or high concentrations of salts scanners. The detail in which the anatomy of the
will increase the CT number ofbody fluids. Measure- knee can be demonstrated, and the clinical signifi-
ment of the CT number of a region of interest in an cance ascribed to its various structures, has improved
image must therefore be considered only a guide to significantly.
its composition. At an extreme not met in clinical
practice, but potentially relevant to research, the CT
number of ice at O°C (approximately -80 RU) is lower
than that of fat (the CT number of which increases as
it cools). Specimens scanned straight from the freez-
er may look quite different to that expected! (WHITE-
HOUSE et al. 1993).
The visual impression of the density of a region of
interest is influenced by the window and level settings
of the image, the calibration of the display and the
densities in the surrounding part of the image. Par-
ticularly within bone, the surrounding high density
of bone can give a lytic lesion the visual impression
of a lower density than actually exists. Consequently,
measurement rather than estimation of any region of
interest is essential; recording an image in which the
CT numbers of important regions of interest are mea-
sured is a useful addendum to the hard copy.
The window width and level are calibrated contrast
and brightness settings for image display. All scanners Fig. 3.5. Early lytic phase Paget's disease ofbone causing subtle
reduction in cortical density in the distal femur. Coarsening of
have pre-set buttons allowing different window/level
the trabecular pattern in the condyles is visible on a "standard"
combinations to be instantly applied. These com- bone window but the reduction in cortical density of the femo-
monly have settings deemed appropriate for bone, ral shaft requires a narrower window for clear demonstration.
lung, brain etc. but typically the bone setting is aimed (Courtesy of Dr. A. Horrocks, Wythenshawe Hospital, UK)
30 R. W. Whitehouse
3.4.1 3.4.2
Anatomy Immobilisation
A detailed knowledge of the appearances of the knee Although eT of the knee is a rapid procedure, immo-
and surrounding structures in all imaging planes is bilisation may be necessary to prevent movement
necessary for adequate interpretation. Knowledge of artefacts, partieularly in children. Sandbags, Velcro
anatomieal structures not easily or consistently dem- straps and stieking tape will usually suffice. Even
onstrated on eT is still needed to assess the likeli- better immobilisation is achieved routinely in trauma
hood of their involvement by any pathology whieh is patients by the pIaster of Paris cast or backslab they
demonstrated. The anatomy of the region has been are usually fitted with. Scanning through a pIaster
covered in other chapters. Selected eT images are cast does not significantly interfere with image quali-
included here for comparative purposes (Fig. 3.6). tywhilst the immobilisation achieved is usually excel-
For the clearest depietion of artieular surfaces and lent, such that a temporary cast is also worth consid-
fractures, images perpendicular to the plane of the ering for occasional patients inadequately immobil-
artieulation or fracture are usually best (vide infra), ised by other methods.
whilst for tendons, ligaments and menisci an imaging
plane perpendicular to the long axis of the structure
is useful. This limits the value of eT for the menisci
and tibiofemoral joint as direct imaging perpendieu-
lar to these structures is rarely feasible (unless iso-
metrie voxels from thin-section spiral eT are avail-
able, allowing high-quality reformats).
/20
4
~14
35
11 /19 12
51 71
18 10
81
8 ....13 "15
91
b
Fig.3.6a,b
Computed Tomography (CT) and CT Arthrography 31
I
25
....- 20
29
31 ~14
21_ 4.
4. ~241
51_
4
71 - 23 - - - 32
30
11 12
81
10 35 19
91 -15
18 13
"
27
Ar 16
I
14
/
k
\,;
11
10 15 12
5 I
71 13
91 /
---'"
e
/
25
\17 .... 20
Fig.3.6a-f. Anatomy of the knee on
26
selected transverse 2-mm CT sec-
21,
"'" 4
/
22
/241
tions, performed after single-con-
trast arthrography. Line drawings of
51, 23
each section identify structures as
enumerated: 1, Quadriceps tendon;
71- 12
2, vastus medialis; 3, vastus later-
35 19
91
alis; 4, femur; 5, sartorius; 6, long
18
27 saphenous vein; 7, gracilis; 8, semi-
membranosus; 9,semitendinosus; 10,
popliteal vein; 11, popliteal artery;
12, biceps femoris; 13, tibial nerve;
14, iliotibial band; 15, common pero-
f neal nerve; 16, suprapatellar bursa;
17, patella; 18, medial head of gas-
71.~
trocnemius; 19, lateral head of gas-
trocnemius; 20, lateral patellar reti-
Y 31 36 naculum; 21, medial collateralliga-
ment; 22, anterior cruciate liga-
ment; 23, posterior cruciate ligament;
24, popliteus; 25, patellar tendon;
26, medial synovial plica; 27, short
saphenous vein; 28, geniculate ves-
sels; 29, lateral meniscus; 30, medial
meniscus; 31, tibia; 32, lateral collat-
eral ligament; 33, fibular head; 34,
proximal tibiofibular joint; 35, plan-
taris; 36, tibialis anterior. Postscript
ct' indicates tendon
32 R. W. Whitehouse
3.5.2
Knee Morphology and Surgery
3.5.3.2
Dynamic Cl of the Patellofemoral Joint
dense lesions such as pigmented villonodular syno- popliteal vessels (ISHIKAWA et al.1999) and othervas-
vitis 1 may have characteristic appearances on cr cular lesions (Fig. 3.14). Similarly, softtissue enhance-
(CHEN et al. 1999; LIN et al. 1999). cr is unreliable for ment in masses or synovium can be demonstrated
follow-up scanning of the resection site of soft tissue but timing is critical, with peak enhancement being
sarcoma (HUDSON et al. 1985). later, less marked and more variable in onset than in
Contrast enhancement can be used with volume the abdomen. cr scanner software which pre-scans
rendering to demonstrate arte rial and graft stenosis at low mA to detect the onset of enhancement and
and obstructions after vascular surgery down to the triggers the study at that point may have a role.
Fig. 3.15. A pair of meniscal ossicles are demonstrated on CT arthrography using twin spiral acquisition with I-mm collimation,
a pitch of land reconstruction at 0.5-mm increments. (Courtesy of Dr. S. Bianchi, Höpital Cantonal, Geneva; reproduced from
MARTINOLl et al. 2000, with permission)
3.6.2
Technique
version associated with anterior knee pain. Clin Orthop tibial tunnellocation and the anterior aspect of the poste-
339:152-155 rior cruciate ligament insertion. Arthroscopy 13:465-473
Eckstein F, Adam C, Sittek H, Becker C, Milz S, Schulte E, Reiser Mughetti M, De-Masi M, Teodorani A, Gagliardi S (1998) Falsa
M, Putz R (1997) Non-invasive determination of cartilage immagine con Tomografia Computerizzata di lesione de!
thiekness throughout joint surfaces using magnetie reso- menisco mediale dei ginocchio. Studio retrospettivo in 400
nance imaging. J Biomech 30:285-289 casi. Radiol Med Torino 96:310-312
Eckstein F, Schnier M, Haubner M, Priebsch J, Glaser C, Engl- Muhle C, Brossmann J, Heller M (1999) Kinematie CT and MR
meier KH, Reiser M (1998) Accuracy of cartilage volume imaging of the patellofemoral joint. Eur RadioI9:508-518
and thiekness measurements with magnetie resonance Nagamine R, Miura H, Inoue Y, Tanaka K, Urabe K, Okamoto
imaging. Clin Orthop 352:137-148 Y, Nishizawa M, Iwamoto Y (1997) Malposition of the tibial
Froe!ieh JJ, Ishaque N, Saar B, Regn J, Walthers EM, Mauer- tubercle during flexion in knees with patellofemoral arthri-
mann F, Klose KJ (1999) Steuerung von perkutanen Biop- tis. Ske!etal Radio126:597 -60 1
sien mittels CT-Fluoroskopie. [Control of percutaneous Nagao N, Tachibana T, Mizuno K (1998) The rotational angle
biopsywith CT fluoroscopy.] Rofo Fortsehr Geb Rontgenstr in osteoarthritie knees. Int Orthop 22:282-287
Neuen Bildgeb Verfahr 170:191-197 Nawfel RD, Judy PF, Silverman SG, Hooton S, Tuncali K,
Goldberg SN, Keogan MT, Raptopoulos V (2000) Percutaneous Adams DF (2000) Patient and personne! exposure during
CT-guided biopsy: improved confirmation of sampling site CT fluoroscopy-guided interventional procedures Radiol-
and needle positioning using a multistep technique at CT ogy 216:180-184 (and comments pages 9-10)
fluoroscopy. J Comput Assist Tomogr 24:264-266 Pretorius ES, Fishman EK (1999) Volume-rendered three-
Gray SD, Kaplan PA, Dussault RG (1997) Imaging of the knee. dimensional spiral CT: musculoske!etal applications. Radio-
Current status. Orthop Clin North Am 28:643-658 graphies 19:1143-1160
Hangody L, Kish G, Karpati Z, Udvarhe!yi I, Szigeti I, Be!y M Radermacher K, Portheine F, Anton M, Zimolong A, Kaspers
(1998) Mosaieplasty for the treatment of artieular carti- G, Rau G, Staudte HW (1998) Computer assisted orthopae-
lage defects: application in clinieal practiee. Orthopedies die surgery with image based individual templates. Clin
21:751-756 Orthop 354:28-38
Haubner M, Eckstein F, Schnier M, Losch A, Sittek H, Becker Sans N, Morera-Maupome H, Galy-Fourcade D, Jarlaud T, Chi-
C, Kolem H,Reiser M,Englmeier KH (1997) A non-invasive avassa H, Bonnevialle P, Giron J, Railhac JJ (1999) Resec-
technique for 3-dimensional assessment of artieular carti- tion percutanee sous controle tomodensitometrique des
lage thiekness based on MRI, part 2. Validation using CT osteomes osteoides. Suivi a moyen terme de 38 cas. [Per-
arthrography. Magn Reson Imaging 15:805-813 cutaneous resection under computed tomography guid-
Hodge JC (1999) Percutaneous biopsy of the musculoskeletal ance of osteoid osteoma. Mid-term follow-up of 38 cases.]
system: a review of 77 cases. Can Assoc Radiol J 50:121-125 J RadioI80:457-465
Hsieh J (1997) Analysis of the temporal response of computed Sati M, deGuise JA, Drouin G (1997) Computer assisted knee
tomography fluoroscopy. Med Phys 24:665-675 surgery: diagnosties and planning ofknee surgery. Comput
Hudson TM, Schakel M, Springfie!d DS (1985) Limitations of Aided Surg 2:108-123
computed tomography following excisional biopsy of soft Schutzer SF, Ramsby GR, Fulkerston JP (1986) The evaluation
tissue sarcomas. Skeletal Radiol13:49-54 of patellofemoral pain using computerized tomography: a
Ishikawa M, Morimoto N, Sasajima T, Kubo Y (1999) Three- preliminary study. Clin Orthop 204:286-293
dimensional computed tomographie angiography in lower SilvermanSG, Tuncali K, Adams DF, Nawfe! RD, Zou KH,
extremity revascularization. Surg Today 29:243-247 Judy PF (1999) CT fluoroscopy-guided abdominal inter-
Lin J, Jacobson JA, Jamadar DA, Eilis JH (1999) Pigmented vil- ventions: techniques, results, and radiation exposure. Radi-
lonodular synovitis and related lesions: the spectrum of ology 212:673-681
imagingfindings.AJR 172:191-197 Stevens MA, EI-Khoury GY, Kathol MH, Brandser EA, Chow S
Lomasney LM, Griffiths AA, Tobin JP, Rabin SI, Cronin TR (1999) Imaging features of avulsion injuries. Radiographies
(1999) Comparison of 4 computed tomographie imaging 19:655-672
protocols for lateral tibial plateau fractures. Can Assoc Van de Berg BC, Lecouvet FE, Poilvache P, Dubuc JE, Bedat
RadiolJ 50: 11 0-116 B, Maldague B, Rombouts JJ, Malghem J (2000) Dual-detec-
Malghem J, Vande-berg BC, Lebon C, Lecouvet FE, Maldague tor spiral CT arthrography of the knee: accuracy for detec-
BE (1998) Ganglion cysts of the knee: artieular communiea- tion of meniscal abnormalities and unstable meniscal tears.
tion revealed by delayed radiography and CT after arthrog- Radiology 216:851-857
raphy.AJR 170:1579-1583 Voto SJ, Cook AJ, Weiner DS, Ewing JW,Arrington LE (1990) Treat-
Mann TA, Black KP, Zanotti DJ, Barr M, Teater T (1999) The ment of osteoid osteoma by computed tomography guided
natural his tory of the intercondylar notch after notchplasty. excision in the pediatrie patient. J Pediatr Orthop 10:510-513
Am J Sports Med 27:181-188 Walker C, Cassar-Pullicino VN, Vaisha R, McCall IW (1993)
Martino F, De Serio A, Macarini L, Rizzo A, Laforgia R, Roton- The patello-femoral joint - a critieal appraisal of its geo-
do A,Ange!elli G (1998) Ultrasonographyversus computed metrie assessment utilizing conventional axial radiography
tomography in evaluation of the femoral trochlear groove and computed arthro-tomography. Br J RadioI66:755-761
morphology: a pilot study on healthy, young volunteers. Eur Watson JT, Coufal C (1998) Treatment of complex lateral
RadioI8:244-247 plateau fractures using Ilizarov techniques. Clin Orthop
Martinoli C, Bianchi S, Spadola L, Garcia J (2000) Multimodal- 353:97-106
ity imaging assessment of meniscal ossicle. Skeletal Radiol Whitehouse RW, Economou G, Adams JE (1993) The influence
29:481-484 of temperature on quantitative computed tomography:
McGuire DA, Hendricks SD, Sanders HM (1997) The rela- implications for mineral densitometry. J Comput Assist
tionship between anterior cruciate ligament reconstruction Tomogr 17:945-951
4 Magnetic Resonance Imaging
CHRISTINE HERON and ANDREW HINE
4.1
Introduction
4.2
Following the introduction of magnetic resonance General Considerations
(MR) imaging of the knee into clinical practice there
was an exponential rise in publications documenting Many factors contribute to the production of opti-
its efficacy. Research into this area of musculoskele- mum images. In addition to the hardware employed,
tal MR imaging has now plateaued and the present these include the signal to noise ratio (SNR), con-
emphasis is on the fine tuning of certain aspects of trast to noise ratio (CNR), resolution, pulse sequenc-
es employed and scan time. The field strengths of
magnets vary and are described as low, medium and
C.HERON,MD high. Low field strength magnets operate at 0.2 T and
Consultant Radiologist, St. George's Hospital, Blackshaw Road,
below. Medium strengths are between 0.3 and 0.9 T
London, SW17 OQT, UK
A.HINE,MD but are usually 0.5- T magnets. High fields are 1-1.5 T.
Consultant Radiologist, Central Middlesex Hospital, Acton The field strength influences a number of factors that
Lane, London, NW10 5NS, UK are particularly important in relation to the SNR.
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
42 C. Heron and A. Hine
The SNR is influenced by both operator-dependent field of view dimensions x slice thickness
voxeI = - - - - - - - - - - - - - -
and non-operator-dependent factors. The non-oper- matrixsize
ator-dependent factors include the field strength of
the magnet and the intrinsic molecular structure of The matrix size is determined by the number of
the tissue being examined. The molecular structure frequency encodings and/or phase encodings. Low
of the tissue determines the density of protons and frequency and phase encoding numbers result in a
the Tl, T2 and T2* relaxation times. Operator-depen- coarse matrix with large voxels (assuming a fixed
dent factors include the type of coil, the field of square field of view). Large voxels have a higher SNR
view, the number of acquisitions, sampling band- than small voxels because they contain more pro-
width, matrix size, slice thickness and the TR, TE and tons to produce signals. The SNR is influenced by any
flip angle. Changes in TR influence the degree of Tl change in parameter which alters the voxel volume.
weighting and hence the signal attenuation. In T2- It can be appreciated, for instance, that doubling the
Magnetic Resonance Imaging 43
field of view results in a doubling in the length of two Reducing the sampling bandwidth results in less high
sides of the voxel and hence a fourfold increase in frequency noise being sampled and consequently, as
the SNR, whereas doubling the slice thickness dou- the signal remains unaltered, there is an increase
bles the length of only one side of the voxel and in the SNR. However, it is necessary to increase the
hence only doubles the SNR. Both matrix size and sampling time as the sampling bandwidth decreases
number of acquisitions affect the SNR. Doubling the to prevent any loss of resolution. This results in an
matrix size in both the phase encoding and frequen- increase in the minimum TE that can be employed.
cy encoding directions for a fixed field of view results The choice of pulse sequence influences the SNR. In
in the SNR being reduced by a factor of 2 square spin echo sequences all of the longitudinal magneti-
root of 2, but the scan time is doubled because there sation is converted into transverse magnetisation, but
are twice as many phase encoding steps to sampie in gradient echo pulse sequences only a proportion
(Fig. 4.3). Doubling the number of acquisitions dou- of the longitudinal magnetisation is converted into
bles the scan time but only increases the SNR by a transverse magnetisation as in this case the flip angle
maximum of square root of 2 (about 1.4). is less than 90°. As a result of this, the signal is gener-
In the frequency encoding axis the sampling band- ally greater in spin echo sequences. In summary, SNR
width determines the range of frequencies sampled. can be increased by using a large field of view, a coarse
44 c. Heron and A. Hine
the field of view has the largest effect on the voxel Assuming that the frequency encoding is double the
volume. Halving the field of view results in a four- phase encoding, the field of view in the phase direc-
fold decrease in the voxel volume, with correspond- tion will be half that in the frequency direction. The
ing improvement in the resolution (Fig. 4.4). Decreas- pixels remain square and the spatial resolution is
ing the slice thickness and therefore reducing the maintained.
voxel size also increases the spatial resolution. Whilst It may be necessary to increase the field of view in
pixels which are square result in the best spatial reso- the phase direction in order to obtain coverage. This
lution, it is possible for pixels to be rectangular. This will increase scan time, and the SNR from the smaller
may arise when the field of view remains square but square pixels will be lower than that from the rectan-
an uneven matrix is chosen. The frequency number gular pixels. Typically such systems permit the selec-
of the matrix is usually the highest and the phase tion of rectangular pixels which keep the field of view
number is entered to adjust the scan time and res- square and increase the SNR in the phase direction
olution. When the phase number is less than the with adequate coverage. The phase encoding number
frequency number, the spatial resolution is reduced and scan time are maintained and the pixels are
along the phase axis. In some systems the pixels rectangular in the phase direction. As each pixel is
remain square irrespective of the matrix selected. larger, the SNR increases and the spatial resolution is
a b
c d
By adjusting the imaging parameters it is possible while solid structures are of lower signal intensity
to influence the weighting of a gradient echo than in conventional spin echo images.
sequence. Tl-weighted gradient echo images are pro-
duced by employing a large flip angle, short TR and 4.2.5.6
short TE. T2* -weighted images result from a small Echo Planar Imaging
flip angle and a TR which is relatively long in order
to permit sufficient recovery of the longitudinal mag- Echo planar imaging (EPI) is a further fast imaging
netisation of the protons. Gradient echo images have technique in which multiple gradient echo es are
scan times which are typically significantly shorter acquired for each excitation. In effect, EPI is the gra-
than spin echo images. The main disadvantage is dient echo equivalent of fast spin echo imaging. The
their susceptibility to magnetic field inhomogene- shortest echo plan ar images can be obtained in less
ities. than 100 ms using single-shot techniques. There is
Gradient echo sequences are frequently obtained only one excitation pulse and all of the echoes are
in volume acquisitions. This is a particularly advan- obtained from this. Movement artefacts do not arise
tageous technique in MR imaging of the knee. It can with these very fast imaging times but spatial reso-
be used for visualising very small structures and for lution and SNR can pose problems. These may be
reformatting images of structures which do not lie improved by employing multi-shot techniques in
in a single anatomical plane, such as ligaments. An which, after each excitation pulse, several echo es are
entire volume of tissue is imaged and the chosen obtained. Movement artefact is more problematic in
slice thickness may be as low as 1 mm. Thinner slices multi-shot techniques but can be reduced by shorten-
result in longer scan times, as illustrated in the equa- ing the echo train length. At present, EPI is not used
tion below: routinely in the knee. In the future it may have a role
in dynamic studies of the knee during movement and
Scan time = TR x number of excitations x phase encoding in MR fluoroscopy for guiding interventional pro ce-
number x slice encoding number dures.
Using the fast spin echo technique, proton density and 4.2.5.7.1
T2-weighted images can be obtained in a fraction of Fat Saturation
the time required for conventional spin echo sequenc-
es. For each 90 0 pulse, multiple echoes are obtained In the fat saturation technique an RF pulse with the
as opposed to one. This is achieved by rapidly apply- same resonance frequency as fat is added to each
ing multiple 1800 pulses with different phase encoding slice-selection RF pulse. This is followed by a homo-
gradients for a set TE. The different phase encodings geneity spoiling gradient pulse which dephases the
result in multiple echoes (up to 32), thereby potential- lipid signal.
ly reducing imaging time by 32-fold. Fast spin echo The dephasing results in no signal emanating from
images are susceptible to blurring and edge artefacts the fat in the MR images. The advantages of this tech-
but these can be minimised by using a short echo nique are that fat only is suppressed and it may be
train length and a long effective TE. As a result of the used with any imaging sequence. It is useful for post-
frequent refocussing pulses applied in the fast spin contrast scanning, e.g. following intravenous contrast
echo technique, fat appears of higher signal intensity, enhancement and MR arthrography. The fat satura-
48 c. Heron and A. Hine
4.2.5.7.2
Inversion Recovery
water protons become partially saturated and they whieh some contrast diffuses into the joint. In order
therefore emit less signal. In this way tissues which to facilitate this process, the knee should be actively
contain bound water demonstrate lower signal inten- exercised following the intravenous injection of con-
sity on magnetisation contrast images. The co nt rast trast. Whilst certain success has been described with
generally resembles that of T2-weighted images and this technique (SUH et al. 1996), it fails to distend the
considerable signal is lost in solid tissues but this joint with fluid, which is the most valuable aspect of
does not occur in fluid or adipose tissue. the procedure. The employment of indirect arthrog-
raphy is therefore likely to remain limited. The direct
injection of contrast into the knee joint is a straight-
4.2.6 forward procedure. It is possible for the experienced
Image Display radiologist to perform this in the MR room without
the need for radiographie screening but screening
At the present time MR images are generally report- guidance may be preferred. Either a medial or a lat-
ed from hard copy. Some centres re-image the menis- eral approach is employed. The posterior border of
ci using greater magnification and special narrow the patella is palpated and the ne edle inserted imme-
window settings to enhance meniscal pathology. This diately posterior to this at amid point between the
practice is time consuming and costly and has not superior and inferior patellar poles. Free flow of fluid
been shown to offer significant advantages. Cost sav- confirms correct needle placement. The preferred
ings may be made by reporting investigations from a contrast agent differs between centres. The majority
work station. This also permits manipulation of the use a dilute gadolinium solution (004 cc of gadolini-
images and reformatting of three-dimensional datas- um added to 100 cc of saline). If the procedure is to be
ets at the time of reporting. Limited hard copy of rele- performed under radiographie screening guidance,
vant images can be produced for clinical use. Report- part of the saline solution may be replaced by iodin-
ing times are generally longer using this technique. ated contrast.Alternatively, iodinated contrast may be
Three-dimensional images have been obtained in one injected initially in order to check that the needle is
plane and images reformatted in additional planes correct1y positioned and then the infusate is changed
prior to reporting of the scans (WIESLANDER et al. for the gadolinium solution. Approximately 50 ml of
1998). Whilst this practiee reduces imaging time, the contrast agent is injected into the knee joint. As the
time required for reformatting and reporting in this use of intra-articular gadolinium has not received
study was greater than the imaging time which it universal acceptance by pharmaceutical regulatory
saved. bodies, some centres inject normal saline into the
joint. Following the injection of a gadolinium-based
contrast agent, Tl-weighted images with fat satura-
4.2.7 tion are employed. When normal saline is used as the
MR Arthrography contrast agent, T2-weighted or gradient echo images
with T2* weighting are employed (Fig. 4.6).
The introduction of MR imaging largely obviated the
need for conventional arthrography of the knee. Cer-
tain conditions, however, are better appreciated in
the presence of fluid within the joint and there is a
limited role for MR arthrography. This includes the
assessment of the postoperative meniscus, the eval-
uation of osteochondral lesions and demonstration
of loose bodies (PEH and CASSAR-PULLICINO 1999;
GRAINGER et al. 2000).
Adecision to perform MR arthrography is based
on the findings on a non-contrast scan and the need
to further assess one of the above abnormalities in a
patient with no joint effusion. MR arthrography may
be either direct or indirect.
Indirect arthrography is performed by injecting Fig.4.6. MR arthrogram employing normal saline. Fat-satu-
gadolinium intravenously and scanning the knee rated T2-weighted axial image showing a detached portion of
after an interval of approximately 40 min during hyaline cartilage on the medial facet of the patella
50 C. Heron and A. Hine
The invasive nature of MR arthrography should Coronal images are essential for the evaluation of
not unduly limit the use of this technique, which the collateralligaments and axial images supplement
most patients find acceptable (ROBBINS et al. 2000). these. The patellofemoral joint is best assessed on
The complication rate from MR arthrography should axial images but can also be visualised in the sagittal
be very low. The risk of the introduction of infection plane. Traditionally, MR imaging of the knee has been
should, however, be borne in mind. BROSSMANN et performed with the knee fully extended and posi-
al. (1996) compared MR imaging, MR arthrography tioned in 10° to 15° of external rotation. In this posi-
with both saline and gadolinium and CT and CT tion the anterior cruciate ligament is aligned paral-
arthrography in the detection of osseous and carti- lel with sagittally orientated images. This permits the
laginous loose bodies in the knees of cadavers. MR fulilength of ligament to be assessed on one sagittal
arthrography with gadolinium was the most accu- section. There are, however, different ways of achiev-
rate technique, followed by MR arthrography with ing this aim. Positioning of sagittal images may be
saline. The best sequences for the detection of loose determined by visualising the alignment of the ante-
bodies on both unenhanced scans and images fol- rior cruciate ligament on axial or coronal images
lowing saline arthrography were a T2-weighted spin and planning the sagittal images in this orientation
echo sequence and a spoiled GRASS sequence. (Fig. 4.8). Alternatively, images of the anterior cruci-
It is very likely that, as MR software continues to ate ligament can be reformatted from three-dimen-
improve, the indications for MR arthrography will sional datasets. In centres using thin slice thickness-
decrease. es for the sagittal images, the anterior cruciate lig-
ament is typically visualised throughout its length
on a single sagittal seetion even with the knee in
4.2.8 the neutral position. Obliquely orientated images are
Patient Positioning and Imaging Planes therefore not necessary.
a b
Fig.4.7. a T2* gradient echo sagittal image of the knee showing a tear of the posterior horn of the medial meniscus.
b The coronal image demonstrates the displaced portion of the medial meniscus (arrow) in this bucket handle tear
Magnetic Resonance Imaging 51
Fig.4.9a-d. Evaluation of
meniscal degeneration. a
T2* gradient echo sagit-
tal sequence shows high
signal intensity degenera-
tive change in the poste-
rior third of the medial
meniscus. b Spin echo
Tl-weighted sequence of
the same meniscus is
less sensitive for the
detection of the meniscal
degeneration. c Fast spin
echo fat -saturated proton
density sequence is also
sensitive for the meniscal
degeneration. d Fast spin
echo fat-saturated T2-
weighted sequence fails
to show the meniscal
degeneration
ments in all three planes (ROYCHOWDHURY et al. 1997). T2* -weighted gradient echo images for the detection
The requirement to include the fulliength of the ante- of subtle signal change within the ligament (Fig. 4.10).
rior cruciate ligament on a single sagittal image has Proton density-weighted fast spin echo images with
already been indicated. Several studies have investi- fat suppression are also useful in the assessment of
gated the value of scanning the anterior cruciate liga- both the cruciate and collateralligaments. In princi-
ment in some degree of flexion (NAKANISHI et al.1997; pIe, inversion recovery images provide excellent con-
NUTSU et al. 1996, 1998; PEREIRA et al. 1998). It has trast between the injured ligament, adjacent joint fluid
been demonstrated that this generally aids assessment and soft tissues. In practice, however, this sequence is
of the integrity of the ligament. As with the menisci, sometimes of limited value as motion artefact from
conventional spin echo sequences provide reasonably adjacent vascular structures may impair visualisation
accurate assessment. Tl-weighted spin echo images of the ACL. Inversion recovery images in the coronal
visualise the anatomy of the non-injured ligament plane, however, are of value in the assessment of the
well, but in the presence of a joint effusion the integrity collateralligaments.
of the cruciate ligaments may be difficult to assess. For
this reason, proton density and T2-weighted images
are preferred because they provide contrast between 4.2.11
the ligament and joint fluid. T2*-weighted gradient Hyaline Cartilage
echo images are sensitive to signal intensity changes
within the ligament resulting from injury, and they Numerous researchers have investigated the value
also accurately assess the integrity of the ligament. The of MR imaging in the assessment of hyaline car-
anterior cruciate ligament is well visualised on fast tilage abnormalities using a wide variety of pulse
spin echo images (HA et al. 1998; MUNK et al. 1997). sequences. The search for a technique which accu-
T2-weighted fast spin echo images, particularly with rately assesses hyaline cartilage has been intensified
fat suppression, provide excellent contrast between recently by the further refinement of treatments
the ligament and adjacent joint fluid and are especial- for hyaline cartilage disease. Surgical interventions
ly useful for assessing the integrity of the ligament include shaving and repair of the hyaline cartilage
(Fig. 4.10). They do not appear to be as sensitive as whilst drug therapies aim to modify degenerative
a b
Fig.4.10a, b. Assessment of the cruciate ligaments. a Fast spin echo T2-weighted fat-saturated image clearly delineates the
anterior cruciate ligament and shows increased signal intensity and partial disruption of the fibres of the posterior cruciate
ligament. Note also the high signal intensity bone bruising in the patella. b Gradient echo T2* image of the same knee also
demonstrates the anterior cruciate ligament weil and the partial tear of the posterior cruciate ligament. The signal change
within the posterior cruciate ligament is more conspicuous on the gradient echo T2* image. The bone bruising in the patella
is not detected with this sequence. however
54 C. Heron and A. Hine
change in articular cartilage. Probably of greater the two images acquired (one from the first echo and
importance is the knowledge that a significant the other from the second echo). By optimising the
number of patients presenting with internal derange- imaging parameters it is possible to produce images
ment of the knee will be found to have abnormalities which have the highest contrast between intermedi-
of hyaline cartilage as the sole cause for their symp- ate signal hyaline cartilage (from the first echo) and
toms. In addition, it is known that the outcome of high signal fluid (from the second echo) (Figs.4.11b,
treatment for other causes of internal derangement 4.12). Additional contrast is obtained by filtering the
such as meniscal tears is influenced by the state of second echo image to eliminate noise before adding
the articular cartilage. Accurate information relating it to the first echo image (HARDY et al. 1996; RUEHM
to the articular cartilage is therefore of considerable et al. 1998). The sequence is not,however,provided by
importance to many clinical decisions. all MR manufacturers.
The requirement in terms of MR imaging is for a Fat-saturated proton density and T2-weighted fast
technique which provides high resolution and sharp spin echo images have been used for hyaline carti-
contrast between hyaline cartilage and the adjacent lage imaging (KAWAHARA et al. 1998; BREDELLA
bone, soft tissues and fluid which may be present et al. 1999). The results using these sequences are
within the joint. Histologically, hyaline cartilage has comparable to those employing gradient echo and
three layers (a superficial, a transitional and a deeper DESS sequences with intrasubstance cartilage damage
radial zone). Although MR images demonstrating shown more clearly (Fig. 4.11). The abilityoffast spin
layers have been reported by several researchers echo techniques to visualise hyaline cartilage better
(DISLER et al. 1996, 2000; ERICKSON et al. 1996; ERICK- than conventional spin echo techniques is thought to
SON and PROST 1997; FRANK et al. 1997; KNEELAND be due to fast spin echo images providing magnetisa-
2000; MCCAULEY and DISLER 1998; RUBENSTEIN et ti on transfer contrast in addition to T2 contrast (YAO
al. 1993, 1997; UHL et al. 1998; WALDSCHMIDT et al. et al. 1996).
1997; XIA et al. 1997), they do not appear to corre- As has already been indicated, MR arthrography
spond exacdy to the histologicallayers. As yet it is enhances hyaline cartilage visualisation. Gadolini-
not clear to what extent truncation arte fact and the um-chelate contrast agents have been used intrave-
"magie angle" phenomenon are responsible for the nously and injected direcdy into the knee to identify
appearance of layers in MR images of hyaline carti- regions of hyaline cartilage degeneration (BASHIR
lage (ERICKSON et al. 1996; FRANK et al. 1997; RUBEN- et al. 1997). These agents diffuse into the surface of
STEIN et al. 1997; WACKER et al. 1998). cartilage in which the proteoglycan matrix is disrupt-
Hyaline cartilage is of low signal intensity on Tl- ed but the cartilage is grossly intact. This has been
weighted spin echo sequences. It is therefore indis- reported to enhance visualisation of early degenera-
tinguishable from joint fluid and this sequence is tive changes on MR images.
not of value in hyaline cartilage assessment. Articu- There have been several reports of the measure-
lar cartilage also appears as low signal intensity on ment of hyaline cartilage volume using MR imaging
conventional T2-weighted images but in this case (Dupuy et al. 1996; ECKSTEIN et al. 1996, 1998,2000;
joint fluid is of high signal intensity and an arthro- HYHLIK-DuRR et al. 2000; KSHIRSAGAR et al. 1998;
gram effect is produced. This sequence, however, PIPLANI et al. 1996; SITTEK et al. 1996; STAMMBERGER
suffers from inadequate differentiation between et al. 1999). The present techniques are labour inten-
hyaline cartilage and cortical bone and a lack of sive as they require predominandy manual delinea-
sensitivity to hyaline cartilage defects and areas of tion of hyaline cartilage and it seems unlikely that
inhomogeneity. they will be widely adopted until the process can be
Gradient echo images offer considerable advantag- automated. In general, the measurements appear to
es over conventional spin echo techniques. Using a fat- be accurate and reproducible.
saturated T2* -weighted three-dimensional spoiled Studies using short echo time projection recon-
gradient -echo sequence, hyaline cartilage is of high struction MR imaging, and measurements of T2
signal intensity and contrasts sharply with adjacent relaxation time of articular cartilage, have demon-
structures of low signal intensity (Fig.4.11a). High strated T2 lengthening at sites of early cartilage
sensitivity and specificity rates for the detection of degeneration (BROSSMANN et al. 1997; GOLD et al.
hyaline cartilage defects have been reported using 1998; MOSHER et al. 2000). Their accuracy in hyaline
this sequence (DISLER et al. 1996; RECHT et al. 1996b; cartilage assessment has been shown to be superior
WANG et al. 1999). The dual echo in the steady state to that of fat-suppressed three-dimensional spoiled
(DESS) sequence produces one averaged image from GRASS images and magnetisation transfer contrast
Magnetic Resonance Imaging 55
Fig. 4.11a-d. Comparison of sequences for the demonstration of patellar hyaline cartilage. a T2* gradient echo sequence with fat
saturation. b DESS sequence. c Fat-saturated fast spin echo proton density sequence. d Fat-saturated fast spin echo T2-weighted
sequence. The hyaline cartilage is clearly differentiated from the joint fluid on all sequences but the intrasubstance cartilage
damage is seen more readily on the fat-saturated fast spin echo sequences
images. Images with diffusion weighting have also SNR of these images means that they are unlikely
shown promise in the evaluation of hyaline cartilage to have clinical applications. Magnetisation transfer
(BURSTEIN et al. 1993; FRANK et al. 1999). As sodium techniques have been described above. It was thought
nuclei are abundant in articular cartilage, sodium- that the presence of increased unbound water pro-
based MR imaging has been employed. The sodium tons in areas of degenerate hyaline cartilage would be
nuclei attach to proteoglycan macromolecules. These visualised using this technique. The present results
are reduced in degenerative disease, and sodium- suggest that magnetisation transfer images are not
based MR studies have been used to detect early comparable to other sequences for the evaluation of
degenerative change (REDDY et al. 1998). The poor hyaline cartilage (ADLER et al. 1996).
56 C. Heron and A. Hine
4.3
MR Protocol for
Routine Examination of the Knee
4.4
MR Protocols for
Specific Clinical Problems
4.2.12
Bone 4.4.1
Synovium and Soft Tissue Masses
The spectrum ofbony abnormalities which affect the
knee ranges from common findings such as degener- Normal synovium is not visible on non-contrast-
ative change and bone bruising due to trauma, to rare enhanced MR images, while the thickened inflamed
primary bone tumours. The majority of bony prob- synovium enhances following intravenous gadolini-
lems generate an increased signal intensity on fat- um DTPA. Thickened synovium can sometimes be
suppressed images, and it is therefore important to appreciated on non-enhanced images but is not reli-
include images with fat suppression in at least one ably demonstrated (RAND et al. 1999a). Synovial
plane. Fast STIR and T2-weighted images with fat thickening and synovial plicae mayaIso be more con-
suppression are comparable to conventional sequenc- spicuous following MR arthrography as the joint dis-
es (ARNDT et al. 1996). In fractures and avascular tension facilitates visualisation of synovial surfaces.
necrosis the lesion can be further evaluated using Tl- Soft tissue masses should be evaluated with a range of
weighted spin echo sequences. If there is difficuhy Tl-weighted and T2-weighted images with the addi-
determining whether alesion is solid or cystic, Tl- tion of fat suppression in order to permit char-
weighted spin echo sequences, supplemented if nec- acterisation and determine their extent. Tl-weight-
essary by Tl-weighted sequences following intrave- ed images following intravenous contrast determine
nous gadolinium, are employed. enhancement characteristics.
Magnetic Resonance Imaging 57
4.4.3 4.4.5
Tendons MR Imaging of the Postoperative Knee
MR imaging may be undertaken to evaluate acute The most frequent indications for performing post-
injuries or chronic conditions arising in tendons operative MR imaging in the knee are to assess the
around the knee joint. The patellar tendon followed menisci following previous resection or repair and the
bythe quadriceps insertion are most frequently inves- anterior cruciate ligament following reconstruction.
tigated. If the symptoms are thought to be localised Whatever the nature of the previous surgery to the
to the tendon, initial imaging using ultrasound is rec- knee, there is a likelihood of metal artefact on subse-
ommended. MR imaging may be employed for fur- quent MR images. For this reason, spin echo and fast
ther evaluation or may detect abnormalities which spin echo images are generally preferred to gradient
have not been suspected clinically. A 3-inch surface echo images. In the assessment of the postoperative
coil positioned directly over the tendon and the use meniscus, proton density-weighted and T2-weighted
of a high-field system provide optimal visualisation. fast spin echo images with fat saturation produce ade-
Images are obtained in the sagittal and axial planes. A quate visualisation (L1M et al. 1999). The postoper-
12-cm field of view is recommended. Preferred pro- ative meniscus may be difficult to assess as the crite-
tocols include Tl-weighted spin echo sequences, ria for the diagnosis of a meniscal tear in an unoper-
T2-weighted fast spin echo imaging with fat satura- ated meniscus no longer apply. When doubt exists, MR
tion, T2* -weighted gradient echo imaging and STIR arthrography may be of value (SCIULLI et al. 1999).
58 c. Heron and A. Hine
MR imaging of the anterior cruciate ligament fol- and will be pointing in the same direction. With the
lowing reconstruction is employed to determine the increasing lapse of time following excitation, this dif-
integrity of the ligament (IRIZARRY and RECHT 1997; ference in precession results in the two protons being
SCHATZ et al.1997). It is also of value to assess wheth- out of phase. They subsequently return to being in
er the ligament is optimally positioned and whether phase and, depending upon the TE employed, imag-
there is impingement on the ligament in addition to ing may be undertaken when they are either in or out
the appearances of the intercondylar region follow- ofphase.
ing notchplasty (MAY et al. 1997; RECHT et al. 1996a; In the case of a voxel which contains both fat and
TOMCZAK et al. 1997). Furthermore, it may reveal water (at a fat/water interface), out of phase protons
other complications such as a cyelops lesion (BRAD- will partially cancel each other and result in a low
LEY et al. 2000) or fibrosis. Images in the sagittal, signal intensity line. In-phase protons within a simi-
coronal and axial planes (MURAKAMI et al. 1998) lar voxel result in a high signal intensity line. These
employing Tl-weighted and fast spin echo proton differences may be exploited to enhance the visual-
density- and T2-weighted images with fat saturation isation of bone marrow disease using out of phase
can be used to address questions relating to the images. Chemical shift artefact can be minimised by
position and integrity of the ligament. Tl-weighted the application of fat suppression. Alternatively, selec-
images with fat saturation following intravenous gad- tive water excitation techniques may be employed but
olinium are used to assess soft tissue problems such are not available on all systems.
as cyelops lesions and fibrosis, and to obtain infor- The magie angle phenomenon arises in structures
mation about the degree of revascularisation of the which are composed of ordered parallel fibres. In the
reconstructed ligament. knee this ineludes ligaments, tendons and the menis-
ci. The posterior cruciate ligament is the most com-
monly affected structure in the knee. The parallel
orientation of fibres normally causes an increase in
4.5 the T2 dephasing. However, when the fibres are posi-
Artefacts tioned at 55° to the main magnetic field, the acceler-
ated loss of T2 magnetisation is decreased and struc-
Artefacts may be broadly divided into two categories. tures which are usually of low signal intensity dem-
There are those resulting from the inherent nature of onstrate an increase in signal (Fig.4.13) (ECHIGO et
the MR scanning technique and those which relate to al. 1999; OLEAGA and KRESSEL 1990; PETERFY et al.
patient factors such as movement. 1994). The effect is maximal on images with relatively
Chemieal shift artefact is one of the better known short TEs.
artefacts resulting from the physical characteristics of The truneation artefaet can arise at sites where
MR imaging and it is more conspicuous at high field there is a significant contrast difference between two
strengths (HOOD et al. 1999). In spin echo sequences, adjacent structures. In the knee this applies at the
chemical shift artefact arises because hydrogen pro- boundary between hyaline cartilage and the menisci
tons within fat precess at a different frequency from (tuRNER et al. 1991). As a result of Fourier transfor-
hydrogen protons within water. The hydrogen pro- mation of acquired data, the boundary is not accu-
tons in fat are surrounded by electrons which cause rately represented and aseries of alternating high
the proton to precess at a lower frequency than water- and low signal intensity lines are seen parallel to the
based hydrogen protons. The precessional frequency boundary. These artefacts are most marked when
determines the position that the proton is assigned a low phase encoding matrix is used in the supe-
along the frequency-encoding gradient. The assigned ro-inferior direction. Increasing the phase encoding
position along the frequency-encoding gradient matrix and applying it in the anteroposterior direc-
therefore cannot be an accurate spatial representa- tion should resolve uncertainty.
tion at the site of fat/water interfaces. Chemical shift A further artefact which can simulate a meniscal
artefact results in black and white lines appearing at tear or loose bodywithin a joint is the vaeuum phenom-
the interface. enon. When there is a small pocket of air between the
Gradient echo images produce an alternative type articular surfaces this is seen as a region of low signal
of chemical shift artefact. It also arises as a result of intensity. On gradient echo images the low signal focus
the differences in precessional frequency for protons increases in size due to blooming but it appears smaller
embedded in fat compared with those in water. Imme- on spin echo images. Further scanning following repo-
diately after excitation the protons precess together sitioning which redistributes the gas should remove
Magnetie Resonance Imaging 59
.AIo_ _ _ _... b
the artefact. Inadvertent injection of air bubbles at MR Metal objects in the knee produce their own local
arthrography will generate similar artefacts. magnetic fields and may therefore markedly distort the
In the past "wrap around" artefact could be prob- MR image. Even small metal fragments resulting from
lematic in knee MR imaging. This artefact arises bone drilling can cause significant artefact which is
when signal from tissues outside of the imaging more pronounced on gradient echo images (Fig. 4.14).
region are assigned a point within it. Wrap around Artefacts from patient motion (Fig.4.15) may be
arte fact has now been virtually eliminated by the minimised by positioning the patient comfortably and
use of local surface coils and also by the use of applying padding within rigid coils in order that the
a technique known as "over sampling". The latter potential for movement is reduced. Artefact from vas-
involves the acquisition of data from a large area cular structures, for example in the popliteal fossa,
but presentation of data from only the region of may be troublesome but altering the frequency and
interest with resultant elimination of wrap around phase encoding directions can ensure that regions of
artefact. particular interest are not obscured (Fig. 4.16).
in artieular cartilage; in vivo assessment with delayed appearance with fat -suppressed three-dimensional spoiled
Gd(DTPA)-enhanced imaging. Radiology 205:551-558 gradient-recalled sequences. Radiology 201:260-264
Bradley DM, Bergman AG, Dillingham MF (2000) MR imaging Escobedo EM, Hunter JC, Zink-Brody GC et al (1996) Useful-
of cyelops lesions.AJR 174:719-726 ness of turbo spin-echo MR imaging in the evaluation of
Bredella MA, Tirman PF, Peterfy CG et al (1999) Accuracy of meniscal tears: comparison with a conventional spin-echo
T2-weighted fast spin-echo MR imaging with fat saturation sequence. AJR 167:1223-1227
in detecting cartilage defects in the knee: comparison with Eustace S, Hentzen P,Adams J et al (1999) Comparison of con-
arthroscopy in 130 patients. AJR 172: 1073-1 080 ventional and turbo spin-echo Tl-weighted MR imaging in
Brossmann J, Preidler K-W, Daenen B et al (1996) lmaging acute knee trauma. AJR 172: 1393-1395
of osseous and cartilaginous intraartieular bodies in the Frank LR, Brossmann J, Buxton RB et al (1997) MR imaging
knee: comparison of MR imaging and MR arthrography truncation artifacts can create a false laminar appearance
with CT and CT arthrography in cadavers. Radiology in cartilage. AJR 168:547-554
200:509-517 Frank LR, Wong EC, Luh W-M et al (1999) Artieular cartilage
Brossmann J, Frank LR, Pauly JM et al (1997) Short echo time in the knee: mapping of the physiologie parameters at MR
projection reconstruction MR imaging of cartilage: com- imaging with a local gradient coil - preliminary results.
parison with fat -suppressed spoiled GRASS and magnetiza- Radiology 210:241-246
tion transfer contrast MR imaging. Radiology 203:501-507 Gold GE, Thedens DR, Pauly JM et al (1998) MR imaging of
Burstein D, Gray ML, Hartman AL et al (1993) Diffusion of artieular cartilage of the knee: new methods using ultra-
small solutes in cartilage as measured by nuelear magnetie short TEs. AJR 170: 1223-1226
resonance (NMR) spectroscopy and imaging. J Orthop Res Grainger AJ, Elliott JM, Campbell RSD et al (2000) Direct
11:465-478 MR arthrography: a review of current use. Clin Radiol
Cheung LP, Li KC, Hollett MD et al (1997) Meniscal tears of the 55:163-176
knee: accuracy of detection with fast spin-echo MR imag- Ha TPT,Li KCP, Beaulieu CF et al (1998) Anterior cruciate liga-
ing and arthroscopie correlation in 293 patients. Radiology ment injury: fast spin-echo MR imaging with arthroscopie
203:508-512 correlation in 217 examinations.AJR 170:1215-1219
Crues JV III, Mink J, Levy TL et al (1987) Meniscal tears of the Hardy PA, Recht MP, Piraino D et al (1996) Optimisation of a dual
knee: accuracy of MR imaging. Radiology 164:445-448 echo in the steady state (DESS) free-precession sequence for
Delfaut EM, Beltran J, Johnson G et al (1999) Fat suppression imaging cartilage. J Magn Reson Imaging 6:329-335
in MR imaging: techniques and pitfalls. Radiographies Hauger 0, Frank LR, Boutin RD et al (2000) Characterization
19:373-382 of the "Red Zone" of knee meniscus: MR imaging and his-
Disler DG, McCauley TR, Kelman CG et al (1996) Fat-sup- tologie correlation. Radiology 217: 193-200
pressed three-dimensional spoiled gradient -echo MR imag- Hood MN, Ho VB, Smirniotopoulos JG et al (1999) Chemieal
ing ofhyaline cartilage defects in the knee: comparison with shift: the artifact and elinieal tool revisited. RadioGraphs
standard MR imaging and arthroscopy AJR 167:127-132 19:357-371
Disler DG, McCauley TR, Ratner LM et al (1997) In-phase Hyhlik-Durr A, Faber S, Burgkart R et al (2000) Precision of
and out-of-phase MR imaging ofbone marrow: predietion tibial cartilage morphometry with a coronal water-excita-
of neoplasia based on the detection of co existent fat and tion MR sequence. Eur Radioll0:297-303
water. AJR 169: 1439-1447 lrizarry JM, Recht MP (1997) MR imaging of the knee liga-
Disler DG, Recht MP, McCauley TR (2000) MR imaging of ments and the postoperative knee. Radiol Clin North Am
artieular cartilage. Skeletal Radio129:367-377 35:45-76
Dupuy DE, Spillane RM, Rosol MS et al (1996) Quantification Jaramillo D, Laor T, Mulkern R (1994) Comparison between fast
of artieular cartilage in the knee with three-dimensional spin-echo and conventional spin-echo imaging of normal
MR imaging. Acad Radio13:919-924 and abnormal musculoskeletal structures in children and
Echigo J, Yoshioka H, Takahashi H et al (1999) Signal intensity young adults. luvest Radio129:803-811
changes in anterior cruciate ligament autografts: relation Kawahara Y, Uetani M, Nakahara N et al (1998) Fast spin-echo
to magnetic field orientation. Acad Radio16:206-210 MR of the articular cartilage in the osteoarthrotic knee.
Eckstein F, Gavazzeni A, Sittek H et al (1996) Determination Correlation of MR and arthroscopie findings. Acta Radiol
of knee joint cartilage thickness using three-dimensional 39:120-125
magnetie resonance chondro-crassometry (3D MR-CCM). Kneeland JB (2000) Magnetie resonance imaging of articular
Magn Res Med 36:256-265 cartilage. Semin Roentgeno135:249-255
Eckstein F, Westhoff J, Sittek H et al (1998) In vivo reproduc- Koskinen SK, Yla-Outinen H, Aho HJ et al (1997) Magnetiza-
ibility of three-dimensional cartilage volume and thiekness tion transfer and spin lock MR imaging of patellar cartilage
measurements with MR imaging. AJR 170:593-597 degeneration at 0.1 T.Acta Radio138:1071-1075
Eckstein F, Stammberger T, Priebsch J et al (2000) Effect of Kowalchuk RM, Kneeland JB, Dalinka MK et al (2000) MRI
gradient and section orientation on quantitative analysis of of the knee: value of short echo time fast spin-echo using
knee joint cartilage. J Magn Reson lmaging 11:161-167 high performance gradients versus conventional spin-echo
Ekman EF, Pope T, Martin DF et al (1994) Magnetie resonance imaging for the detection of meniscal tears. Skeletal Radiol
imaging of iliotibial band syndrome. Am J Sports Med 29:520-524
22:851-854 Kshirsagar AA, Watson PJ, Tyler JA et al (1998) Measurement
Eriekson SJ, Prost RW (1997) Laminar structures on MR of localized cartilage volume and thiekness of human knee
images of artieular cartilage: reply. Radiology 204:16-18 joints by computer analysis of three-dimensional magnetie
Eriekson SJ, Waldschmidt JG, Czervionke LF et al (1996) Hya- resonance images. lnvest Radio133:289-299
line cartilage: truncation artifact as a cause of trilaminar Lim PS, Schweitzer ME, Bhatia M et al (1999) Repeat tear of
62 C. Heron and A. Hine
postoperative meniscus: potential MR imaging signs. Radi- lateral meniscus on short -TE MR images of the knee. AJR
ology 210:183-188 163:149-154
Listerud J, Einstein S, Outwater E et al (1992) First principles Peterson DM, Carruthers CE, Wolverton BL et al (1999) Appli-
of fast spin echo. Magn Reson Q 8:199-244 cation of a birdcage coil at 3 Tesla to imaging of the human
Loredo R, Hodler J, Pedowitz R et al (1999) Posteromedial knee using MRI. Magn Reson Med 42:215-221
corner of the knee: MR imaging with gross anatomie cor- Pfirrmann CW, Zanetti M, Romero J et al (2000) Femoral troch-
relation. Skeletal RadioI28:305-311 lear dysplasia: MR findings. Radiology 216:858-864
Masciocchi C, Barile A, Satragno L (2000) Musculoskeletal Piplani MA, Disler DG, McCauley TR et al (1996) Artieular
MRI: dedieated systems. Eur RadioI1O:250-255 cartilage volume in the knee: semiautomated determina-
May DA, Snearly WN, Bents R et al (1997) MR imaging findings tion from three-dimensional reformations of MR images.
in anterior cruciate ligament reconstruction: evaluation of Radiology 198:855-859
notchplasty.AJR 169:217-222 Powers CM, Shellock FG, Pfaff M (1998) Quantification of
McCauley TR, Disler DG (1998) MR imaging of artieular car- patellar tracking using kinematie MRI. J Magn Reson Imag-
tilage. Radiology 209:629-640 ing 8:724-732
Mosher TJ, Dardzinski BJ, Smith MB (2000) Human artieular Rand T, Imhof H, Czerny C et al (1999a) Discrimination
cartilage: influence of aging and early symptomatie degen- between fluid, synovium, and cartilage in patients with
eration on the spatial variation of T2 - preliminary findings rheumatoid arthritis: contrast enhanced spin echo versus
at 3 T. Radiology 214:259-266 non-contrast-enhanced fat-suppressed gradient echo MR
Muhle C, Brossmann J, Heller M (1996) Kinematie MRI of imaging. Clin Radiol 54: 107-11 0
the knee using a specially designed positioning deviee. J Rand T, Imhof H, Turetschek K et al (1999b) Comparison of
Comput Assist Tomogr 20:522-525 low field (0.2 T) and high field (1.5 T) MR imaging in the
Muhle C, Ahn JM, Yeh L et al (1999a) Iliotibial band frietion differentiation of torned from intact menisci. Eur J Radiol
syndrome: MR imaging findings in 16 patients and MR 30:22-27
arthrographie study of six cadaverie knees. Radiology Recht MP, Piraino DW, Applegate G et al (1996a) Complica-
212:103-110 tions after anterior cruciate ligament reconstruction: radio-
Muhle C, Brossmann J, Heller M (1999b) Kinematie CT and MR graphie and MR findings. AJR 167:705-710
imaging of the patellofemoral joint. Eur RadioI9:508-518 Recht MP, Piraino DW, Paletta GA et al (1996b) Accuracy
Munk PL, Hilborn MD, VelletAD et al (1997) Diagnostie equiv- of fat-suppressed three-dimensional spoiled gradient-echo
alence of conventional and fast spin echo magnetie res- FLASH MR imaging in the detection of patellofemoral
onance imaging of the anterior cruciate ligament of the artieular cartilage abnormalities. Radiology 198:209-212
knee. Aust RadioI4l:238-242 Reddy R, Insko E, Noyszewski EA et al (1998) Sodium MRI
Murakami Y, Sumen Y, Ochi M et al (1998) MR evaluation of human artieular cartilage in vivo. Magn Reson Med
of human anterior cruciate ligament autograft on oblique 39:697-701
axial imaging. J Comput Assist Tomogr 22:270-275 Reeder JD, Matz SO, Becker L et al (1989) MR imaging of
Nakanishi K, Horibe S, Shiozaki Y et al (1997) MRI of normal the knee in the sagittal projection: comparison of three-
anterior cruciate ligament (ACL) and reconstructed ACL: dimensional gradient-echo and spin-echo sequences. AJR
comparison of when the knee is extended with when the 153:537-540
knee is flexed. Eur Radiol 7:1020-1024 Robbins MI, Anzilotti KF Jr, Katz LD et al (2000) Patient
Nishimura G, Yamato M, Tamai K et al (1997) MR findings in perception of magnetie resonance arthrography. Skeletal
iliotibial band syndrome. Skeletal RadioI26:533-537 RadioI29:265-269
Niitsu M, Ikeda K, Fukubayashi T et al (1996) Knee extension Roychowdhury S, Fitzgerald SW, Sonin AH et al (1997) Using
and flexion: MR delineation of normal and torn anterior MR imaging to diagnose partial tears of the anterior cruci-
cruciate ligaments. J Comput Assist Tomogr 20:322-327 ate ligament: value ofaxial images.AJR 168:1487-1491
Niitsu M, Ikeda K, Itai Y (1998) Slightly flexed knee position Rubenstein JD, Kim JK, Morava-Protzner let al (1993) Effects
within a standard knee coil: MR delineation of the anterior of collagen orientation on MR imaging characteristies of
cruciate ligament. Eur RadioI8:113-115 bovine artieular cartilage. Radiology 188:219-226
Niitsu M, Nakai T, Ikeda K et al (2000) High-resolution MR Rubenstein J, Recht M, Disler DG et al (1997) Laminar struc-
imaging of the knee at 3 T. Acta Radiol4l :84-88 tures on MR images of articular cartilage. Radiology
Oleaga L, Kressel H (1990) High signal intensity in the pos- 204:15-16
terior cruciate ligament on sagittal gradient -echo images: Rubin DA, Kneeland JB, Listerud J et al (1994) MR diagnosis of
normal variant. Radiology 177:295 meniscal tears of the knee: value offast spin-echo vs con-
Peh WC, Cassar-Pullicino VN (1999) Magnetie resonance ventional spin-echo pulse sequences. AJR 162:1131-1135
arthrography: current status. Clin RadioI54:575-587 Ruehm S, Zanetti M, Romero J et al (1998) MRI of patellar artie-
Peh WC, Chan JH, Shek TW et al (1999) The effect of using ular cartilage: evaluation of an optimized gradient echo
shorter echo times in MR imaging of knee menisci: a study sequence (3D-DESS). J Magn Reson Imaging 8:1246-1251
using a porcine model. AJR 172:485-488 Schatz JA, Potter HG, Rodeo SA et al (1997) MR imaging of ante-
Pereira ER, Ryu KN, Ahn JM et al (1998) Evaluation of rior cruciate ligament reconstruction. AJR 169:223-228
the anterior cruciate ligament of the knee: comparison Sciulli RL, Boutin RD, Brown RR et al (1999) Evaluation of
between partial flexion true sagittal and extension sag- the postoperative meniscus of the knee: a study compar-
ittal oblique positions during MR imaging. Clin Radiol ing conventional arthrography, conventional MR imaging,
53:574-578 MR arthrography with iodinated contrast material, and
Peterfy CG, Janzen DL, Tirman PFJ et al (1994) 'Magic angle' MR arthrography with gadolinium-based contrast mate-
phenomenon. A cause of increased signal in the normal rial. Skeletal RadioI28:508-514
Magnetic Resonance Imaging 63
Sittek H, Eckstein F, Gavazzeni A et al (1996) Assessment of Wacker FK, Bolze X, Felsenberg D et al (1998) Orientation-
normal patellar cartilage volume and thiekness using MRI: dependent changes in MR signal intensity of artieular car-
analysis of currently available pulse sequences. Skeletal tilage: manifestation of the "magie angle" effect. Skeletal
RadioI25:55-62 RadioI27:306-310
Stabler A, Glaser C, Reiser M (2000) Musculoskeletal MR: knee. Waldschmidt JG, Rilling RJ, Kajdacsy-Balla AA (1997) In vitro
Eur Radioll0:230-241 and in vivo MR imaging of hyaline cartilage: zonal anat-
Stammberger T, Eckstein F, Englmeier KH et al (1999) Deter- omy, imaging pitfalls, and pathologie conditions. Radio-
mination of 3D cartilage thiekness data from MR imaging: graphies 17:1387-1402
computational method and reproducibility in the living. Wang SF, Cheng HC, Chang CY (1999) Fat-suppressed three-
Magn Reson Med 41:529-536 dimensional fast spoiled gradient-recalled echo imaging:
Suh JS, Cho JH, Shin KH et al (1996) Chondromalacia of the a modified FS 3D SPGR technique for assessment of patel-
knee: evaluation with a fat-suppression three-dimensional lofemoral joint chondromalacia. Clin Imaging 23: 177 -180
SPGR imaging after intravenous contrast injection. J Magn White LM, Schweitzer ME, Johnson WJ et al (1996) Role of T2-
Reson Imaging 6:884-888 weighted fast-spin-echo imaging in the diagnosis of menis-
Tomczak RJ, Hehl G, Mergo PJ et al (1997) Tunnel placement cal tears. J Magn Reson Imaging 6:874-877
in anterior cruciate ligament reconstruction: MRI analysis Wieslander SB, Rappeport ED, Lausten GS et al (1998) Multi-
as an important factor in the radiologieal report. Skeletal planar reconstruction in MR imaging of the knee. Compar-
RadioI26:409-413 ison with standard sagittal and coronal images. Acta Radiol
Turner DA, Rapoport MI, Erwin WD, McGould M et al (1991) 39:116-119
Truncation artefact: a potential pitfall in MR imaging of the Xia Y,Farquhar T,Burton-Wurster N et al (1997) Origin of car-
menisci of the knee. Radiology 179:629-633 tilage laminae in MRI. J Magn Reson Imaging 7:887-894
Tyrrell RL, Gluckert K, Pathria M et al (1988) Fast three-dimen- Yao L, Gentili A, Thomas A (1996) Incidental magnetization
sional MR imaging of the knee: comparison with arthros- transfer contrast in fast spin-echo imaging of cartilage. J
copy. Radiology 166:865-872 Magn Reson Imaging 6:180-184
Uhl M, Ihling C, Allmann KH et al (1998) Human artieular Yu JS, Salonen DC, Hodler J et al (1996) Posterolateral aspect
cartilage: in vitro correlation of MRI and histologie find- of the knee: improved MR imaging with a coronal oblique
ings. Eur RadioI8:1123-1129 technique. Radiology 198:199-204
5 Ultrasound
STEFANO BIANCHI, CARLO MARTINOLI and JEAN GARCIA
CONTENTS 5.2
5.1 Introduction 65
US Scanning Technique
5.2 US Scanning Technique 65
5.3 Anterior Aspect 66 A detailed reference note from the clinician with an
5.3.1 Suprapatellar Region 66 indication as to the specific structures to be inves-
5.3.2 Patellar Region 68 tigated (patellar tendon, popliteal space, peroneal
5.3.3 Infrapatellar Region 69
5.4 Medial Aspect 70
nerve etc.) and a presumptive clinical diagnosis must
5.5 Lateral Aspect 71 first be obtained for all patients. Focusing on a lim-
5.6 Posterior Aspect 71 ited area of the knee joint reduces the duration of
5.6.1 Internal Region 72 the examination and allows an accurate assessment
5.6.2 Middle Region 72 of the structures examined. For example, in a patient
5.6.3 Lateral Region 73
References 74
with a history of localised pain over the cranial por-
tion of the patellar tendon occurring during sport-
ing activity, US is needed to confirm the diagnosis
of patellar tendonitis and to accurately evaluate the
5.1 size of the lesion. In this case, a thorough examina-
Introduction tion of the popliteal fossa is useless and not required.
In comparison to computed tomography (CT) and
The refinement of broadband linear array transduc- magnetic resonance (MR) imaging, US is inferior
ers has increased the capability of ultrasound (US) in the evaluation of most intra-articular structures.
to evaluate the normal structures of the musculo- Since the clinician may ignore these limitations,
skeletal system as well as to detect and characterise knowledge of the presumptive clinical diagnosis is
subtle pathological changes (MARTINOLl et al. 1999). important to avoid vain examinations. On the other
Although many knee diseases that affect the menisci hand, US is more efficient and cost-effective in the
and cruciate ligaments are not accurately imaged by assessment of superficial structures and can be con-
US, a variety of conditions, including tendon diseas- sidered the modality of choice in patients with para-
es, vascular and nerve lesions, joint disorders and articular lesions.
para-articular cysts, can be accurately assessed by The patient history and arecent radiographie eval-
this diagnostic modality (GROBBELAAR and BOUF- uation are routinely obtained before the US examina-
FARD 2000). The purpose of this chapter is to describe tion. The availability of standard radiographs can be
the scanning technique of US examination and to essential for the correct interpretation of difficult US
illustrate the normal US anatomy of the knee. images of dis orders that are obvious on radiographs.
To give an example, the US appearance of a bony
exostosis can be misleading and its diagnosis time
S. BIANCHI, MD
consuming, whereas it is readily evident on standard
Division of Radiodiagnosis and Interventional Radiology,
Hopital Cantonal Universitaire, 24 rue Micheli -du -Crest, 1211 radiographs. On the other hand, given the diagnosis
Geneva 14, Switzerland radiographically, US can assess the thickness and
C. MARTINOLI, MD regularity of the cartilaginous cap of the exostosis
Istituto di Radiologia, Universita di Genova, Largo Rosanna and can be useful in the follow-up of the lesion (MAL-
Benzi 1, 16100 Genoa, Italy GHEM et al. 1992).
J. GARCIA, MD
Professor, Division of Radiodiagnosis and Interventional Radi-
Routine examination of the knee can be accom-
ology, Hopital Cantonal Universitaire, 24 rue Micheli-du-Crest, plished with broadband linear array transducers
1211 Geneva 14, Switzerland working at a frequency band range of 5-13 MHz.
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
66 S. Bianchi et al.
In the evaluation of the popliteal fossa, the lower which can mimic pathological changes (FORNAGE
band frequencies can help in imaging the deeper 1987). Knee flexion of nearly 23°, obtained by plac-
structures, whereas higher frequencies (10-13 MHz) ing a small pillow beneath the popliteal space, leads
improve the assessment of the most superficial struc- to adequate stretching of the tendons of the extensor
tures, such as the patellar tendon and the peroneal mechanism (quadriceps and patellar tendons) and
nerve at the level of the fibular head. Stand-off pads avoids the possible artefacts related to tendon con-
are usually not required. Colour and power Doppler cavity (BIANCHI et al. 1994). Alternatively the patient
US are useful in the evaluation of the popliteal ves- can be asked to contract the quadriceps muscle,
sels, in the assessment of the neovasculature of para- which leads to a rectilinear appearance of both ten-
articular masses and in depicting synovial or tendon dons.
hyperaemia. The use of an extended-field-of-view The anterior aspect can be divided into three
technique yields a panoramic view of the region regions: the suprapatellar, the patellar and the infrapa-
scanned and, more importantly, makes the interpre- tellar.
tation of the US images easier for the referring physi-
cian.
The patient is examined in the recumbent position. 5.3.1
In selected cases, the decubitus position may be help- Suprapatellar Region
ful. In the assessment of intra-articular loose bodies,
for example, sonograms obtained in different positions The suprapatellar region includes the quadriceps
(recumbent, standing) can help in showing the dis- tendon, suprapatellar synovial recess, suprapatellar
placement of the loose bodies, confirming their intra- fat pad and prefemoral fat, the distal femoral metaph-
articular location (BIANCHI and MARTINOLl 1999). ysis and the trochlea (Figs. 5.1, 5.2).
Different degrees of knee flexion and extension of the
joint can be obtained to assess normal or pathologi-
cal structures. Evaluation of the trochlear cartilage can
easily be performed using anterior transverse sono-
grams obtained with the knee forcefully flexed (MAR-
TINO et al. 1998). Assessment of the cartilage of the
posterior aspect of both condyles must be accom-
plished with the patient supine and the knee in full
extension. Examination of the contralateral knee is
usually not essential and is not routinely performed,
but can be helpful in order to compare the patholog-
ical side with the contralateral one in selected cases.
The opportunity to perform a dynamic examination
is a particular advantage of USo Different structures,
including ligaments and tendons, can be examined at
rest, during stress manoeuvres or during active muscle
contraction. Comparison of the US findings obtained
at rest and during movement is useful and usually pro-
vides supplementary information to the morphologi-
cal features.
5.3
Anterior Aspect
b
The anterior aspect of the knee is evaluated by exam-
ining the patient in the supine position. As a general
Fig. S.la, b. Anterior aspect. Suprapatellar region. a Probe posi-
rule, tendons must be evaluated while stretched and tioning for longitudinal examination. b Corresponding sono-
with the probe parallel to them. This can avoid arte- gram. F, Femur; P, patella; arrowheads, quadriceps tendon; SF,
factual hypoechogenicity due to tendon anisotropy, suprapatellar fat; SR, synovial recess; PF, prefemoral fat
Ultrasound 67
a b
Fig. 5.2a, b. Anterior aspect. Suprapatellar region. Trochlear cartilage. a Probe positioning for transverse examination. b Cor-
responding sonogram. F, Femur; asterisks, trochlear cartilage
The quadrieeps muscle is located in the anterior the presence of the so-called suprapatellar synovial
thigh and is formed by the vastus lateralis, the vastus plica, whieh, if an effusion is present, can be imaged
medialis (whieh are respectively located externally at USo On longitudinal sonograms the plica appears as
and internaIly), the superfidally located rectus femo- an oblique curvilinear structure showing a posterior-
ris and the vastus intermedius (deeply located in inferior concavity. The plica, which rarely has clinieal
the midline of the thigh). The vastus lateralis and relevance, must be differentiated from a postinflam-
medialis join in the midline to insert in a common matory fibrous septum. In normal conditions, on both
tendon. The rectus femoris inserts in a thin superfi- longitudinal and transverse sonograms the suprapa-
dal tendon while the vastus intermedius inserts in tellar recess appears as a hypoechoie structure owing
a deep thiek tendon. The quadriceps tendon is then to the overlap of the anterior and posterior synovial
formed by the overlaying of three different tendons: membrane. A variable amount of fluid can be appre-
the superfidal tendon of the rectus femoris, the inter- dated on US as an anechoie area located inside the
mediate common tendon of the vastus lateralis and bursa. US detection of intra-articular fluid appears
medialis, and the deep tendon from the vastus inter- to be more accurate and reproducible than clinieal
medius. This multilayered appearance is readily evi- detection (HAUZEUR et al. 1999). Care must be taken
dent at US and can be weIl imaged on longitudinal when evaluating the presence of a joint effusion by
sonograms (BIANCHI et al. 1994). While the superfi- USo Small effusions can be detected on suprapatellar
dal tendon is always weIl seen, differentiation of the longitudinal sonograms only at dynamie examination
intermediate and the deep tendon can be more dif- obtained during isometrie contraction of the quadri-
ficult. The possibility of evaluating the internal struc- ceps. This can be explained by the cranial displace-
ture of the quadriceps tendon to discriminate the ment of the fluid induced by displacement of Hoffa's
different components has practical value since this fat pad against the femoral condyles and by tighten-
allows differentiation between complete tears, which ing of the posterior fasda. Additionally, when fluid
are treated by surgery, and partial tears involving is present in very small amounts, it can be imaged
only one or two tendons, whieh may be amenable only on transverse sonograms in the lateral or medial
to conservative treatment. Transverse sonograms can pouch of the suprapatellar recess. It must be stressed,
help in assessing the extension of the tear in the coro- however, that a modest amount of intra-artieular fluid
nal plane. is normal in asymptomatie subjects and that corre-
The suprapatellar synovial recess or suprapatellar lation with clinieal data and with the contralateral
bursa lies posterior to the quadriceps tendon and knee is mandatory. When analysis of synovial fluid is
suprapatellar fat pad and anterior to the prefemoral essential, for example to rule out an infection or gout,
fat. In uterine life the synovial recess is completely US-guided aspiration can be easily performed (WANG
separated from the artieular cavity by a septum. A SC et al. 1999). Utilisation ofUS for this purpose leads
perforation of the septum normally occurs at the end to careful selection of the site of puncture, particu-
of the 5th foetal month and allows free communiea- larly in small effusions, and confirms the intra-artic-
tion between the two spaces. In a small percentage of ular positioning of the needle tip. Usually US-guided
cases, incomplete resorption of the septum leads to arthrocentesis is less painful than blind puncture.
68 S. Bianchi et al.
a b
Fig. S.3a, b. Anterior aspect. Patellar region. Medial patellar retinaculum. a Probe positioning for transverse examination.
b Corresponding sonogram. F, Femur; P, patella; arrowheads, medial patellar retinaculum
Ultrasound 69
the mid third of the tendon for anterior cruciate lig- tissue. A synovial bursa can be located between the
ament reconstruction. This aspect must not be con- two components but cannot be demonstrated by US
fused with a longitudinal tendon split (ADRIANI et in the normal state. Collateralligament injuries can
al. 1995). Internal hypoechoic areas can be found be diagnosed with US (LEE et al. 1996; MATHIEU et
in asymptomatic elite athletes (COOK et al. 1998). al.1997). Dynamic US images obtained during valgus
This has implications for clinicians managing ath- stress improve the assessment of ligament integrity
letes with anterior knee pain. Calcifications and ossi- (DE FLAVIIS et al. 1988). Images are obtained both at
fication of the distal portion of the tendon can be rest and during stress manoeuvres and documented.
found in asymptomatic subjects and are usuallyrelat- Then the distance between the tibia and femur can be
ed to sequelae of Osgood-Schlatter disease. Deep to easily measured and compared with the contralateral
the patellar tendon, between it and the anterior por- knee. Although the possibility of evaluating menis-
tion of the condyles, the intracapsular Hoffa's pad cal tears with US has been reported (GERNGROSS and
appears as an adipose structure containing internal SOHN 1992), there is now general agreement that the
fibrous septations. sensitivity and specificity of US for this purpose are
inferior to those of MR imaging. On the other hand,
US can readily detect and assess meniscal cysts (PEET-
RONS et al.1990; RUTTEN et al.1998), which are always
5.4 associated with meniscal tears. When meniscal ossicle,
Medial Aspect which represents a vestigial structure, is suspected on
standard radiographs, US can confirm the diagnosis.
The medial aspect of the joint is examined by asking The ossicle appears as a regular hyperechoic structure
the patient to externally rotate the leg. The medial embedded within the posterior horn of the medial
collateralligament, internal femorotibial joint space, meniscus (MARTINOLI et al. 2000a). The pes anseri-
medial meniscus and pes anserinus tendons are eval- nus tendons (sartorius, gracilis and semitendinosus
uated (Fig. 5.5). The medial collateralligament com- tendons) insert in the anteromedial portion of the
prises two portions: the superficial portion, which tibial metaphysis. The different tendons can be distin-
originates from the superior aspect of the medial guished by US only at a higher level. As the tendons
condyle and inserts on the internal tibial metaphy- approach the distal insertion inferiorly, they blend
sis, and the deep component, which links the internal together and cannot be differentiated. Different syno-
meniscus to the femur (meniscofemoral ligament) vial bursae located among the tendons and between
and to the tibia (meniscotibialligament). On US, both them and the tibia lower the friction among these
superficial and deep ligament components appear structures. In normal conditions these bursae cannot
as regular, hyperechoic laminae separated by a be demonstrated by US, whereas they are readily evi-
hypoechoic area related to fat and loose connective dent when distended by an effusion in bursitis.
a _ _ _ _
b
Fig. 5.5a, b. Medial aspect. Medial collateral ligament. a Probe positioning for longitudinal examination. b Corresponding
sonogram. F, Femur; T, tibia; M, medial meniscus; medial collateral ligaments: arrows, superficial portion; arrowheads, deep
portion
Ultrasound 71
5.5 5.6
Lateral Aspect Posterior Aspect
The lateral aspect of the joint is examined by asking The the patient is then asked to lie prone with the
the patient to internally rotate the leg. The lateral col- knee extended, to allow examination of the posteri-
lateral ligament, external femorotibial joint space, lat- or region. The US examination is performed in the
eral meniscus and distal aspect of the iliotibial band transverse and sagittal planes, starting with evalua-
are evaluated (Fig. 5.6). The lateral collateralligament tion of the intern al region, followed by assessment of
appears as a cord-like hyperechoic structure which is the middle and lateral regions.
located at the posterolateral aspect of the joint and
inserts into the peroneal head and lateral femoral
condyle. US of the lateral meniscus shares the same 5.6.1
limitations as US of the medial meniscus. Examin- Internal Region
ing the knee in forceful flexion makes meniscal cysts
more apparent. The distal tract of the iliotibial band The structures that can be assessed by US are the
is found at the anterior aspect of the lateral face. It proximal portions of the tendons of pes anserinus,
appears as a thin fibrillar structure that inserts in the semimembranosus tendon and the gastrocne-
a tibial tubercle located at the anterolateral aspect mius-semimembranosus synovial bursa (Fig. 5.7).
of the tibial epiphysis (BONALDI et al. 1998). The The sartorius is mainly composed of muscle fibres
more distal portion of the band normally widens and is located internally. The gracilis tendon is locat-
just before its insertion into the tibia. This normal ed posterior to the sartorius and, as its name implies,
appearance must not be misdiagnosed as localised it is the thinnest of the internal tendons. The cranial
tendinopathy. The standard US examination must portion of the semitendinosus tendon is located pos-
also include transverse sonograms of the proximal terior to the semimembranosus muscle, then, more
tibiofemoral joint and of the antero-external com- caudally, it lies behind the semimembranosus tendon.
partment of the leg since this is the most frequent The large semimembranosus tendon lies in a more
location of intramuscular ganglia. These ganglia orig- lateral position and inserts by its direct tendon on
inate from the proximal tibiofemoral joint, extend the posteromedial aspect of the tibial epiphysis. The
inside the anterior tibialis or peroneus muscle and indirect tendon cannot be detected at USo Between
can cause compression of the peroneal nerve (BIAN- the tendon of the semimembranosus and the medial
CHI et al. 1995a). head of the gastrocnemius is located the gastroc-
nemius-semimembranosus synovial bursa, which is
not normally depicted by USo In young subjects the
bursae do not communicate with the knee joint, while
Fig. 5.6a, b. Lateral aspect. Iliotibial band. a Probe positioning for longitudinal examination. b Corresponding sonogram.
T, tibia; arrowheads, iliotibial band
72 s. Bianchi et al.
a ....._ __
b
a L - - - -.......- -
and the hyaline eartilage ean be assessed. The lateral Grassi W, Lamanna G, Farina A et al (1999) Sonographie imag-
head of the gastroenemius muscle is smaller than the ing of normal and osteoarthritie cartilage. Semin Arthritis
Rheum 28:398-340
medial head. Its tendon may eontain the fabella, a ses-
Grobbelaar N, Bouffard JA (2000) Sonography of the knee, a
amoid bone that appears as a eurvilinear hypereeho- pietorial review. Semin illtrasound CT MR 21:231-274
ie strueture showing adefinite posterior shadowing. Hauzeur JP, Mathy L, De Maertelaer V (1999) Comparison
Care must be taken not to eonfuse the fabella with an between clinieal evaluation and ultrasonography in detect-
intra-artieular loose body. The intratendinous loea- ing hydrarthrosis of the knee. J Rheumato126:2681-2683
Jozwiak M, Pietrzak S (1998) Evaluation of patella position
tion and the fixed position during dynamie examina-
based on radiologie and ultrasonographie examination:
tion exclude an intra-articular fragment. Differentia- comparison of the diagnostie value J Pediatr Orthop
tion from a eapsular ealcifieation is best aehieved by 18:679-682
lateral radiographs. Learch TJ,Braaton M (2000) Lipoma arborescens: high-resolution
ultrasonographie findings. J illtrasound Med 19:385-389
Lee JI, Song IS, Jung YB et al (1996) Medial collateralligament
injuries of the knee: ultrasonographie findings. J Ultra-
sound Med 15:621-625
References Malghem J, Vande Berg B, Noel H et al (1992) Benign osteo-
chondromas and exostotie chondrosarcomas: evaluation
Adriani E, Mariani PP, Maresca G et al (1995) Healing of the of cartilage cap thiekness by ultrasound. Skeletal Radiol
patellar tendon after harvesting of its mid-third for ante- 21:33-37
rior cruciate ligament reconstruction and evolution of the Martino F, De Serio A, Macarini L et al (1998) illtrasonogra-
unclosed donor site defect. Knee Surg Sports Traumatol phy versus computed tomography in evaluation of the fem-
Arthrose 3:138-143 oral trochlear groove morphology: a pilot study on healthy,
Bianchi S, Martinoli C (1999) Detection of loose bodies in young volunteers. Eur Radio18:244-247
joints. Radiol Clin North Am 37:679-690 Martinoli C, Bianchi S, Derchi LE (1999) illtrasound of tendon
Bianchi S, Zwass A, Abdelwahab IF et al (1994) Diagnosis of and nerves. Radiol Clin North Am 37:691-711
tears of the quadrieeps tendon of the knee: value of sonog- Martinoli C, Bianchi S, Spadola L et al (2000a) Multimodality
raphy.AJR 162:1137-1140 imaging assessment of meniscal ossicle. Skeletal Radiol
Bianchi S, Zwass A, Abdelwahab IF et al (1995a) Sonographie 29:481-484
evaluation of intramuscular ganglia. Clin Radiol Martinoli C, Bianchi S, Gandolfo N et al (2000b) US of nerve
50:235-236 entrapments in osteofibrous tunnels of the upper and lower
Bianchi S, Zwass A, Abdelwahab IF et al (1995b) Sonographie limbs. Radiographies 20 [Spec]:SI99-S217
evaluation of lipohemarthrosis: clinieal and in vitro study. Mathieu P, Wybier M, Busson J et al (1997) The medial col-
J illtrasound Med 14:279-282 lateral ligament of the knee. Ann Radio140: 176-181
Bonaldi VM, Chhem RK, Drolet R et al (1998) Iliotibial band Peetrons P,Allaer D, Jeanmart L (1990) Cysts of the semilunar
frietion syndrome: sonographie findings. J illtrasound Med cartilages of the knee: a new approach by ultrasound imag-
17:257-260 ing. A study of six cases and review of the literature. J Ultra-
Coari G, lagnocco A, Zoppini A (1995) Chondrocalcinosis: sound Med 9:333-337
sonographie study of the knee. Clin RheumatolI4:511-514 Ptasznik R, Feller J, Bartlett J et al (1995) The value of sonog-
Cook JL, Khan KM, Harcourt PR et al (1998) Patellar tendon raphy in the diagnosis of traumatic rupture of the anterior
ultrasonography in asymptomatie active athletes reveals cruciate ligament of the knee. AJR 164: 1461-1463
hypoechoie regions: a study of 320 tendons. Vietorian Rutten MJ, Collins JM, van Kampen A et al (1998) Meniscal
Institute of Sport Tendon Study Group. Clin J Sport Med cysts: detection with high-resolution sonography. AJR
8:73-77 171 :491-496
De Flaviis L, Nessi R, Leonardi M et al (1988) Dynamie ultra- Silvestri E, Martinoli C, Derchi LE et al (1995) Echotexture
sonography of capsulo-ligamentous knee joint traumas. J of peripheral nerves: correlation between US and histo-
Clin Ultrasound 16:487-492 logie findings and criteria to differentiate tendons. Radiol-
Fornage BD (1987) The hypoechoie normal tendon. A pitfall. J ogy 197:291-296
illtrasound Med 6:19-22 Starok M, Lenchik L, Trudell D et al (1997) Normal patellar
Gebhard F, Authenrieth M, Strecker W et al (1999) illtrasound retinaculum: MR and sonographie imaging with cadaverie
evaluation of gravity induced anterior drawer following correlation.AJR 168:1493-1499
anterior cruciate ligament lesion. Knee Surg Sports Trau- Wang SC, Chhem RK, Cardinal E et al (1999) Joint sonography.
matol Arthrose 7:166-172 Radiol Clin North Am 37:653-668
Gerngross H, Sohn C (1992) illtrasound scanning for the diag- Wang TG, Wang CL, Hsu TC et al (1999) Sonographie evalua-
nosis of meniscal lesions of the knee joint. Arthroscopy tion of the posterior cruciate ligament in amputated speci-
8:105-110 mens and normal subjects. J illtrasound Med 18:647-653
Clinical Applications
6 Congenital and
Developmental Abnormalities of the Knee
KARL ]OHNSON and A. MARK DAVIES
CONTENTS 6.1
6.1 Introduction 77 Introduction
6.2 Alignment Abnormalities 77
6.2.1 Normal Development 77 Congenital abnormalities that affect the knee
6.2.2 Genu Valgum (Knock Knees) 78 include intrinsic joint disorders, bone and carti-
6.2.3 Genu Varum (Bow Legged) 79
6.3 Hyperextension Deformities of the Knee 80
lage diseases and soft tissue disorders. The abnor-
6.3.1 Normal Development 80 mality may be isolated to the knee, be unilateral or
6.3.2 Genu Recurvatum 80 bilateral and symmetrical or asymmetrical. Alter-
6.3.3 Congenital Subluxation/Dislocation 80 natively, the knee abnormality may be part of a
6.4 Congenital Ankylosis of the Knee 81 wider syndrome of disorders. The identification
6.5 Tibial/Fibular Bowing and Pseudo-arthrosis 81
6.6 Limb Deficiencies 82 of an abnormality within the knee should raise
6.6.1 Fibula Hemimelia 82 the suspicion of further musculoskeletal anoma-
6.6.2 Tibial Hemimelia 82 lies elsewhere.
6.6.3 Femoral Absence 83
6.7 Femoral Bifurcation/Duplication 83
6.8 Discoid Meniscus 83
6.9 Congenital Absence of the Anterior Cruciate Ligament 83
6.10 Arthrogryposis Multiplex 83
6.11 Transplacental Acquired Infections 84 6.2
6.11.1 Rubella 84 Alignment Abnormalities
6.11.2 Cytomegalovirus 84
6.11.3 HIV Infection 84
6.11.4 Congenital Syphilis 84
6.12 Neurofibromatosis 85 6.2.1
6.13 Haematological Disorders 85 Normal Development
6.13.1 Haemophilia 85
6.13.2 Thalassaemia 86 The terms ge nu valgum and ge nu varum describe
6.14 Diaphyseal Aclasis 86
6.l5 Osteogenesis Imperfecta 86
the relationship of the tibia with respect to the
6.16 Modelling Deformities 86 femur. Genu valgum indicates abduction of the tibia
6.16.1 Erlenmeyer Flask Deformity 87 at the knee joint and genu varum, adduction. The
6.16.2 Epiphyseal Irregularity 88 angle of varus is calculated at the intersection
6.16.3 Metaphyseal Widening and Splaying 89 of lines drawn along the mid shaft of the femur
6.17 Patella 92
6.17.1 Patellar Abnormalities 92 and tibia in neutral position. Following birth and
6.17.2 Absent Patella 92 through childhood the normal pattern of develop-
6.18 Congenital Dislocation of the Patella 92 ment is to progress from a position of varum to
6.19 Fabella 92 one of valgum and then return to near neutral by
6.20 Popliteal Pterygium Syndrome 92 adolescence (SHOPFNER and COIN 1969; FOREMAN
6.21 Congenital Bone Dysplasias 93
References 93 and ROBERTSON 1985). In the newborn the average
angle of varus is 17°; this decreases to an average of
K. JOHNSON, MD 9° by 1 year, and by 2 years there is an average 2° of
Consultant Paediatric Radiologist, Princess of Wales, Bir- valgus. At 3 years of age, when the child is walking,
mingham Children's Hospital, Steelhouse Lane, Birmingham, the average angle of valgus is 11 0, which reduces to
B46NH, UK
6° by the age of 13 (Fig. 6.1). In the adult there is
A.M. DAVIES, MD
Consultant Radiologist, MRI Centre, Royal Orthopaedic a mild degree of femorotibial varus (SALENIUS and
Hospital, Bristol Road, Birmingham, B31 2AP, UK VANKKA 1975).
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
78 K. Johnson and A. M. Davies
mentation of unsuspected physeal bars can be helpful than congenital dislocation of the hip (BENSAHEL
for surgical planning, and assessment of the physeal et al. 1989; NIEBAUER and KING 1960). The deformi-
growth plate with 3D reformatting further aids surgi- ty commonly involves both knees and usually there
cal intervention (CRAIG et al. 1999). are other associated congenital abnormalities. These
The differential diagnosis for genu varum includes abnormalities include hip dislocation, arthrogrypo-
unresolved and severe physiological bowing, rickets, sis multiplex, elbow dislocation and club-foot defor-
osteogenesis imperfecta and other congenital bone mity (CURTIS and FISHER 1969; BELL et al. 1987).
dysplasias. In the majority of bone dysplasias, the An isolated form of anterior tibial displacement with-
genu valgum is the result of abnormal ossification out hyperextension has been described (CURTIS and
and bone remodelling. FISHER 1970).
Lateral radiographs of the knee should be obtained
in both flexion and extension to demonstrate the
irreducibility of the anterior dislocation (Fig.6.4).
6.3 Anterior radiographs mayaiso demonstrate rotatory
Hyperextension Deformities of the Knee and lateral movement. Serial radiographs will show
a delay in the ossification of both the femoral and
6.3.1 the tibial epiphysis (CARLSON and O'CONNOR 1976).
Normal Development Other radiological findings include patella baja, patel-
la alta, genu valgum and hypoplasia of the lateral
The angle of extension of the knee joint is measured femoral condyle (BENSAHEL et al. 1989).
from the intersection of lines drawn along the shaft Arthrography and MR imaging demonstrate flat-
of the femur and the tibia. Hyperextension of the tening of the inferior portion of the femoral con-
knee joint of up to 20° is a normal variant at birth. dyles. Fibrous contractures of the quadriceps muscles
This hyperextension is completely stable, there is no have been demonstrated and the lack of a suprapatel-
limitation of flexion and the integrity of the joint lar pouch is believed to be a bad prognostic indica-
and surrounding structures is maintained. A wide tor.
spectrum of hyperextension deformities have been
described, ranging from persistence of the normal
physiological hyperextension of the infant into child-
hood to complete dislocation (LAuRENcE 1967). 6.4
Congenital Ankylosis of the Knee
6.6
Limb Deficiencies
6.6.1
Fibular Hemimelia
6.11.2
Cytomegalovirus
6.11.3
HIV Infection
6.12
Neurofibromatosis
6.13
Haematological Disorders
6.13.1
Haemophilia in a swollen joint. In the advanced stages there is carti-
lage loss and articular destruction which will be seen
The most common disorder is haemophilia, which is as loss of joint space and peri-articular erosions.
X-linked and occurs in 1 in 10,000 male births. Hae- MR imaging is able to delineate hypertrophic pannus
mophilia is due to a deficiency of factor 8, a constituent from fresh haemorrhage and is more sensitive in
of the clotting cascade; there is also factor 9 deficiency detecting early subcortical cysts and cartilage destruc-
(Christmas disease) and factor 11 deficiency. tion (HERMANN et al.1992). Gadolinium enhancement
Soft tissue haemorrhage around the knee joint of the synovium is less intense than with juvenile idio-
is relatively common and will cause distortion and pathic arthritis, probably due to the hypovascular con-
obliteration of fat and fascial planes on radiographs. nective tissue and haematological degradation prod-
MR imaging is able to accurately localise the site ucts in the synovium (NAGELE et al.1995).
and soft tissue compartment of this haemorrhage. Another feature of the disease is haemophilic
Repeated soft tissue bleeding can lead to myositis pseudotumours, which are lytic expansile lesions that
ossificans. occur in the ilium, femur and tibia. They may be
Later radiographic signs are of peri-articular osteo- intra-osseous, where they are seen as a well-defined
porosis and accelerated maturation, with epiphyse al medullary lucency with a thin sclerotic rim, or sub-
overgrowth and widening of the intercondylar notch periosteal, where there is extensive subperiosteal
(Fig. 6.9). These signs reflect the localised hyperaemia re action and resorption of cortex with a soft tissue
and synovial proliferation caused by repeated haem- component. Discrimination of pseudotumours may
arthroses. Radiographs are not able to differentiate be improved by MR imaging as it will detect any
between fresh haemorrhage and increased synovium blood degradation products (GAARY et al. 1996).
86 K. Johnson and A. M. Davies
6.13.2
Thalassaemia
6.14
Diaphyseal Aclasis
Fig.6.10. A 14-year-old boywith diaphyseal aclasis who clini-
Diaphyseal aclasis is one of the multiple pedunculat-
cally has multiple subcutaneous bony protuberances. An AP
ed bony exostoses that commonly occur around the radiograph of the knee shows multiple bony exostoses which
faster growing joints. There is a familial preponder- are directed away from the joint
ance, with some cases being autosomal dominantly
inherited with variable penetrance. The exostoses
appear within the first 10 years of life, are often bilat- LENCE et al. 1979; SPRANGER et al. 1982). All the
eral, can be symmetrical and are assodated with types have reduced bone mineral density and are
short stature (Fig. 6.10). The bony prominence points prone to repeated fracturing which causes metaph-
away from the nearest joint and is covered with a car- yseal irregularity, epiphyseal widening, widespread
tilage cap to produce a lobulated outline on radio- bone remodelling and deformity and excessive callus
graphs (SHOGRY and ARMSTRONG 1990). formation (Fig. 6.11). Traumatic fragmentation of the
Around the knee joint the exostosis may causes physes may lead to "popcorn" calcification of the
localised pain and swelling due to the development epiphysis.
of a tendonitis or bursitis. Nerve compression effects Type II is invariably lethal. Distinguishing features
may occur, as can pseudo-aneurysm formation and among the other types are the age at presentation, the
venous compression within the popliteal fossa. Malig- propensity to fracture, the state of dentition and the
nant change can occur; the rate has been reported to colour of the sclera.
be between 1% and 25% (WILLMS et al. 1997; WUIS-
MAN et al. 1997; VOUTSINAS and WYNNE-DAVIES
1983).
6.16
Modelling Deformities
6.16.1.2
Craniometaphyseal Dysplasia
6.16.1.3
Otopalatal Digital Syndrome
6.16.1
Erlenmeyer Flask Deformity
metaphyseal widening of the small bones of the intercondylar noteh. On a lateral radiograph of the
hand. The condition is associated with frontal and knee the irregularity is usually seen posterior to the
facial hypoplasia, cleft palate and conductive deaf- intercondylar eminen ce.
ness (GALL et al. 1972).
6.16.2.2
6.16.1.4 Chondrodysplasia Punctata
Osteodysplasty (Melnick-Needles Syndrome)
Autosomal recessive and dominant forms of chon-
These patients do not usually present until early drodysplasia punctata are recognised, and the disor-
infancy and nearly all reported cases have been der occurs in 1 in 110,000 births. There is small punc-
female (KAUFMAN 1973). As the child grows there is tate calcification of varying size in the epiphysis that
involvement of alliong bones, which develop ribbon occurs before the ossification centres. The dominant
waviness, cortical irregularity and S-shaped bowing form (Conradi-Hunerman) involves the long bones
(MELNICK and NEEDLES 1966). and the spine. It is non-Iethal and there is asymmetri-
cal shortening of the limbs (Fig. 6.13).
6.16.1.5
Gaucher's Disease
6.16.1.6
Niemann-Pick Disease
Fig.6.13. AP radiograph of the knee in chondrodysplasia
Niemann-Pick disease occurs due to abnormal accu- punctata. There are stippled epiphyses, with multiple calcified
punctata of varying size
mulation oflipid in the body. Skeletal features are sim-
ilar to those of Gaucher's disease, except that epiphy-
seal necrosis is not seen (LACHMAN et al. 1973). The recessive form shows symmetrical rhizomelic
shortening of the limbs with mild spinal involve-
ment. There are facial and ocular abnormalities and
6.16.2 death usually occurs within 2 years. There is diaph-
Epiphyseallrregularity yseal thickening and metaphyseal splaying, humeral
involvement being more severe than in the femur
6.16.2.1 (SHEFFIELD et al. 1976).
Normal Variant
6.16.2.3
Irregular ossification of the epiphyses around the Hypothyroidism
knee joint is common, occurring in approximately
80% of children of about 4 years of age. The changes Congenital hypothyroidism is usually due to absence
may be unilateral or bilateral and are more frequent of functioning thyroid tissue. There is delayed skel-
in girls. The irregularity is most often seen on the etal maturation in both the appearance and the ossi-
lateral aspect of the knee but very rarely involves the fication of the epiphysis. The epiphyses are fragment-
Congenital and Developmental Abnormalities of the Knee 89
ed and stippled. The stippling is coarser with larger mal facies and hypotonia. There are widespread skull,
fragments than in chondrodysplasia punctata (CHEW ehest, gastrointestinal and cardiac anomalies. There
1991). is metaphyseal splaying of the long bones and irregu-
lar stippling of the epiphyses.
6.16.2.4
Multiple Epiphyseal Dysplasia 6.16.2.9
Zellweger (Cerebrohepatorenal) Syndrome
Multiple epiphyseal dysplasia is the commonest con-
genital skeletal dysplasia, and invariably involves the Zellweger syndrome is an autosomal recessive con-
proximal femoral epiphysis if epiphyses around the knee dition with muscle hypotonia, hepatomegaly, renal
are affected. There is shortening of the limbs and irregu- cysts and facial and cerebral dysgenesis. There is stip-
lar mottled calcification of the epiphyses. The epiphyseal pling of the epiphysis. Death occurs in early infancy
irregularity may cause premature degenerative disease (BAROLETTI et al.1978).
(SPRANGER 1976; VAN MOURIK et al. 2001).
6.16.2.10
6.16.2.5 Transplacental Acquired Infections
Spondylo-epiphyseal Dysplasia
See Sect. 6.11.
There are five subcategories of spondylo-epiphyseal
dysplasia (SED): SED congenita (mild coxa vara), SED
congenita (severe coxa vara), SED tarda (autosomal 6.16.3
recessive and dominant), SED tarda (x-linked) and SED Metaphyseal Widening and Splaying
tarda (with progressive arthropathy) (WYNNE-DAVIES
et al. 1985). There is wide variation in the presentation 6.16.3.1
of the disorder, with varying severity. The epiphyseal Rickets
appearance can varyfrom normal to severe fragmenta-
tion and there also may be some metaphyseal irregular- Deficiency of vitamin D or its metabolie derivatives
ity (SPRANGER and LANGER 1970). causes failure in bone mineralisation and in cartilage,
causing riekets (RUSSELL and HILL 1974). Such defi-
6.16.2.6 ciency may be dietary or be due to familial vitamin
Fetal Warfarin Syndrome D resistance (x-linked hypophosphataemia). The fea-
tures manifest first in the rapidly growing bone ends,
Fetal warfarin syndrome results from maternal inges- particularly around the wrist, knee and proximal
tion of warfarin, whieh causes bleeding in the primor- humerus. There is widening of the growth plate, cup-
dial cartilage in the fetus. At birth there is stippling ping, fraying and irregularity of the distal metaphyses
of the unossified epiphysis, most commonly seen in and generalised loss ofbone density (Fig. 6.14). Over-
the proximal femur and calcaneum. The appearances all skeletal maturation may be delayed (SWISCHUK
disappear by 1 year of age (JOHNSON 1979). and HAYDEN 1979).
6.16.2.7 6.16.3.2
Trisomy18 Trisomy21
See Sect. 6.16.2.8.
Trisomy 18 is more commonly seen in females. The
child is usually hypotonie with abnormal facies. The 6.16.3.3
child does not usually survive beyond 6 months. There Hypophosphatasia
are associated cardiac and renal anomalies. The epiph-
yses are stippled (FRANCESHINI et al. 1974). Hypophosphatasia is an autosomal recessive condi-
tion due to a deficiency of serum and alkaline phos-
6.16.2.8 phatase, with increased urinary phospho-ethanol-
Trisomy21 amine. There are four forms that decrease in the
severity of symptoms with increasing age. In the neo-
Trisomy 21 is the commonest chromosomal defect, natal form, death usually occurs within 6 months of
classieally presenting with mental retardation, abnor- birth. There is severe hypotonia, and radiologieally
90 K. Johnson and A. M. Davies
6.17 6.18
Patella Congenital Dislocation of the Patella
dislocation and loss of congruity of the articulation A high-Iying patella (patella alta) is associated
with the femur, but no assessment of the ability of with contractures of the quadriceps muscles, which
ultrasound to detect mild patellar subluxation has can be seen as abnormal signal on MR imaging, patel-
been published (MILLER et al.1998). lar subluxation and Sinding-Larsen-Johansson dis-
In the skeletally mature knee, the lateral radio- ease.An abnormally low-Iying patella (patella baja) is
graph can be used to assess the normal position of associated with achrondroplasia and poliomyelitis.
the patella. The position is obtained from the ratio On the axial projection there is some dispute over
of the length of the patellar tendon to the diagonal the variation in angulation that can be considered
length of the patella (INsALL and SALVATI 1971), the to be within normal limits. This is particularly true
normal ratio being approximately 1, i.e. the diagonal if radiographs and axial CT measurements are com-
length of the patella should equal the length of the pared (MARTINEZ et al. 1983). On the axial projec-
patellar tendon. tion, lines are drawn along the patellofemoral sulcus.
In the skeletally immature knee this ratio is inac- The angle of intersection is known as the sulcus
curate owing to the underestimation of the size of the angle. The average value is 1420 on plain radiographs
unossified patella. A separate ratio has been derived (BRATTSTRÖM 1964) and 121 0 on CT (STANDFORD et
for use in the immature skeleton (KOSHINO and SUGI- al. 1988). An abnormally shallow sulcus is associated
MOTO 1989) using the midpoint of the epiphyseallines. with an increased incidence of patellar subluxation
The length between the midpoints of the distal femoral (BRATTSTRÖM 1964).A congruence angle is obtained
and proximal tibial epiphyseal growth plates is calcu- from two lines: one bisecting the sulcus angle and
lated; this is the femoral-tibial distance (FT). The patel- one from the sulcus angle to the apex of the patella.
lar axis (PT) is measured from the tibial midpoint This congruence angle should be less than 160 on
to the centre of the patella. A PT/FT ratio of 0.9-1.3 plain radiographs (MERCHANT et al. 1974) and no
is considered normal. A deviation of more than 20% more than 100 on CT (STANDFORD et al. 1988)
from normal is considered significant (Fig. 6.15). (Fig.6.16).
Fig.6.15. Demonstration of the size of the patella in ossified Fig.6.16. Skyline views of the patella. The left knee dem-
skeleton from a lateral radiograph of the knee. The midpoint onstrates the sulcus angle. This is the angle of intersection
of the growth plate of both the tibia and the femur is calcu- of lines drawn along the femoral condyles. The right knee
lated. A connecting line is drawn between these two points demonstrates the congruence angle. This angle is formed
(this is the femoral-tibialline, PD. A line from the midpoint from a line bisecting the sulcus angle and another drawn
of the patella to the midpoint of the tibia is drawn (this is the from the apex of the sulcus angle through the inferior point
patella-tibialline, PD. The PT1FT ratio is calculated; a ratio of of the patella. The congruence angle should be less than
0.9-1.3 is considered normal 16°
92 K. Johnson and A. M. Davies
anterior subluxation of the knee. J Bone Joint Surg James W, Moule B (1966) Hypophosphatasia. C!in Radiol
51A:255-269 17:368-376
Curtis BH, Fisher RL (1970) Heritable congenital tibiofemoral Johansson E, Aparisi T (1983) Missing cruciate ligament in
subluxation. Clinieal feature as surgieal treatment. J Bone congenital short femur. J Bone Joint Surg 65A:ll09-1115
Joint Surg 52A:1104 Johnson JF (1979) Case report. Coumadin embryonopathy.
Dalinka MK, Coren G, Hensinger R et al (1974) Arthrography Skeletal RadioI3:244-246
in Blount's disease. Radiology 113:161-164 Kaufmann KJ (1973) Osteodysplasia (Melniek-Needles). Semin
Dietz WH, Gross WL, Kirkpatriek JA (1982) Blount disease RoentgenoI8:182-186
(tibia vara): Another skeletal disorder associated with child- Klatte EC, Franken EA, Smith JA (1976) The radiographie spec-
hood obesity. J Pediatr 101:735-737 trum in neurofibromatosis. Semin Roentgenol11:17-33
Ducou Le Point H, Mousselard H, Rudelli A et al (1995) Ko JY, Shih CH, Wenger DR (1999) Congenital dislocation of
Blount's disease: Magnetie resonance imaging. Pediatr the knee. J Pediatr Orthop 19(2):252-259
Radiol25: 12-14 Koshino T, Sugimoto K (1989) New measurement of patella
Dugdale TW, Renshaw TS (1986) Instability of the patel- height in the knees of children using the epiphyseal line
Iofemoral joint in Down's syndrome. J Bone Joint Surg midpoint. J Pediatr Orthop 9:216-219
68A:405-413 Kujala UM, Osterman K, Kurmanu M (1989) Patellofemoral
Faure C, Petrel P (1968) Hereditary onycho-osteodysplasia. relationship in recurrent patellar dislocation. J Bone Joint
Ann Radiol11:376-388 Surg 71A:788-792
Foreman KA, Robertson WW (1985) Radiographie measure- Kuhns LR, Slovis T, Hernandez R et al (1977) Knee matura-
ment of infantile tibia vara. J Pediatrie Orthop 5:452-455 tion as a differentiating sign between congenital rubella
Frantz CH, O'Rahilly R (1961) Congenital skeletallimb defi- and cytomegalovius infection. Pediatr RadioI6:36-38
ciencies. J Bone Joint Surg 43A:1202-1224 Lachman R, Crocker A, Schulman J et al (1973) Radiologieal
Franceshini P, Fabris C, Ponzone A et al (1974) Skeletal alter- findings in Niemann-Piek disease. Radiology 108:659-664
ations in Edwards disease (trisomy 18 syndrome). Ann Lalain JJ, Lerat JL, Moyen B et al (1987) Les luxations congeni-
RadioI17:361-367 tales de rotule. Apport du scanner-indications chirurieales.A
Gaary E,Grolin JB,Jaramillo D (1996) Pseudotumour and arthrop- propos de 22 cases Rev Chir Orthop 73 [SuppI2]:152-156
athy in the knees of a haemophiliac. Skeletal RadioI25:85-87 Laor T, Jaramillo D, Hoffer FA et al (1996) MR imaging in con-
Gall JC, Stern AM, Poznanski AK et al (1972) Oto-palatal- genitallower limb deformities. Pediatr RadioI26:381-387
digital syndrome. Comparison of c1inieal and radiologieal Laurence M (1967) Genu recurvatum congenitum. J Bone Joint
manifestations in males and females. Am J Hum Genet Surg49B:121-134
24:24-36 Laville JM, Chau E, Willemen L, Kohler R, Garin C (1999)
Girdwood TG, Gibson WJA, MacIntosh TF (1969) Craniome- Blount's disease: c1assification and treatment. J Pediatr
taphyseal dysplasia congenita - Pyle's disease in a young Orthop 8B:19-25
child. Br J RadioI42:299-303 Lawson JP (1985) Symptomatie radiographie variants in
Green JP, Waugh W (1968) Congenitallateral dislocation of the extremities. Radiology 157:625-631
patella. J Bone Joint Surg 50B:285-289 Levine AM, Drennan JC (1982) Physiologieal bowing and tibia
Green WT (1975) Painful bipartite patella: areport of three vara: the metaphyseal-diaphyseal angle in the measurement
cases. Clin Orthop 110: 197-200 of bowleg deformities. J Bone Joint Surg 64A: 1158-1163
Harty MP, Markowitz RZ, Rustien RM et al (1994) Imaging Loder RT, Schaffer JJ, Bardenstein MB (1991) Late-onset tibia
features of human immunodeficiency virus infection in vara. J Paediatr Orthop 11: 162-167
infants and children. Semin RoentgenoI29:303-314 MacEwan DW, Dunbar JS (1958) Radiology study of physiolog-
Henderson Re, Lechner CT, DeMasai RA et al (1990) Variability ie knock knees in childhood. J Can Assoc RadiollO:59-63
in radiographie measurement of bowlegged deformity in Macnab I (1952) Recurrent dislocation of the patella. J Bone
children. J Pediatr Orthop 10:491-494 Joint Surg 34A:957-967
Hermann G, Gilbert MS, Abdelwahab IF (1992) Hemophilia: MacPherson RI (1974) Craniodiaphyseal dysplasia, a disease
evaluation of musculoskeletal involvement with CT, Sonog- or group of diseases? J Can Assoc RadioI25:22-33
raphy and MR imaging.AJR 158:119-123 Martinez S, Korokobin M, Fondren FB, Hedlund LW, Goldner JL
Heyman SH, Herndon CH (1949) Congenital posterior angula- (1983) Diagnosis of patellofemoral malalignment by com-
tion of the tibia. J Bone Joint Surg 41A:476-488 puted tomography. J Comput Assist Tomogr 7:1050-1053
Heselson NG, Raad MS, Hamersma H et al (1979) The radio- Malumed J, Hudanieh R, Collins M (1999) Congenital absence
logieal manifestation of the metaphyseal dysplasia (Pyle's of the anterior and posterior cruciate ligaments in the
disease). Br J RadioI52:431-440 presence of bilateral absent patellae. Am J Knee Surg
Heyman CH, Herndon CH, Heiple KG (1959) Congenital pos- 12:241-243
terior angulation of the tibia with talipes calcaneus. A Melniek JC, Needles CF (1966) An undiagnosed bone dyspla-
long-term report of eleven patients. J Bone Joint Surg sia; a 2 family study of 4 generations and 3 generations. AJR
41A:476-88 97:39-48
Hodson P, Goldblatt J, Beiehton P (1976) Non-neuropathic Merchant AC, Mercer RL, Jacobsen RH, Cool CR (1974) Roent-
Gaucher disease presenting in infancy. Br J Radiol genographie analysis of patellofemoral congruence. J Bone
49:1018-1022 Joint Surg 56A:1391-1396
Insall J, Salvati E (1971) Patella position in the normal knee Miller TT, Shapiro MA, Schultz E et al (1998) Sonography of
joint. Radiology 101:101-112 patella abnorma!ities in children.AJR 171:739-742
Insall J, Goldberg V, Salvati E (1972) Recurrent dislocations and Moseley JE (1974) Skeletal change in the anaemias. Semin
the high riding patella. C!in Orthop 88:67-69 RoentgenoI9:169-184
Congenital and Developmental Abnormalities of the Knee 95
Nagele M, Bruning R, Kunze V et al (1995) Hemophilic arthrop- Spranger J, Langer LO, Wiedman HR (1974) Bone dysplasias.
athy of the knee joint. Static and dynamic Gad-DPTA An atlas of constitutional disorders of skeletal develop-
enhanced MRI. Eur RadioI3:547-552 ment. Saunders, Philadelphia
Newell RLM, Durbin FC (1976) The aetiology of congenital Spranger J, Cremin B, Beighton P (1982) Osteogenesis imper-
angulation of tubular bones with constriction of the med- fecta congenita: features and prognosis of a heterogeneous
ullary canal and its relationship to congenital pseudoar- condition. Pediatr RadioI12:21-27
throsis. J Bone Joint Surg 58B:444-447 Spranger JW, Langer CO (1970) Spondylepiphyseal dysplasia
Niebauer JJ, King DE (1960) Congenital dislocation of the knee. congenital. Radiology 94:313-322
J Bone Joint Surg 42A:207-225 Stanford W, Phelan J, Kathol MH et al (1988) Patellofemoral
Ogden JA, McCarthy SM, Jokl P (1982) The painful bipartite joint motion: evaluation by ultrafast computed tomogra-
patella. J Pediatr Othop 2:263-269 phy. Skeletal RadioI17:487-492
Pappas AM (1983) Congenital abnormalities of the femur and Stanisavljevic S, Zemenick G, Miller D (1976) Congenital irre-
related lower extremity malformations: classification and ducible, permanent lateral dislocation of the patella. Clin
treatment. J Pediatr Orthop 3:45-60 Orthop 116:190-199
Pappas AM (1984) Congenital posteromedial bowing of the Stark JE, Siegel MJ, Weinberger E, Shaw DW (1995) Discoid
tibia and fibula. J Pediatr Orthop 4:525-531 menisci in children. MR feature. J Comput Assist Tomogr
Rathgeb JM, Ramsey PL, Cowell HR (1974) Congenital kypho- 19:608-611
scoliosis of the tibia. Clin Orthop 103:178-190 Stern M (1964) Persistent congenital dislocation of the patella.
Resnick CS, Young JWR, Levine AM et al (1990) Osteofibrous dys- Report of a case. J Int Coll Surg 41:654-656
plasia (ossifying fibroma) of the tibia. Skeletal Radio119:217 Swischuk LF, Hayden CK (1979) Rickets. A roentgenographic
Russell JGB, Hill LF (1974) True fetal rickets. Br J Radiol scheme for diagnosis. Pediatr RadioI8:203-208
47:732-734 Thomas NP, Jackson AM, Aichroth PM (1985) Congenital
Ryan JR, Perrin JC, Morewa LG (1978) Congenital synostosis absence of the anterior cruciate ligament. J Bone Joint Surg
of the knee. Clin Orthop 135:34-35 67B:572-575
Sachdev M, Berry K, Chawla S (1982) Osseous manifestations van Mourik JBA, Weerdenburg JP, Verhar JAN (2001) Magnetic
in congenital syphilis. A study of 55 cases. Clin Radiol resonance imaging of multiple epiphyseal dysplasia (type
33:319-323 2) of the knee. Radiography 7:61-64
Salenius P,Vankka E (1975) The development of the tibiofemo- Voutsinas S, Wynne-Davies R (1983) The frequency of malig-
ral angle in children. J Bone Joint Surg 57 A:259-269 nant disease in diaphyseal aclasis and neurofibromatosis. J
Sharrad WJW (1976) Problems of childhood. Knock knees and Med Genet 20:345-349
bow legs. Br Med J i:826-827 Walker J, Rang M, Daneman A (1991) Ultrasonography of
Shaw JCC (1988) Copper deficiency and non-accidental injury. the unossified patella in young children. J Pediatr Orthop
Arch Dis Child 63:448-455 11:100-102
Sheffield LJ, Danks DM, Mayne V et al (1976) Chrondysplasia Weiner D, MacNab I, Turner M (1977) The fabella syndrome.
punctata - 23 cases of a mild and relatively common vari- Clin Orthop 126:213-215
ant. J Paediatr 89:916-923 Willms R, Hartwig CH, Bohm P, Sell S (1997) Malignant trans-
Shogry MEC,Armstrong P (1990) Reactive bursa formation sur- formation of a multiple cartilaginous exostosis - a case
rounding an osteochondroma. Skeletal RadioI19:465-467 report. Int Orthop 21:133-136
Shopfner CF, Coin CG (1969) Genu varus and valgus in chil- Wolfgang GL (1984) Complex congenital abnormality of the
dren. Radiology 92:723-732 bone extremities: femoral bifurcation, tibial hemimelia, and
Sillence DO, Senna, Danks M (1979) Genetic heterogeneity in diastasis of the ankle. Case report and review of the litera-
osteogenesis imperfecta. J Med Genet 16: 101-116 ture. J Bone Joint Surg 66A:453-458
Silverman JM, Mink JH, Deutsch AL (1989) Discoid menisci of Wuisman PIJM, Jutte PC, Ozaki T (1997) Secondary chondro-
the knee: MR imaging appearance. Radiology 173:351-354 sarcoma in osteochondromas: medullary extension in 15 of
Södergard J, Ryoppy S (1990) The knee in arthrogryposis mul- 45 cases. Acta Orthop Scand 68:396-400
tiplex congenital. J Pediatr Orthop 10:177-182 Wynne-Davies R, Hall CM, Apley AG (1985) Atlas of skeletal
Spranger J (1976) The epiphyse al dysplasias. Clin Orthop Relat dysplasias. Churchill Livingstone, Edinburgh
Res 114:46-60
7 Non-accidentallnjury
STEPHEN CHAPMAN
CONTENTS toma and long bone fraetures and stated that the
injuries "were either not observed or were denied
7.1 Introduction 97
when observed" (CAFFEY 1946). Further reeognition
7.2 Anatomy and Physiology of
the Developing Skeleton 97 of the eondition followed and in 1961 KEMPE et al.
7.3 Patterns of Skeletal Injury 98 eoined the powerful term "the battered ebild" for a
7.3.1 Subperiosteal New Bone Formation 98 multidisciplinary meeting of the Ameriean Aeademy
7.3.2 Metaphyseal Fractures 98 of Pediatries (KEMPE et al. 1962). Within 5 years,
7.3.3 Epiphyseal Separations 100
every US state had passed a ebild abuse reporting
7.3.4 Shaft Fractures 100
7.4 Fracture Healing 100 law.
7.4.1 Radiological Changes in the Soft Tissues 101 Greater publie and medieal awareness of the eon-
7.4.2 Response of the Periosteum to Injury 102 dition, the inereasingly broad definition of abuse to
7.4.3 Fracture Line Definition 102 include sexual abuse and emotional maltreatment,
7.4.4 Calius Formation 102
greater emphasis on thorough investigation and, in
7.4.5 Remodelling 102
7.5 The Distal Femur 103 the United States, a mandated reporter system, have
7.5.1 Shaft Fractures 103 eombined to inerease the number of reported eases of
7.5.2 Metaphyseal Fractures 103 ebild abuse and negleet (RICHMAN 2000). Data for the
7.6 The Proximal Tibia 103 United States reeord that in 1974 there were 60,000
7.7 The Proximal Fibula 103 abused or negleeted ehildren (KRUGMAN 1997); in
7.8 Differential Diagnosis 103
7.8.1 Normal Variants 103 1997 the figure was nearly 3.2 million (WANG 1997).
7.8.2 Birth Trauma 104 While a substantial number (possibly as high as 65%;
7.8.3 Osteogenesis Imperfecta 104 WANG 1997) of these reported eases of ebild abuse
7.8.4 Rarer Bone Dysplasias with Bone Fragility 105 are found to be unsubstantiated after investigation,
7.8.5 Temporary Brittle Bone Disease 105 the figures do reveal areal inerease, whieh refleets
7.8.6 Other Differential Diagnoses 105
References 106 the inereasing social and financial strains on families,
both rieh and poor.
Radiologists have a eritieal role in identifying ehil-
dren who have been injured and who may be injured
further if the eorreet diagnosis of non-aeeidental
7.1 injury is not made and the ehild returned to the eare
Introduction of the perpetrator. The radiologist's roles are listed in
Table 7.1.
The reeognition of ehild abuse as an evil, as opposed
to a parental right, is relatively modern. The first clear
medieal deseription of the abuse of ehildren was in
1860, when AMBROISE TARDlEU deseribed the soft 7.2
tissue, skeletal and intraeranial injuries suffered by 32 Anatomy and Physiology
Parisian ebildren (TARDlEU 1860). In 1946, CAFFEY of the Developing Skeleton
deseribed the association between subdural haema-
At birth the diaphyses or shafts of the tubular bones
are eomposed of fetal or woven bone that lacks haver-
S. CHAPMAN, MD
sian systems. However, as periosteal-mediated appo-
Consultant Paediatric Radiologist, Birmingham Children's sitional bone formation and remodelling enlarge the
Hospital, Steelhouse Lane, Birmingham, B4 6NH, UK overall diameter of the shaft and the width of the
Table 7.1. Roles of the radiologist KOGUTT et al. 1974; LAUER et al. 1974; MERTEN et al.
1983; RYAN et al. 1977). EBBIN et al. (1969) did not
To identify the radiologieal findings
To differentiate abnormalities from normal variation in the find any fractures in children over the age of 5 years.
growing skeleton AKBARNIA et al. (1974), in aseries of 217 abused
To differentiate traumatie from non-traumatie abnormalities infants, found that 78% of fractures occurred below
To suggest a mechanism of causation for traumatie lesions the age of 3 years and 50% occurred below 1 year of
To date the injury
age.
To identify or suggest an underlying skeletal abnormality
whieh would predispose the child to fractures, either (a)
when there are obvious radiographie abnormalities or (b)
when the mechanism seems to be appropriate but the 7.3.1
degree of force is less than would have been expected to Subperiosteal New Bone Formation
cause the injury
line passes peripherally towards the cortex it deviates line is wider there will be alucent line extending
away from the physis to undercut a thieker peripher- across the metaphysis (Fig. 7.4); the fracture line may
al segment that encompasses the subperiosteal bone be visible in only one plane. If the radiographie pro-
collar (Fig. 7.2). Thus, there is separation of a disk of jection is such that only the thicker peripheral seg-
bone, whieh is thin centrally and thicker peripher- ment is demonstrable, then the resulting radiograph-
ally (Fig. 7.3). The radiographie appearance is typi- ie finding is the corner fracture (Figs. 7.1, 7.5a). When
cal, but radiographically variable. When the separat- the entire fracture is visible and viewed obliquely, the
ed ossified disk is very thin, the radiograph is normal, disk appears to be further separated from the adja-
although histologieal examination of the resected cent metaphysis and the bucket-handle appearance
bone may demonstrate the lesion. When the fracture results (Fig. 7.2).
Haemorrhage at the fracture line and beneath the
adjacent periosteum is usually minimal or absent.
SNBF is usually lacking and sclerosis along the frac-
ture margins is inconspicuous. When healing is com-
plete, there may be no radiographie evidence of the
previous injury (Fig.7.5). Dating of these injuries
is difficult. During healing, hypertrophie cartilage
may extend into the metaphysis, producing focal
lucencies, indentations or broader scalloping of the
metaphyseal margin (Fig. 7.6).
Fig. 7.2. A large bucket-handle fracture of the proximal tibia with Fig.7.4. A metaphyseal fracture of the proximal tibia, which is
a markedly thickened peripheral segment on the medial side visible only as a thin radiolucent band
100 S.Chapman
a
7.3.3
Epiphyseal Separations
7.3.4
Fig. 7.5a, b. A metaphyseal fracture of the distal femur. a In the
Shaft Fractures
acute situation there is the typieal appearance of corner frac-
tures. These did not disturb the periosteal attachment and, with-
out any significant haemorrhage, there was no radiographie In general, shaft fractures are not specific for abuse.
evidence of bony healing. b A follow-up film 8 weeks later is The most common sites are the femur and humerus
entirely normal with no evidence of the previous injury followed by the tibia and forearm (KING et al. 1988;
KowAL-VERN et al. 1992; THOMAS et al. 1991; WOR-
LOCK et al. 1986).
7.4
Fracture Healing
Table 7.2. Timing of radiological changes in paediatric fractures (modified from O'CONNOR and COHEN 1998). NB The evolution
of fracture healing is more rapid in the young infant compared with the older child and is influenced by lack of immobilisation
and repetitive trauma
Resolution of soft tissue changes (depends on severity of injury) 2-5 days 4-10 days 10-21 days
Subperiosteal new bone formation 4-10 days 10-14 days 14-21 days
Loss of fracture line definition and appearance of soft callus 10-14 days 14-21 days
Hard callus 14-21 days 21-42 days 42-90 days
RemodeHing 3 months 1 year 2 years - growth plate closure
Fig. 7.7a, b. Femoral shaft fracture. a In the acute situation, 1 h after the fracture was sustained, the soft tissue planes are still
weH preserved. b The foHow-up film at 10 days demonstrates that the muscle/fat interfaces have become ill defined because of
extension of blood from the fracture site. The fracture line is now wider and less weH defined
102 S.Chapman
periosteal new bone is dependent on the age of the widens (Fig. 7.7b). There are no objective criteria to
child: the younger the child, the earlier it appears. In estimate fracture healing using this sign but it is not
the infant it can be seen radiographically as early as 4 apparent radiologically before 10 days and reaches a
days after injury, but an interval of7-14 days is more peak at 2-3 weeks.
usual. CUMMING (1979), in a study of birth-related
fractures, concluded that the absence of periosteal
new bone 11 days after birth should suggest the pos- 7.4.4
sibility of abuse. Beyond the neonatal period, perios- Callus Formation
teal reaction delayed longer than 21 days may be a
consequence of poor nutrition or vitamin D deficien- With the production of osteoid and its subsequent
cy. The amount of subperiosteal new bone increases calcification and ultimate transformation into bone
with repeated trauma and continuing subperiosteal there is an increase in density along the fracture line.
haemorrhage (leading to exuberant callus forma- With impacted fractures and other fractures where
tion), but chronic repetitive trauma may be of such the periosteum remains intact or little disturbed, this
severity as to destroy periosteal new bone as it devel- may be the only evidence of injury and it is, there-
ops, with resultant failure of the usual sequence of fore, a less valuable sign than subperiosteal new bone
changes with healing. and fracture line clarity. Again, movement may inhib-
Ultrasonography has been used to demonstrate it or destroy endosteal callus and a fracture line may
subperiosteal haemorrhages and occult long bone remain clearly visible after a bone has united by peri-
fractures (GRAIF et al. 1988) before radiographs osteal new bone. Endosteal soft callus is first visu-
confirm the presence of bony injury and may be a alised as an ill-defined increase in density and begins
useful supplementary investigation when attempt- to form soon after the first appearance of periosteal
ing to date an acute injury. Radionuclide bone scan- new bone, i.e. at about 10-14 days. When lamellar
ning is popular in some centres but its role is bone bridges the fracture, about 1 week later, the
controversial. False-negative rates range from 0.8% stage of hard callus has begun. The chronology of
(STY and STARSHAK 1983) to 27% (MERTEN et al. callus formation and fracture consolidation is depen-
1983). From the data of these and others (HAASE dent on age.A birth-related fracture may be united at
et al. 1980; ]AUDES 1984; SMITH et al. 1980) we can 1 month, but a similar fracture in an 8-year-old will
conclude that scintigraphy demonstrates superior be united at 8 weeks.
sensitivity for rib fractures, especially at the cos-
tovertebral junction, undisplaced shaft fractures
and subperiosteal haemorrhage. ROSENTHAL et al. 7.4.5
(1976) reported that the earliest fracture demon- Remodelling
strated with radionuclides was 7 h after the injury
and that if a bone scan obtained 3 or more days Estimation of the age of a fracture by this criterion
after injury fails to reveal a focallesion, then a frac- is very difficult because initial deformity, the volume
tu re has been excluded. A positive scan steadily of callus produced and the age of the child are major
becomes less so as the age of a fracture increases, variables. In the young infant with a stable undis-
but there is no definite time course. The pitfalls of placed fracture, the remodelling process may be com-
imaging with a bone-seeking isotope are that sym- plete at 3 months, while in the older child with an
metrical metaphyseal injuries adjacent to normal- angular deformity or a markedly displaced fracture,
ly "hot" epiphyseal growth plates, some vertebral remodelling may continue for up to 2 years.
body fractures, and many skull fractures may be
undetectable.
7.5
7.4.3 The Distal Femur
Fracture Une Definition
The overall prevalence of femoral fractures in abused
Fresh fractures, including metaphyse al fractures, have children with fractures is 20% (AKBARNIA et al. 1974;
sharply defined margins. With the development of GALLENO and OPPENHEIM 1982; HERNDON 1983;
an osteoclastic response to necrotic bone, the frac- O'NEILL et al. 1973; ROSENBERG and BOTTENFIELD
ture ends become less well defined and the fracture 1982).
Non-accidental Injury 103
7.5.1 7.6
Shaft Fractures The Proximal Tibia
Femoral shaft fractures have been estimated to The tibia is a common site of fracture in non-acci-
account for 1.6% of all fractures in children (LANDIN dental injury, accounting for 7%-18% of the total
1983) (Fig. 7.7). There is no association between the (AKBARNIA et al. 1974; GALLENO and ÜPPENHEIM
morphological features of the fracture and the likeli- 1982; HERNDON 1983; KING et al. 1988; LEvENTHAL
hood of abuse (REx and KAY 2000). Although there is et al. 1993; LODER and BOOKOUT 1991; WORLOCK
some disagreement in the published literature, trans- et al. 1986). Metaphyseal fractures considerably out-
verse and oblique fractures seem to be the most number shaft fractures; indeed, the proximal tibial
common type, both in abuse and in accidental injury; metaphysis is the most common site for this partic-
spiral fracture is not suggestive of abuse (KING et ular fracture. There are no specific points relating
al. 1988; LODER and BOOKOUT 1991; SCHERL et al. to metaphyseal fractures at this site (KLEINMAN and
2000). MARKS 1996).
Age, however, is a significant factor in determining
the likelihood that a femoral shaft fracture is due to
abuse. ANDERSON (1982) reviewed 122 femoral frac-
tures in 117 patients. In the group under 13 months
of age, 15 (83%) of 18 had been abused; of 24 chil- 7.7
dren under 2 years of age, 19 (79%) had been abused. The Proximal Fibula
BEALS and TuFTS (1983) suggested that as many as
30% of femoral fractures in children less than 4 years Metaphyseal fractures are less common in the proxi-
of age were the result of a non-accidental injury. mal fibula than in the proximal tibia. When present,
THOMAS et al. (1991) found that 60% of femoral frac- there is usually an associated metaphyseal fracture of
tures in children less than 1 year of age were due to the proximal tibia.
abuse; above the age of 2 years, 90% were accidental.
In their study of 47 femoral fractures, REX and KAY
(2000) noted that 13 (93%) of the 14 inflicted frac-
tures occurred in children less than 1 year of age.
This general body of opinion contrasts with the study 7.8
of WELLINGTON and BENNET (1987) in which there Differential Diagnosis
were only 14 fractures (36%) due to abuse in 39 chil-
dren under 1 year of age; furthermore, only two were 7.8.1
considered definite abuse and 12 were suspected. Normal Variants
7.8.1.1
7.5.2 Physiological Subperiosteal New Bone Formation
Metaphyseal Fractures
Radiographically, physiological subperiosteal new
The distal femur is a common site for metaphyseal bone formation presents as a thin well-defined densi-
fractures and approximately 25% of these fractures ty,2 mm or less in thickness, separate from the cortex
are to be found at this site (KLEINMAN and MARKS of the bone and restricted to the diaphysis (Fig.7.8)
1998; LODER and BOOKOUT 1991; WORLOCK et al. (KLEINMAN and KWON 1998). The tibia is more com-
1986). They can be bilateral or unilateral and are monly affected than the femur. The changes are most
commonly associated with fractures of the ipsilateral often bilateral and symmetrical and are unusual
proximal tibial metaphysis. above the age of 6 months (SHOPFNER 1966).
104 S.Chapman
7.8.2
Birth Trauma
7.8.3
Osteogenesis Imperfecta
King J, DiefendorfD,Apthorp J et al (1988) Analysis of 429 frac- Paterson CR, McAilion S, Miller R (1983) Osteogenesis imper-
tures in 189 battered children. J Pediatr Orthop 8:585-589 fecta with dominant inheritance and normal sclerae. J Bone
Kleinman PK (1998) Diagnostie imaging of child abuse, 2nd Joint Surg [Br] 65:35-39
edn. Mosby, St Louis, p 22 Paterson CR, McAllion S, Shaw J (1987) Clinieal and radiologi-
Kleinman PK, Kwon DS (1998) Differential diagnosis IV: cal features of osteogenesis imperfecta type IV A. Acta Pae-
normal variants. In: Kleinman PK (ed) Diagnostie imaging diatr Scand 76:548-552
of child abuse, 2nd edn. Mosby, St Louis, pp 225-228 Paterson CR, Burns J, McAllion SJ (1993) Osteogenesis imper-
Kleinman PK, Marks SC Jr (1995) Relationship of the sub- fecta: the distinction from child abuse and the recognition
periosteal bone collar to the metaphyseallesions in abused of a variant form. Am J Med Genet 45: 187 -192
infants. J Bone Joint Surg [Am] 77:1471-1476 Rex C, Kay PR (2000) Features of femoral fractures in nonac-
Kleinman PK, Marks SC Jr (1996) A regional approach to the cidental injury. J Pediatr Orthop 20:411-4l3
classie metaphyseallesion in abused infants: the proximal Riehman HA (2000) From a radiologist's judgement to public
tibia. AJR 166:421-426 policy on child abuse and neglect: what have we wrought?
Kleinman PK, Marks SC Jr (1998) A regional approach to Pediatr RadioI30:219-228
the classie metaphyseallesion in abused infants: the distal Rosenberg N, Bottenfield G (1982) Fracture in infants: a sign
femur. AJR 170:43-47 of child abuse.Ann Emerg Med 11:178-180
Kleinman PK, Marks SC, Blackbourne B (1986) The metaphy- Rosenthal L, Hili RO, Chuang S (1976) Observations of the
seallesion in abused infants: a radiologie-histopathologie use of 99ffiTc-phosphate imaging in peripheral bone trauma.
study. AJR 146:895-905 Radiology 119:637-641
Kleinman PK, Belanger PL, Karellas A et al (1991) Normal Ryan MG, Davis AA, Oates RK (1977) One hundred and
metaphyseal radiologie variants not to be confused with eighty-seven cases of child abuse and neglect. Med JAust
findings of infant abuse. AJR 156:781-783 2:623-628
Kogutt MS, Swischuk LE, Fagan CJ (1974) Patterns of injury Scher! SA, Miller L, Lively N et al (2000) Accidental and
and significance of uncommon fractures in the battered nonaccidental femur fractures in children. Clin Orthop
childsyndrome.AJR 121:143-149 376:96-105
Kowal-Vern A, Paxton TP, Ros SP et al (1992) Fractures Shopfner CE (1966) Periosteal bone growth in normal infants.
in the under-3-year-old age cohort. Clin Pediatr (Phila) A preliminary report. AJR 97:154-163
31:653-659 Sillence DO, Senn A, Danks DM (1979) Genetic heterogeneity
Krugman RD (1997) Child protection poliey. In: Helfer ME, in osteogenesis imperfecta. J Med Genet 16: 101-116
Kempe RS, Krugman RD (eds) The battered child. Univer- Smith FW, Gilday DL, Ash JM et al (1980) Unsuspected costo-
sity of Chieago Press, Chieago vertebral fractures demonstrated by bone scanning in the
Landin LA (1983) Fracture patterns in children. Analysis of child abuse syndrome. Pediatr Radioll 0: 103-1 06
8,682 fractures with special reference to incidence, etiol- Smith R (1995) Osteogenesis imperfecta, non-accidental injury
ogy, and secular changes in a Swedish urban population and temporary brittle bone disease. Arch Dis Child
1950-1979. Acta Orthop Scand SuppI202:1-109 72:169-176
Lauer B, Broeck ET, Grossman M (1974) Battered child syn- Sty JR, Starshak RJ (1983) The role of bone scintigraphy in
drome: review of l30 patients with controls. Pediatries the evaluation of the suspected abused child. Radiology
54:67-70 146:369-375
Leventhal JM, Thomas SA, Rosenfield NS, et al. (1993) Frac- Taitz LS (1987) Child abuse and osteogenesis imperfecta. Br
tures in young children. Distinguishing child abuse from Med J 295:1082-1083
unintentional injuries. Am J Dis Child 147:87-92 Tardieu A (1860) Etude medieo-legale sur les seviee et mauvais
Loder RT, Bookout C (1991) Fracture patterns in battered chil- traitments exerces sur des enfants. Ann Hyg Publ Med Leg
dren. J Orthop Trauma 5:428-433 l3:361-398
Matthew MO, Ramamohan N, Bennet C (1998) Importance of Thomas SA, Rosenfield NS, Leventhal JM et al (1991) Long-
bruising associated with paediatrie fractures: prospective bone fractures in young children: distinguishing accidental
observational study. Br Med J 317:1117-1118 injuries from child abuse. Pediatries 88:47l-476
Merten DF, Radkowski MA, Leonidas JC (1983) The abused Wall (1995) Re AB (Child abuse: expert witnesses). FLR
child. A radiologie al reappraisal. Radiology 146:377-381 1:181-200
O'Connor JF, Cohen J (1998) Dating fractures. In: Kleinman PK Wang C-T (1997) Current trends in child abuse reporting and
(ed) Diagnostie imaging of child abuse, 2nd edn. Mosby, St fatalities: the results of the Annual Fifty State Survey. Center
Louis, pp 168-177 for Child Abuse Prevention Research, Prevent Child Abuse
Ogden JA (1990) Injury to the immature skeleton. In: Toulouki- Ameriea, Chieago
an RJ (ed) Pediatrie trauma. Mosby Year Book, St Louis, Wardinsky TD (1995) Genetic and congenital defect condi-
pp 399-439 tions that mimie child abuse. J Farn Pract 41:377-383
Ojima K, Matsumoto H, Hayase T et al (1994) An autopsy Wardinsky TD, Vizcarrondo FE, Cruz BK (1995) The mistaken
case of osteogenesis imperfecta initially suspected as child diagnosis of child abuse: a three-year USAF medieal centre
abuse. Forensie Sci Int 65:97-104 analysis and literature review. Milit Med 160:15-20
O'Neill JA Jr, Meacham WF, Griffin PP et al (1973) Patterns Wellington P, Bennet GC (1987) Fractures of the femur in
of injury in the battered child syndrome. J Trauma childhood. Injury 18:103-104
l3:332-339 Wor!ock P, Stower M, Barbor P (1986) Patterns of fractures
Paters on CR, McAilion SJ (1987) Child abuse and osteogenesis in accidental and non-accidental injury in children: a com-
imperfecta. Br Med J 295:1561 parative study. Br Med J 293: 100-102
8 Bone Trauma
PRUDENCIA N. M. TYRRELL and VICTOR N. CASSAR-PULLICINO
CONTENTS 8.1
Introduction
8.1 Introduction 109
8.2 Indications for Radiography in
Suspected Knee Injury 110 Trauma to the knee is a common injury and the
8.3 Extra-articular Fractures 11 0 mainstay of its initial imaging assessment is still
8.3.1 Supracondylar Fractures 110 the conventional radio graph. Multiple ligamentous
8.3.2 Distal Femoral Physeal Fractures 110 structures contribute to stability of the knee, and
8.3.3 Fractures of the Proximal Fibula 112
8.3.4 Proximal Tibial Physeal Fractures 112
demonstration of a fracture should alert the radiolo-
8.3.5 Fractures of the Patella 113 gist to the possibility of concomitant injury to these
8.4 Intra-articular Fractures 115 soft tissue structures. This chapter deals specifically
8.4.1 Femoral Condylar Fractures 115 with bone trauma, and that which can usually be eval-
8.4.2 Fractures of the Tibial Plateau 116 uated on conventional radiographs alone, although
8.4.3 Fractures of the Tibial Spine and
Intercondylar Eminence 117
when indicated computed tomography (CT) and
8.4.4 Avulsion of the Tibial Tuberosity 118 magnetic resonance (MR) imaging may be required
8.4.5 Osgood-Schlatter Lesion 119 to determine the full extent of the injury.
8.4.6 Sinding-Larsen-Johansson Disease 121 Some types of bone injury are specificaily associat-
8.4.7 Marginal Avulsion Fractures 121 ed with certain age groups and knowledge of this is
8.4.8 Osteochondral Fractures 122
8.5 Dislocation 122
helpful, particularlywhen the conventional radiographs
8.5.1 Knee 122 appear normal. During growth the abundant cartilagi-
8.5.2 Patellar Dislocation 124 nous structures (physes and apophyses) are the weakest
8.5.3 Proximal Tibiofibular Joint 125 links in the bone-joint-tendon-ligament complex, and
8.6 Fatigue Fractures 125 fractures in the vicinity of the knee have peculiar char-
8.6.1 Stress Fractures 125
8.6.2 Insufficiency Fractures 125
acteristics in this age group. In the older person, partic-
8.7 Conclusion 126 ularlya female with osteoporosis, fractures of the lateral
References 126 tibial plateau may be easily overlooked. These can be
highlighted by induding internal and external oblique
views when the injury is suspected.
There are important neurovascular structures dose-
ly related to the distal femur, primarily posteriorly and
laterally. The possibility of associated damage to these
structures is paramount when evaluating bone trauma
to the knee joint at anyage (e.g. knee dislocation in the
adult and physeal fractures in the child with epiphyseal
separation). Concomitant bony injuries, particularly at
the hip and ankle joints, should also be considered in
association with fractures of the knee.
P.N.M. TYRRELL, MD This chapter is designed to allow the reader to
Consultant Radiologist, Department of Diagnostic Imaging, quickly refer to the pertinent information about a
The Robert Jones and Agnes Hunt Orthopaedic and District specific injury. It will address the variety of injuries
Hospital, Oswestry, Shropshire, SYlO 7AG, UK that can occur about the knee in both the child
v'N. CASSAR-PULLQQICINO, MD
Consultant Radiologist, Department of Diagnostic Imaging,
and the adult with reference to the mechanism of
The Robert Jones and Agnes Hunt Orthopaedic and District injury, patterns of fracture and dassification, with
Hospital, Oswestry, Shropshire, SY10 7AG, UK brief mention of related management issues.
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
llO P. N. M. Tyrell and V. N. Cassar-Pullicino
11 111
IV v
affects the medial condyle. This is the most growth arrest. SH II injuries have a high association
common undisplaced spontaneously reduced with subsequent growth disturbance. This retarda-
epiphyse al injury at the knee. It may be difficult tion of growth usually occurs on the side opposite
to appreciate on initial examination. A tear of the to the metaphyseal fracture since this is the point
anterior cruciate ligament (ACL) can be associated of maximal injury to and separation of the physis.
with this type of fracture. The radiographic signs of premature closure usually
- SH type IV consists of a vertical fracture line become evident within 6 months after injury.
extending down through the metaphysis, across
the physis and through the epiphysis to the articu-
lar surface. The fracture usually involves the lateral 8.3.3
metaphysis extending into the intercondylar notch Fractures of the Proximal Fibula
with separation of the lateral femoral condyle.
- SH type V is physeal impaction and has a high Fractures of the head and neck of the fibula rarely
association with consequent growth retardation. occur in isolation and are usually associated with
These are often found in association with frac- fracture of the lateral tibial condyle, ligamentous
tures of the proximal tibia. injuries of the knee or fracture of the ankle.
Fractures of the head of the fibula are of three
Avulsion injuries from the edge of the physis types:
including a portion of the perichondrium and 1. Direct impaction
attached bone may occur when there is avulsion of a 2. Valgus injuries, usually associated with fracture of
collateralligament. the lateral tibial condyle and medial collateralliga-
Separate from the Salter Harris classification of ment injury
physeal injury, injuries can be described relative to 3. Varus injuries in which there is avulsion of the sty-
the type of displacement, and this in turn is asso- loid process of the fibula at the site of insertion of
ciated with the mechanism of injury. Anterior dis- the lateral collateralligament and biceps femoris
placement of the epiphysis is due to a hyperextension tendon. These can be associated with injury to the
injury to the knee. This is associated with a risk of common peroneal nerve. There mayaiso be signif-
neurovascular damage posteriorly. Medial or lateral icant damage to the lateral capsule and ligamen-
displacement also can occur in association with SH tous structures (Fig. 8.3).
type II injury.
In children, because the physis is radiolucent, When injuries of the head and neck of the fibula
appreciation of injury is determined by observation occur a little more distally, they are often associated
of widening, displacement or adjacent bony disrup- with external rotational injury at the ankle.
tion (OZONOFF 1979). Standard anteroposterior (AP)
and lateral radiographs may overlook injury and
oblique views may be required. In an infant, only the 8.3.4
centre of the distal femoral epiphysis is ossified. This Proximal Tibial Physeal Fractures
should be in line with the long axis of the femur on
both AP and lateral projections. Comparative views The Salter Harris classification applies to this area
with the opposite knee in cases of doubt may be help- (SALTER and HARRIS 1963). In the radiographic eval-
ful. The radiolucent line representing the physis on uation of this type of fracture it is paramount to
the AP projection usually measures 3-5 mm until remember that even minimally displaced fractures
adolescence. Stress views mayaiso be necessary, and may have been severely displaced before recoiling to
can help differentiate between ligamentous and phy- a nearly reduced location as seen on the radiographs.
seal injury (ROGERS 1992). Most physeal injuries here are types I and II. Sixty-
In general terms, SH I and II injuries can be six percent of SH type II injuries are displaced. The
reduced by closed me ans. This can fail, however, often metaphyseal fracture is usually medial and the sepa-
owing to a sleeve of periosteum interposed within the ration lateral, resulting in a valgus deformity. In SH
physis, in which case open reduction becomes nec- type III injury, the fracture usually involves the later-
essary (SPONSELLER and BEATY 1996). SH III and al epiphysis and the medial collateralligament (MeL)
IV almost always require open reduction and inter- is torn. In SH type IV injuries, the medial or lateral
nal fixation to minimise disruption of the articular tibial condyle can be affected. This type of injury
surface and to decrease the likelihood of premature is particularly associated with avulsion of the ACL
Bone Trauma 113
..
Fig.8.3. a AP radiograph demonstrating an avulsion fracture
of the styloid process of the fibula. b Varus stress radiograph
highlights the fractured fragment from the styloid process of
the fibula and demonstrates marked widening of the lateral
compartment consistent with capsular and ligamentous dis-
c
ruption. c Coronal STIR sequence confirms the lateral capsular
disruption. Note also the displaced lateral meniscus
114 P. N. M. Tyrell and V. N. Cassar-Pullicino
Fractures of the patella can be associated with with the verticallimb extending into the intercondy-
fractures of the femoral shaft, condyles, proximal lar notch. The fractures are usually associated with
tibia and posterior dislocation of the hip, as in a dash- high impaction and may be comminuted and dis-
board injury. placed with resultant loss of congruity of joint
Small ossicles of bone adjacent to the distal pole, surfaces. Fractures confined to one condyle are usu-
if scale-like and closely apposed to the antero-infe- ally obliquely oriented. The mechanism of injury is
rior surface of the patella, may represent an accesso- thought to be axial loading with a varus or valgus
ry ossification centre. Fragmentation or elongation of force. There is usually significant associated soft tissue
the distal pole associated with patella alta in a child injury due to disruption of ligamentous attachments.
with cerebral palsy indicates long-standing extensor
mechanism stress. Abipartite patella, differentiated
from a fracture by the characteristic location of the
cleft and the smooth sclerotic margins, is best seen
on the AP view (Fig. 8.8).
Closed management is usually recommended for
non-displaced fractures. Operative treatment, howev-
er, is indicated for fractures with more than 3 mm of
articular displacement or for fractures that disrupt the
extensor mechanism (SPONSELLER and BEATY 1996).
8.4
Intra-articular Fractures
8.4.1
Femoral Condylar Fractures
These fractures can often be visualised on AP - Type I is a split fracture of the lateral tibial condyle.
and lateral projections but oblique views (in varying This is usually a vertical fracture of the condyle
degrees of internal and external rotation) are some- with no depression and tends to occur in young
times required. These can demonstrate the degree of adults.
separation of fracture fragments, which is important - Type 11 is a split fracture of the lateral tibial con-
in the planning of treatment. CT is also particularly dyle with depression. It occurs as a result of a lat-
good in this area, since the relationship of the frac- eral bending force with axialloading and there is
ture line to the intercondylar notch can be identified usuallyan associated tear of the MCL.
and related to the cruciate ligament attachment sites. - Type III is a depressed fracture of the lateral tibial
In this type of fracture, because of anticipated exten- plateau.
sive soft tissue injury, MR imaging is often valuable - Type IV is a fracture of the medial tibial plateau.
to confirm the ligaments involved and to facilitate the There is a high association with soft tissue injury
management of the injury. Non-operative treatment as a result of high impact, related to varus and
is confined to patients with incomplete or non-dis- axialloading.
placed fractures. When the fracture is displaced, then - Type V is a bicondylar fracture with varying
open reduction and internal fixation to achieve joint degrees of articular depression and displacement
congruity and stabilisation are indicated. (these fractures may be T- or Y-shaped).
- Type VI is a bicondylar fracture with metaphyseal!
diaphyseal dissociation, and usually results from a
8.4.2 high impact injury.
Fractures of the Tibial Plateau
A vertical fracture through the posterior margin
Fractures of the tibial plateau occur due to a com- of the condylar surface mayaiso occur.
bination of valgus and compression forces generated In the conventional radiographic evaluation of the
by impaction of the femoral condyles against the normal tibia, the condylar region slopes downwards
tibial plateau. The lateral tibial condyle and plateau by approximately 15° posteriorly, and therefore in
are weaker than the medial plateau, and fractures are the standard AP projection the tibial plateau is not
thus more likely to occur here. Also the femoral con- seen tangentially. Condylar/plateau fractures, par-
dyles are stronger than the tibial plateau, and thus ticularly depressed fractures involving the anterior
with impaction injury the plateau is more vulnerable. plateau margin, can be overlooked on account of
Varus stresses and adduction of the tibia occur much this. Fractures of the lateral tibial plateau are most
less frequently than valgus stress; hence the greater common in the elderly. They may occur from
propensity for injury to involve the lateral condyle. pedestrian/automobile accidents, but are more usu-
A number of classifications of fractures of the ally due to twisting falls. In undisplaced fractures,
tibial plateau have been used. That given here is the the fracture line may be obliquely oriented and not
one proposed by SCHATZKER et al. (1979) (Fig. 8.9). well seen on standard AP and lateral projections. An
11 111
IV V VI
AP projection with 15 0 caudad angulation produces posterolateral aspect of the tibial condylar region.
a tangential view of the tibial plateau and facilitates Some cruciate ligament injuries can be associated with
observation of depressed fractures. The depression pure hyperextension and this mechanism of injury is
may be gentle or marked, with bone fragments being associated with bruising ofboth the medial and the lat-
driven deep into the condyle (often associated with a eral condyle. Ligamentous injury occurs in 10%-12%
vertical split fracture) (Fig. 8.lOa, b). Minimally dis- of plateau fractures and is more likely due to the valgus
placed fractures of the lateral tibial condyle are often mechanism of injury rather than axial loading. The
seen only in the oblique view. ACL and MCL are most likely to be injured. The MCL
Conventional tomography was previously used in injury most often occurs in association with a split or
the assessment of tibial condylar and plateau fractures, split-depression type of fracture of the lateral tibial
but has been superseded by CT. CT can be carried out plateau. Where an injury of the MCL occurs in associa-
through pIaster casts, sagittal and coronal (Fig.8.lDc) tion with valgus stress, there may be an avulsion of the
reconstructions can be obtained, and accurate mea- MCL as opposed to a tear. This avulsion can be associ-
surements of the degree of depression of the plateau ated with a flake fracture from the medial aspect of the
and the degree of separation of fragments can be medial femoral condyle.
achieved. CT does not replace conventional radiogra-
phy in the initial assessment of tibial plateau fractures,
but can act as an adjunct, together with 3D reconstruc- 8.4.3
tion, in improving the accuracy of the classification Fractures of the Tibial Spine and
of the fracture (WICKY et al. 2000). This assists the Intercondylar Eminence
surgeon and results in improved management of the
injury. The management of tibial plateau fractures is Fractures of the tibial spine and intercondylar emi-
controversial and beyond the scope of this review. Ulti- nence are more common in children than in adults.
mately, the aim of treatment, as with fractures on the The injury is most commonly seen in conjunction
femoral side of the joint, is to achieve, as far as is pos- with high-energy plateau fractures and is frequently
sible, a stable knee which is mobile, painless and satis- avulsion in nature, typically occurring at the site
factorily aligned, with a minimal risk of post-traumat- of origin of the ACL. In older children and adults,
ic osteoarthritis (SPONSELLER and BEATY 1996). the ACL is usually tom within its substance without
When there has been marked soft tissue injury to an associated avulsion fracture. The mechanism of
the knee, such as with an ACL tear, there is often asso- injury is thought to be hyperextension combined
ciated bone bruising/microtrabecular injury which is with strong rotational forces. There is an incidence of
only detectable on the subsequent MR imaging study associated ligamentous injuries and this is greater in
(Fig. 8.11). Such injury not infrequently occurs in the adults than children.
a b
The classification of these fractures as proposed Avulsion of the posterior tibial eminen ce can
by MEYERS and McKEEVER (1959) is related to the occur, at the tibial insertion site of the posterior cru-
degree of displacement of the eminence (Fig. 8.12). ciate ligament (PCL) (TORISU 1977; Ross and CHES-
- Type I: A horizontal fracture line at the base of the TERMAN 1986). This is, however, rare.
anterior portion of the tibial spine. Elevation of For non-or minimally displaced fractures (types I
only the anterior edge of the fracture. and II), management is usually non-operative. Open
- Type II: The eminence is elevated, angulated and reduction and internal fixation is required for type
hinged posteriorly. III displaced fractures.
- Type IIIA: The entire eminen ce has been avulsed
from the parent bone.
- Type IIIB: The entire eminence has been avulsed, 8.4.4
and is elevated and rotated. Avulsion of the Tibial Tuberosity
Fractures may be viewed on standard AP or lateral Developmentally the tibial tubercle passes through
views (Figs.8.13, 8.14), but a notchltunnel projec- four stages (SPONSELLER and BEATY 1996):
tion gives a better view of the intercondylar region 1. The cartilaginous stage, in which no ossification
and oblique views may be helpful. In children, the centre in the cartilaginous anlage of the tibial
avulsed fragment may be mainly unossified cartilage tubercle is present.
with only a very small thin ossified fragment. The 2. The apophyseal stage, in which the ossification
margins of the bones should be scrutinised for fine centre for the tubercle appears (occurs at 8-12
avulsion fractures and oblique views may be helpful years in females and 9-14 years in males).
in addition. MR imaging will allow evaluation of 3. The epiphyseal stage, in which the secondary ossi-
associated soft tissue injuries. fication centres of the tubercle and the proximal
tibial epiphysis coalesce to form a tongue of bone
continuous between the tubercle and proximal
11 lilA
tibial epiphysis.
4. The final bony stage, in which the epiphysealline
is closed between the fully ossified tuberosity and
the tibial metaphysis.
the tubercle, the surrounding perichondrium and ele extends into the proximal tibial epiphysis and
the adjacent periosteum. Avulsion can be associated the knee joint.
with sudden acceleration or deceleration of the knee These fractures can be further classified according
extensor mechanism. to the presence or absence of displacement.
WATSON-JONES (l955) described three types of
avulsion fracture of the tibial tuberosity. OGDEN and On a standard AP radiograph, the tubercle lies just
co-workers (l980) later refined this, describing three lateral to the midline of the tibia and therefore the
types depending on the distance of the fracture from best profile is obtained with the tibia slightly inter-
the distal tip of the tubercle (Fig. 8.15). nally rotated (Fig. 8.16).
- Type I: The separation through the distal portion - Type I fractures are usually treated with closed
of the physis under the tubercle breaks up proxi- reduction. Open reduction and internal fixation is
maHy through the secondary ossification centre of required for type II and III injuries.
the tubercle. Acute avulsion of the tibial tubercle differs from the
- Type II: The separation extends anteriorly through Osgood-Schlatter lesion (OSGOOD 1903; SCHLAT-
the area bridging the ossification centres of the TER 1903), which is a more chronic lesion but is feit
tibial tubercle and the proximal tibial epiphysis. to represent one end of the spectrum of avulsive
- Type III: A fracture/separation of the entire tuber- phenomena at this site.
120 P. N. M. Tyrell and V. N. Cassar-Pullicino
Fig. 8.14. a AP projection demonstrating a fracture of the intercondylar eminence with extension into the lateral tibial condyle.
b The externaioblique projection better demonstrates the fractured eminence. c The internal oblique projection highlights
extension of the fracture into the lateral tibial condyle
11 111
Q
t
t Fig. 8.15. Line diagram of avulsion fractures of the tibial tuber-
osity (types I-III). See text for details
8.4.5 8.4.6
Osgood-Schlatter Lesion Sinding-Larsen-Johansson Disease
The Osgood-Schlatter lesion occurs when the tibial SINDING-LARSEN (1921) and ]OHANSSON (1922)
tubercle is in the apophyseal stage and when the sec- independently described a condition seen most
ondary ossification centre has appeared. The carti- commonly in adolescents that consists in pain, ten-
lage overlying the ossification centre anteriorly and derness and soft tissue swelling over the lower pole
posteriorly can resist tension forces better than the of the patella associated with bony fragmentation.
bone of the ossification centre. Usually no single The lesion is likely due to a traction phenomenon
acute traumatic event is responsible for the devel- in which repeated minor trauma at the proximal
opment of the Osgood-Schlatter lesion; rather repeat- attachment of the patellar tendon, initially com-
ed normal stresses or overuse (chronic tug) produce mencing as inßammation, is followed by calcifica-
minor disruption of the bony apophysis (OGDEN and tion or ossification (MEDLAR and LYME 1978), or
SOUTHWICK 1976). On radiographs, best taken with in wh ich patellar fracture or avulsion produces one
a soft tissue technique, the edges of the distal patel- or more distinct ossification sites. Similar findings
lar ligament are blurred and there may be tiny ßake are seen in athletes with «jumper's knee" (HECKMAN
ossific fragments either anterior or superior to the and ALKIRE 1984).
main ossification centre; these fragments, however, The radiograph demonstrates small bony frag-
can be difficult to differentiate from normal mul- ments adjacent to the distal surface of the patella with
ticentric ossification of the tubercle. Later in the overlying soft tissue swelling. The bone fragments
course, the tiny displaced fragments may enlarge and may eventually coalesce and become incorporated
may fail to unite and remain as separate ossicles into the patella. Ultrasound (DE FLAVIIS et al. 1989)
detached from the mature tubercle. The smooth scle- and MR imaging may help to confirm the diagnosis.
rotic margin of these ossicles indicates that they The pathogenesis of this dis order is similar to that
are long-standing in nature, and not acute. On CT, of Osgood-Schlatter disease, and the two conditions
increased width and decreased attenuation of the have been described in association (TRAVERSO et al.
tendon is observed, and on MR imaging there is 1990).
enlargement and increased signal intensity of the
tendon. Deep and superficial infrapatellar bursae are
often noted. Bone abnormalities are seen less fre- 8.4.7
quently. On MR imaging an ossicle may be seen, and Marginal Avulsion Fractures
on occasion a matching defect can be seen in the
tibial tubercle, suggesting that a bone fragment has Avulsion fractures occur at characteristic sites relat-
detached from the tuberosity. Sometimes, this ossicle ed to ligamentous attachments and, because of these
may fuse. On MR imaging, decreased signal intensity attachments, they usually do not migrate. There is
has been reported in the marrow of the tibial tuberos- usually an associated cortical irregularity of the
ity and tibial epiphysis on Tl-weighted images, with parent bone although this may be very subtle, and not
increased signal intensity on T2-weighted images, always detectable on plain radiographs.
implying the presence of marrow oedema (ROSEN- The most common avulsion is from the anterior
BERG et al. 1992). These appearances contrast with tibial eminence, at the site of attachment of the ACL
those of acute fracture-separation of the tubercle (LILEY and BAxTER 1990) (Fig.8.17). Very rarely,
physis, caused by violent contraction of the quadri- there may be an avulsion from the posterior aspect
ceps mechanism, which tends to occur during the of the intercondylar eminen ce at the site of attach-
epiphyse al stage of development of the tubercle, when ment of the PCL (TORISU 1977; Ross and CHESTER-
the tubercle ossification centre has coalesced with MAN 1986). A cortical fragment may be avulsed from
that of the proximal tibial epiphysis .. The manage- the femoral attachment site of the MCL. If there is
ment of Osgood-Schlatter lesion is symptomatic and injury to the MCL at this site, without cortical avul-
supportive. sion, a linear focus of calcification may later develop.
This may be referred to as a Pellegrini-Stieda lesion
(Fig. 8.18). The lateral collateralligament and biceps
tendon may avulse from their insertion on the head
of the fibula, resulting in avulsion of the styloid pro-
cess.
122 P. N. M. Tyrell and V. N. Cassar-Pullicino
a b
8.4.8
Osteochondral Fractures
Fig.8.19. a AP radiograph demonstrating a subtle fracture Hne through the lateral aspect of the lateral tibial condyle (arrow).
There is also an osteochondral defect of the lateral femoral condyle. b Lateral projection again shows the lateral femoral condyle
impaction (osteochondral) injury. c Axial CT image confirms the fracture through the lateral aspect of the lateral tibial condyle
and d also shows a fracture fragment of the posterior aspect of the medial femoral condylar margin. e Sagittal MR Tl-weighted
image reveals the impaction (osteochondral) injury of the lateral femoral condyle. f Coronal MR Tl-weighted image reveals
the impaction injury of the lateral femoral condyle and also injury to the lateral aspect of the lateral tibial condyle. g Sagittal
Tl-weighted MR image also reveals a disrupted ACL in its posterior aspect. h Coronal T2-weighted (gradient echo) image
beautifully demonstrates the Segond fracture at the lateral margin of the lateral tibial condyle. The lateral meniscus is disrupted.
The lateral aspect of the intercondylar notch is empty, this being indicative of a concomitant injury to the ACL. (The lateral
femoral condylar impaction injury is consistent with this)
124 P. N. M. Tyrell and V. N. Cassar-Pullicino
Manco LG, Schneider R, Pavloc H (1983) Insufficiency frac- Schlatter C (1903) Verletzurgen des Schnabelformigen Fortsatz-
tures ofthe tibial plateau.AJR 140:1211-1215 es der Oberen Tibiaepiphyse. Beitr Klin Chir 38:874-887
Medlar RC, Lyme ED (1978) Sinding-Larsen-Johansson dis- Seinsheimer F III (1980) Fractures of the distal femur. Clin
ease. Its etiology and natural history. JBJS 60A:I113-1116 Orthop 153:169-179
Meyers MH, McKeever FM (1959) Fracture of the intercondylar Sinding-Larsen MF (1921) A hitherto unknown affection of
eminence of the tibia. J Bone Joint Surg 41A:209-222 the patella in children. Acta Radioli: 171
Montgomery JB (1987) Dislocation of the knee. Orthop Clin Spritzer CE, Courneya DL, Burk DL Jr et al (1997) Medial reti-
North Am 18:149-156 nacular complex injury in acute patellar dislocation: MR
Müller ME, Allgower M, Schneider R, Willenegger H (1979) findings and surgical implications. AJR 168: 117-122
Manual of internal fixation. Springer, Berlin Heidelberg Sponseller PD, Beaty JH (1996) Fractures and dislocations about
NewYork the knee. In: Rockwood CA Jr, Wilkins KE, Beaty JH (eds)
Neer CS, Grantham SA, Shelton ML (1967) Supracondylar frac- Fractures in children,4th edn. Lippincott-Raven, Philadelphia
ture ofthe adult femur. J Bone Joint Surg 49A:591-613 Stewart MJ,Sisk TD, Wallace SL Jr (1966) Fractures ofthe distal
Ogden JA,Southwick WO (1976) Osgood-Schlatter's disease and third of the femur. J Bone Joint Surg 48A:784-807
tibial tuberosity development. Clin Orthop 116:180-189 Stiell IG, Greenberg GH, Wells GA et al (1995) Derivation of
Ogden JA, Tross RB, Murphy MJ (1980) Fractures of adecision rule for the use of radiography in acute knee
the tibial tuberosity in adolescents. J Bone Joint Surg injuries. Ann Emerg Med 26:405-413
62A:205-215 Stiell IG, Greenberg GH, Wells GA et al (1996) Prospective vali-
Osgood RB (1903) Lesions of the tibial tubercle occurring dation of adecision rule for the use of radiography in acute
during adolescence. Boston Med Surg J 148: 114-117 knee injuries. JAMA 275:611-615
Ozonoff MB (1979) Pediatric orthopaedic radiology. Saunders, Stiell IG, Wells GA, Hoag RH et al (1997) Implementation of
Philadelphia the Ottawa knee rule for the use of radiography in acute
Rogers LF (1992) The knee and shafts of the tibia and fibula. knee injuries. JAMA 278:2075-2079
In: The radiology of skeletal trauma. Churchill Livingstone, Tigges S,Pitts S,Mukundan S Jr et al (1999) External validation
NewYork of the Ottawa knee rules in an urban trauma center in the
Rorabeck CH, Bobechko WP (1976) Acute dislocation of the United States. AJR 172: 1069-1 071
patella with osteochondral fracture: a review of 18 cases. J Torisu T (1977) Isolated avulsion fracture of the tibial attach-
Bone Joint Surg 58B:237-240 ment of the posterior cruciate ligament. J Bone Joint Surg
Rosenberg ZS, Kauelblum M, Cheung YY et al (1992) Osgood- 59 A:68-72
Schlatter lesion: fracture or tendinitis? Scintigraphic, CT Traverso A, Baldari A, Catalari F (1990) The coexistence of
and MR imaging features. Radiology 185:853-858 Osgood-Schlatter's disease with Sinding-Larsen-Johans-
Ross AC, Chesterman PJ (1986) Isolated avulsion of the tibial sons disease. Case report in an adolescent soccer player. J
attachment of the posterior cruciate ligament in childhood. Sports Med Phys Fitness 30:331-333
J Bone Joint Surg 68B:747 Virolainen H, Visuri T, Kuusela T (1993) Acute dislocation of
Salter RB, Harris WR (1963) Injuries involving the epiphyseal the patella: MR findings. Radiology 189:243-246
plate. J Bone Joint Surg 45A:587-622 Watson-Jones R (1955) Fractures and joint injuries, 4th edn.
Schatzker J, Tile M (1987) The rationale of operative fracture Livingstone, Edinburgh
care. Springer, Berlin Heidelberg New York Wicky S, Blaser PF, Blanc CH et al (2000) Comparison between
Schatzker J, McBroom R, Bruce D (1979) The tibial plateau standard radiography and spiral CT with 3D reconstruc-
fracture: the Toronto experience. 1968-1975. Clin Orthop tion in the evaluation, classification and management of
138:94-104 tibial plateau fractures. Eur RadioI1O:1227-1232
9 The Menisci
JAVIER BELTRAN and STEVEN SHANKMAN
CONTENTS 9.2
Technical Considerations
9.1 Introduction 129
9.2 Technical Considerations 129
9.2.1 Pulse Sequences 129 MR imaging of the knee is performed using a trans-
9.2.2 Artifacts 131 mit-receive general-purpose extremity surface coil
9.2.3 Post Processing 132 manufactured by most companies. Quadrature and
9.2.4 Field Strength 133 phase array coils are also available. Increased spatial
9.3 Normal Anatomy 133
9.3.1 Normal MR Imaging Appearance:
resolution and decreased S/N ratio are significant
Variations and Pitfalls 134 advantages of these coils. The knee is placed in full
9.4 Meniscal Trauma 139 extension in a neutral position. In general, small field
9.4.1 Significance of Signal Alterations 140 of view (FOV) of 14-16 cm and 3-4 mm slice thick-
9.4.2 Terminology and Signs of a Tear 142 ness is recommended.
9.5 Meniscal Cysts, Ossicles, and Discoid Meniscus 147
9.6 Postoperative Meniscus 147
MR images are obtained in the axial, sagittal, and
9.7 Accuracy 149 coronal planes. The sagittal images are frequently ori-
References 149 ented following the longitudinal axis of the lateral
femoral condyle in order to obtain good visualization
of the anterior cruciate ligament. The images should
cover a field of view extending from the suprapatellar
region, including the distal portion of the quadriceps
9.1 tendon, to the proximal tibia. An extended field of view
Introduction may be necessary to assess patients with suspected
proximal quadriceps tendon tears or tears involving
MR imaging has revolutionized our ability to pic- the distal insertion of the medial collateralligament.
ture the soft tissue structures of the musculoskeletal Radial imaging, with multiple oblique planes ori-
system. Increased soft tissue contrast coupled with ented in a radial fashion centered over each side of the
multiplanar slice capability allows us to visualize tibial plateau, with extra sections oriented along the
the muscles, tendons, ligaments, cartilage, and bone axis of the anterior cruciate ligament, have been pro-
marrow in a way that is unprecedented. Although the posed as an alternative to orthogonal plane imaging.
knee is a common site for all dis orders occurring This technique can also be used with 2DFT gradient
in and about the joints, the large majority of cases echo pulse sequences and provides images similar
requiring MR imaging are traumatic in nature. The to those obtained with conventional arthrography.
following discussion focuses on MR imaging tech- Although this technique was described a decade ago
niques, normal MR imaging anatomy of the menisci (QUINN et al. 1992), it has not gained popular accep-
and meniscal pathology. tance, probably because of the extra time it takes to
set up the planes of section and the distortion of the
anatomy visualized in multiple oblique sections.
J. BELTRAN, MD
Chairman and Clinical Professor of Radiology, Department of
Radiology, Maimonides Medical Center, 4802 Tenth Avenue, 9.2.1
Brooklyn, NY 11219, USA Pulse Sequences
S. SHANKMAN, MD
Vice Chairman and Pro gram Director, Department of Radiology,
Maimonides Medical Center, 4802 Tenth Avenue, Brooklyn, NY Different institutions use different pulse sequences
11219, USA for knee imaging, mostly based on conventional or
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
130 J. Beltran and S. Shankman
fast spin echo techniques (FSE). It is convenient to CHEUNG et al. (1997) evaluated double FSE sagittal
obtain a proton density-weighted and T2-weighted images for the detection of meniscal tears in a group
set of images in one plane (e.g., 2,000120-80 TR/TE), of 293 patients, and obtained a sensitivity and spec-
frequently the sagittal. Proton density images are ificity similar to those of conventional SE. In their
probably the most sensitive for detection of menis- imaging protocol they included TR of 4,000 ms, an
cal tears. The axial plane can be imaged using Tl- ETL of 8 and an effective TE "as short as possible", in
weighted pulse sequences (e.g., 500/20 TR/TE) or the range of 18 ms. They also concluded that FSE is
alternativelyT2* -weighted gradient echo techniques, an alternative to conventional SE for knee imaging.
which allow good visualization of the articular car- More recently, EUSTACE et al. (1999) compared
tilage of the patellofemoral joint (e.g., 400/9120,TR/ Tl-weighted conventional and turbo SE (TSE) or
TE/flip angle). The same T2* sequence can be used hybrid rapid relaxation enhancement (RARE) pulse
in the coronal plane. Coronal proton density fat-sup- sequences in patients with acute knee trauma (60
pressed high-resolution images provide high sensi- menisci). Their rationale was to take advantage of
tivity for bone marrow lesions in addition to menis- the fast imaging time that can be obtained with TSE
cal tears. (1 min 20 s), as compared with SE (4 min 10 s), in
The value of FSE pulse sequences versus conven- an acute trauma situation. Their results indicate that
tional spin echo (SE) techniques has been discussed both techniques are similarly accurate.
extensively in the radiological literature (RUBIN et Three-dimensional Fourier transform (3DFT)
al. 1994; ESCOBEDO et al. 1996; WHITE et al. 1996; volume acquisition techniques are popular owing to
CHEUNG et al. 1997; EUSTACE et al. 1999). RUBIN et software improvements and availability of worksta-
al. (1994) evaluated 129 menisci in 66 consecutive tions. These techniques are performed using gradi-
patients with suspected meniscal tears using conven- ent echo pulse sequences with short TR/TE (e.g.,
tional and FSE sequences. They concluded that FSE 18/9/30, TR/TElflip angle). The advantages of 3DFT
showed only 65% of meniscal tears shown on conven- data acquisition include improved signal-to-noise
tional SE images and therefore did not recommend ratio and reconstruction of true contiguous slices of
FSE for evaluation of meniscal tears. These authors about 1 mm in infinite planes of section, with mini-
attributed the blurring artifact of FSE to the low sen- mal distortion. When used in the axial plane, 3DFT
sitivity of this technique. Blurring artifact in FSE gradient echo sequences are capable of displaying the
sequences is related to the use of short effective TE, menisci in more than one consecutive image. This
and it is related to the attenuation of the T2 decay is helpful for certain types of meniscal tears such as
that occurs at the edges of the K-space when the later radial tears. These techniques became popular fol-
echo es of an echo train chain are used to collect the lowing the introduction of clinical MR applications
high-spatial-frequency data (MELK I et al. 1991). Tis- for musculoskeletal imaging. HAGGAR et al. (1988)
sues with short T2 and pulse sequences using short used 3DFT GRASS (gradient recalled acquisition in
effective TE, large ETL and small acquisition matrices the steady state) sequences for the evaluation of
will contribute to the blurring effect of FSE sequenc- meniscal tears in a group of 35 patients. They com-
es. Decreased signal intensity of the menisci sec- pared the accuracy of this technique with convention-
ondary to increased magnetization transfer that al SE imaging, using 3-mm partitions, and observed
occurs with FSE sequences mayaIso contribute to the equal detection of meniscallesions with both tech-
decreased accuracy of FSE sequences for the detec- niques.
tion of meniscal tears (LISTERUD et al. 1992). Similar Different forms of fat suppression are frequently
results were published by WHITE et al. (1996). These used for the evaluation of musculoskeletal lesions
authors compared the accuracy of T2-weighted FSE (MARTy-DELFAuT and BELTRAN 1999). In the knee,
and proton density SE imaging in the detection of the main value of fat suppression has been in the
meniscal tears in 152 menisci and found FSE to be study of bone contusions (LAL et al. 2000) and carti-
slightly less accurate that proton density SE. lage lesions (TRATTNIG et al. 1998), and for this pur-
On the other hand, ESCOBEDO et al. (1996) evaluat- pose, fat suppression is often added to the arma-
ed 80 menisci with conventional SE and FSE sequenc- mentarium of pulse sequences for general evaluation
es and obtained the same sensitivity and similar of knee disorders. For the assessment of meniscal
specificity for both sequences for the detection of lesions, fat suppression techniques are mostly used
meniscal tears. They concluded that proton density in combination with direct or indirect MR arthrog-
FSE imaging with a short ETL is a reasonable sub- raphy, to enhance the contrast between the injected
stitute for conventional SE imaging of the knee. Gd-DTPA and the adjacent pericapsular fat.
The Menisci 131
9.2.2
Artifacts
9.2.4
Field Strength
eral margin of the medial meniscus is more firmly eral aspect of the medial femoral condyle. The liga-
attached to the joint capsule and to the tibial plateau ments of Humphry and Wrisberg arise together, the
itself, the latter via the coronary ligament. As with the former coursing anterior to the posterior cruciate lig-
lateral side, the capsule does not extend anteriorly ament (PCL) and the latter posterior to the PCL.
into the joint near the posterior central attachment
site.
The transverse ligament connects the anterior 9.3.1
horns of both menisci. Its thickness varies from 1 to Normal MR Imaging Appearance:
4 mm. It arises at the most anterior-superior portion Variations and Pitfalls
of the lateral meniscus and crosses in front of the
tibial attachment of the anterior cruciate ligament, Sagittal images best show the anterior and posterior
merging with the superior portion of the posterior horns of the medial and lateral menisci (Fig.9.5),
aspect of the anterior horn of the medial meniscus. while coronal images best show the meniscal bodies
The meniscofemoralligaments are inconstant, 3- (Fig. 9.6). Thin-section axial images allow good visu-
to 4-mm fibrous bands arising from the posterior alization of the menisci, especially their free edges
horn of the lateral meniscus and attaching to the lat- (Fig.9.7).
QX!l
a '
_ -'1::._.... f
Sagittal sections show the posterior horn of the rior horn; it may appear as a round dot anterior to
medial meniscus as an isosceles triangle, the sides the meniscus and should not be confused with a tear
nearly twice as long as the base. The anterior horn of (Fig. 9.8) (HERMAN and BELTRAN 1988). The anterior
the medial meniscus is about one-half the width of horn of the lateral meniscus is in fact an uncommon
the posterior horn, appearing more as an equilateral site for a meniscal tear.
triangle. It may vary in its appearance, sometimes The anterior horn of the lateral meniscus mayaiso
showing an almost rounded configuration. It sits on demonstrate a striated or speckled increased signal
the extreme edge of the anterior tibia, the transverse intensity pattern on proton density images. This is
ligament joining it more superiorly and somewhat another normal variation, which may represent dense
posteriorly. This junction may create an "arrowhead" collagenous fibers of the anterior cruciate ligament
appearance, pointing posteriorly. The lower portion intertwined with the fibrocartilage of the meniscus
represents the anterior horn attachment site, and itself (Fig. 9.9). SHANKMAN et al. (1997) demonstrat-
the upper portion represents the transverse ligament ed this speckled pattern of the anterior horn of the
junction (Fig. 9.8). lateral meniscus in 22 MR imaging studies of the
On the lateral side, the transverse ligament attach- knee in patients in whom arthroscopic surgery per-
es to the most anterior-superior aspect of the ante- formed for other lesions showed no evidence of a tear
136 J. Beltran and S. Shankman
in this area. The anterior and posterior horns of the posed between the body of the medial meniscus and
lateral meniscus are of about equal size, the anterior the capsule.
being slightly smaller; both appear as isosceles trian- Posteriorly, coronal sections show the posterior
gles. horns as flat bands. On the lateral side, the popliteus
The posterior horn of the lateral meniscus differs tendon courses upward and laterally at 45° (Figs. 9.12,
from the medial in that the popliteus tendon is inter- 9.13). Synovium extends superior and inferiorly
posed between the periphery of the meniscus and its around the tendon through the opening in the cap-
attachment to the capsule. The resulting gap should sule and appears as increased signal intensity, linear
not be mistaken for a tear (HERMAN and BELTRAN
1988). The ligament ofWrisberg is usually seen at its
origin at the superior margin of the posterior horn.
It may appear as a round dot adjacent to the superior
aspect of the posterior horn, and it, too, should not
be confused with a tear (Fig. 9.10). VAHEY et al. (1990)
analyzed and correlated 109 MR knee examinations
with the arthroscopic findings. They found that the
meniscofemoral ligament produced the appearance
of a pseudotear of the posterior horn of the lateral
a
meniscus in 39% of cases.
The more peripher al sagittal images show the
bodies of the medial and lateral menisci, although
not optimally. On both sides, the menisci appear as
flat bands. On the lateral side, the more central slices
show a "bow tie" configuration owing to the sm aller
radius of the curvature. Volume averaging of the cap-
sule and menisci on extreme peripheral slices may be
confused with a tear of the meniscal body (Fig. 9.11)
b
(HERMAN and BELTRAN 1988). This normal increased
signal mayaiso represent truncation artifact. Paral-
lel signal lines are produced in the phase encoding
direction at edges where there is a large, abrupt tran-
sition in tissue signal intensity.
Coronal sections at the midportion of the knee
best show the bodies of both menisci. They appear
triangular in shape, the lateral slightly larger than
the medial. The capsular attachment on the medial
side is incorporated into the tibial or medial collat-
c
eral ligament. A small amount of fat may be inter-
b
Fig.9.14. Red-white zone. Sagittal Tl fat-saturated image
Fig. 9.13a, b. Normal popliteus tendon. Sagittal proton den- through the posterior horn of the medial meniscus dem-
sity images showing the popliteus tendon (arrows) coursing onstrating the peripheral portion of the meniscus, which is
through the posterior horn of the lateral meniscus. Also evi- hyperintense (red zone; straight arrow) in relation to the
dent is the transverse ligament anteriorly fibrous central portion (black zone; curved arrow)
The Menisci 139
confused with a meniscal tear. Another potential pitfall strength, and continuously improving pulse sequence
is due to the presence of a vacuum phenomenon that design. As research in the field of treatment of
mimics a meniscal tear (SHOGRY and POPE 1991). meniscal injuries progresses, MR imaging faces new
Normal anatomical meniscal variants are rare. The diagnostic challenges including accurate distinction
most frequent is the discoid meniscus, discussed between degenerative and non degenerative meniscal
below. Another variant is the so-called meniscal tears, postoperative evaluation of the meniscus, espe-
ftounce. Sagittal images obtained through the body cially in those patients treated with advanced tech-
of the lateral or medial menisci occasionally dem on- niques such as meniscal repair and meniscal trans-
strate a folded configuration or S-shaped fold, termed plant, assessment of the added information provided
the meniscal ftounce by Yu et al. (1997) (Fig. 9.15). bythe intra -articular or intravenous administration of
This fold occurs in the absence of a tear and it does Gd-DTPA,and the evaluation of associated injury,e.g.,
not increase the prevalence of meniscal tears. osteochondral, capsular, and ligamentous lesions.
Classically, meniscal tears are etiologically charac-
terized as traumatic or degenerative. Acute traumat-
ic tears are found in the young, athletic population.
Tears found in the older population generally occur
at sites of meniscal degeneration (SMILLIE 1970a,
b), and are often asymptomatic (BoDEN et al. 1992;
KORNICK et al. 1990). Treatment of acute traumatic
meniscal injuries is focused on preserving as much
meniscal tissue as possible, thus minimizing the risk
of development of premature osteoarthritis. Current
trends in the treatment of even large meniscal tears
involve primary meniscal repair (NEwMAN et al.
1993; HENNING et al. 1991; DANDY et al. 1990; GIL-
LQUIST and MESSNER 1993; JACKSON 1968; MCGINTY
et al. 1977). The accurate preoperative evaluation of
not only the presence of a meniscal tear but also of its
configuration becomes of paramount importance in
treatment planning. On the other hand, small, min-
imally symptomatic te ars may be treated conserva-
tively, since the long-term prognosis seems to be
Fig. 9.15. Meniscal flounce. Sagittal proton density image dem- highly dependent on the amount of residual meniscal
onstrating an undulating portion of the body of the meniscus tissue rather than its biomechanical integrity (NOBLE
(arrow)
and HAMBLEN 1975; FERRER-RoCA and VILALTA
1980).
Meniscal tears occur as a result of stress placed
through the substance of the meniscus, particularly
the posterior horns (SEEDHOLM 1979; SEEDHOLM
9.4 and HARGREAVES 1979). The most important stress
Meniscal Trauma factors contributing to the development of a menis-
cal tear include differences in the friction coefficient
Suspected meniscal te ars are one of the most fre- between the superior and inferior articular surfaces
quent indications for musculoskeletal MR imaging. of the meniscus, and the nature of the biomechanical
Most health care professionals dealing with patients forces acting in normal joint motion. The combina-
presenting clinically with knee pain in whom the tion of these factors leads to the development of tears
possibility of a meniscal tear is considered, agree that of the microstructure of the fibrocartilage, which
MR imaging provides a fast and accurate preopera- in turn causes molecular changes and chondrocyte
tive assessment of the knee. Technical developments depletion (NOBLE and HAMBLEN 1975; SEEDHOLM
in the last decade in both high- and low-field sys- 1979; SEEDHOLM and HARGREAVES 1979).
tems have significantly improved the ability of MR
to detect meniscal injuries. These improvements
include better surface coil design, increased gradient
140 J. Beltran and S. Shankman
-"--_--I d
The Menisci 141
intensity. These are the distinctive features of a tear, that when MR scans show grade I or 11 signal chang-
with very high correlation with arthroscopy. The cor- es, a meniscal tear is unlikely.
relation between the histological findings and signal In another study, DILLON et al. (1990) evaluated
changes was validated by a study of 12 knee spec- the clinical significance of grade 11 meniscal abnor-
imens from autopsies or above-knee amputations malities in aseries of 365 patients of whom 44 under-
published by STOLLER et al. (1987). went arthroscopic evaluation. These authors subdi-
CRUES et al. (1986, 1987) evaluated the relation- vided grade 11 meniscal abnormalities into grade
ship between signal changes in the menisci and the IIA (linear abnormal signal not contacting with the
presence of tears demonstrated arthroscopically in a meniscal surface), grade IIB (abnormal signal con-
series of 277 menisci in 144 knees, with arthroscopy tacting with the surface seen only in one image) and
or arthrotomy correlation. They found that 89% of grade IIC (extensive wedge-shaped signal abnormal-
menisci exhibiting grade I or 11 changes were normal ity not in contact with the surface, the "meniscus
at surgery, and that 94% of menisci with grade III MR within a meniscus" appearance). Arthroscopy dem-
signal changes had tears. They concluded that MR onstrated 3% te ars in group IIA, 0% in group IIB and
imaging can separate surgically significant from non- 50% in group IIC. They concluded that meniscal tears
significant meniscallesions, and that it is useful for should be diagnosed only if contact is seen in more
preoperative screening of suspected meniscal tears. that one image and that many group IIC lesions may
In a very early study, REICHER et al. (1986) evaluated have tears. DE SMET et al. (1993a) also found that if
the accuracy of MR imaging in a group of 49 patients, the signal abnormality within a meniscus is contact-
correlating with subsequent arthroscopy. In their ing one articular surface, a te ar is usually present, but
grading system they added a grade IV, representing if the extension into the articular surface is seen in
a gross distortion of the normal shape, truncation, only one section, a tear is less likely.
or a large focus or line of increased signal within DILLON et al. (1991) also performed a prospective
the meniscus. About 80% of meniscal abnormalities study of grade 11 meniscal abnormalities without
graded III or IV had tears. The false-positive find- associated lesion of the anterior cruciate ligament,
ings involved the posterior horns of the menisci, the to ascertain whether these lesions progress to com-
site of most false-negative arthroscopic examinations plete tears. The study group consisted of 27 menisci.
(QUINN and BROWN 1991). The negative predictive On follow-up over 3 years, six decreased in size, 18
value of MR imaging was 100%. remained unchanged and two lesions disappeared.
DE SMET et al. (1994) analyzed the cause of errors The authors concluded that most grade 11 lesions are
of MR imaging in meniscal tears in aseries of 400 stable.
MR examinations performed for suspected meniscal In a further attempt to establish the value of the
te ars in which the accuracy was 90%. In this group grading system for meniscal lesions, KORNICK et
they found 70 cases in which the MR diagnosis did al. (1990) reviewed the MR imaging knee examina-
not agree with the surgical findings. They found that tions of 64 asymptomatic volunteers in the 2nd to
40% were unavoidable errors (false-positive or false- 8th decades of life, and analyzed these for meniscal
negative diagnoses that could not be avoided even abnormalities. They found grades I, II and III in all
in retrospective examination), 39% were related to decades, with a prevalence of 25% as early as the 2nd
equivocal MR imaging findings and 21 % were due to decade, and increasing sharply with age. They con-
interpretation errors. In this series, 6% of the menis- cluded that there is a baseline prevalence of meniscal
cal tears could not be identified, even retrospective- signal in the asymptomatic population.
ly. Unavoidable false-positive diagnoses due to healed
tears or tears missed at arthroscopy were an infre-
quent problem in their study, occurring in only 1.5% 9.4.2
of the 400 knees evaluated. Terminology and Signs of a Tear
KAPLAN et al. (1991) evaluated the significance
of high signal in the meniscus that does not clearly Although the grading system for meniscal signal
extend to the surface (grades land 11) in aseries of abnormalities has been helpful in correlating the MR
142 consecutive patients undergoing MR examina- imaging patterns with histological abnormalities, it
tions of the knee. The prevalence was 14% (20 cases). is often found in clinical practice that interobserver
of these, 13 showed no evidence of tear at arthros- and intraobserver variability is a common problem
copy or arthrotomy and one showed internal degen- with any type of classification scheme. Furthermore,
eration on histological examination. They concluded the report of the presence of a grade I or 11 "lesion"
142 J. Beltran and S. Shankman
a c
tear. In the lateral meniscus this assessment is more An extensive vertieal or oblique tear with central
difficult. The posterior horn of the lateral meniscus displacement of the free edge is referred to as a buck-
should cover the most medial portion of the lateral et-handle tear. The handle represents the displaced
tibial plateau. Failure to do so is also suggestive free margin towards the intercondylar notch, and
of a radial tear in this location. A secondary sign the peripheral non-displaced fragment represents the
of a radial tear is the presence of subluxation of bucket (Fig. 9.20). The tear normally starts in the pos-
the meniscus, in the absence of osteoarthritis, as terior horn of the meniscus and extends anteriorly
described by TUCKMAN et al. (1994). in a longitudinal fashion towards the anterior horn
"Parrot beak" tears may be considered a variation (SINGSON et al. 1991). Compressive forces by the fem-
of a radial tear in which a portion of the meniscus oral condyle produce central displacement of the
adjacent to the tear becomes slightly displaced and its fragment of meniscus. The tear may be asymmetrie
shape resembles the beak of a parrot (Fig.9.19). and spare the anterior horn.
They are best seen on coronal or thin-section axial Bucket-handle te ars are clinicaHy significant
images. On sagittal views, the "bow tie" configuration because they produce mechanical locking or diffi-
of the lateral meniscus may be disrupted centraHy culty in fuHy extending the knee joint. The medial
(STOLLER et al. 1987; DE SMET et al. 1993a, b). meniscus is more frequently involved. Bucket-handle
144 J. Beltran and S. Shankman
truncated or shortened meniscus on coronal images sign was present in 44% of medial and 29% of lateral
(Fig. 9.20c). The "double posterior cruciate ligament" tears. The fragment in the intercondylar notch was
sign described by SINGSON et al. (1991) and WEISS present in 66% of medial and 43% of lateral menis-
et al. (1991) has also been found to be highly sen- cal tears. The authors concluded that MR imaging is
sitive. This sign refers to the presence of an anteri- sensitive in detecting large bucket-handle te ars, but
orly located low signal intensity band representing the sensitivity decreased when the te ar was small.
the displaced meniscal fragment in the intercondy- Another sign of a bucket-handle tear described by
lar notch, above the tibial spine, in an orientation HELMS et al. (1998) is the so-called absent bow tie
more or less parallel to the posterior cruciate liga- sign. The body of the meniscus normally measures
ment. This sign is seen in the sagittal plane, and in 9-12 mm in width; thus sagittal images through the
aseries of 18 consecutive bucket-handle te ars pub- body of the meniscus using a section thickness of
lished by SINGSON et al. (1991) it was present in all of 4-5 mm should reveal the bow-tie configuration of the
them (Fig. 9.20b). body of the meniscus in two consecutive sections. In a
HARAMATI et al. (1993) described the "flipped bucket-handle tear, the body of the meniscus is short-
meniscus" sign. This sign refers to the displacement ened and therefore is seen in only one section. In their
of a meniscal fragment into the anterior compart- series, HELMS et al. (1998) found this sign to be present
ment. According to the original description, the frag- in 32 (97%) of33 cases ofbucket-handle tears.
ment remains attached to the rest of the meniscus at Longitudinal, vertically oriented tears with dis-
two points. On MR imaging, the anterior horn and placement of a portion of meniscal tissue are referred
the flipped meniscal fragment offer an appearance of to as vertical flap tears. They can be considered as
an unusually enlarged horn or a double anterior horn a bucket-handle tear that has progressed to one of
(Fig. 9.21) (RuPF et al. 1998). The ipsilateral posterior the horns of the meniscus, resulting in a fragment of
horn is decreased in size or absent. This sign can be meniscal tissue attached by only one point to the rest
seen in bucket-handle tears involving the medial or of the meniscus. The fragment can become displaced
the lateral menisci. medially, anteriorly, or posteriorly. If the fragment of
In a retrospective evaluation of 39 arthroscopical- meniscal tissue becomes displaced anteriorly, it may
ly proven bucket-handle tears, WRIGHT et al. (1995) be flipped over the anterior horn of the meniscus,
found that the "double posterior cruciate" ligament producing the same MR imaging findings as have
sign was seen in 53% of the medial and in none of been described above in relation to the bucket-han-
the lateral bucket-handle tears. The flipped meniscus dIe tear. If the tear extends to the remaining point of
a b
Fig. 9.21a, b. Flap meniscal tear. a Diagrammatic representation of a flipped fragment of meniscus anterior to the anterior horn.
b Flipped meniscus sign. Sagittal proton density image showing the displaced fragment of the posterior horn (arrow) of the
lateral meniscus projecting anterior to the anterior horn
146 J. Beltran and S. Shankman
attaehment of the flap, then a free meniseal frag- found two superiorly displaeed flap tears. In a retro-
ment oeeurs. MR imaging ean demonstrate the mor- speetive evaluation of 236 MR imaging examinations
phology of the vertieal flap tears and the loeation of showing displaeed meniseal tears, LECAS et al. (2000)
the fragment. No series are available in the literature found 11 inferiorly displaeed flap tears. The fragment
assessing the aeeuraey of MR imaging in the diagno- lay inferomedial to the tibial plateau and deep in rela-
sis of this type of tear. tion to the medial eollateralligament.
Another type of flap tear oeeurs when a horizontal Meniseoeapsular separation is defined as a detaeh-
tear produees a fragment of meniseal tissue that may ment of the peripheral portion of the meniseus from
beeome displaeed superiorly or inferiorly (Fig. 9.22) its eapsular attaehment. These lesions are uneom-
(RUFF et al. 1998). These tears are less frequent than mon. They are treated nonoperatively if they are iso-
the bueket-handle tears but in one series aeeounted lated, or a meniseal repair is done if there is an asso-
for 19% of symptomatie meniseal injuries (DANDY ciated ligamentous injury (PRICE and ALLEN 1978;
and ]ACKSON 1975). They usually oeeur in the medial STONE 1979; HAMBERG et al. 1983; STRAND et al.
meniseus. Superior displaeement of the flap is more 1985). The results of this eonservative management
frequent than inferior displaeement. In one series are good beeause the tear oeeurs in the red zone of
of 25 displaeed meniseal injuries, RUFF et al. (l998) the meniseus, where vaseularization allows for heal-
ing. The MR imaging signs of meniseoeapsular sepa-
ration have been summarized by RUBIN et al. (l996).
In medialiesions, displaeement of the meniseal edge
from the tibial margin and fluid between the eapsule
and the meniseus are evident, while the identifieation
of a fascicle tear suggests laterallesions (Fig.9.23).
In a prospeetive evaluation of 52 eases diagnosed to
have meniseoeapsular injuries, the positive predie-
tive value of these signs was low, with poor eorrela-
tion with arthroseopie findings (RUBIN et al. 1996).
9.5
Meniscal Cysts, Ossicles, and
Discoid Meniscus
Meniseal eysts result from the aeeumulation of syno-
vial fluid within the meniseus following a meniseal
tear. A parameniseal eyst represents an extension of
Fig.9.22a, b. Flap meniscal tear. a Diagrammatic representa- the meniseal eyst into the surrounding soft tissues
tion of a fiap tear of the medial meniscus. The fragment is
displaced superiorly. b Coronal proton density fat-saturated
(TYSON et al. 1995; LEKTRAKUL et al. 1999). They are
image demonstrating a vertically displaced fragment of the more often seen in association with lateral meniseal
medial meniscus (arrow), adjacent to the PCL tears, but medial meniseal tears mayaiso develop
The Menisci 147
parameniscal cysts. Their size varies from a few milli- A discoid meniscus refers to a meniscus, almost
meters to several centimeters and they produce pain. always the lateral one, that is not C-shaped but rather
Large cysts are palpable at the level of the joint line. disk-like in configuration, covering most of the tibial
The meniscal tear associated with the cyst is gen er- plateau to varying degrees. It is prone to tearing and
ally horizontal in configuration and extends to the is usually seen in children and adolescents. They may
periphery of the meniscus (Fig.9.24). Occasionally be asymptomatic and incidentally noted. Although
they may be found adjacent to the posterior cruciate seen in youngsters, it is believed that they are more
ligament, simulating a posterior cruciate ganglion developmental than congenital in that the fetal menis-
cyst (LEKTRAKUL et al. 1999). cus never assurnes such a shape.
Meniscal ossicles are rare lesions found in young The diagnostic criteria of discoid meniscus on MR
individuals (MARTINOLl et al. 2000). Their origin is imaging include continuity of meniscal tissue with
controversial. Some researchers believe that they are the anterior and posterior horns in three or more
vestigial structures but others propose a post-trau- 5-mm contiguous sagittal sections and/or a trans-
matic etiology (BERG 1991; RICHMOND and SARNO verse width greater than 14 mm (ARAKI et al. 1994;
1988). In general they are asymptomatic and they are SILVERMAN et al. 1989) (Fig. 9.25).
typically found in the posterior horn of the medial
meniscus, near the tibial attachment (SCHNARKOWS-
KI et al. 1995). Since they are calcified, they can be
identified on radiographs and are often mistaken
for loose bodies. On MR imaging they are seen as a
focal area of intrameniscal high signal intensity on
Tl-weighted images owing to the presence of fatty
marrow within the ossicle, and are surrounded by a
low signal intensity margin representing the cortex
a
of the ossicle.
a
9.6
Postoperative Meniscus
_ _ _ _-li b
Buckwalter KA, Braunstein EM, Janizek DB et al (1993) MR meniscal tears.Arch Orthop Trauma Surg 120(112):14-16
imaging of meniscal tears: narrow versus conventional Eriekson SJ, Cox IH, Hyde JS et al (1991) Effect of tendon ori-
window width photography. Radiology 187:827-830 entation on MR imaging signal intensity: a manifestation
Bui-Mansfield LT, Youngberg RA, Warme W et al (1997) Poten- of the "magie angle" phenomenon. Radiology 181:389-392
tial cost savings of MR imaging obtained before arthros- Eriekson SJ, Prost TW, Timins ME (1993) The "magie angle"
copy of the knee: evaluation of 50 consecutive patients. AJR effect: background physies and clinieal relevance. Radiol-
169:913-918 ogy 188:23-25
Chan PS, Kneeland JB, Gannon FH et al (1998) Identification Escobedo EM, Hunter JC, Zink-Brody GC et al (1996) Useful-
of the vascular and avascular zones of the human meniscus ness of turbo spin-echo MR imaging in the evaluation of
using magnetie resonance imaging: correlation with histol- meniscal tears: comparison with a conventional spin-echo
ogy. Arthroscopy 14:820-823 sequence. AJR 167:1223-1227
Cheung LP, King CPL, Hollett MD et al (1997) Meniscal tears Eustace S, Hentzen P, Adams J et al (1999) Comparison of con-
of the knee: accuracy of detection with fast spin-echo MR ventional and turbo spin-echo Tl-weighted MR imaging in
imaging and arthroscopie correlation in 293 patients. Radi- acute knee trauma. AJR 172:1393-1395
ology 203:508-512 Ferrer-Roca 0, Vilalta C (1980) Lesion of the meniscus.
Cotten A, Delfaut E, Demondion X et al (2000) MR imaging Part 1. Macroscopie and histologie findings. Clin Orthop
of the knee at 0.2 and 1.5 T: correlation with surgery. AJR 146:301-307
174:1093-1097 Fischer SP, Fox JM, DelPizzo W et al (1991) Accuracy of diag-
Crues JB, Stoller DW (1993) The menisci. In: Mink Jet al (eds) noses from magnetie resonance imaging of the knee. J Bone
Magnetie resonance imaging of the knee. Raven, New York, Joint Surg [Am] 73:2-10
pp 91-140 Franklyn PD, Lemon RA, Barden HS (1997) Accuracy of imag-
Crues JV, Mink J, Levy TL et al (1986) Meniscal tears of the ing the menisci on an in-office, magnetie resonance imag-
knee. AJR 47:293-30 ing extremity system. Am J Sports Med 25:383-388
Crues JV, Mink J, Levy TL et al (1987) Meniscal tears of the Fullerton G, Cameron I, Ord V (1985) Orientation of tendons
knee: accuracy of MR imaging. Radiology 164:445-448 in the magnetie field and its effect on T2 relaxation times.
Crues JV, Ryu R, Morgan FW (1990) Meniscal pathology: Radiology 155:433-435
the expanding role of magnetie resonance imaging. Clin Gillquist J, Messner K (1993) Long-term results of meniscal
Orthop 252:80-87 repair. Sports Med Arthrosc Rev 1: 159-163
Czervionke LF, Czervionke JM, Daniels DL et al (1988) Glashow JL, Katz R, Schneider M et al (1989) Double-blind
Characteristie features of MR truncation artefacts. AJR assessment of the value of magnetie resonance imaging in
151:1219-1228 the diagnosis of anterior cruciate and meniscal lesions. J
Dandy DJ (1990) The arthroscopie anatomy of symptomatie Bone Joint Surg [Am] 71:113-119
meniscallesions. J Bone Joint Surg [Br] 72:628-633 Graf B, Docter T, Clancy W Jr (1987) Arthroscopie meniscal
Dandy DJ, Jackson RW (1975) Meniscectomy and chondro- repair. Clin Sports Med 6:525-535
malacia of the femoral condyle. J Bone Joint Surg [Am] Haggar AM, Froelieh JW, Hearshen et al (1988) Meniscal abnor-
57:1116-1119 malities of the knee: 3DFT fast-scan grass MR imaging.AJR
DeHaven KE (1985) Rationale for meniscus repair or excision. 150:1341-1344
Clin Sports Med 4:267-273 Hamberg P,Gillquist J, Lysholm J (1983) Suture of new and
DeSmet AA, Norris MA, Yandow DR et al (1993a) MR diagnosis old peripheral meniscus tears. J Bone Joint Surg [Am]
of meniscal tears of the knee: importance of high signal in 65:193-197
the meniscus that extends to the surface.AJR 161:101-107 Haramati N, Staron RB, Rubin S et al (1993) The flipped menis-
DeSmet AA, Norris MA, Yandow DR et al (1993b) Diagnosis cus sign. Skeletal RadioI22:273-277
of meniscal tears of the knee with MR imaging: effect of Hauger 0, Frank LR, Boutin RD et al (2000) Characterization
ob server variation and sampIe size on sensitivity and speci- of the "red zone" of knee meniscus: MR imaging and histo-
ficity. AJR 160:555-559 logic correlation. Radiology 217:193-200
DeSmet AA, Tuite MJ, Norris MA et al (1994) MR diagnosis Haven KE,Gillquist J (1989) Non-operative treatment of menis-
of meniscal tears: analysis of causes of errors. AJR cal tears. J Bone Joint Surg [Am] 71:811-822
163:1419-1423 Helms CA, Laorr A, Cannon DW (1998) The absent bow tie
Deutsch AL, Mink JH, Fox JM et al (1990) Peripheral sign in bucket-handle tears of the menisci in the knee. AJR
meniscal tears: MR findings after conservative treatment or 170:57-61
arthroscopie repair. Radiology 176:484-488 Hendler RC (1984) Arthroscopie meniscal repair; surgieal tech-
Dillon EH, Pope CF, Lynch K (1990) The clinieal significance nique. Clin Orthop 190:163-169
of stage 2 meniscal abnormalities on magnetic resonance Henning CE, Lynch MA (1985) Current concepts of meniscal
knee images. Magn Reson Imaging 8:411-415 function and pathology. Clin Sports Med 4:259-265
Dillon EH, Pope CF, Jokl P et al (1991) Follow-up of grade Henning CE, Yearout KM, Vequist SW et al (1991) Use of the
2 meniscal abnormalities in the stable knee. Radiology fascia sheath coverage and exogenous fibrin clot in the
181:849-852 treatment of complex meniscal tears. Am J Sports Med
Disler DG, Kattapuram SV, Chew FS et al (1993) Meniscal tears 19:626-631
of the knee: preliminary comparison of three-dimensional Herman LJ, Beltran J (1988) Pitfalls in MR imaging of the knee.
MR reconstruction with two-dimensional MR imaging and Radiology 167:775-781
arthroscopy. AJR 160:343-345 Jackson JP (1968) Degenerative changes in the knee after men-
Elvenes J,Jerome CP,Reikeras 0 et al (2000) Magnetie resonance iscectomy. Br Med J 2:525-527
imaging as a screening procedure to avoid arthroscopy for Jackson DW, Jennings LD, Maywoon RM et al (1988) Magnetic
The Menisci 151
resonance imaging of the knee. Am J Sports Med 16:29-38 versus high-field MR imaging of the knee: a comparison of
Kaplan PA, Nelson NL, Garvin KL et al (1991) MR of the knee: signal behavior and diagnostie performance. Eur J Radiol
the significance of high signal in the meniscus that does 19:132-138
not clearly extend to the surface. AJR 156:333-336 Peterfy CG, Janzen DL, Tirman PFJ et al (1994) «Magie-angle»
Kersting-Sommerhoff B, Gerhardt P et al (1995) MR of the phenomenon: a cause of increased signal in the normal
knee joint: first results of a comparison of 0.2 T specialized lateral meniscus on short-TE MR images of the knee. AJR
system and 1.5 T high field strength magnet. Rofo Fortschr 163:149-154
Geb Rontgenstr Neuen Bildgeb Verfahr 162:390-395 Poehling GG, Ruch DS, Chabon SJ (1990) The landscape of
Kinnunen J, Bondestam S, Kivioja A et al (1994) Diagnostie meniscal injuries. Clin Sports Med 9:539-549
performance of low field MR imaging in acute knee inju- Potter HG, Rodeo SA, Wickiewiez TL et al (1996) MR
ries. Magn Reson Imaging 12:1155-1160 imaging of meniscal allografts: correlation with clinieal
Korniek J, Trefelner E,McCarthy Set al (1990) Meniscal abnor- and arthroscopie outcomes. Radiology 198:509-514
malities in the asymptomatie population at MR imaging. Priee CT, Allen WC (1978) Ligament repair in the knee with
Radiology 177:463-465 preservation of the meniscus. J. Bone Joint Surg [Am]
Lal NR, Jamadar DA, Doi K, Newman JS et al (2000) Evaluation 60:61-65
of bone contusions with fat -saturated fast spin -echo pro- Quinn SF, Brown TF (1991) Meniscal tears diagnosed with
ton-density magnetie resonance imaging. Can Assoc Radiol MR imaging versus arthroscopy: how reliable a standard is
J 51(3):182-185 arthroscopy? Radiology 181:843-847
Lecas LK, Helms CA, Kosarek FJ et al (2000) Inferiorly dis- Quinn SF, Brown TR, Szumowski (1992) Menisci of the knee:
placed flap tears of the medial meniscus: MR appearance radial MR imaging correlated with arthroscopy in 259
and clinieal significance.AJR 174:161-164 patients. Radiology 185:577-580
Lektrakul N, Skaf A, Heh L et al (1999) Pericruciate meniscal Raunest J, Ob erle K, LoehnertJ et al (1991) The clinieal value of
cysts arising tears of the posterior horn of the medial magnetie resonance imaging in the evaluation of meniscal
meniscus: MR imaging features that simulate posterior cru- disorders. J Bone Joint Surg [Am] 73:11-16
ciate ganglion cysts.AJR 172:1575-1579 Reieher MA, Hartzman S, Duckwiler GR et al (1986) Meniscal
Listerud J, Einstein S, Outwater E et al (1992) First principles injuries: detection using MR imaging. Radiology
of fast spin echo. Magn Reson Q 8: 199-244 159:753-757
Lotysch M, Mink J, Crues JV et al (1986) Magnetie resonance Riehmond JC, Sarno RC (1988) Posttrauma intracapsular
imaging in the detection of meniscal injuries. Magn Reson bone fragments; association with meniscal tear. AJR
Imaging 4:185 150:159-160
Lufkin RB, Pusey E, Stark DD et al (1986) Boundary artefact due Riel KA, Reinish M, Kersting-Sommerhoff B et al (1997) 0.2
to truncation errors in MR imaging. AJR 147:1283-1287 Tesla magnetie resonance imaging of internaliesions of the
Lynch MA, Henning C, Glick KR Jr (1983) Knee joint surface knee joint: a prospective arthroscopieally controlled clini-
changes, long-term follow-up meniscus te ar treatment cal study. Knee Surg Sports Traumatol Arthrosc 7:37-41
in stable anterior cruciate ligament reconstructions. Clin Rubenstein JD, Kim JK, Morava-Protzner I et al (1993) Effects
Orthop 172:148-153 of collagen orientation on MR imaging characteristies of
Mackenzie R, Keene GS, Lomas DJ et al (1995) Errors at knee bovine cartilage. Radiology 188:219-226
magnetie resonance imaging: true or false? Br J Radiol Rubin DA, Kneeland JB, Listerud J et al (1994) MR diagnosis of
68(814):1045-1051 meniscal tears of the knee: value of fast spin-echo vs. con-
Martinoli C, Bianchi S, Spadola L et al (2000) Multimodality ventional spin-echo pulse sequences. AJR 162:1131-1135
imaging assessment of meniscal ossicle. Skeletal Radiol Rubin DA, Britton CA, Towers JD et al (1996) Are MR imaging
29:481-484 signs of menisco-capsular separation valid? Radiology
Marty-Delfaut E, Beltran J (1999) Fat saturation: MRI tech- 201:829-836
niques and pitfalls. Radiographies 19:373-382 Ruff C, Weingardt JP, Russ PD et al (1998) MR imaging patterns
McGinty JB, Guess LF, Marvin RA (I977) Partial or total of displaced meniscus injuries of the knee. AJR 170:63-67
meniscectomy: a comparative analysis. J Bone Joint Surg Ruwe PA, Wright J, Randall RL et al (1991) Can MR
59:763-766 imaging effectively replace diagnostie arthroscopy? Radiol-
Melki PS, Mulkern RV, Schwartz RB et al (1991) Comparing ogy 181:335-339
the FAISE method with conventional dual-echo sequences. Ryu KN, Kim IS, Kim EJ, Ahn JW et al (1998) MR imaging of
J Magn Reson Imaging 1:319-326 tears of discoid lateral menisci.AJR 171:963-967
Metcalf RW (1991) Arthroscopie meniscal surgery. In: McGinty Schnarkowski P, Tirman PFJ, Fuchigami KD et al (1995) Menis-
JB (ed) Operative arthroscopy. Raven, New York, cal ossicle: radiographie and MR imaging findings. Radiol-
pp 227-236 ogy 196:47-50
Miller DB (1988) Arthroscopie meniscus repair. Am J Sports Seedholm BB (1979) Transmission of the load in the knee joint
Med 16:315-320 with special reference to the role of the menisci. Part I.
Mink JH, Levy T, Crues JV (1988) Tears of the anterior cruciate Anatomy, analysis, and apparatus. Eng Med 8:207-219
ligament and menisci of the knee: MR imaging evaluation. Seedholm BB, Hargreaves DJ (1979) Transmission of the load
Radiology 167:769-774 in the knee with special reference to the role of the menisci.
Newman AP, Daniels AU, Burks RT (1993) Principles and deci- Part 11. Experimental results, discussion and conclusion.
sion making in meniscal surgery. Arthroscopy 9:33-51 Eng Med 8:220-228
Noble J, Hamblen DL (1975) The pathology of the degenerate Shakespeare D, Rigby H (1983) The bucket-handle tear of the
meniscus lesion. J Bone Joint Surg [Br] 57:180-186 meniscus. J Bone Joint Surg [Br] 65B:383-387
Parizel PM, Dijkstra HAJ, Geenen PJ et al (1995) Low-field Shankman S, Beltran J, Melamed E et al (1997) Anterior horn of
152 J. Beltran and S. Shankman
the lateral meniscus: another potential pitfall in MR imag- planar imaging: comparison with conventional3D fat-sup-
ing of the knee. Radiology 204:181-184 pressed gradient echo sequence and correlation with his-
Shellock FG, Stone K, Crues JV III (1998) Assessment of tology. J Comput Assist Tomogr 22:8-14
the knee using an extremity MR imaging system: diagnos- Tuckman GA, Miller WJ, Remo JW et al (1994) Radial tears of
tie finding compared to arthroscopy (abstract). Radiology the menisci: MR findings.AJR 163:395-400
209(P):192 Turner DA, Rapoport MI, Erwin WD et al (1991) Truncation
Shogry MEC, Pope TL (1991) Vacuum phenomenon simulating artefact: a potential pitfall in MR imaging of the menisci of
meniscal or cartilaginous injury of the knee at MR imag- the knee. Radiology 179:629-633
ing. Radiology 180:513-515 Tyson LL, Daughters TC Jr, Ryu RKN et al (1995) MR imaging
Silverman JM, Mink JH, Deutsch AL (1989) Discoid menisci appearances of meniscal cysts. Skeletal Radiol 24:421-424
of the knee: MR imaging appearance. Radiology Vahey TN, Bennett HT, Arrington LG, Shelbourne KD, Ng J
173:351-354 (1990) MR imaging of the knee: pseudotear of the lateral
Singson RD. Feldman F, Staron R et al (1991) MR imaging of meniscus caused by the meniscofemoral ligament. AJR
displaced bucket-handle tear of the medial meniscus. AJR 154:11237-1239
156:121-124 Vahlensieck M, Sommer T, Textor J et al (1998) Indirect
Smillie IS (1970a) Clinieal features of internal derangements MR arthrography: techniques and applications. Eur Radiol
relative to the menisci. In: Injuries of the knee joint, 4th edn. 8:232-235
Livingstone, London, pp 70-97 Warren RF (1985) Arthroscopie meniscus repair. Arthroscopy
Smillie IS (1970b) Surgieal pathology of the menisci. In: 1:170-172
Injuries of the knee joint, 4th edn. Livingstone, London, Weiss KL, Morehouse HT, Levy IM (1991) Sagittal MR imag-
pp 39-69 ing of the knee: a low-signal band parallel to the posterior
Smith DK, Totty WG (1990) The knee after partial meniscec- cruciate ligament caused by a displaced bucket-handle tear.
tomy: MR imaging features. Radiology 176:141-144 AJR 156:117-119
Stoller DW (1989) Magnetie resonance imaging in orthopae- Wennogle CF, Chang PS (1998) Diagnostie accuracy of a
dies and rheumatology. Lippincott, Philadelphia, p 99 low field strength extremity-only MR scanner in internal
Stoller DW, Martin C, Crues JV et al (1987) Meniscal derangement of the knee: correlation with arthroscopy and
tears: pathologie correlation with MR imaging. Radiology 1.5 Tesla MR imaging (abstract). Radiology 209(P):192
163:731-735 White LM, Schweitzer ME, Johnson WJ et al (1996) The role
Stone RG (1979) Peripheral detachment of the menisci of the of T2 weighted fast-spin-echo imaging in the diagnosis of
knee. Orthop Clin North Am 10:643-657 meniscal tears. J Magn Reson Imaging 6:874-877
Strand T, Engesaeter LB, Molster AO (1985) Meniscus repair in Wood MR, Henkelman RM (1985) Truncation artefacts in mag-
knee ligament injuries.Acta Orthop Scand 56:130-132 netie resonance imaging. Magn Reson Med 2:517-526
Suarez-Almazor ME, Kaul P, Kendall CJ et al (1999) The cost- Wright DH, DeSmet AA, Norris M (1995) Bucket-handle
effectiveness of magnetie resonance imaging for patients tears of the medial and lateral menisci of the knee: value
with internal derangement of the knee. Int J Technol Assess of MR imaging in detecting displaced fragments. AJR
Health Care 15:392-405 165:621-625
Trattnig S, Huber M, Breitenseher MJ et al (1998) Imaging Yu JS, Cosgarea AJ, Kaeding CC, Wilson D (1997) Meniscal
artieular cartilage defects with 3D fat-suppressed echo flounce MR imaging. Radiology 203:513-515
10 The Cruciate and Collateral Ligaments
CHARLES P. Ho
10.2.1
MR Imaging Technique
10.2.2
The Normal ACL Fig. 10.1 a, b. Intact ACL. a The entire length and tibial attach-
ment of the ACL (arrows) are best evaluated on slightly oblique
sagittal images oriented along the course of the ACL. b The
The dense organized collagen bundles of ligaments
proximal femoral attachment of the ACL (arrows) is best seen
with sparse cellular elements have very few water mol- on transverse images. Proton density (long TR, short TE)
ecules and relatively mobile hydrogen proton nuclei to turbo spin echo (TSE) images
The Cruciate and Collateral Ligaments 155
10.2.3
The Injured ACL
Fig.l0.4. Proximal eomplete tear of the ACL. Fig.lO.6. Distal avulsion of the ACL. Sagittal T2-weighted
Sagittal T2-weighted image revealing a dis- image demonstrating a 1.5-em avulsion fragment of the ante-
erete defeet in proximal third of the ligament, rior tibial eminenee (arrow) at the ACL attaehment, with about
filled with high-signal fluid and hemorrhage I-ern elevation of the fragment. Note also the minimally dis-
(arrow) plaeed fraeture of the posterior tibial eminenee (open arrow)
and the posterior eapsule tearing (curved arrow)
10.3 with the ACL, the bands of the PCL typically are
Posterior Cruciate Ligament not distinguished as separate bands at MR imaging,
as no macroscopic intervening fat tissue and signal
The posterior cruciate ligament (PCL) originates at are normally seen. Rather, the ligament is seen as a
its proximal attachment at the medial femoral con- homogeneously low/black signal smooth ligament.
dyle along the anterior superior medial aspect of the Because of its relative thickness, the entire length as
intercondylar notch, and shows a smooth posterior well as the proximal and distal attachments of the
bowed course to its distal attachment along the pos- PCL generally may be seen on one or more adjacent
terior tibial eminence in the typically imaged supine sagittal images even when the sagittal images are
extension position (Fig.lO.lO). Like the ACL, it con- angled slightly obliquely along the orientation of the
sists of two functional bands. The posteromedial ACL as described above. The short diameter cross-
band is tightened/stressed first in extension, while the section of the PCL should also be evaluated proximal-
anterolateral band is stressed first in flexion. Unlike lyon the coronal images and distally on the transverse
a b
10.7
Conclusion
tionized the capabilities and role of diagnostic imag- Ho CP, Marks PH, Steadman JR (1999) MR imaging of knee
ing in the evaluation of the ligamentous structures anterior cruciate ligament and associated injuries in skiers.
MRI Clin North Am 7: 117-l30
about the knee, as these ligamentous structures may
Kaplan PA, Walker CW, Kilcoyne RF et al (1992) Occult frac-
now be directly imaged and accurately and precisely ture patterns of the knee associated with anterior cruciate
evaluated. Comprehensive examination of the liga- ligament tears: assessment with MR imaging. Radiology
mentous structures as weH as the remainder of the 183:835-838
knee at MR imaging is best achieved by using images Kim MG, Kim BH, Choi JA et al. (2001) Intra-articular ganglion
cysts of the knee. Eur Radiolll :834-840
in aH three of the orthogonal planes of the knee, as
Lee SH, Petersilge CA, Trudell DJ et al (1996) Extrasynovial
each plane is important and best for seeing specific spaces of the cruciate ligaments: anatomy, MR imaging and
portions of these structures. diagnostic implications. AJR 166: 1433
Miller TT, Gladden P, Staron RB et al (1997) Posterolateral sta-
bilizers of the knee: anatomy and injuries assessed with MR
References imaging. AJR 169:1641
Roychowdhury S, Fitzgerald SW, Sonin AH et al (1997) Using
Brandser EA, Riley MA, Berbaum KS et al (1996) MR imaging MR imaging to diagnose partial tears of the anterior cruci-
of anterior cruciate ligament injury: independent value of ate ligament: value ofaxial images.AJR 168:1487
primary and secondary signs. AJR 167:121 Rubin DA, Kettering JM, Towers JD et al (1998) MR imaging
Delzell PB, Schils JP, Recht MP (1996) Subtle fractures ab out of knees having isolated and combined ligament injuries.
the knee. Innocuous-appearing yet indicative of significant AJR 170:1207
internal derangement. AJR 167:699 Schweitzer ME, Tran D,DeelyDM et al (1995) Medial collateral
Goldman AB, Pavlov H, Rubenstein D (1998) The Segond frac- ligament injuries: evaluation of multiple signs, prevalence
ture of the proximal tibia: a small avulsion that reflects and location of associated bone bruises, and assessment
major ligamentous damage. AJR 151: 1163-1167 with MR imaging. Radiology 194:825-829
Ha TPT, Li KCP, Beaulieu CF et al (1998) Anterior cruciate liga- Tyrrell PNM, Cassar-Pullicino VN, McCall IW (2000) Intra-
ment injury: fast spin-echo MR imaging with arthroscopic articular ganglion cysts of the cruciate ligaments. Eur
correlation in 217 examinations.AJR 170:1215 Radioll0:1233-1238
11 The Postoperative Knee 1:
Menisci, Cruciate Ligaments, Cartilage
VICTOR N. CASSAR-PULLICINO and s. N. J. ROBERTS
CONTENTS 11.2
SurgicalOverview
11.1 Introduction 165
11.2 Surgical Overview 165
11.2.1 Instability 165 From an orthopaedic surgeon's viewpoint, the clin-
11.2.2 Meniscal Tears 170 ical presentation of knee problems may be divided
11.2.3 Chondral and Osteochondral Damage 171 into three categories: pain, too much movement
11.3 Imaging Perspective 172 and too Httle movement. The commonest intra-artic-
11.3.1 Cruciate Ligament Reconstruction 172
11.3.2 The Postoperative Meniscus 183
ular causes of pain after an acute injury has settled
11.3.3 Articular Cartilage Assessment 188 are chondral and osteochondral damage along with
11.3.4 Extra-articular Structures 193 damage to the menisci. Too much (or the wrong kind
11.4 Conclusion 195 of) movement is caused by an inadequacy of either
References 195
the static (ligaments) or dynamic (muscle) restraints
to knee movements. By far the commonest surgical
cause of an unstable knee is a deficiency of the ante-
rior cruciate ligament (ACL). "Not enough move-
11.1 ment" may either be an ever-present stiffness in the
Introduction knee caused by either intra-articular or extra-articu-
lar pathology, or an intermittent catching or locking
Accurate interpretation of the postoperative status which is likely to be caused either by pateHofemoral
of the knee cannot be achieved without a working subluxation or by a mobile intra-articular fragment
knowledge of the utilised surgical procedures and such as a loose body or a meniscal tear.
techniques, as weH as familiarity with the expected
imaging appearances of the structures that have been
repaired, resected or replaced. A thorough under- 11.2.1
standing of these two areas is therefore a prerequi- Instability
site to the identification and correct interpretation of
recurrent or new problems in the postoperative knee. The knee is stabilised by four major ligaments: the
This chapter will focus primarily on the postopera- medial and lateral collaterals, and the anterior and
tive assessment of the knee following surgery to the posterior cruciates. There are important secondary
menisci, cruciate ligaments and articular cartilage. restraints and there has been considerable recent
surgical interest in the role of the soft tissue com-
plex at the posterolateral corner. The wedge-shaped
menisci also contribute to knee stability, particularly
anteroposteriorly.
It is generaHy, although not universally, accepted
that even complete tears of isolated extra-articular
V. N. CASSAR-PULLlCINO, MD ligaments often heal weH without surgery, and the
Consultant Radiologist, Department of Diagnostic Imaging, functional outcome is better without than with sur-
The Robert Jones and Agnes Hunt Orthopaedic and District gical repair or reconstruction. This has been shown
Hospital, Oswestry, Shropshire, SYlO 7AG, UK most clearly in the case of isolated tears of the medial
S. N. J. ROBERTS, MA, FRCS (Orth)
Consultant Orthopaedic and Sports Injury Surgeon, The
coHateralligament. The intra-articular cruciates heal
Robert Jones and Agnes Hunt Orthopaedic and District Hos- less weH, but undoubtedly a low-energy"peel off" of
pital, Oswestry, Shropshire, SY10 7AG, UK the ACL's femoral origin may leave a viable stump
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
166 V. N. Cassar-Pullicino and S. N. J. Roberts
which is able to re-attach itself to an area adjacent to When the decision to operate has been made, the
its anatomical origin or to the posterior cruciate liga- surgeon has a number of choices to make regarding
ment (PCL), allowing it to provide a degree of func- the type of graft, the positioning of the graft and its
tion in anteroposterior stabilisation. fixation to bone at either end.
Partly for this reason, but also because of the
varying capacity for individuals to rehabilitate their 11.2.1.1
dynamic stabilisers with proprioceptive and neuro- Choice of Graft
muscular training, not all tom ACLs will lead to sig-
nificant functional deficit. A number of attempts have Various prosthetic grafts that were popular in the
been made to devise scoring systems in an attempt to 1980s have very largely been abandoned, particularly
predict which individuals will manage well with reha- in young patients, as the failure rate in the medium
bilitation, and who will be unable to cope with the level term was unacceptably high. They may still have a
of activity that they demand without ligament recon- role in the older low-demand patient. Allografts, such
struction. However, universal agreement on these sys- as patellar tendon or Achilles tendon, have the advan-
tems has not been achieved. They have been based on tage of not causing any donor site morbidity, but have
the amount and level of activity demanded, but there is been shown to incorporate more slowly. They there-
a very poor correlation between the degree of antero- fore either delay an individual's return to sport or fail
posterior laxity and the likelihood of reconstruction if a slower rehabilitation is not adhered to.
being necessary. In the United Kingdom at present, it The common choices of autograft are the patellar
would appear reasonable to suggest that reconstruc- tendon (BPTB), hamstring STG tendons (semi-ten-
tion of the ACL in most, if not all patients should only dinosus and gracilis) and,less commonly, the quad-
follow failure of an adequate attempt at non-operative riceps tendon. The central one-third of the patellar
management. The exception to this would be if there tendon is most commonly used (although the medial
is a bony avulsion, since in this situation near anatom- third has also been recommended). This produces a
ical repair as opposed to reconstruction can be per- tendon approximately 10 mm in diameter and allows
formed. the surgeon to harvest a bone block from both patella
If adequate function cannot be achieved in the and tibial tuberosity in continuity with this, making
knee following rehabilitation, surgical reconstruction an overall BPTB graft 10 or 11 cm in length (Fig.11.1).
using a graft must be performed. The indication for The patellar tendon is broader proximally than dis-
surgical reconstruction of the ACL is therefore insta- tally, so if a harvest is taken parallel with the ten-
bility. If there is significant pain aside from giving don's fibres, the bone block proximallywill be slightly
way episodes, then an alternative diagnosis should larger than that distally, often producing a lO-mm-
be sought, i.e. meniscal or chondral pathology. Con- diameter cylindrical graft from the patella and a
versely, radiological evidence of degenerative change 9-mm bone block from the tibia. This is convenient
or chronological age is not, of itself, a contra-indica- surgically since it allows the graft to be inverted and
tion to operative stabilisation of an unstable knee if it the smaller block passed through the lO-mm-diam-
is the instability which is the predominant complaint. eter tibial tunnel into the 9-mm-diameter femoral
tunnel more easily. It also recreates to some extent The graft is chosen according to the surgeon's
the anatomy of the slightly convergent natural ACL, training and preference, there being very little differ-
which has a broader tibial than femoral footprint. ence in published outcomes. There is probably more
During the operation, a small amount of bone graft morbidity associated with a patellar tendon graft (in
is commonly taken from the tibia to fill the patellar particular anterior knee pain), but the advantage of
donor site in an attempt to minimise postoperative bone to bone fixation may outweigh this disadvan-
anterior knee pain. tage in selected patients.
Hamstring tendons are harvested through a short
incision over their insertion into the tibia. The sar- 11.2.1.2
torius fascia is split and elevated, revealing the graci- Positioning of the Graft
lis and semi-tendinosus tendons. Both the STG ten-
dons are taken in their entirety using a tendon strip- The optimum position of the tibial attachment site of
per. They are then folded in half, giving a four-strand the neo-ACL is in the centre of the tibial plateau, and
graft with an ultimate strength of over 4,000 N, a tunnel is drilled with a jig from the anteromedial
approximately twice that of the torn cruciate which tibial cortex, entering the knee joint at this point. A
it is used to replace. Despite being 30% stronger, the blind-ended femoral tunnel is drilled either through
diameter of this graft is typically rather smaller than the tibial tunnel or using the medial arthroscopy
the patellar tendon. This is because it has a more portal. The starting point for this femoral tunnel is
circular cross-section than the rather Hat patellar critical to the success of the procedure, and the most
tendon, and therefore contains more collagen. Reha- common surgical error contributing to its failure.
bilitation allows recovery of function of the harvest- The entry point of the tunnel needs to be as far poste-
ed tendon within a year. Imaging reveals formation of rior as possible in the roof of the intercondylar notch,
a neo-tendon along the line of the harvested tendon, allowing a minimal 2 mm of femoral cortex behind
but with a rather more proximal insertion. In the case it to prevent the tunnel blowing out into the popli-
of the patellar tendon, repair of the defect is so good teal fossa. The tunnel cannot be too far back without
that revision reconstructions have been performed "blowing out", and some surgeons even pass the graft
re-harvesting the same central third as was used for right "over the top" - outside the femur - without
the primary surgery. This is not currently recom- drilling a tunnel at all. This accepts a position which
mended owing to the inferior quality of the scar is a little too far back, but ensures that the error of
tissue. placing it too far forward cannot be made (Fig.l1.2).
Fig. 11.2a, b. Anterior cruciate ligament reconstruction using the "over the top" technique
168 V. N. Cassar-Pullicino and S. N. J. Roberts
a b
Fig. 11.3. Tl- (a) and T2-weighted (b) sagittal images showing the use of a biodegradable screw fixing the bone block in the
tibial tunnel
The role of "notchplasty" (the removal of bone clinical responses during the degradation process
from the edges of the intercondylar notch) is con- (LAJTAI et al. 1999; WARDEN et al. 1999). The screw is
troversial. In cases of long-standing ACL deficiency, passed alongside the bone block, filling the tunnel and
there is often osteophytic encroachment of the space, compressing the block against the side wall (Fig.ll.4).
leaving inadequate room for graft positioning. This There are a number of alternative fixation tech-
is not the case in early reconstructions, but surgeons niques more commonly employed when a hamstring
usually find it helpful to remove a little bone from STG graft is used. If a hamstring tendon is used, it
the medial border of the lateral femoral condyle to is placed in exactly the same mann er as described
ensure that there is no graft impingement, and also above, but a number of techniques are used for fixa-
to aid visualisation of the femoral tunnel position- tion since many surgeons feel that interference screw
ing. The rate of ACL rupture has been correlated with fixation is less satisfactory in the absence of a block
the measured notch width (probably as a result of of bone attached to either end of the graft.
the correlation between a congenitally narrow notch Pre-operative imaging in these cases excludes an
and a congenitally small and weak ACL), and some anatomical abnormality and secondary pathology. A
patients will require a degree of notchplasty to allow postoperative AP and lateral radiographs of the knee
the passage of an adequate graft. confirms a satisfactory position of the graft and its
fixation (Fig. 11.5).
11.2.1.3
Fixation Devices 11.2.1.4
Posterior Cruciate Ligament
After tensioning the BPTB graft, it is typically fixed
using an interference screwwhich may be either metal- The PCL is three tim es stronger than the ACL and
lic or bio-absorbable polymer - usually polylactic acid perhaps 8-10 tim es less commonly ruptured. There
(Fig. 11.3). These biodegradable screws do not create is even more debate as to which PCL ruptures require
any artefact, allowing optimal assessment of the graft, reconstruction than is the case for the ACL, but it is
its tunnels and fixation status. As the bio-absorbable undoubtedly true that some patients with a truly iso-
screws are not easily visible radiographically, they are lated PCL rupture manage to continue with little or
best assessed by computed tomography (CT) and mag- no functional deficit even in professional sport. It has
netic resonance imaging (MRI). They have been shown increasingly been recognised that combined injuries,
to be safe and effective for fixation of bone blocks particularly of the PCL with the posterolateral corner
during ACL reconstruction, with no reported adverse complex, have a less satisfactory natural history, and
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 169
~-,- _ _ _~ b
Fig.ll.4. AP (a) and lateral (b) radiographs after an initial BPTB ACL reconstruction using metal interference screws which are
placed within the tibial and femoral tunnels. Following failure of this reconstruction, arevision hamstring (STG) reconstruction
was carried out with removal of the interference screws (c, d)
many surgeons are now advocating early surgery in a few millimetres of it. If this technique is used, the
these cases. graft needs to be passed through the tibial tunnel and
Broadly, the surgical technique for PCL reconstruc- forwards over the posterior tibial plateau - a right
tion is similar to that for ACL reconstruction, although angle known as the "killer curve" - to allow it to head
the tibial tunnel needs to enter the knee joint in the forward across the knee joint to its insertion into the
popliteal fossa, and the popliteal artery is at signifi- medial femoral condyle. In order to avoid this acute
cant risk from this. The tibial tunnel is drilled directly angulation, it has been suggested that an "in -lay" tech-
towards the popliteal artery and exits the tibia within nique should be used which involves a surgical expo-
170 V. N. Cassar-Pullicino and S. N. J. Roberts
a b
Fig. 11.5. AP (a) and lateral (b) views showing the harvest sites within the patella and tibial tuberosity clearly delineated
as surgical defects, along with the bone plugs that are located within the tibial and femoral tunnels, respectively. Note that
biodegradable screws have been used for fixation
sure of the popliteal fossa and a bone block being laid lescent human meniscus only the peripheral third of
on the posterior tibial cortex. This reduces the angle the meniscus has a blood supply and it is therefore
through which the graft must bend to enter the knee only tears in this region which are usually considered
joint (BERG FELD et al. 2001). There is debate as to the to be repairable. The vascularity of either side of the
optimum location of the femoral insertion into the tear can be assessed arthroscopically without tour-
medial femoral condyle, and considerable interest in niquet (Fig. 11.6), and typically tears with a blood
the use of two femoral tunnels, with part of the graft supply on at least one side (described as "red on
passed into each, in an attempt to recreate the mul- white") have a high rate ofhealing (over 85%).
tiple bundles of the natural PCL. A very wide range of techniques have been used
to repair meniscal tears. Initially, various techniques
of suture were used, either passing from the outside
11.2.2 in (with a knot on the intra-articular surface of the
Meniscal Tears meniscus) or passing from the inside out through
curved needles with sutures tied over the capsule.
The "C" -shaped fibrocartilaginous meniscus converts These latter inside-out techniques require small skin
a nearly Hat tibial plateau into something of a socket incisions to enable safe knot placement, exduding
to accommodate the convex femoral condyle. Com- neurovascular entrapment in the suture. More recent-
pressive forces are converted into a "hoop-stress" so ly"all inside" techniques have been developed using
that most of the resisted force in the meniscus is cir- bio-absorbable darts, arrows (barbed devices passed
cumferential rather than radial. Most of the collagen across the tear to hold the edges together), staples
in the meniscus runs circumferentially rather than and cannulated screws. These avoid both the skin
radially, confirming this, and it is for this reason that incision and the risk of damage to neurovascular
longitudinal splits of the meniscal body such as those structures and appear to have success rates similar to
which produce bucket-handle tears may be repaired the more invasive techniques.
successfully. Radial tears, which inevitably cut across It has been noted that the highest success rates
most of the stress-bearing collagen fibres, do not have been achieved with meniscal repair in conjunc-
have a high rate of healing. In the adult or even ado- tion with cruciate ligament reconstruction, and it has
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 171
Red White
on white on white Fig. 11.6. A diagram out-
lining the blood supply of
the meniscus and the "red"
and "white" zones
been suggested that this is because of the haemar- cedures have included abrasion arthroplasty, autol-
throsis which occurs postoperatively. This may pro- ogous chondrocyte transplant, micro-fracture tech-
vide healing factors to the poorlyvascularised healing nique, articular cartilage transplant, fresh osteo-
zone. Some surgeons have also attempted to improve chondral allografts and transfer of osteochondral
the local environment by inserting and suturing a autograft plugs. Penetration of the subchondral bone
fibrin clot across the repair, but this is now not com- (sometimes known as the Pridie procedure) has been
monlyused. used for over 50 years in an attempt to allow granu-
When meniscal repair is impossible, the arthros- lation tissue, and subsequently a fibrocartilaginous
copist needs to carry out a partial meniscectomy. scar, to form in the chondral defect. This may be done
Loss in function of the meniscus varies with the either by drilling or by the very similar micro-frac-
amount of meniscus removed, and so if repair is not ture technique, which involves driving a pick through
considered suitable, or proves unsuccessful, the sur- the subchondral bone in an attempt to produce a
geon tries to leave as much functional meniscus as tunnel with a rough as opposed to a smooth margin,
possible. There is a correlation between the amount allowing better clot adherence. No attempt is made to
of meniscus removed and the risk of osteoarthritis reconstruct the hyaline cartilage with this technique,
in the medium term. The amount removed is usually and it is known that the fibrocartilaginous scar which
recorded as a proportion of the radial width of the is produced is not as durable as the articular cartilage
meniscus. it replaces, and that it lacks cohesive attachment to
the underlying bone.
Osteochondral autografts (OATS) became popular
11.2.3 in the early 1990s as a me ans of filling symptomatic
Chondral and Osteochondral Damage defects. The technique involves harvesting a cylindri-
cal plug from a supposedly less important part of the
Articular cartilage has no blood supply, no nerve knee's articular surface. A plug is taken perpendicu-
supply, no lymphatic drainage and a very low meta- lar to the articular surface including articular carti-
bolie rate. It has little or no capacity to regenerate. lage, subchondral bone and a length of subchondral
Symptomatic defects in the articular surface are cancellous bone. Various manufacturers now pro-
therefore quite difficult to treat and a number of tech- duce sets of instruments which allow for plugs of an
niques have been used. appropriate length to be taken, varying in diameter
Cartilage repair procedures have been utilised with from 4 mm upwards. A slightly under-sized cylinder
varying degrees of success over the years. The pro- is drilled in the recipient site and the graft is press-
172 V. N. Cassar-Pullicino and S. N. J. Roberts
fitted. If multiple plugs are used, they are combined Considerable efforts have been made to produce
alongside each other, making a "mosaic" (hence the a non-invasive outcome assessment of the quality
term "mosaicplasty"), and careful selection and tes- and quantity of regenerated hyaline cartilage in the
sellation of multiple sizes may allow the triangular defect, particularly using MRI sequences. Although
defects between the plugs to be minimised. This tech- little is known of the structural and biological organ-
nique mayaIso be employed using larger allograft isation of the repair tissue, recent biopsy studies
donors (Fig. 11.7). indicate that within 12 months ACI successfully pro-
Since the mid 1980s, autologous chondrocyte cul- duces replacement cartilage tissue consisting of a
ture techniques have been developed whereby tissue deeper hyaline cartilage-like zone, with an upper
engineering hyaline cartilage may be introduced into a zone which is more fibrocartilaginous in nature
defect. This was pioneered in Sweden and is known as (RICHARDSON et al. 1999).
autologous chondrocyte implantation (ACI) or autolo-
gous chondrocyte transplantation. The classical tech-
nique involves a two-stage procedure. Firstly, a small
biopsy is taken of the articular cartilage, usually from 11.3
the lateral supracondylar ridge. This is minced, digest- Imaging Perspective
ed and cultured in the laboratory for approximately 3
weeks before the second stage procedure is performed. Although the imaging options do include convention-
The second procedure involves an open arthrotomy al radiography and CT, the assessment relies heavily
and, after debridement of the defect margins, a peri- on MRI. In view of this it is crucial that one under-
osteal patch usually taken from the anterior tibia is stands the biological processes that take place under
shaped and sutured over the defect to the surround- normal and pathological conditions in the three areas
ing cartilage like the skin over ablister. A watertight identified: cruciate ligament reconstruction, menis-
seal is attempted and often supplemented using fibrin cal resection or repair and surgery to the articular
glue before the suspension of cultured chondrocyte is cartilage. Although the ideal objective is to use MRI
injected under the patch. Excellent long-term results as the method of choice for assessing the problem-
have been reported in a cohort up to 10 years, but there atic postoperative knee in preference to arthroscopy,
are few results from randomised trials. there are potential pitfalls to correct interpretation.
There has been a steady increase in the referral rates
for MRI of the postoperative knee, which has resulted
in a gradual increase in the required knowledge for
accurate assessment. At the same time, improvements
in MR technology, new sequences, dynamic assess-
ment and MR arthrography are all helping to increase
the specificity of MRI in these postoperative states.
11.3.1
Cruciate Ligament Reconstruction
the graft material merely acts as a scaffold, wh ich In PCL reconstructions the ligamentisation process
allows it to become invaded by synovium, resulting in as identified by MRI is similar, but it takes much
neo-vascularisation and eventually neo-ligamentisa- longer, requiring an average of 24 months. MR stud-
tion (Fig. 11.8). In stage 1 (1-3 months postopera- ies of tunnel placement in PCL reconstruction have
tively), proliferating soft tissue is seen around the shown that improper femoral tunnel location and
graft, which retains its typical tendinous low signal clinical signs of instability are statistically significant.
MR characteristics. In stage 2 (3-9 months postop- The tibial tunnel is usually well placed, but proper
eratively), the graft becomes progressively hyperin- location of the femoral tunnel seems to be more criti-
tense, while in stage 3 (12 months postoperatively), cal owing to the lack of specific anatomicallandmarks
complete ligamentisation usually occurs (Fig. 11.9). at surgery (MARIANI et al. 1999). Contrast-enhanced
a b
Fig.l1.8. Coronal (a) and sagittal (b) Tl-weighted images of a well-placed BPTB ACL reconstruction which is fixed by biodegrad-
able screws. Note the low signal of the graft
a b
Fig.ll.9. Coronal (a) and sagittal (b) Tl-weighted images 9 months after ACL reconstruction, showing ligamentisation
174 V. N. Cassar-Pullicino and S. N. J. Roberts
studies, although interesting, are not required in the SMITH 1991; DREZ etal.I99l}.Insuchinstances,apart
successfully trans plan ted ligament, and should be from the assessment of the cruciate graft integrity, a
reserved for instances of suspected graft impinge- further search needs to be made to exelude associ-
ment where the internal changes within the liga- ated meniscal, chondral and osteochondral injury. In
ment are seen to enhance. There is no requirement cases where there is elearly elinical evidence of resid-
for routine MRI of the reconstructed cruciate liga- ual instability, imaging will be required to assess
ment as elinical assessment with conventional radio- the postoperative bony and soft tissue anatomy in
graphs is adequate. When the postoperative course preparation for arevision of a failed cruciate liga-
or surgical outcome is deemed unsatisfactory, either ment reconstruction. Although MRI is on ce again
by the patient or by the surgeon, MRI is required the mainstay of this investigation, CT mayaiso be
(Figs. 11.1 0, 11.1l). There is often a low correlation required in the full assessment of the bone contours
between the surgeon's evaluation and the patient's and tunnels.
perception of knee stability and function.
Knee stiffness and residual knee instability are 11.3.1.1
the two most common indications for postoperative Imaging Technique
imaging in the early postoperative period, usually
within the first 6 months and uncommonly up to 12 Conventional radiographs of the postoperative knee
months. Knee stiffness, which is usually heralded by are obtained before MR assessment to provide infor-
the failure to achieve full knee extension, requires the mation concerning the location of the femoral and
exelusion of roof impingement of the graft, arthrofi- tibial tunnels, the method of fixation, the type of graft
brosis and the "cyelops" lesion. Residual instability in material used, exelusion of radi opaque loose bodies
the first 6 months is usually the result of poor surgical and localisation of the patella. The mainstay of the MR
technique, a failure of graft incorporation or a tear investigation is the acquisition of a full assessment
of the graft, which can result from aggressive reha- of the knee to inelude the articular surfaces, menisci
bilitation. Failure which occurs more than 1 year after and other ligaments, along with detailed assessment
surgery is usually due to trauma, which occurs after of the neo-ligament. Conventional spin echo or fast
resumption of full activity. Traumatic re-injury has spin echo techniques are commonly used to evaluate
been reported in about 5%-10% of the athletic popu- the knee postoperatively. Gradient echo images are
1ation (JOHNSON et al. 1996). These patients usually not ideally suited owing to the sensitivity to metal
present with a new haemarthrosis following an injury artefacts, which are quite common in this type of
and elinically have an increased laxity. It needs to operation. Sagittal proton density-weighted (TR-TE,
be stressed, however, that during the first year after 2600/16) and T2-weighted turbo spin echo (TSE,
reconstruction, the graft strength and stiffness are 2600/98) images, along with coronal Tl-weighted spin
30%-50% of normal and excessive loads can lead echo sequence (500/12), coronal fat -suppressed proton
to plastic deformation and elongation (CLANCY and density-weighted (TRTE 2600/16) and coronal T2-
Fig.11.10. Tl- (a) and T2-weighted (b) sagittal sequences showing widening of the tibial tunnel, fragmentation of the biodegrad-
able screw and bone block, as weil as the formation of an aseptic fluid collection anterior to the tibia
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 175
weighted turbo spin echo (TSE 2600-98) sequences not routinely used, however, and is probably of lim-
are obtained, along with an axial T2-weighted fast ited clinical value except in cases of suspected graft
spin echo sequence. Although probably not required impingement. Recently, contrast-enhanced sequen-
routinely, oblique coronal and sagittal images can also tial MRI studies, performed over a 2-year period in
be obtained along the course of the graft using one of successful ACL reconstructions, have shown a 9%
the previous images as the scout, which allow visuali- quantitative enhancement 2 weeks postoperatively,
sation of the entire graft in its intra-articular portion an average 50% enhancement between 12 and 52
on a single image (Fig. 11.12). weeks postoperativelY and an average 65% enhance-
The process of successful biological incorporation ment at 1 year, with a reduction to 25% enhancement
of autografts, allografts and synthetie grafts includes after 76 weeks (VOGL et al. 2001). Low-field "niehe"
graft necrosis, revascularisation, cellular repopula- magnets and more recently "open" magnets allow
tion, collagen deposition and matrix remodelling the dynamic assessment of the knee and intra-artieu-
(CORSETTI and JACKSON 1996; ARNOCZKY et al. 1994). lar structures during joint function. Where available,
Assessment of this "ligamentisation" process can in these can add a further dimension to the status of the
part be done using intravenous gadolinium-DTPA to neo-ligament and in particular, help establish graft
assess the degree of enhancement of the graft and failure or impingement. Although MR arthrography
the peri-ligamentous tissues. Gadolinium-DTPA is has an increasing role in the assessment of the post-
176 V. N. Cassar-Pullicino and S. N. J. Roberts
c d
Fig.ll.12. The use of oblique sagittal (a, b) and oblique coronal images (c, d) obtained using Tl-weighted sequences to optimise
visualisation of the intra-articular portion of the ACL reconstruction
operative knee and certainly depicts the outline of effect on the postoperative result of ACL reconstruc-
the neo-ligament optimally, it is in the assessment of tion and ideally it should be positioned in the postero-
the meniscus and articular cartilage that this tech- medial portion of the original footprint of the ACL.
nique is of most use. The tunnel should be seen as parallel, but also posteri-
or to Blumensaat's line when the knee is in full exten-
11.3.1.2 sion, so as to eliminate the potential for graft impinge-
Successful Neo-ligament ment. Conventional radiography can diagnose several
potential complications of the arthroscopic autolo-
Conventional radiographs in the immediate postop- gous bone-patellar tendon-bone (BPTB) technique
erative period after ACL reconstruction confirm the used for ACL reconstructions. These include patellar
near-anatomical placement of the tibial and femoral fracture, migration of the bone plug and hardware
tunnels, the correct placement and orientation of the failure (GRAF and UHR 1988; MANASTER et al. 1988).
fixation devices utilised and the absence of significant Furthermore, if the lateral radiograph shows that the
bony debris within the joint, and identify abnormali- intra-articular opening of the tibial tunnel is anterior
ties related to the harvest site with particular refer- to the roof of the notch (Blumensaat's line), graft
ence to the patella. Ideal femoral tunnel placement is impingement between the graft and the anterior edge
as posterior in the notch as possible without violation of the intercondylar notch can be expected.
of the posterior cortical wall (blow-out). Placement Over the last 10 years or so there has been a grad-
of the tibial tunnel is now known to have a profound ual increase in the understanding of the MR changes
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 177
within the neo-cruciate ligament, which are tempo- length of the graft. The increased T2 signal is never
rally related and reflect the evolving biological pro- as bright as fluid, and residual continuous low-signal
ces ses that take place within a successful graft recon- fibres can be identified traversing the length of the
struction. After the initial pessimism towards the graft on these T2-weighted images. Arecent study
potential role of MRI, and at times conflicting reports, (JANSSON et al. 2001) has shown that on T2-weight-
there is now general agreement on the expected ed and STIR images the peri-ligamentous tissue is
appearances during the early period of the graft, typically seen as intermediate signal intensity streaks
during its incorporation and ligamentisation, and along and within the graft. In their study, JANSSON
after its maturation at 2 years. et al. found that 10 of their 20 cases showed interme-
In the immediate postoperative period the graft diate signal intensity within the intra-articular por-
(BPTB or STG) is visualised as a broad band oflow tion of the graft on proton density- and Tl-weighted
signal intensity, approximately equalling that of the images. Seven of these ten cases, however, had a ham-
patellar tendon. It is low signal on all pulse sequences string tendon autograft, and indeed previous studies
in the first 3 months (MINK et al.1993; HOWELL et al. concerning the maturation of STG tendon autografts
1991a). Richly vascularised peri-ligamentous tissue have demonstrated a variety of imaging findings on
is seen to cover the graft within 1 month after sur- postoperative MRI, including a high signal intensity
gery (HOWELL et al. 1991b, 1995). After the initial 3- area within the intra-articular portion of the graft
to 4-month postoperative period, some unimpinged as well as the graft within the tibial bone tunnel
intact grafts remain low in signal intensity during the (MURAKAMI et al. 1998, 1999). These articles show
first 3 years after surgery (HOWELL et al. 1991 b). How- that the maturation of the bone-tendon interface in
ever, a number of articles have shown that between the tibial bone tunnel establishes itself earlier than
3 and 9 months after surgery, as the ligamentisation the maturation within the intra-articular segment of
process progresses and evolves, the graft increases the graft.
in signal intensity and at times may become indis- After 8-12 months the ligamentisation process
tinguishable from surrounding tissues on the proton should have evolved and maturation of the graft is to
density- and Tl-weighted images (RAK et al. 1991; be expected irrespective of its origin (BPTB or ham-
YAMATO and YAMAGISHI 1992; SCHATZ et al. 1997; string autograft). There is histological evidence that
CASSAR-PULLICINO et al. 1994; TuITE and DE SMET by 12 months after autogenous ACL reconstruction,
1996; STOCKLE et al. 1998). Owing to the revasculari- the graft resembles the intact ACL (JOHNSON 1993).
sation of the graft, the cellular incorporation and the MURAKAMI et al. (1999) also showed that after 12
synovial proliferation in this postoperative period, months the entire hamstring autograft appeared as
the MR signal is expected to alter on the Tl- and a low signal intensity bundle on the proton density-
proton density-weighted images. At MRI this liga- weighted spin echo images at 0.2 T. There is, there-
mentisation process is more likely to be seen in fore, general agreement that the ACL autograft matu-
STG autografts rather than BPBT repairs, most likely ration should be complete 12 months after surgery
because peri-ligamentous tissue develops more read- and should be depicted as low signal intensity on the
ily around each of the four strands of the hamstring proton density-weighted images.
autograft than around the single BPBT strand (JANS- The normally aligned neo-ligament in its intra-
SON et al. 2001). This biological activity during the articular course should lie parallel and immediately
3- to 9-month postoperative period is also reflected posterior to the roof of the intercondylar notch. The
in the appearances of the stable unimpinged graft on tibial tunnel should also be parallel and posterior
the T2-weighted images. In instances where the graft to the intercondylar roof (Blumensaat's line) as seen
appears as a low Tl and proton density signal, the on the sagittal images. The anterior wall of the tibial
appearances of a low T2 signal are to be expected. tunnel ideally should He just posterior to Blumen-
However, when the revascularisation process takes saat's line so that the opening of the tibial tunnel
place, evidence of ligamentisation is provided by an is between 42% and 50% of the anteroposterior dis-
increased T2 signal within the graft and peri-liga- tance from the anterior edge of the tibia (HOWELL et
mentous tissues. This increased signal can be dif- al. 1991a; CASSAR-PULLICINO et al. 1994; MANASTER
fuse, but can also occur in a segmental distribution et al. 1988). The intra-articular portion of the graft
(CHEUNG et al. 1992; YAMATO and YAMAGISHI 1992). should enter the femoral tunnel at the posterosupe-
These authors showed that in some instances, full- rior margin of the intercondylar notch. The position
thickness revascularisation of the graft occurs while of the femoral tunnel ideally should be at the intersec-
at times there is a non-uniform distribution along the tion of the posterior femoral cortex with the poste-
178 V. N. Cassar-Pullicino and S. N. J. Roberts
rior physeal scar, as seen on the sagittal MR sequence sity- or Tl-weighted images within the stable asymp-
(MANASTER et al. 1988). On the coronal MR images tomatic cruciate reconstructions (SCHICK et al. 1995).
the femoral tunnel should be at the 11 o'dock posi- It is probably unlikely that the increased signal seen
tion in the right knee and the 1 0' dock position in the in some instances in unimpinged stable asymptomat-
left knee. The position of this femoral tunnel is crit- ic ACL grafts is due to the so-called magic angle phe-
ical in ensuring near isometry, which is a prerequi- nomen on that can occur in tendons and ligaments
site for the maintenance of a constant length and ten- (ERICKSON et al. 1993).
sion of the graft throughout the range of flexion and Posterior cruciate ligament reconstructions are
extension of the knee. Significant deviation from this carried out less often than ACL reconstructions. The
femoral tunnellocation will prevent isometry of the reconstruction is indicated particularly in athletes
graft, rendering the knee eventually unstable. The and in those who demonstrate symptomatic instabil-
tibial tunnel on the coronal MR images should open ity. The tom PCL is also ideally reconstructed using a
on the intercondylar eminen ce of the tibia in the mid- patellar tendon graft, although this is not always POS-
line. sible in view of the increased length that is required
Artefacts at the time of performance of MRI are and technical difficulties in a PCL reconstruction. On
to be expected. The presence of metal interference MRI the intact graft has been shown to be uniformly
screws within the femoral and tibial tunnels will gen- oflow signalintensity on Tl-weighted images (MUNK
erate a varying degree of artefact, which predudes et al. 1992), and the stable neo-ligament, unlike the
complete assessment of the tunnels. There is usually, ACL, remains similar in appearance to the patellar
however, no interference in the assessment of the tendon after surgery, exhibiting low signal on all MR
intra-articular component of the graft. Non-metal pulse sequences (TUITE and DE SMET 1996). The fem-
fixation devices allow better assessment of fixation oral tunnel should open just anterior to the femoral
and the graft within the tunnels. It is not unusual to insertion of the native PCL while the tibial tunnel
have microscopic fragments of metal within the joint should open in the joint in the mid-line within the
and these will give rise to a varying degree of MR small depression which lies just posterior to and
susceptibilityartefacts (Fig.l1.13). The gradient echo below the articular surface (MANASTER et al. 1988).
sequences are more vulnerable to this type of arte- This is a developing field in achallenging area of
fact. MR sequences obtained on more modern scan- reconstructive surgery of the knee and recently use of
ners utilising shorter echo times can give rise to focal double bundles has been recommended (RACE and
areas of increased signal intensity on the proton den- AMIS 1998).
incongruence at the patellofemoral and tibiofibular stable neo-ligament can have increased signal on the
articulations (JOHNSON 1993). proton density-weighted and short TE MR images.
Biological failure presents as recurrent instability Partial tears of the neo-ligament can also occur: they
in a patient without a history of a new episode of are highlighted by areas of increased fluid-like signal
trauma, and in whom there is no evidence of a on the T2-weighted images, affecting only part of the
surgica1!technical error in the placement of the graft. graft substance with residual intact continuous fibres
Biological failure results when the ligamentisation still present. MAYWOOD et al. in 1993 correcdy iden-
process is inadequate and as a result the graft scaf- tified such partial tears of the neo-ligament using
fold is easily elongated and rendered biomechanical- oblique sagittal T2-weighted images along the plane
ly unsound and prone to damage. A number of fac- of the graft. However, CHEUNG et al. and YAMATO
tors affect the ligamentisation of the graft, most of and YAMAGISHI in 1992 demonstrated a wavy con-
which cannot be assessed by MRI; these include type tour with peripheral high T2 signal in asymptomatic
of graft, biomechanical loads during rehabilitation, stable grafts within the first year after surgery, which
fixation and host response. MRI in graft rupture could have been misinterpreted as partial graft tears
shows an absence of the expected continuity of the byMRI.
intact graft fibres. Apart from discontinuity, absence Apart from biological, mechanical and structural
of the ligament within the osseous tunnels and failure of the graft, joint laxity can also arise from
marked bowing of the ligament are further signs of improper placement of the osseous tunnels, improp-
graft failure. These signs are useful within the first er sizing of the graft length and loosening or fracture
3 months of surgery, or after maturation has taken of the fixation with particular reference to the bone
place 1 year after surgery. On T2-weighted images, plug of the BPTB autograft. These possibilities can
increased signal intensity iso-intense to fluid is pres- also be assessed adequately by MRI.
ent within the intra-articular portion of the graft In summary, at any postoperative stage, a graft
(MUNK et al. 1992; YAMATO and YAMAGISHI 1992; exhibiting continuity of structure and a homoge-
TUITE and DE SMET 1996); this is the most specific neous low signal intensity from the tibial tunnel
sign of an acute complete tear of the graft (Fig.l1.16). across the intra-articular component to the femoral
As has been previously described, the T2-weighted tunnel is indicative of an intact graft. After the first
sequence is very important in differentiating the year the graft is also expected to have matured and
revascularisation between the 3- and the 12-month should exhibit primarily low signal. In the interven-
postoperative period as the intact and unimpinged ing period, particularly between the third and ninth
182 V. N. Cassar-Pullicino and S. N. J. Roberts
a b
Fig.l1.l6. Tl- (a) and T2-weighted (b) sagittal images showing a clear complete failure of the ACL graft
months postoperatively, increased signal intensity peri-ligamentous cysts can form within the tunnels,
may be seen as anormal finding owing to the revas- presumably arising from invagination of synovium,
cularisation and ligamentisation process with an with synovial fluid within the bonytunnels (MURAKA-
intact unimpinged graft. However, increased signal MI et al.1999). Although unusual, occasionally one also
intensity demonstrated on Tl- and proton density- identifies areas of bone marrow oedema around the
weighted sequences in such instances can be indica- enlarged bony tunnel; this may signify a true osteo-
tive of disrupted grafts as well as intact, but impinged lytic process, possibly secondary to stress shielding
grafts. It is for this reason that the integrity of the at this site. Demonstration of this abnormal tunnel
ACL graft cannot be determined on the basis of by MRI is an important consideration when planning
simply the proton density- or Tl-weighted images. revision surgery (FAHEY and INDELICATO 1994).
T2-weighted images are essential. Without the T2-
weighted images, proton density- and Tl-weighted
images may lead to a 50% false-positive diagnosis of 11.3.2
a graft rupture (JANSSON et al. 2001). The Postoperative Meniscus
Fig. 1l.17a, b. There is impingement on the anterior aspect of the graft by an osteophyte arising from the intercondylar roof,
which was treated by a notchplasty. Also note the widening of the tibial tunnel
Fig. 1l.18a, b. Two examples in the coronal and sagittal plane showing a
trumpet-shaped tibial tunnel with abnormal widening in the region of the
intercondylar eminence. Note the healed Segond fracture in a
184 V. N. Cassar-Pullicino and S. N. J. Roberts
is carried OUt. Only the unstable meniscal fragment er, it has been replaced by MR arthrography, as weIl as
is removed and in particular, the torn portion of the dynamic assessment of the knee to enhance the accu-
meniscus that is protruding into the joint is excised racy rate by diminishing the degree of uncertainty in
completely. The surgeon attempts to shave, fashion distinguishing the expected postoperative appearanc-
and trim the residual meniscal rim from anterior es from pathological meniscal states.
to posterior in an attempt to make it as perfectly There are two fundamental MR criteria for diag-
smooth as possible to reduce the risk of problems nosing a meniscal tear in the unoperated knee: dem-
postoperatively. More recently, meniscal transplanta- onstration of an increased internal signal on a short
tion and meniscal replacement, e.g. using collagen TE image unequivocally contacting an articular sur-
meniscus implant, have been introduced as options face of the meniscus, and an abnormal meniscal
in instances of significant complex meniscal tears shape. However, although these two criteria, if pres-
(POTTER et al. 1996). ent, will have an accuracy of over 90%, they cannot
Over the last decade, as arthroscopic refinements be applied with the same degree of confidence in the
and concepts of meniscal tear treatment have evolved, postoperative state. Following meniscal debridement,
there has been a gradual increase in the knowledge of rasping, partial meniscectomy or meniscal repair,
the expected MR appearances after meniscal resection conventional MRI of the postoperative meniscus has
and repair. Familiaritywith these expected MR appear- been shown to be unreliable (EGGLI et al. 1995;
ances is essential before embarking on an assessment BRONSTEIN et al. 1992; TSAI et al. 1992), with accu-
of the postoperative meniscus as they underlie the cor- racies ranging between 38% and 80%. This is in
rect interpretation of the MR appearances (Fig.11.19). part because abnormal meniscal morphology is to be
Initially, conventional MRI was used in the assess- expected after a previous tear which has been treat-
ment of the postoperative meniscus. Gradually,howev- ed by partial meniscectomy. Furthermore, when a
a D...._ _ __ ....._ _
b
_ ____ _ _ d
c
Fig. 11.19. Tl- and T2-weighted coronalimages (a, b) and Tl- and T2-weighted sagittal images (c, d) showing a meniscal "ghost"
on the Tl-weighted sequences due to the presence of fluid following a complete meniscectomy
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 185
partial meniscectomy is performed, the hyperintense finding on the short TE sequences (Tl and proton
intra-substance (grade I or II) signal intensity that density) and can be confused with a meniscal tear.
was present within the meniscus pre-operatively can Access to the pre-operative MRI, the arthroscopic
be converted into a seemingly grade III signal inten- findings and surgical operative details, use of MR
sity simulating a meniscal tear (Figs. 11.20, 11.21). arthrography, and dynamic assessment using an
It is more likely than not that, following injury, open MR magnet all help in reducing the uncertain-
a meniscus will never return to its normal pre-inju- ty, enabling correct interpretation of postoperative
ry signal intensity on Tl- or proton density-weight- meniscal findings.
ed images (ARNOCZKY et al. 1994; MUELLNER et al.
1999). This means that even after a successful menis- 11.3.2.1
cal repair, the healed meniscus may still show a Post-meniscectomy Assessment
hyperintense signal within its substance at the site of
the repaired tissue, which will appear as an expected Recurrent symptoms following meniscal surgery can
be evaluated by conventional MRI, conventional
arthrography, MR arthrography or arthroscopy. As
Signal Conversion previously indicated, the grade III signal intensity
alterations seen on conventional MRI in a post-
operative meniscus on short TE images (Tl and
proton density) are unreliable indicators of a re-tear.
I. Pre-op Although demonstration of a grade III signal inten-
sity on T2-weighted sequences on conventional MRI
is a good predictor of a recurrent or residual tear, one
needs to remember that the sensitivity is only about
60% (FARLEY et al. 1991). Indeed, FARLEY et al. con-
cluded that conventional arthrography is more useful
than conventional MRI in the assessment of the post-
operative meniscus.
Multiple authors have stressed the importance of
11. Post-op knowing the extent of meniscal resection as this
affects the accuracy of conventional MRI in the
assessment of the meniscus after partial meniscec-
tomy (Fig. 11.22). Although, without question, this is
___ ___ All OlWthlt2001
very valuable information at the time of reporting
~_~_~
a b
Fig.ll.21. Tl- (a) and T2-weighted (b) coronal sequences following partial meniscectomy. The meniscus has been "converted",
so the grade II intrinsic changes can be misinterpreted as a grade III tear. No tear was identified at arthroscopy
b
and TOTTY (1990), there are some reasonable guide- nificant number of false-positive and false-negative
lines that one can follow. When less than 25% of the diagnoses of recurrent tears. The authors coined the
meniscus has been resected, it can be evaluated using term "signal conversion", whereby an apparent grade
the MRI criteria employed in the pre-operative state. III signal can be generated following arthroscopic
In these studies virtually 90% of the menisci exhib- resection in a meniscus that demonstrated a grade
iting grade III signal were shown by arthroscopy to I or II signal abnormality originally. In view of this
have suffered a re-tear. When the partial meniscec- signal conversion, the authors proposed two criteria
tomy is more extensive (more than 30%), the stan- which can be used to distinguish a re-tear from a
dard MRI criteria used in the unoperated meniscus stable postoperative meniscus. In the first place, dem-
become significantly less accurate. DEUTSCH et al. onstration of a displaced meniscal fragment is to be
(1992) divided the more extensive meniscectomy regarded as a highly specific but insensitive finding
patients (more than 30%) into two groups, depend- in patients with recurrent meniscal tear. Secondly, if
ing on whether they had had a resection of less than the grade III signal abnormality seen on the short
or more than 75% of the meniscus. In the group with TE images also corresponds to an area of high signal
less than 75% meniscal resection, neither the pres- on the T2-weighted images, then it can be reason-
ence of internal grade III signal contacting the menis- ably assumed that a re-tear is present (Fig. 11.23).
cal surface nor abnormalities of meniscal s~ape were Although the demonstration of this linear high T2
as accurate in predicting a recurrent meniscal tear as signal within the meniscus is probably the most
in the unoperated state, with the generation of a sig- common criterion that is used in the diagnosis of
a b
Fig. 11.23. Tl- (a) and T2-weighted (h) coronal images showing a re-tear following a previous partial meniscectomy. Note the
associated early degenerative changes
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 187
recurrent tear on conventional MR images, one needs the short TE images at a site which is known to
to remember that it is only present in 33% of tears in have been normal on the pre-operative MRI, then it
non-operated menisci (MINK et al. 1998). In a com- can be safely conduded that this represents a new
parative assessment with arthrography and conven- meniscal tear. Similar results and condusions were
tional MRI, FARLEY et al. (1991) showed that the reached by ApPLEGATE et al. in 1993, but on ce again
T2 demonstration of a meniscal tear was the most these authors highlighted the important role of MR
common finding, while ApPLEGATE et al. (1993) used arthrography in distinguishing a healed meniscal
this criterion in post-meniscectomy patients who had repair from one that has re-tom. In a comparative
a joint effusion, and reported a 56% sensitivity and study utilising conventional MRI and conventional
90% specificity for diagnosis of re-tom menisci. In arthrography in evaluating meniscal repair, VAN
the group of patients in whom more than 75% of TROMMEL et al. in 1998 reported a small series of
the meniscus has been removed, conventional MRI patients in whom the MR technique was more accu-
has an accuracy rate of 85% with arthroscopy. The rate than arthrography in discriminating partial or
specificity for a re-tear of a meniscus when abnormal complete healing. These authors used fast spin echo
signal iso-intense with fluid is seen on T2-weighted fat-suppressed sequences to evaluate the meniscal
images dearly suggests that there is a valuable role repair, but only eight cases had arthroscopic confir-
for MR arthrography in the assessment of the post- mation of their findings. Like FARLEY et al. in 1991,
operative meniscus following partial meniscectomy. VAN TROMMEL et al. (1998) stressed the importance
ApPLEGATE et al. (1993) performed MR arthrograms of a line of high T2 signal intensity iso-intense to
in their postoperative patients and increased the sen- fluid within the meniscus as indicating incomplete
sitivity for meniscal re-tears from 56% to 90%, while healing of a meniscal repair. In a somewhat larger
maintaining a specificity of dose to 90%. series of 20 patients, LIM et al. (1999) carried out
a retrospective study which compared five specific
11.3.2.2 findings on conventional MRI with second-Iook
Post-meniscal Repair arthroscopy in patients evaluated for possible repeat
tears of the postoperative meniscus. The authors
The potential for healing of meniscal te ars exists found that the two signs with the highest correlation
particularly within the vascularised outer third of were (a) signal extending to the meniscal surface
the meniscus. Small stable peripheral tears iden- on a proton density-weighted image and (b) fluid
tified, for example, at the time of ACL reconstruc- intensity signal on a T2-weighted image extending
tion are treated completely conservatively and they into the meniscus. As found in previous studies,
heal spontaneously. There have been marked refine- postoperative abnormalities in meniscal morpholo-
ments in the per arthroscopic surgical techniques gy were of little value in predicting recurrent tears.
to repair peripherallarge unstable tears. As in the
post-meniscectomy state, assessment by conven- 11.3.2.3
tional MRI of the repaired meniscus can be diffi- MR Arthrography
cult. Morphologically the meniscus is more likely
to appear normal, but the healed te ar will still pro- Studies evaluating the postoperative meniscus by
duce a grade III signal appearance on the short TE conventional MRI have shown the high specificity of
images. Second-Iook arthroscopy has shown that the a grade III signal within the meniscus which demon-
meniscus has healed and is indeed stable (DEUTSCH strates signal intensity that is iso-intense with fluid
et al. 1990) in these circumstances. FARLEY et al. on the T2-weighted sequence. It follows, therefore,
(1991) also showed a grade III signal abnormality that direct MR arthrography utilising dilute gado-
on proton density- and Tl-weighted images in 27 linium DTPA (Gd-DTPA) with saline in a ratio of
of 29 cases of repaired menisci, conduding that the approximately 1:250 will increase the accuracy of
presence of signal contacting the articular surface MRI. Recently, there has been an ever-increasing
of the meniscus is not a reliable predictor for recur- interest in developing the role of indirect MR
rent or residual meniscal tear. Demonstration of a arthrography following the intravenous injection
high signal iso-intense with fluid on T2-weighted of Gd-DTPA in assessing the postoperative knee.
images predicted a recurrent meniscal tear, arepair ApPLEGATE et al. (1993) showed that 66% of postop-
that had not healed or a new te ar with a sensitivity erative patients presenting with a new knee pain
of 60% and a specificity of 92% (FARLEY et al. 1991). had a recurrent meniscal te ar confirmed by arthros-
Clearly if a grade III signal abnormality is seen on copy. The greatest value of MR arthrography lies in
188 V. N. Cassar-Pullicino and S. N. J. Roberts
Fig.l1.25. Advanced osteoarthritic changes demonstrated in the medial compartment following meniscectomy on Tl- (a) and
T2-weighted (b) sagittal sequences
ing, which encourages bleeding in the hope that the lar cartilage surface, while dynamic intravenous stud-
resultant blood clot will stimulate the development ies can help to monitor the results of treatment in
of a fibrocartilage repair. TUITE and DE SMET (1996) instances of autogenous osteochondral transplanta-
reported that the MR appearance of this fibrocar- tion.As surgical techniques become more refined and
tilage exhibits a higher signal on the T2-weighted acceptable, it is likely that there will be a growing
images when compared with the meniscal fibrocarti- demand for non-invasive methods of monitoring the
lage. More recently, cartilage repair procedures have results of treatment. However, it remains to be seen
been developed and are being constantly refined. The whether MR in its monitoring role does affect the
two most commonly employed are autologous chon- outcome of these therapeutic procedures, or whether
drocyte transplantation and osteochondral autograft it is useful in directing the treatment of overgrowth
transplantation (OATS, mosaicplasty). Different sur- of the transplanted cartilage before this becomes
gical techniques are employed in achieving these symptomatic.
repair procedures, placing unique demands on the MR studies following autologous chondrocyte
choice of MR pulse sequences postoperatively, and transplantation show a variable degree of artefact on
creating problems in the correct interpretation of the the surface of the graft, which interferes with accurate
postoperative MR images. The optimum sequences assessment of the thickness and surface contour of
for depiction of contour and intrinsic cartilage detail the repair tissue (Fig. 11.26). Fat-suppressed proton
should be utilised. In addition, in some instances MR density-weighted fast sequences are very useful in
arthrography can provide better detail of the articu- minimising this artefact, allowing adequate morpho-
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 191
Fig. 11.26. Autologous chondrocyte transfer in the medial femoral condyle. Note good thickness at the operated site with the
presence of small metal artefacts on the surface
logical assessment (RECHT et al. 2001). Study of the Osteochondral autogenous graft transfers have
ACI graft should include an estimation of the volume recently been evaluated using conventional radiog-
of graft that fills the original defect, surface contour raphy, CT and MRI with dynamic Gd-DTPA intrave-
restoration, status of subchondral bone and neigh- nous enhancement (VERSTRAETE et al. 2000). In addi-
bouring bone marrow oedema (Fig. 11.27). Sequen- tion, the postoperative MR appearances have also
tial MRI studies allow the identification of partial been correlated with the postoperative clinical status
and complete graft delamination with or without the (SANDERS et al. 2001). In about 50% of instances radi-
use of MR arthrography (Fig. 11.28) (ALPARSLAN et ography detects the donor and recipient sites, while
al. 2001). MRI also reliably detects graft hypertrophy. CT and MRI virtually show all the grafts cleady at the
This can occur by extension into the intercondylar recipient sites. The size of the graft, its relationship
notch, or when overgrowth of the graft occurs, usu- with the subchondral bone and cartilage contour can
ally at the edges with the native cartilage, particu- be evaluated. Subchondral alignment occurs in less
lady in the non-weight-bearing surfaces. The rele- than 40% of grafts, while cartilage surface contour
vance and interpretation of signal changes within the alignment is perfect in 80% of cases. As the osteo-
repaired cartilage remains a topic of active research. chondral "plug" is taken from a different part of
At present MRI is unable to determine accurately the the joint surface it usually has a different thickness
tissue type (s) within the graft. of articular cartilage. At the time of placement the
a b
Fig.l1.27a, b. Satisfactory reconstitution of the medial femoral condylar cartilage following autologous chondrocyte transplan-
tation. Note the heterogeneous signal within the cartilage bed
192 V. N. Cassar-Pullicino and S. N. J. Roberts
Fig. 1l.28a, b. Partial delamination at the site of previous successful autologous chondrocyte transplantation
emphasis is to ensure cartilage contour continuity, and early revascularisation at 4 weeks with improvement
for this reason it is not surprising that there is often a at 6 weeks. The development of oedema within the
discrepancy in the alignment of the subchondral bone graft along with this enhancement pattern and even-
of the plug with the recipient area (Fig. 11.29). CT is tual return of normal fatty marrow signal at 1 year
better at showing the impaction status of the osseous indicates normal graft incorporation. After an initial
component of the plug, its incorporation and the revascularisation phase the graft undergoes a phase
formation of peripherally located resorption cysts. of resorption which in turn is followed by an incor-
MRI shows bone marrow oedema around the plug in poration phase highlighted by osseous remodelling
the first few days, which increases between 3 and 9 and repair (EINHORN 1995).
months following surgery and then gradually disap- In 20% of OATS cases, graft protuberance or
pears. The transplanted osteochondral plug retains depression ranging between 1 and 2 mm can be
its normal fatty marrow signal intensity on MR stud- expected. This does not appear to change with time
ies done within 2 weeks, but at 4 weeks heteroge- on subsequent MR investigations, nor does it appear
neous signal is noted, highlighted by decreased Tl to have any bearing on the short-term clinical out-
and increased T2 appearances. At the 6th week post- come.Arthroscopic studies in the follow-up of osteo-
operatively, the plugs show a uniformly decreased chondral plug transfer have reported a fibrocartilage-
signal intensity on Tl-weighted sequences and a uni- like tissue which fills the gaps between the transplant -
formly increased T2 signal intensity. At 5 months ed osteochondral plugs, improving the congruity of
after surgery the plugs on ce again appear heteroge- the joint surface (HANGODY et al. 1997). Short-term
neous as there is a gradual return of the normal fatty follow-up MR studies and CT arthrography,however,
marrow signal, usually commencing within the cen- have not confirmed this finding, and it may be a phe-
tral portion of the plug. Successfully transplanted nomen onthat develops with time. The osteochondral
plugs demonstrate normal fatty marrow signal 1 year surface at the donor sites appears irregular on MRI,
after the surgery (Fig. 11.30). Dynamic Gd-DTPA with very litde alteration on follow-up examinations.
enhancement of the marrow surrounding the graft There is usually no or very poor reconstitution of
is intense starting at 2 weeks following surgery, and bone here, with at best only a thin fibrocartilaginous
gradually decreases with time so that by 6-9 months tissue within the crater of the defects. Donor site
postoperatively there is minimal to no enhancement bone marrow oedema is generally less conspicuous
in the surrounding marrow. This enhancement pat- than graft site oedema. A variable degree of micro-
tern is thought to parallel the revascularisation of the metal artefact can be identified in some instances at
plug, which is initially avascular and demonstrates both the donor and the recipient site.
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 193
Fig. 11.30. Multiple osteochondral grafts have been implanted within the
patella (a). The donor sites seen in (b) do not show any evidence of any sig-
nificant reconstitution of the osteochondral surface. A sagittal Tl-weighted
image (c) 1 year after plug transfer shows a good cartilage surface and a
return of the fatty signal within the patella. Axial images (d, e) show the
multiple mosaicplasty appearances following surgery
b c
d e
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 195
11.3.4.3 11.3.4.4
DonorSites Patelia/Pateliofemoral Joint
Alterations oeeur within the soft tissues where har- Patellar maltraeking, assessed by axial imaging (CT
vesting for ACL repair has taken plaee. Donor site or MRI) in static and kinematic studies, is not uncom-
regeneration of the semi-tendinosus and gracilis mon postoperatively. Residual lateralisation of the
tendons following their transeetion has been shown patella often persists despite primary repair of the
to oeeur clinieally and objeetively. MRI has shown medial soft tissue struetures for acute patellar dislo-
that the tendons do re-grow, inserting as a diffuse eation, or after a generous lateral retinaeular release
fan-shaped attaehment into the medial popliteal for patellar maltracking. Less often the patella can
fascia, and they are also thought to resurne their drift medially postoperatively as a result of an over-
funetion (CROSS et al. 1992). Following BPTB ACL zealous lateral release, which can be clinically signifi-
reeonstruetions, the immediate post-harvest sono- cant. Patellar maltraeking can also be seen after BPTB
graphie and MR appearanee of the residual tendon harvesting, but it may have existed prior to the sur-
is abnormal. Adefeet at the site of rem oval of the gery. It is eommon to see a low-Iying patella (patella
eentral one-third of the tendon is easily diseern- baja) in the first 6 months after a BPTB procedure,
ible, especially sonographically. The tendon dem- with areturn to the normal position by 12 months
onstrates diffuse thickening with inereased Tl and (MINK et al. 1993). Transversely orientated stress frac-
T2 MR signals for up to 12 months. In the asymp- tures have been reported and rarely complete avul-
tomatie healed state the tendons remain thiekened sive fractures of the patella can also eomplicate the
but the low MR signal eharaeteristies of the normal harvesting procedure if a large bone plug has signifi-
tendon should return. Sonographically, the abnor- cantly weakened the patellar bone strength.
mal eehogenicity of the tendon diminishes after 1
year and the margins of the defeet beeome less dis-
tinet. Ongoing symptoms may be due to patellar
tendonitis, or rarely rupture of the patellar tendon, 11.4
both of whieh require ultrasound or MRI for diag- Conclusion
nosis.
Studies utilising ultrasound (WILEY et al. 1997), This chapter has particularlyfocussed on cruciate lig-
MRI and histology of tendon biopsies at the pre- ament surgery and assessment following such repair,
viously harvested site within the patellar tendon meniscal surgery, and the assessment of articular car-
have shown that at 2 years the defeet of the eentral tilage disorders and their treatment. There are other
one-third of the tendon beeomes indistinguishable intra-articular eauses of ongoing symptoms, includ-
from the rest of the normal tendon (NIXON et al. ing loose bodies, osteoneerosis of the femoral and
1995; ADRIANI et al. 1995; KARTUS et al. 2000). tibial condyles and reflex sympathetie dystrophy, all
In about 25% of eases there is persistenee of the of which have been eovered elsewhere in the book.
tendon defeet (LIU et al. 1996; BERNICKER et al.
1998). In addition, histologieal assessment at the
bone insertion points shows an absenee of the
References
normal fibroeartilage zone at the end of the repaired
tendon. These morphological changes have also Adriani E, Mariani PP, Maresea G, Santori N (1995) Healing
been shown in animal studies (PROCTOR et al. of the patellar tendon after harvesting of its mid-third for
1997), wh ich indicate that the repaired portion of anterior eruciate ligament reeonstruetion and evolution of
the tendon does not have the same ultra-struetural the unclosed donor site defeet. Knee Surg Sports Traumatol
Arthrose 3:l38-143
and biomeehanical properties as the normal tendon
Allen PR, Denham RA, Swan AV (1984) Late degenerative
by the second year of the postoperative period. In ehanges after meniseeetomy. Faetors affeeting the knee
the most recent study of 19 patients using ultra- after operation. J Bone Joint Surg Br 66:666-671
sound, MRI and histology, the authors eoncluded Alparslan L, Minas T, Winalski CS (2001) Magnetic resonanee
that reharvesting of the healed patellar tendon at imaging of autologous ehondroeyte implantation. Semin
Ultrasound CT MR 22:341-351
least up to 2 years after the primary harvest is not
Applegate GR, Flannigan BD, Tolin BS, Fox JM, Dei Pizzo
recommended for revising ACL reeonstructions as W (1993) MR diagnosis of reeurrent tears in the knee:
this may be biomeehanically unsound (KARTUS et value of intraarticular eont material. AJR Am J Roentgenol
al. 2000). 16:821-825
196 V. N. Cassar-Pullicino and S. N. J. Roberts
Arnoczky SP, Cooper TG, Stadelmaier DM, Hannafin JA (1994) Eggli S, Wegmuller H, Kosina J, Huckell C, Jakob RP (1995)
Magnetie resonance signals in healing menisci: an experi- Long-term results of arthroscopie meniscal repair.An anal-
mental study in dogs. Arthroscopy 10:552-557 ysis of isolated tears. Am J Sports Med 23:715-720
Bellelli A, Adriani E, Margheritini F, Camillieri G, Della Rocca Einhorn TA (1995) Enhancement of fracture-healing. J Bone
C, Mariani pp (1999) Synovial healing in reconstructed Joint Surg Am 77:940-956
cruciate ligaments. Our personal experience compared in Eriekson SJ, Prost RW, Timins ME (1993) The "magie angle"
single interventions and combined reconstruction (in Ital- effect: background physies and clinieal relevance. Radiol-
ian). Radiol Med (Torino) 98:454-461 ogy 188:23-25
Bergfeld JA, McAliister DR, Parker RD, Valdevit AD, Kambie Fahey M, Indelieato PA (1994) Bone tunnel enlargement after
HE (2001) A biomechanieal comparison of posterior cruci- anterior cruciate ligament replacement. Am J Sports Med
ate ligament reconstruction techniques. Am J Sports Med 22:410-414
29:129-136 Fairbank TJ (1948) Knee joint changes after meniscectomy. J
Bernieker JP, Haddad JL, Lintner DM, DiLiberti TC, Bocell JR Bone Joint Surg Br 30:664-670
(1998) Patellar tendon defect during the first year after ante- Farley TE, Howell SM, Love KF, Wolfe RD, Neumann CH
rior cruciate ligament reconstruction: appearance on serial (1991) Meniscal tears: MR and arthrographie findings after
magnetie resonance imaging. Arthroscopy 14:804-809 arthroscopie repair. Radiology 180:517-522
Bolano LE, Grana WA (1993) Isolated arthroscopie partial Graf B, Uhr F (1988) Complications of intra-artieular anterior
meniscectomy. Functional radiographie evaluation at five cruciate reconstruction. Clin Sports Med 7:835-848
years. Am J Sports Med 21:432-437 Hangody L, Kish G, Karpati Z, Szerb I, Udvarhelyi I (1997)
Bronstein R, Kirk P, Hurley J (1992) The usefulness of MRI Arthroscopie autogenous osteochondral mosaieplasty for
in evaluating menisci after meniscus repair. Orthopaedies the treatment of femoral condylar artieular defects. A pre-
15:149-152 liminary report. Knee Surg Sports Traumatol Arthrosc
Cassar-Pullicino VN, McCail IW, Strover AE (1994) MRI of 5:262-267
the knee following prosthetie cruciate ligament reconstruc- Hanks GA, Gause TM, Sebastianelli WJ, O'Donneli CS, Kalenak
tion. Clin RadioI49:89-99 A (1991) Repair of peripheral meniscal tears: open versus
Cassidy RE, Shaffer AJ (1981) Repair of peripheral meniscus arthroscopie technique. Arthroscopy 7:72-77
tears. A preliminary report. Am J Sports Med 9:209-214 Harter RA, Osternig LR, Singer KM, James SL, Larson RL, Jones
Cheung Y, Magee TH, Rosenberg ZS, Rose DJ (1992) MRI of DC (1988) Long-term evaluation of knee stability and func-
anterior cruciate ligament reconstruction. J Comput Assist tion following surgieal reconstruction for anterior cruciate
Tomgr 16:134-137 ligament insufficiency. Am J Sports Med 16:434-443
ClancyWG Jr, Smith L (1991) Arthroscopie anterior and poste- Howell SM, Clark JA (1992) Tibial tunnel placement in anterior
rior cruciate ligament reconstruction technique. Ann Chir cruciate ligament reconstructions and graft impingement.
GynaecoI80:141-148 Clin Orthop 283:187-195
Corsetti JR, Jackson DW (1996) Failure of anterior cruciate Howell SM, Berns GS, Farley TE (1991a) Unimpinged and
ligament reconstruction: the biologie basis. Clin Orthop impinged anterior cruciate ligament grafts: MR signal
325:42-49 intensity measurements. Radiology 179:639-643
Coupens SD, Yates CK, Sheldon C, Ward C (1992) Magnetie Howell SM, Clark JA, Blasier RD (1991b) Serial magnetie res-
resonance imaging evaluation of the patellar tendon after onance imaging of hamstring anterior cruciate ligament
use central one-third for anterior cruciate ligament recon- autografts during the first year of implantation. A prelimi-
struction. Am J Sports Med 20:332-335 nary study.Am J Sports Med 19:42-47
Covall DJ, Wasilewski SA (1992) Roentgenographie changes Howell SM, Knox KE, Farley TE, Taylor MA (1995) Revas-
after arthroscopie meniscectomy: five year follow-up in cularization of a human anterior cruciate ligament graft
patients more than 45 years old. Arthroscopy 8:242-246 during the first two years of implantation. Am J Sports Med
Cross MJ, Roger G, Kujawa P,Anderson IF (1992) Regeneration 23:42-49
of the semitendinosus and gracilis tendons following their Howell SM, Gittins ME, Gottlieb JE, Traina SM, Zoellner TM
transection for repair of the anterior cruciate ligament. Am (2001) The relationship between the angle of the tibial
J Sports Med 20:221-223 tunnel in the coronal plane and loss of flexion and anterior
Dandy DJ, Jackson RW (1975) Meniscectomy and chondro- laxity after anterior cruciate ligament reconstruction. Am J
malacia of the femoral condyle. J Bone Joint Surg Am Sports Med 29:567-574
57:1116-1119 Jansson KA, Karjalainen PT, Harilainen A, Sandelin J, Soila K,
DeHaven KE (1985) Meniscus repair in the athlete. Clin Orthop Tallroth K,Atonen HJ (2001) MRI of anterior cruciate liga-
198:31-35 ment repair with patellar and hamstring tendon autografts.
Deutsch AL, Mink JH, Fox JM, Arnoczky SP, Rothman BJ, Skeletal RadioI30:8-14.
Stoller DW, Cannon WD Jr (1990) Peripheral meniscal tears: Johnson LL (1993) The outcome of a free autogenous semiten-
MR findings after conservative treatment or arthroscopie dinosus tendon graft in human anterior cruciate reconstruc-
repair. Radiology 176:484-488 tive surgery: a histologieal stydy. Arthroscopy 9: 131-142
Deutsch AL, Mink JH, Fox JM, Friedman MJ, Howell SM (1992) Johnson DL, Bealle D (1999) Meniscal allograft transplanta-
The postoperative knee. Magn Reson Q 8:23-54 tion. Clin Sports Med 18:93-108
Drez DJ Jr, DeLee J, Holden JP, Arnoczky S, Noyes FR, Johnson DL, Either DB, Vanarthos WJ (1996) Herniation of
Roberts TS (1991) Anterior cruciate ligament reconstruc- the patellar fat pad through the patellar tendon defect
tion using bone-patellar tendon-bone allografts. A biologi- after autologous bone-patellar tendon-bone anterior cruci-
cal and biomechanieal evaluation in goats. Am J Sports Med ate ligament reconstruction. A case report. Am J Sports Med
19:256-263 24:201-204
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 197
Kannus P, Jarvinen M (1987) Conservatively treated tears of Race A, Amis AA (1998) PCL reconstruction. In vitro biome-
the anterior cruciate ligament. Long-term results. J Bone chanical comparison of "isometrie" v single and double-
Joint Surg Am 69:1007-1012 bundled "anatomie" grafts. J Bone Joint Surg Br 80:173-179
Kartus J, Rostgard-Christensen L, Movin T, Lindahl S, Ejerhed Rak KM, Gillogly SD, Schaefer RA, Yakes WF, Liljedahl RR
L, Karlsson J (2000) Evaluation of harvested and normal (1991) Anterior cruciate ligament reconstruction: evalua-
patellar tendons: a reliability analysis of magnetic res- tion with MR imaging. Radiology 178:553-556
onance imaging and ultrasonography. Knee Surg Sports Rangger C, Klestil T, Gloetzer W, Kemmler G, Benedetto KP
Traumatol Arthrosc 8:275-280 (1995) Osteoarthritis after arthroscopie partial meniscec-
Kitziger KJ, DeLee JC (1990) Failed partial meniscectomy. Clin tomy. Am J Sports Med 23:240-244
Sports Med 9:641-660 Recht MP, Piraino DW, Cohen MA, Parker RD, Bergfeld JA
Lajtai G, Humer K, Aitzetmuller G, Unger F, Noszian I, Orthner (1995) Localized anterior arthrofibrosis (cyclops lesion)
E (1999) Serial magnetic resonance imaging evaluation of after reconstruction of anterior cruciate ligament: MR
a bioabsorbable interference screw and the adjacent bone. imaging findings. AJR Am J RoentgenoI165:383-385
Arthroscopy 15:481-488 Recht M, Bobie V, Burstein D, Disler D, Gold G, Gray M, Kramer
Lim PS, Schweitzer ME, Bhatia M, Giuliano V, Kaneriya PP, J, Lang P, McCauley T, Winalski C (2001) Magnetie res-
Senyk RM, Oliveri M, Johnson W, Amster B, Parker L (1999) onance imaging of artieular cartilage. Clin Orthop 391
Repeat tear of postoperative meniscus: potential MR imag- (Suppl):S379-S396
ing signs. Radiology 210:183-188 Riehardson JB, Caterson B, Evans EH, Ashton BA, Roberts S
Liu SH, Hang DW, Gentili A, Finerman GA (1996) MRI and (1999) Repair of human artieular cartilage after implan-
morphology of the insertion of the patellar tendon after tation of autologous chondrocytes. J Bone Joint Surg Br
graft harvesting. J Bone Joint Surg Br 78:823-826 81: 1064-1068
Manaster BJ, Remley K, Newman AP,Mann FA (1988) Kneelig- Roos H, Lauren M, Adalberth T, Roos EM, Jonsson K, Lohman-
ament reconstruction: plain film analysis. AJR Am J Roent- der LS (1998) Knee osteoarthritis after meniscectomy: prev-
genoI150:337-342 alence of radiographie change after twenty-one years com-
Mariani PP, Adriani E, Bellelli A, Maresca G (1999) Magnetic pared with matched controls. Arthritis Rheum 41:687-693
resonance imaging of tunnel placement in posterior cruci- Sanders TG, Mentzer KD, Miller MD, Morrison WB, Campbell
ate ligament reconstruction. Arthroscopy 15:733-740 SE, Penrod BJ (2001) Autogenous osteochondral "plug"
Maywood RM, Murphy BJ, Uribe JW, Hechtman KS (1993) Eval- transfer for the treatment of focal chondral defects: post-
uation of arthroscopic anterior cruciate ligament recon- operative MR appearance with clinieal correlation. Skeletal
struction using magnetic resonance imaging. Am J Sports RadioI30:570-578
Med 21:523-527 Schatz JA, Potter HG, Rodeo SA, Hannafin JA, Wiekiewiez TL
Mink JH, Deutsch AL (1989) Magnetic resonance imaging of (1997) MR imaging of anterior cruciate ligament recon-
the knee. Clin Orthop 244:29-47 struction. AJR Am J RoentgenoI169:223-228
Mink JH, Reicher MA, Crues JV III, Deutsch AL (1993) MRI of Schiek F, Dammann F, Lutz 0, Claussen CD (1995) Adapted
the knee, 2nd edn. Raven Press. New York techniques for clinieal MR imaging of tendons. MAGMA
Mink JH, Levy T, Crues JV 3rd (1998) Tears of the anterior 3:103-107
cruciate ligament and menisci of the knee: MR imaging Schimmer RC, Brulhart KB, Duff C, Glinz W (1998) Arthroscopie
evaluation. Radiology 167:769-774 partialmeniscectomy: a 12-year follow-up and two-step eval-
Muellner T, Egkher A, Nikolic A, Funovics M, Metz V (1999) uation of the long-term course. Arthroscopy 14: 136-142
Open meniscal repair: clinical and magnetie resonance Sciulli RL, Boutin RD, Brown RR, Nguyen KD, Muhle C, Lektr-
imaging findings after twelve years. Am J Sports Med akul N, Pathria MN, Pedowitz R, Resniek D (1999) Evalu-
27:16-20 ation of the postoperative meniscus of the knee: a study
Munk PL, Vellet AD, Fowler PJ, Miniaci T, Crues JV 3rd (1992) comparing conventional arthrography, conventional MR
Magnetie resonance imaging of reconstructed knee liga- imaging, MR arthrography with iodinated contrast mate-
ments. Can Assoc Radiol J 43:411-419 rial, and MR arthrography with gadolinium based contrast
Murakami Y, Sumen Y, Ochi M, Fujimoto E, Adachi N, Ikuta material. Skeletal RadioI28:508-5 14
Y (1998) MR evaluation of human anterior cruciate liga- Smith DK, Totty WG (1990) The knee after partial meniscec-
ment autograft on oblique axial imaging. J Comput Assist tomy: MR imaging features. Radiology 176:141-144
Tomogr 22:270-275 Stockle U, Hoffmann R, Schwedke J, Lubrieh J, Vogl T, Sudkamp
Murakami Y, Sumen Y, Ochi M, Fujimoto E, Deie M, Ikuta Y NP, Haas N (1998) Anterior cruciate ligament reconstruc-
(1999) Appearance of anterior cruciate ligament autografts tion: the diagnostie value of MRI. Int Orthop 22:288-292
in their tibial bone tunnels on oblique axial MRI. Magn Tsai CL, Liu CN, Lim AC (1992) Meniscal repair with autog-
Reson Imaging 17:679-687 enous periosteum and fibrin adhesive system (in Chinese).
Nixon RG, SeGall GK, Sax SL,Cain TE, Tullos HS (1995) Recon- Zhonghua Yi Xue Za Zhi (Taipei) 49: 170-176
stitution of the patellar tendon donor site after graft har- Tuite MJ, De Smet AA (1996) MR of the postoperative knee.
vest. Clin Orthop 317:162-17l Top Magn Reson Imaging 8:2-14
Potter HG, Rodeo SA, Wickiewicz TL, Warren RF (1996) MR Van Trommel MF, Simonian PT, Potter HG, Wiekiewiez TL
imaging of meniscal allografts: correlation with clinical (1998) Arthroscopic meniscal repair with fibrin clot of
and arthroscopie outcomes. Radiology 198:509-514 complete radial tears of lateral meniscus in the avascular
Proctor CS, Jackson DW, Simon TM (1997) Characterization of zone. Arthroscopy 14:360-365
the repair tissue after removal of the central one-third of Verstraete KL, Huysse WCJ, Steines D, Lang P, Declercq D,
the patellar ligament. An experimental study in goat model. Rousseaux M (2000) Assessment of osteochondral auto-
J Bone Joint Surg Am 79:997-1006 grafts of the knee with CT, conventional MRI, dynamie
198 V. N. Cassar-Pullicino and S. N. J. Roberts
contrast-enhanced and cartilage thiekness images. Sev- cruciate ligament reconstruction.Arthroscopy 15:474-480
enth Annual Meeting, European Society of Musculoskeletal Whitman TL, Diduch DR (1998) Transient posterior knee pain
Radiology, Leiden with the meniscal arrow. Arthroscopy 14:762-763
Vogl TJ, Schmitt J, Lubrieh J, Hochmuth K, Diebold T, Dei Tredi- Wiley JP, Bray RC, Wiseman DA, Elliott PD, Ladly KO, Vale
ci K, Sudkamp N (2001) Reconstructed anterior cruciate LA (1997) Serial ultrasonographie imaging evaluation of
ligaments using patellar tendon ligament grafts: diagnostie the patellar tendon after harvesting its central one third
value of contrast-enhanced MRI in a 2 year follow-up regi- for anterior cruciate ligament reconstruction. J Ultrasound
men. Eur Radiolll: 1450-1456 Med 16:251-255
Warden WH, Friedman R, Teresi LM, Jackson DW (1999) Mag- Yamato M, Yamagishi T (1992) MRI of patellar tendon ante-
netie resonance imaging of bioabsorbable polylactic acid rior cruciate ligament autografts. J Comput Assist Tomogr
interference screws during the first 2 years after anterior 16:604-607
12 The Postoperative Knee 2:
Arthroplasty, Arthrodesis, Osteotomy
THOMAS H. BERQUIST
_ _ _......
b
a ~ ~~_:.:..-;
Fig. 12.1. a Photograph of porous-coated condylar knee system with all poly-
ethylene patellar component (1), metal tibial tray (2) with cutout for posterior
cruciate ligament (arrow) and femoral component (3) with two condylar pegs
(sm all arrows). Standing AP (b) and lateral (c) radiographs of the left knee
demonstrating a cemented condylar knee arthroplasty. The patellar compo-
nent is all polyethylene (1); the tibial tray (2) is metal with a polyethylene
spacer (arrow). The two condylar pegs (arrows) are visible on the femoral
c
component (3), indicating the condylar design
cruciate ligament is resected. A central tibial post and patients, has largely been abandoned in favor of tita-
box in the femoral component (Fig. 12.2) provide sta- nium- and cobalt-based alloys. Cobalt-chromium
bility in this semieonstrained system (BERQUIST 1995; or cobalt-chromium-molybdenum alloys and tita-
COYTE et al.1999). nium-aluminum-vanadium alloys are most com-
Constrained prostheses are used in patients with monlyused in conjunction with ultra-high molecu-
inadequate soft tissue support or bone loss or for revi- lar weight polyethylene artieular surfaces (BERQUI-
sion arthroplasty. Linked or nonlinked designs are ST 1995; BRYAN and PETERS 1973; FRIEDMAN et al.
available (Fig. 12.3). Fully hinged designs (Fig. 12.3) 1993).
frequently fail. Component design varies for tibial, femoral, and
patellar implants. Tibial components may be sym-
metric (Fig. 12.4) or anatomie (Fig. 12.5) (medial
12.2.3 and lateral polyethylene inserts configured to the
Component Selection size of the tibial condyle). Tibial trays may be metal
backed with varying thickness of inserts (Figs. 12.4,
Considerable research has gone into developing 12.5) or all polyethylene (Fig. 12.6). Components
the most appropriate alloys and plastic for joint may be designed for ingrowth, in which case screws
implants. Stainless steel, though effective in elderly can be used for additional fixation or for cement
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 201
Fig. 12.3. a Frontal and b lateral photographs of a hinged Guepar prosthesis. c AP radiograph of a loose Guepar prosthesis with
cement fracture and lucency (arrows) due to toggling of the components
fixation (Fig. 12.7). Augmentation sterns or wedges Femoral eomponents may be porous coated to
ean be used when there is bone loss (Fig. 12.8). allow bone ingrowth or they may be designed for
Systems are also available for unieompartmental eement fixation. Symmetrie and anatomie designs
arthroplasty (ApEL et al. 1991; BERQUIST 1995; are available, similar to the tibial eomponents. Aug-
STUART 1991). mentation sterns are available when there is bone loss
202 Th. H. Berquist
Fig.12.4. Nonanatomie polyethylene spacers with the same size Fig.12.5. a Photograph of Genesis tibial traywith artieular surface
and configuration for both condyles for anatomie polyethylene insert. b Polyethylene insert is asym-
metrie to match the anatomie configuration of the condyles
a b
Fig.12.6. AP (a) and lateral (b) radiographs of a cemented posterior stabilized knee arthroplasty. The tibial component is
polyethylene and appears lucent (arrows) compared with the opaque cement
or for revision procedures (BERQUIST 1995; COYTE Patellar surface replacement is still most commonly
et al. 1999). There are fewer loosening problems with performed to reduce pain and prevent later complica-
femoral components than with tibial ones (BERQUIST tions (BARRACK et al.1997; BERQUIST 1995; MERKow
1995; COLLIER et al. 1991). et al. 1985). Patellar components may be all polyethyl-
Patellar resurfacing may not be required. Recent ene or metal backed (Fig. 12.9). Most surgeons prefer
reports suggest that the results of patellar resur- cemented, all polyethylene inserts (see Fig.12.11)
facing and retention of the patella are similar. How- (BAUER and SCHELS 1999; BERQUIST 1995).
ever, future repair is still required in 10% of patients Unicompartmental arthroplasty has regained some
(BARRACK et al.1997). support in recent years (BERQUIST 1995). This tech-
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 203
Fig.12.10. AP (a) and lateral (b) radiographs of a medial compartment arthroplastywith porous-coated metal-backed components
12.3 12.3.1
Postoperative Imaging and Complications Complications
Anteroposterior (AP) and lateral radiographs are Complications of total knee arthroplasty are sum-
obtained in the immediate postoperative period to marized in Table 12.1. Certain complications can be
check alignment and component position and to treated conservatively. Loosening, infection, osteoly-
exclude obvious problems such as dislocation or peri- sis, fracture, and instability require more aggressive
prosthetic fracture (BERQUIST 1995; SCHNEIDER et therapy (BAUER and SCHELS 1999; BERQUIST 1995;
al. 1984). When the patient can tolerate a complete BOYD et al. 1993; COYTE et al. 1999; ELLINGSEN and
knee series this should be obtained to provide a base- RAND 1994; GILLS and MILLS 1991; KIM 1990).
line study. It should include standing AP (most prefer
a full-Iength hip to ankle study), lateral, and patellar Table 12.1. Complications of total knee arthroplasty
views. We prefer to fluoroscopically position the AP Wound healing
and lateral views to assure optimal visualization of Infection
the component-bone and bone-cement interfaces. Extensor mechanism dysfunction
Flexion and extension lateral views can be obtained Loosening
Instability
to grade range of motion. Patients are usually restud-
Fractureldislocation
ied at 6 and 12 months. Yearly follow-up is adequate Deep venous thrombosis
after the first year unless symptoms dictate more fre- Pulmonary emboli
quent evaluation. Pes anserinus bursitis
Specific features should be evaluated on each radio- Synovitis
Osteolysis
graphicview. The position of the tibial andfemoral com-
Peroneal nerve palsy
ponents and the femoral-tibial angle should be assessed
on the standing AP view of the knee (Fig.12.11a). The BAUER and SCHELS (1999); BERQuIsT (1995); BoYD et
tibial tray should be at 90° to the tibial axis and cover al. (1993); COYTE et al. (1999); CUCKLER et al. (1991);
ELLINGSEN and RAND (1994); ENGH and AMMEEN (1997);
+85% of the bony condyle. If the tibial tray is >90° it is GILLS and MILLS (1991); IDUSUYI and MORREY (1996); KIM
considered in valgus and if it is <90° it is considered in (1990); SAMBATAKAKIS et al. (1991); WILSON et al. (1990)
varus. The femoral component should be 97-98° to the
femoral axis on the AP view. The femoral-tibial angle
should be at 5-10° valgus (BERQUIST 1995; COLLIER et
al.1991; PATEL et al.1991).
The position of the tibial tray, femoral compo- 12.3.2
nent, and patella should also be assessed on the lat- Extensor Mechanism Dysfunction
eral view. The tibial tray should be at 90° to the tibial
axis (Fig. 12.11 b). If the angle is >90° it is considered Extensor mechanism dysfunction is the most
extended and if <90°, flexed. The femoral axis should common complication of total knee arthroplasty,
be perpendicular to a line along the base of the com- accounting for up to 50% of revision procedures
ponent (Fig. 12.11 b). Patellar height should be about (BARRACK et al. 1997; BERQUIST 1995; MERKow et al.
the same as the patellar tendon length. Motion can 1985). Most complications occur in the first 2 years
also be assessed with flexed and extended lateral after arthroplasty (BERQUIST 1995; BLOEBAUM et al.
views. Approximately 65° of flexion is required to 1998; BOYD et al. 1993). Complications include soft
climb stairs and 70° for anormal gait. Descending tissue imbalance with abnormal patellar position
stairs requires 85° and rising from a sitting position, and/or abutment on the femoral condyle (Fig.12.12),
105° of flexion (BERQUIST 1995; BURGER et al. 1991). dislocation, fracture, loosening (Fig. 12.13), compo-
Lucent zones should be assessed on AP and later- nent wear, and failure or rupture of the patellar
al views. Fluoroscopic positioning is most effective tendon or quadriceps mechanism (Fig.12.l4) (BER-
to achieve optimal evaluation of the bone-cement or QUIST 1995; BLOEBAUM et al. 1998; BOYD et al. 1993;
metal-bone interfaces. MERKow et al. 1985; SAMBATAKAKIS et al. 1991).
The patellar view should be evaluated for symme- Serial radiographs are generally adequate for diagno-
try, patellar tilt, and loosening (Fig. 12.l1c). sis (BERQUIST 1995).
206 Th. H. Berquist
Fig.12.12. Normal (a) and abnormal (b) patellar views. There is abutment
(arrow) due to extensor mechanism imbalance in b
a
12.3.3
Loosening
12.3.4
Infection
c
Infections may be superficial (wound infection) or
deep. The former usually occur in the early postop-
erative period and are less likely to require compo- Fig.12.15. AP (a), lateral (b), and patellar (c) views show
nent removal. lucency about the posterior aspects of the tibial and femoral
Deep infection is a much more serious problem. components (arrows). There is also cement fragmentation
(open arrow in a). The patellarview (c) shows extensor mecha-
Patients usually present with pain and swelling over
nism imbalance with abutment (arrow) laterally
aperiod of months. In large series the incidence of
deep infection ranges from 1% to 19% (BAUER and
SCHELS 1999; BENGSTON and KNUTSON 1991; BER-
QUIST 1995; RAND 1993; WILSON et al. 1990). The
208 Th. H. Berquist
overall incidence at our institution is 1.2% (BER- or sulfur colloid scans combined with indium-lU
QUIST 1995). The incidence of infection also varies labeled leukocytes (BERQUIST 1995; MANASTER 1995;
with the type of implant used. Hinged prostheses PALESTRO et al. 1991). The accuracy of combined
became infected in 19%; stabilized components in technetium and leukocyte scans approaches 96%
5.1 %; resurfacing designs in 2.8%, and unicompart- (Fig.12.17) (BERQUIST 1995; PALESTRO et al. 1991;
mental components in only O. 8%. The incidence of WILSON et al.1990).
infection with revision is also higher (9%) (BERQUIST Infections are most often due to Staphylococcus
1995; RAND 1993; WILSON et al. 1990). aureus (64%), Pseudomonas aeruginosa (12%), Esch-
Radiographs are normal in nearly 75% of patients erichia co li (5%), and an aerobic organisms (6%)
with infection (BERQUIST 1995; MANASTER 1995). (BAUER and SCHELS 1999; BENGSTON and KNUTSON
However, serial radiographs may demonstrate irreg- 1991; BERQUIST 1995; PALESTRO et al.1991). Manage-
ular lucency about the components, signs of loosen- ment varies with patient condition and virulence of
ing, and soft tissue swelling (Fig. 12.16). (BAUER and the organism (BERQUIST 1995; CUCKLER et al. 1991;
SCHELS 1999; BERQUIST 1995; MANASTER 1995). RAND 1993; WILSON et al. 1990). Antibiotic therapy
Radionuclide imaging is useful in patients with without component rem oval may be successful if
suspected infection (BENGSTON and KNUTSON 1991; infection is caused by low-virulence organisms when
KIM 1990; MINIACI et al. 1989; STUART 1991). components are not loose. With more virulent organ-
Most prefer a dual-radionuclide technique employing isms, component removal, debridement, and intra-
technetium-99m methylene diphosphonate (MDP) articular and intravenous antibiotics are indicated
b c
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 209
(Fig.12.17) (BERQUIST 1995; CUCKLER et al. 1991). ponent loosening, or be due to the component
Arthrodesis can also be used in younger patients, type. Fractures are more common with hinged pros-
patients with greater mobility, and those with viru- theses (BERQUIST 1995; HUE and FITZGERALD 1990).
lent infections (CUCKLER et al. 1991; ELLINGSEN and They are most common in the supracondylar region
RAND 1994; NICHOLS et al. 1991; RAND 1993). (Fig. 12.18) but tibial and patellar fractures also occur.
Component fracture (Fig.12.19) is uncommon and
usually related to adjacent bone pathology. Aggres-
12.3.5 sive surgical treatment of fractures is preferred over
Fractures bed rest or long periods of nonambulation (COR-
DEIRO et al. 1990; GHAZAVI et al. 1997).
The incidence of fracture following knee arthroplas-
ty is 1.2-3% (BERQUIST 1995; CORDEIRO et al. 1990;
ENGH and AMMEEN 1997; SCHNEIDER et al. 1984). 12.3.6
Fractures may occur during component insertion. In Instability
this setting, fractures are usually undisplaced and
they can be treated during the arthroplasty (ENGH Ligament instability has been reported in up to 13%
and AMMEEN 1997). Postoperative fractures may of patients following knee arthroplasty (BERQUIST
result from systemic dis orders (osteoporosis) or com- 1995; SAMBATAKAKIS et al. 1991). Preoperative defor-
210 Th. H. Berquist
12.3.7
b Other (omplications
12.3.8
Deep Venous Thrombosis
12.4
Arthrodesis
12.5
Osteotomy
a
Unilateral or unicompartmental osteoarthritis may
be treated with tibial or supracondylar femoral oste-
otomy. The rationale of this technique is to decrease
the load to the involved compartment and transfer
the weight to the uninvolved side of the joint (BER-
QUIST 1995; COVENTRY 1984, 1987).
Osteotomy is typieally considered in younger
patients «65 years old) with unicompartmental pain
(medial or lateral) whieh correlates with radiographie
involvement (BERQUIST 1995; COVENTRY 1984, 1987).
Supracondylar femoralosteotomies are reserved for
disorders originating in the femur with valgus defor-
mity (Fig.12.22) (BERQUIST 1995; COVENTRY 1984,
1987; MINIACI et al. 1989). Contraindieations for
tibial osteotomy include severe peripheral vascular
disease, reduced range of motion, contracture, signif-
ieant bone loss, and ligament instability (BERQUIST
1995; COVENTRY 1984,1987).
Femoral and tibial osteotomies can be performed
with several fixation systems (Figs. 12.22, 12.23)
including blade plates, buttress plates, and screw fixa-
b tion (COVENTRY 1984,1987; MINIACI et al. 1989).
Most studies demonstrate success rates of 85% fol-
lowing osteotomy (MINIACI et al. 1989). However, it is
Fig. 12.21a, b. Failed compression arthrodesis treated with
plate and screw fixation. AP (a) and oblique (b) views show not unusual for joint degeneration to progress. Other
fixation with pin tracts (arrowheads) from previous external complications include delayed union (Fig.12.22) or
fixation nonunion (Fig. 12.24), loss of reduction, infection,
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 213
b,c
a
d e,f
h,i
g
Fig. 12.22a-i. Supracondylar femoral osteotomy. a Standing AP radiograph demonstrates valgus deformity on the right with
lateral compartment arthrosis. AP (b) and lateral (c) postoperative radiographs following varus osteotomy with 90° blade plate
and cortical screws for fixation. AP (d, f, h) and lateral (e, g, i) radiographs at 1 month (d, e), 5 months (f, g), and 10 months
(h, i), showing delayed union
214 Th. H. Berquist
a a
b b
Fig. 12.23a, b. Postoperative tibial osteotomy with L-buttress Fig.12.24. Nonunion demonstrated on AP (a) and lateral (b)
plate and screw fixation. The osteotomy line is weil defined radiographs
(arrows) on AP (a) and lateral (b) radiographs
plasty. 11. Mechanisms of implant failure. Skeletal Radiol Idusuyi OB, Morrey BF (1996) Peroneal nerve palsy after total
28:483-497 knee arthroplasty. J Bone Joint Surg 78A:177-184
Bengston S, Knutson K (1991) The infected knee arthroplasty. Insali JN, Dorr LD, Scott RD, Scott WN (1989) Rationale
A 6 year follow-up of 357 cases. Acta Orthop Scand of the knee society clinieal grading system. Clin Orthop
62:301-311 248:13-14
Berquist TH (1995) Imaging atlas of orthopedie appliances Kim Y (1990) Incidence of deep vein thrombosis after cement-
and prosthesis. Raven, New York less and cemented knee replacement. J Bone Joint Surg
Bloebaum RD, Backus KN, Jensen JW, Scott DF, Hofmann AA 72B:779-783
(1998) Porous-coated metal-backed patellar components in Kirgis A, Albrecht S (1992) Palsy of the deep peroneal
total knee arthroplasty. J Bone Joint Surg 80A:518-528 nerve after proximal tibial osteotomy. J Bone Joint Surg
Boyd AD, Ewald FC, Thomas WH, Poss R, Sledge CB (1993) 74A:1180-1185
Long-term complieations after total knee arthroplasty Manaster BJ (1995) Total knee arthroplasty: post-operative
with or without patellar resurfacing. J Bone Joint Surg radiologie findings AJR 165:899-904
75A:674-681 Merkow RL, Soundry M, Insali JN (1985) Patellar dislocation
Bryan RS, Peters on LFA (1973) Polycentrie total knee arthro- following total knee replacement. J Bone Joint Surg
plasty. Orthop Clin North Am 4:575-585 67A:1321-1327
Burger RR, Basch T, Hopson CN (1991) Implant salvage in Miniaci A, Ballmer FT, Ballmer PM, Jakob RP (1989) Proximal
infected total knee arthroplasty. Clin Orthop 273:105-112 tibial osteotomy: a new fixation deviee. Clin Orthop
Christensen OM, Christiansen TJ, Johansen T (1990) Polyeth- 246:250-259
ylene failure in PCA unieompartmental knee prosthesis. Niehols SJ, Landon GG, Tallos HS (1991) Arthrodesis with dual
Acta Orthop Scand 61:578-579 plates after failed total knee arthroplasty. J Bone Joint Surg
Collier JP, Mayor MB, McNamara JL, Surprenant VA, Jensen 73A: 1020-1 024
RE (1991) Analysis of the failure of 122 polyethylene Oishi CS, Grady-Benson JC, Otis SM, Colwell CW, Walker RH
inserts from uncemented tibial components. Clin Orthop (1994) The clinieal course of distal deep venous thrombosis
273:232-242 after total hip and total knee arthroplasty as determined by
Cordeiro EN, Costa RC, Carazzato JG, Silva JD (1990) Peripros- duplex ultra-sonography. J Bone Joint Surg 76A:1658-1663
thetie fracture in patients with total knee arthroplasties. Palestro CJ, Swyer AJ, Kim AK, Goldsmith SJ (1991) Infected
Clin Orthop 252:182-189 knee prosthesis: diagnosis with In-l11 leukocyte, Tc-99m
Coventry MB (1984) Upper tibial osteotomy. Clin Orthop sulfur coloid, and Tc-99m MDP imaging. Radiology
182:46-52 179:645-648
Coventry MB (1987) Proximal tibial varus osteotomy for osteo- Patel DV, Ferris BD, Aiehroth PM (1991) Radiologie study
arthritis of the lateral compartment of the knee. J Bone of alignment after total knee replacement. Int Orthop
Joint Surg 69A:32-38 15:209-210
Coyte PC, Hawker G, Croxford R, Wright JG (1999) Rates of Rand JA (1993) Alternatives to reimplantation for salvage of
revision knee replacement in Ontario, Canada. J Bone Joint total knee arthroplasty complieated by infection. J Bone
Surg 81A:773-782 Joint Surg 75A:282-288
Cuckler JM, Star AM, Alavi A, Noto RB (1991) Diagnosis and Sambatakakis A, Welton TJ, Newton G (1991) Radiographie
management of infected total joint arthroplasty. Orthop signs of persistent soft tissue imbalance after knee replace-
Clin North Am 22:523-530 ment. J Bone Joint Surg 73B:751-756
Ellingsen DE, Rand JA (1994) Intra-medullary arthrodesis of Schneider R, Abenovoli AM, Soundry M, Insall J (1984)
the knee after failed total knee arthroplasty. J Bone Joint Failure of total condylar knee replacement. Correlation
Surg 76A:870-877 of radiographie clinieal and surgieal findings. Radiology
Engh GA, Ammeen DJ (1997) Perprosthetie fractures adjacent 152:309-315
to total knee implants. J Bone Joint Surg 79A:1100-1113 Silverton C, Rosenberg AO, Barden RM, Sheinkop MB, Galante
Friedman RJ, Black J, Galante JO, Jacobs JJ, Skinner HB (1993) JO (1996) The prosthesis-bone interface adjacent to tibial
The current concepts in orthopedies, biomaterials, and components inserted without cement. J Bone Joint Surg
implant fixation. J Bone Joint Surg 75A:1086-1109 78A:340-346
Ghazavi MT, Stockly I, Yee G, Davis A, Gross AE (1997) Recon- Stuart MJ (1991) Indieations and patient selection. In: Morrey
struction of massive bone defects with allograft in revision BF (ed) Joint replacement arthroplasty. Churehill-Living-
total knee arthroplasty. J Bone Joint Surg 79A:17-25 stone, New York, pp 971-979
Gills ES, Mills DM (1991) Long-term follow-up evaluation of Weiner SD, Scarborough M, VanderGreend R (1996) Resection
1000 consecutive cemented total knee arthroplasties. Clin arthrodesis of the knee with intercalory allograft. J Bone
Orthop 273:66-76 Joint Surg 78A:185-193
Gross TP, Lennon DW (1992) Osteolytie cyst like area associ- Westrieh GH, Hidaka C, Winsor RE (1997) Disengagement of
ated with polyethylene and metal debris after total knee a locking screw from a modular stern in revision total knee
replacement with an uncemented Vitallium prosthesis. J arthroplasty. J Bone Joint Surg 79A:254-258
Bone Joint Surg 74A:I096-1102 Wilson MG, Kelley K, Thornhill TS (1990) Infection as a com-
Hue FC, Fitzgerald RH Jr (1990) Hinged total knee arthroplas- plieation of total knee replacement arthroplasty. J Bone
ty. J Bone Joint Surg 65A:513-519 Joint Surg 72A:878-883
13 Patellar and Quadriceps Mechanism
NIELS EGUND and LEIF RYD
Table 13.1. The Merchant dassification system of patellofemoral It has generally been accepted that radiographie
disorders assessment of the femorotibial joint spaces should be
I. Trauma (conditions caused by trauma performed in the standing, weight-bearing position
in the otherwise normal knee) with the knee in semifiexion. With the patient stand-
A. Acute trauma ing on one leg, this AP or PA radiographie view also
1. Contusion allows assessment of varus or valgus angulations and
2. Fracture subluxation. Clinieal examinations and tests as well
a) Patella as imaging of the patellofemoral joint are in general
b) Femoral trochlea
c) Proximal tibial epiphysis (tuberde)
performed in the non-weight-bearing supine posi-
3. Dislocation (rare in the normal knee) tion; however, it is uncertain and unexplored how
4. Rupture valid these measurements are for performance in the
a) Quadrieeps tendon standing position, in whieh dis orders develop and
b) Patellar tendon
may give rise to pain. Following the monograph of
B. Repetitive trauma (overuse syndromes)
1. Patellar tendinitis ("jumper's knee")
(AHLBACK 1968), the authors of the present chapter
2. Quadrieeps tendinitis were brought up with weight-bearing radiographie
3. Peripatellar tendinitis (e.g., anterior knee pain in the views of the knee including the patellofemoral joint,
adolescent caused by hamstring contracture) clinically as well as in research. Recently, during ongo-
4. Prepatellar bursitis ("housemaid's knee") ing research, we became aware of some major differ-
5. Apophysitis
a) Osgood-Schlatter disease
ences between weight -bearing and non-weight -bearing
b) Sinding-Larsen-Johansson disease imaging, whieh at present may be considered contro-
C. Late effects of trauma versial but will contribute to changes in the under-
1. Post-traumatie chrondromalacia patellae standing and treatment of patellofemoral disorders.
2. Post -traumatie patellofemoral arthritis
3. Anterior fat pad syndrome (post-traumatie fibrosis)
4. Reflex sympathetie dystrophy of the patella
5. Patellar osseous dystrophy
6. Acquired patella infera 13.2
7. Acquired quadrieeps fibrosis Anatomy
II. Patellofemoral dysplasia
A. Lateral patellar compression syndrome The patellofemoral joint consists of the trochlea
1. Secondary chondromalacia patellae (patellar groove, femoral sulcus) of the femur and the
2. Secondary patellofemoral arthritis patellar artieular surfaces (Fig. 13.1). The embryol-
B. Chronie subluxation of the patella ogy of the patella and the femoral trochlea is fascinat-
1. Secondary chondromalacia patellae
2. Secondary patellofemoral arthritis
ing (LANGER 1929), with early essential characteris-
C. Recurrent dislocation of the patella ties of the adult. Initially, the medial and lateral facets
1. Associated fractures are equal in size, but after 23 weeks' gestation the lat-
a) Osteochrondral (intra-artieular) eral facet tends to predominate. Before this time the
b) Avulsion (extra-artieular) lateral facet of the trochlea has a greater transverse
D. Chronie dislocation of the patella
1. Developmental
width and is anteriorly more prominent, although the
2. Acquired patella is positioned distal to and without contact to
the trochlea. During growth and following function,
III. Idiopathic chondromalacia patellae
the shape of the patella may be modified. WIBERG
IV. Osteochondritis dissecans (1941) described three configurations of the patellar
A. Patella joint (Fig.13.2). Wiberg type 11 (Fig.13.1) has a small-
B. Femoral trochlea
er medial facet and represents the most common
V. Synovial plieae (anatomie variants made symptomatie patellar anatomy. WIBERG suggested that type III may
by aeute or repetitive trauma) be correlated to chondromalacia, but this has not
A. Pathologie medial patellar pliea ("shelf") been confirmed. A number of different anatomieal
B. Pathologie suprapatellar plica
variations and ossification variants may be observed
C. Pathologie lateral patellar pliea
(KEATS 1996).
VI. Iatrogenie dis orders
A. Iatrogenie medieal patellar compression syndrome
B. Iatrogenie chronie medial subluxation of the patella
C. Iatrogenie patella infera
Patellar and Quadrieeps Mechanism 219
a b
Fig. 13.la, b. Normal patellofemoral joint in a 24-year-old male. a Axial radiographie view and b trans axial MR sectioning
through the middle of the patella obtained by a proton fat-saturated sequence with the knee in 20° of flexion. The lateral (L)
facets of the patella and the femoral condyle are larger than the medial facets (M), and the depth of the osseous condylar groove
is greater than the cartilaginous. The medial patellar cartilaginous facet is separated from the odd facet (0) by a ridge
a b
Fig. 13.2a, b. Axial radiographie views of two patellofemoral joints. Wiberg type I (a) is rare and the medial (M) and lateral
(L) facets are almost equal in size. Wiberg type III (b) has a large lateral (L) and a small medial (M) facet and is commonly
associated with a large suJcus angle. Type 11, the most common type, is shown in Fig. 13.1
13.2.1
Soft Tissue Stabilizers I :':;!~~:iml~ Rcctus fcmoris
The patella is the turning point of the converting reti- ~II"" __ \'astus mcdialis
nacular structures, which consist of ligaments, mus- \'astus lateralis
cles and joint capsule (Fig. 13.3). With the large
range of motion within the patellofemoral joint, the
joint capsule, with an extensive synovial expansion, Sartorius Ilmdon
is poorly defined and does not contribute to patellar
stabilization.
13.2.1.1
Active Stabilizers of the Patellofemoral Joint
superior third of the anterior surface of the patella, (Fig. 13.4). The orientation of the patellar tendon is
with some fibers continuing into the patellar tendon. parallel to the long axis of the lower extremity. The
The vasti medialis and lateralis originate from the retinacula are composed of superfidal and deep
superomedial and lateral aspect of the femur, respec- layers (Figs. 13.6, 13.7). The thin oblique superfi-
tively, and unite in the midline in asolid aponeuro- dal retinacula link the patella and patellar ligament
sis that inserts into the base and medial and lateral
aspect of the patella. In the most distal portion of
both the vastus medialis and the vastus lateralis there
are separate smaH muscle groups with an oblique
orientation of their muscle fibers, termed the vastus
medialis obliquus (Fig. 13.5) and vastus lateralis
obliquus, respectively. These provide a direct medial
and lateral puH on the extensor mechanism and are
important for patellar balance in the femoral troch-
lea. The vastus intermedius inserts with a thin but
broad tendon into the base of the patella posterior to
the other tendons.
13.2.1.2
Passive Stabilizers
The patellar tendon, the central portion of the Fig. 13.6. Normal MR anatomy of the patellar retinaculum on
quadriceps tendon, and the medial and lateral reti- a Tl-weighted Flash 2D sequence. The superior portion of the
nacula are the passive elements of soft tissue stabi- medial patellofemoralligament (MPFL) has a broad fibrocar-
tilaginous attachment to the odd facet (0) of the patella and
lization. The patellar ligament connects the patella is dorsally attached to the femoral condyle and medial collat-
with the tibial tuberosity and has a length of 4-6 cm, eralligament (MCL) with extension to the fascia of the dorsal
a width of 25-40 mm, and a thickness of 6-8 mm muscles. LPFL, Lateral patellofemoralligament
Patellar and Quadriceps Mechanism 221
a b
Fig. 13.7. Normal MR anatomy of the patellar (a) and infrapatellar (b) portions of the medial and lateral retinacula on a
proton fat-saturated sequence. PCL, Posterior cruciate ligament; ACL, anterior cruciate ligament; LR, lateral synovial recess;
PT, popliteal tendon; BF, biceps muscle and tendon; FCL, lateral collateralligament; S, sartorius muscle. Other abbreviations
are as in Fig. 13.6
medially to the tendon of the sartorius and the deep sured and calculated by numerous authors, with con-
fascia of the leg and laterally to the fascia lata. In the siderable divergence of quantitative values. These
deeper layers, capsular condensations form fibrous values depend on various factors, including body
layers that link the patella to the medial and lateral weight, quadriceps force, angle of flexion, and individ-
femoral epicondyles and the anterior aspect of ual anatomie factors. It is beyond the scope of this
the menisci (Fig. 13.49). They include the medial textbook to go into detail on these considerations, an
and lateral patellofemoralligaments and, below, the overview of which can be obtained from orthopedic
medial and lateral meniscotibial patellar ligaments. textbooks (FULKERSON 1997; SCUDERI 1995). Howev-
In between the medial patellofemoral and the menis- er, the patellofemoral joint re action force may vary
copatellar ligament, the deep transverse retinaculum
courses direct1y from the iliotibial tract to the patella.
It is supposed that the soft tissue stabilizers are stron-
gest on the lateral side. The normal MR and sono-
graphie appearances of the patellar retinaculum were
described in detail by STAROK et al. (1997).
13.3
Biomechanical Considerations
13.3.1
Patellar Tracking and the Q Angle
Roentgen stereophotogrammetric analysis, de- arthrosis: grade 1 represents softening and swelling
scribed below, allows very accurate measurements of the cartilage; grade 2 is cartilage breakdown (fibril-
of motion between rigid bodies in all three planes, lation) of one half inch or less; grade 3 is cartilage
including rotation (SELVIK 1974), and has been breakdown of greater than one half inch; and grade 4
extensively used in orthopedic and biomechanical is erosion of cartilage down to bone (OUTERBRIDGE
research. Using this technique, BLANKEVOORT et al. 1961). This classification is also useful in the grading
(1988) demonstrated that motion patterns of the of joint degeneration at MR imaging (Chap. 3). The
knee are highly susceptible to small changes in the term "chondromalacia patellae:'however,has become
externaiload configuration and specifically that the controversial since it may cover a large number of dis-
"screw horne" mechanism is not an obligatory effect orders leading to patellofemoral pain (RADIN 1979)
of the passive joint characteristics, but a direct result in which loss of patellar cartilage is not documented,
of the externailoads. This supports the observations is not present, or occurs in combination with other
reported by STEIN et al. (1993), who imaged fluoro- abnormalities such as bursitis (pre- and infrapatellar,
scopically the patellofemoral joint of healthy volun- pes anserinus), the plica syndrome, the fat pad syn-
teers in the anteroposterior plane during walking. drome, arthrosis, synovitis, and meniscal tears.
They recorded a uniform pattern of medial excursion The terms "patellar tilt-compression" and "exces-
of the patella relative to the femur, with a sudden shift sive lateral pressure syndrome" (ELPS) were intro-
from lateral to medial, and concluded that contrary duced by FICAT and HUNGERFORD (1977). These
to conventional understanding, the patella deviates conditions are characterized clinically by pain and
medially rather than laterally during walking. radiologically by lateral patellar tilt as evidenced on
axial patellofemoral radiography, CT, and MR imag-
ing (FULKERSON 1997). They are commonly associ-
ated with chondromalacia patellae and, when articu-
13.4 lar manifestations occur, the site of cartilage lesions
Clinieal and Pathologie Considerations are the same. Classically, it is stated that the medial
facet of the patella is the typical and primary site
The multifactorial etiologies of anterior knee pain of cartilage lesions in chondromalacia, particularly
can be related to variants of the anatomy of the patel- about the ridge that separates the medial and odd
la and alterations in the tensile forces of the extensor facets (WIBERG 1941; INSALL et al. 1976). The car-
mechanism applied to the joint surfaces of the patel- tilage changes on the medial patellar facet with a
lofemoral joint, generated during the complex move- supposed deficient contact to the femoral trochlea
ment of the joint. These alterations may be caused (HENCHE et al. 1981) have been attributed to various
by overuse, disuse, and injuries to bone, cartilage etiologic factors. Several authors have pointed out
and supporting soft tissue structures. Although the the tendency for cartilage that is out of contact with
classification of patellofemoral disorders (Table 13.1) other cartilage to undergo surface fibrillation (GOOD-
devised by MERCHANT (1988) provides a useful FELLOW et al. 1976) and lose an appropriate mecha-
framework for the clinical and imaging approach to nism of synovial fluid nutrition (FULKERSON 1983;
diagnosis and treatment, it is feIt that it fails to rec- LAURIN et al. 1978).
ognize and interpret the pathogenesis of the most
common dis orders of the patellofemoral joint.
Patellar articular cartilage presents different modes 13.4.1
of degeneration. Chondrosis signifies a disorder The Request for Imaging - History and Physical
affecting only the articular cartilage and arthrosis, a Examination
dis order affecting all three components of the joint:
cartilage, subchondral bone, and synovial membrane. Accurate, concise clinical evaluation of the patient
Chondromalacia patellae is a term applied to a syn- with a suspected knee disorders is almost invariably
drome of anterior knee pain in adolescents and suggestive of a working diagnosis. Together with the
young adults (ALE MAN 1928; WIBERG 1941) and the clinical information, this working diagnosis forms
pathoanatomical appearances are described as soft- the cornerstone for tailoring the radiological exami-
ening (malacia), edema, and swelling of the carti- nation and, indeed, for the interpretation of images.
lage. A proposed classification of the surgical and Therefore it is essential that the request for imaging
arthroscopic severity of chondromalacia (FULKER- is seen by an experienced radiologist and that the
SON and SHEA 1990) is almost identical to that of radiographer is provided with a precise written
224 N. Egund and L. Ryd
instruetion explaining the examination protocol to the same as observed at gait, whieh ensures an almost
be followed. The elinieal information should inelude true lateral view of the femoral eondyles. We use a
previous relevant imaging findings, trauma, and treat- knee flexion of between 25° and 35°. The rotational
ment, and should eneompass the spine, hips, and femorotibiallaxity in the weight-bearing position is
ankle. Any history or signs of arthritis, and especially ±12°, and therefore the patients are asked to look
sero negative arthritis or spondylarthropathy, may be straight forward. Shortly before exposure, the patient
decisive for seleetion of the type of examination and is asked to plaee his or her weight on the leg being
for differential diagnosis (enthesopathies are eom- examined, simulating a runner's position of the knee
monly misinterpreted). In Seandinavia many depart- before extension. The lateral radiographie view in
ments of radiology routinely report the examination
findings before the patient leaves the department,
whieh provides an opportunity for patient eontaet
and for supplementary radiographie views or imme-
diate ultrasound when relevant. As a result, many
adults and elderly patients admitted for knee eom-
plaints leave our departments with a diagnosis ofhip
joint synovitis and/or arthrosis.
13.5
Routine Radiography
13.5.1
Lateral View
Fig. 13.12. Normal true lateral radiographie view of the knee 13.5.3
obtained in the standing position using the described deviee The Standing Axial View
(metal screws are visible). The tibial plateau is horizontal and
the position is adequate for the anteroposterior projection of
the femorotibial joint space. There is a dorsal slope of the artie-
The teehnique of standing axial radiographie imag-
ular surface of the patella, the position being less optimal for ing of the patellofemoral joint was introdueed by
the axial radiographie view with a vertical beam direction AHLBAcK (1968). His deviee for support of the patella
a b
a b
Fig. l3.14a, b. Sagittal Tl-weighted MR images of the medial and lateral femoral and tibial condyles. The oblique Hne (1) is in
a different image joining the anterior crest of the tibia. Relative to vertical (horizontal) the inclination of the tibia is 14°. The
concave contour of the medial tibial condyles (a) is horizontal and thus has a dorsal slope of 14°. The orientation of the straight
(commonly convex) contour of the lateral tibial condyle (b) is 4° less than that of the medial tibial condyle
13.5.4
The Supine Axial View
Fig. 13.16. Device for knee support and patient position for the
axial, standing radiographie view of the patellofemoral joint.
Relative to the position in Fig. 13.10, the inclination of the lower
leg has been increased, with the foot placed more dorsally, Fig. 13.17. Patient position to obtain the axial radiographie
whieh brings the artieular surfaces of the patella in Fig. 13.12 view with 90° of knee flexion
into alignment with the vertieal plane
RSA provides two advantages. First the accuracy which is inaccessible for radiographie examination
is about one order of magnitude better than con- in the axial plane (Figs. l3.27, l3.28). The rationale
ventional radiography, and second, the system allows for viewing the patellofemoral relationship in mild
perfect characterization of the motion in all 6 degrees flexion is based on the fact that most patellar sublux-
of freedom of motion. Hence, "out-of-plane" motion ations and dislocations occur within the first 20°-30°
can also be analysed. of flexion. Statie eT and MR imaging can be used to
RSA has not previously been utilized in connec- obtain axial images at defined degrees of flexion with
tion with patellar motion proper. BLANKENVOORT et and without contraction of the quadriceps musele
al. (1988) used it to studythe screw-home mechanism, (DELGADO-MARTINS 1979; MARTINEZ et al. 1983b).
whieh is inherently involved in the complex motions of Assessment of patellar tracking by kinematie eT and
the patella. They found that the screw-home motion is MR imaging has been extensively reported (McNAL-
a facultative occurrence, dependent on how the femur LY et al. 2000; Dupuy et al. 1997; SHELLOCK et al. 1989;
and the tibia are positioned rotationally relative to BROSSMANN et al. 1993, 1994) with and without active
one another at an initial point in time. Similarly, KÄR- loading during flexion to extension, and a number of
RHOLM and co-workers have published a number of deviees have been developed for these purposes.
reports on the kinematies of the knee, ineluding rota-
tional motion between the femur and the tibia (KÄR-
RHOLM et al.1994b; UVEHAMMER et al. 2000). In gen- 13.7.1
eral they have found that femorotibial rotation is less Measurements
in an artificial knee than in the normal one. The poten-
tial of RSA has not yet been fully realized with regard Vertical patellar height is important for the biome-
to the patellar joint, although studies are in progress chanies and stability of the extensor mechanism and
in Lund on patients in whom all three bones (femur, patellofemoral joint. Measurements are performed
tibia, and patella) have been marked. on the lateral radiographie view in at least 30° of flex-
ion, at which the patellar tendon is considered under
tension in the supine position, although this has
not been documented. We perform all measurements
13.7 of patellar height from lateral standing views at more
Cl and MR Imaging of Patellar lracking than 20° of flexion. The most widely employed method
for assessing patellar height has been described by
Transaxial imaging of the patellofemoral joint by eT INSALL and SALVARTI (197l) and is based on the
and MR imaging allows assessment of the position length of the patellar tendon divided by the greatest
of the patella above and within the femoral trochlea diagonal length of the patella (Fig. l3.22a). The
over a range of extension and mild flexion of the knee normal ratio is 1.02 with an SD of 0.l3. A ratio of
_ _ __ b,c
a
Fig. 13.22a-c. The most common methods of measuring patellar height: a INSALL and SALVARTI (1971); b BLACKBURNE and
PEEL (1977); c CATON et al. (1982). LL, Length of the patellar tendon; LP, maximum diagonallength of the patella
230 N. Egund and L. Ryd
13.7.2
The Femoral Trochlea and Patellar Tracking
no significant right/left differences (BRATTSTRÖM
The classical work by means of radiography on 1964).
the configuration of the femoral condyles is that of The congruence angle, wh ich measures the rela-
BRATTSTRÖM (1964), who investigated the angles tionship of the V-shaped patellar articular ridge to
and distances of the femoral trochlea relative to the groove of the femoral trochlea (Fig. 13.25), was
the coronal plane of the dorsal aspect of the femo- introduced by MERCHANT et al. (1974). They sug-
ral condyles (Figs. 13.23, 13.24). The normal sulcus gested that any congruence angle greater that + 16°
angle of 142° (SD ±0.6° with no significant right/left is abnormal and may indicate lateral instability of
or sex differences) was obtained from radiographs the patella. Another measure of patellar tracking,
of the femoral trochlea at an angle of approximate- that might be appropriate, is the lateral patellafemo-
ly 25° between the beam and the longitudinal axis ral angle (LAURIN et al. 1978, 1979). On axial radio-
of the femur. With reversed direction of the beam, graphs, normally an angle formed between a line
MERCHANT et al. (1974) obtained similar measure- connecting the anterior aspects of the femoral con-
ments with respect to the sulcus angle, 138° (SD±6°, dyles and a second line joining the lateral facets of
range 126°-150°), and found no significant change the patella opens laterally (Fig. 13.26). In all their
in the shape of the trochlea through the range of 30 patients with recurrent suhluxation, these lines
heam to femur angles from 15° to 75°. In patients were either parallel or opened medially, hut the
with uni- or bilateral patellar displacement, the lines were normal in 90% of patients with chondro-
sulcus angle may be increased to above 150° with malacia. LAURIN et al. (1979) also introduced a mea-
Patellar and Quadriceps Mechanism 231
a b
Fig. 13.25. Measurement of the congruence angle according to MERCHANT et al. (1974), with a the supine (LAURIN) and b the
standing examination technique in the same normal knee. The landmarks a and c are the same as in Fig. 13.24, and b represents
the highest point of the lateral condyle. Line 0 bisects the sulcus angle and the arrow line joins the lowest point of the articular
ridge of the patella (d) (not marked on figure). There is an obvious medial displacement of the patella between figure parts a
and b, but minor differences in the position of (d) will influence the angle measurement
a b
Fig.13.26. Measurements of the normal lateral patellofemoral angle according to LAURIN et al. (1978, 1979), with a the supine
and b the standing examination technique. The line a-b joins the highest points of the femoral condyles and line d-e the lateral
articular surface of the patella. It appears that point b does not correspond to point e. Angles of 0° or with medial opening
(-) are abnormal and indicate lateral tilt
a b
Fig. 13.27a, b. Measurement of medial-lateral patellar displacement according to LAURIN et al. (1979). a and bare obtained in
the supine and the standing position, respectively. Patellar displacement is measured as the distance in millimeters between the
arrow line drawn 90° to line a-b and the medial edge of the patella. The width between points a and b varies with the angle of
tangency of the beam to the femoral condyles and between individuals
sure of lateral patellar displacement (Fig. 13.27). BROSSMANN et al. 1993). Direct and oblique trans-
The congruence angle, the lateral patellofemoral axial imaging of the femoral condyles by CT and
angle, and measurement of lateral patellar displace- MR imaging allows assessment of the transverse
ment have all been used frequently in the assess- axes and lateral articular surface of the patella rela-
ment of patellar instability by means of CT and MR tive to the plane of the posterior femoral condyles
imaging (Dupuy et al. 1997; McNALLY et al. 2000; (Figs. 13.28, 13.29), which is considered a reliable
232 N. Egund and L. Ryd
reference plane (SCHUTZER et al. 1986; BROSSMANN axial radiographie views, NAGAMINE et al. (1999)
et al. 1993, 1994). demonstrated an abnormal lateral position of the
The position of the tibial tuberosity relative to the tibial tuberosity in patients with lateral patellofemo-
sagittal plane through the sulcus of the femoral troch- ral arthrosis in 30° of flexion, and in comparison with
lea (Fig. 13.30) has been measured by CT and MR CT, the abnormal lateral position of the tibial tuber-
imaging in individuals with normal patellar tracking osity remained unchanged at extension (NAGAMINE
and in patients with patellar instability (ANDO et al. et al. 1997).
1993; MUNETA et al. 1994; BEACONSFIELD et al. 1994; Axial radiographie views in the standing weight-
IONES et al. 1995; McNALLY et al. 2000). The consis- bearing position (AHLBÄCK 1968; EGUND 1986) have
tency of the results indieates that a distance exceed- been used widely in clinieal studies in Sweden and also
ing 20 mm may be indieative of severe maltracking. in the assessment of the degenerative patellofemoral
Using a metal marker on the tibial tuberosity and joint in comparison with MR imaging (BOEGAARD et
al. 1998c). Axial radiographie views obtained in both
the standing and the supine position were compared
in 111 knees in 57 patients (EGUND 2001). In 39 knees
with no change in patellar shift there was a mean
lateral tilt of 3° from standing to supine, and in 33
knees with both lateral shift and tilt in the supine
position, the patella returned to complete alignment
in the standingposition (Figs.13.25-13.27, 13.39).AIso
in the standing position and extension or slight knee
flexion any lateral patellar displacement observed at
supine imaging is nullified (Fig. 13.31) (EGUND et
al. 2001). Optimal axial radiographie views with the
direction of the beams tangential to the patellar joint
surface were obtained in the standing position at 15°
of inclination of the lower leg (EGUND 1986), but the
angle had to be decreased in most examinations in
the supine position. This indicates that patellar height
Fig. 13.28. Measurement at CT of the lateral patellofemoral and patellofemoral contact areas are different in the
angle (vO)relative to the femoral condylar plane (CP) standing and the supine position (Fig.13.32).AIso, the
a b
Fig. 13.29a, b. Measurement at CT of lateral patellar displacement according to BROSSMANN et al. a Bisect offset is percentage,
alb + 100, of patella lateral to the projected perpendicular Hne. b Similar to the measurements of LAURIN et al. (Fig. 13.27)
Patellar and Quadriceps Mechanism 233
examination technique used may influence femoro- increased tendency towards lateral shift and dis-
tibial rotation (Fig. 13.32) and thus the position of the placement. Lateral displacement and tilt may both
tibial tuberosity relative to the femoral trochlea. be reduced (Dupuy et al. 1997) or increased (BROSS-
MANN et al. 1994) during active quadrieeps function.
In normal subjects, the tibial tuberosity is posi-
13.7.3 tioned lateral to the sagittal plane through the deep-
Summary of Radiological Measures of Patellar est point of the femoral trochlea at a distance of less
Tracking than 20 mm (BEACONSFIELD et al. 1994), and in most
patients with maltracking this distance is more than
The results reported in the literature on normal patel- 20 mm (JONES et al. 1995; McNALLY et al. 2000).
lar motion are still inconsistent, but in general the In the weight-bearingposition, the axial radiograph-
original results and thoughts of LAURIN et al. (1978) ie view of normal subjects in the standing position
and MERCHANT et al. (1974) have been confirmed by always shows alignment between the medial and lat-
me ans of CT and MR imaging. eral joint surfaces of the patellofemoral joint (EGUND
In the non-loaded, supine position and in asymp- 2001), and this is also true in those with lateral tilt and
tomatic knees, at 30° ofknee flexion and with relaxed displacement when examination is performed in the
quadriceps muscles, the artieular surfaces of the supine position (Figs. 13.25, 13.31). During walking the
patella are aligned with those of the femoral troch- excursion of the patella is medial rather than lateral
lea, but in some subjects there is a slight tendency (STEIN et al. 1993). Between flexion and full extension
towards lateral tilting and displacement. Between when standing on both legs there is an outward rota-
20° of flexion and complete extension there is an tion of the tibia relative to the femur of 5° (SANFRIDS-
234 N. Egund and L. Ryd
a b
Fig.13.31. Transaxial MR images of a 28-year-old male with patellar instability, obtained a supine at 10° of knee flexion and b
standing, with weight-bearing at extension. The normal lateral patellar displacement in the supine position (a) does not occur
in the standing position, even at extension (b)
a b
Fig. 13.32. a Standing, lateral radiographie view of a knee controlled for tibial cortieal sclerosis. The distal portion of the tibia
is not in contact with the plate of the deviee (see Fig. 13.12) and the tibial plateau is not horizontal. Knee flexion is too high
for the anteroposterior view. b A previous lateral view obtained in the supine position. Femoral rotation is almost identieal in
a and b, but there is a large difference in tibial rotation, indieated by the position of the fibula. Also there is a large difference
in the position of the patella between standing (a) and supine (b), as regards both the vertieal height and the artieular surface
orientation (gray line) relative to the tibial surface
SON et al. 2001b), which may not occur in the non- patellar tracking. Visualization of the function, per-
loaded knee (BLANKEVOORT et al.1988). formance and biomechanics of the knee is inadequate
In conclusion, as previously suggested by STEIN et when using supine imaging, including arthroscopy.
al. (1993), standing examination is to be considered There is consequently a need to reevaluate conclu-
the "state of the art" technique for the assessment of sions drawn without the use of views obtained in the
Patellar and Quadriceps Mechanism 235
standing position. This re evaluation should inc1ude The c1inieal features in patients with patellofemoral
the present supposed etiology of exeessive lateral OA are similar to those of other patellofemoral pain
pressure syndrome (FICAT and HUNGERFORD 1977) problems, but many and especially younger patients
and ehondromalacia of the medial patellar eartilage. may appear with swelling and symptoms due to syno-
vitis and effusion with pressure from Baker eysts
and/or rupture of these eysts (Figs. 13.33, 13.34). Pain
is eommonly referred to the anterior medial and/or
13.8 lateral femorotibial joint, but there is eharaeteristie
Osteoarthrosis of the dis ability at stair c1imbing and deseending and sudden
Patellofemoral Joint pain while arising from achair. The eombination of
patellofemoral and femorotibial OA is eommon, and
Symptomatie knee osteoarthrosis (OA) may oeeur in the lack of eorrelation between severity of symptoms
1.6%-9.4% of adults (FELSON 1998), and OA eonfined and stage of radiographie degeneration is well known.
to the patellofemoral joint or eoneomitant with fem- It has to be kept in mind that any eondition involving
orotibial OA is eommon (AHLBÄCK 1968; LEDING- the hip joint with effusion/synovitis may appear with
HAM et al. 1993). These figures are, however, mainly dominating referred pain to the knee.
reeorded from lateral radiographie views and the dis- The method of grading radiographie change of
tribution within the medial and lateral patellofemo- OA was developed by KELLGREN and LAWRENCE
ral faeets has not been established by me ans ofaxial (1957), but the imperfeetion of their eriteria has led
radiographie views. Lateral patellofemoral OA is also to a large number of radiographie grading seal es
eommon in knees with medial femorotibial OA and for eross-seetional and longitudinal epidemiologieal
varus angulation (HARRISON et al. 1994), but the studies. We use a specifie grading system for the knee
eombination oflateral femorotibial and medial patel- joint (Table 13.2) whieh takes into aeeount the gener-
lofemoral OA is rare. Lateral patellofemoral OA may al radiographie features of OA as proposed by KELL-
be associated with an abnormal lateral position of GREN and LAWRENCE (1957), but with a more spe-
the tibial tuberosity, when examined in the supine cifie measurement of joint spaee loss and attrition
position by axial radiographie views and eT (NAGA- as proposed by AHLBÄCK (1968) and EGUND (1986).
MINE et al. 1997, 1999). The main objeetion to the KELLGREN and LAWRENCE
a b
Fig.13.33a, b. Sagittal MR images of the knee in a 36-year-old female with anterior knee pain and swelling. Both the STIR sequence
(a) and the Tl gadolinium-enhanced fat-saturated sequence (b) demonstrate extensive synovitis with Baker cysts. Alteration in
the anterior cruciate ligament commonly and early results in a retraction phenomenon in the proximal tibia (arrowhead). Is this
common condition, with normal radiographs, a result of reactive arthritis or secondary to early osteoarthritis?
236 N. Egund and 1. Ryd
a b
Fig. 13.34a, b. Transaxial MR images of the knee of the same patient as in Fig. 13.33. In the patella (arrow) and the femoral con-
dyle (white arrowhead) there are superficial cartilage defects and in another section fissuring of the cartilage to the subchondral
bone (black arrowhead). Even small cartilaginous lesion can be demonstrated with this proton fat-saturated sequence
Table 13.2. Grading system for the knee joint employed by the authors
(1957) grading system concerns the question of membrane and periosteum (RESNICK and NIWAYA-
whether or not marginal osteophytes (SPECTOR and MA 1995). The sites of marginal osteophytes are
HOCHBERG 1994; BRANDT et al. 1991) represent most commonly far away from the focal cartilagi-
definitive evidence of cartilage degeneration (BoE- nous lesion of OA. Their presence should be consid-
GARD et al. 1998a): this is a subject which can still ered a sign of processes in the synovium and sur-
unite or divide radiologists. Marginal osteophytes rounding soft tissue which may react to mediators
appear as a sign of bony re action at sites where of cartilage damage and repair (EDWARDS 1998; VAN
articular cartilage is continuous with the synovial DEN BERG et al. 1998).
Patellar and Quadrieeps Mechanism 237
13.8.1
Imaging of Patellofemoral Osteoarthrosis
Standing
Standing
In addition to anterior knee pain, many of these
Fig.13.35. Axial radiographie views obtained in the supine and
patients present with swelling of the knee and effu-
standing positions in a 48-year-old female with anterior knee
pain. Arthrosis suspected from the slight joint space reduction sion with synovitis on contrast-enhanced MR imag-
visible only on the standing examination was confirmed at MR ing (Figs. 13.33, 13.34). Our imaging protocol in these
imaging patients comprises:
238 N. Egund and L. Ryd
13.9
Fig. 13.38. Axial radiographie views of lateral patellar arthro-
sis visualized in both the supine and the standing position.
Traumatic and Overuse Conditions
There is medial displacement of the patella between supine
and weight-bearing Extensor mechanism injuries can occur at all ages
but they are most commonly seen in younger patients
and in association with sports activities. A distinction
must be drawn between acute and chronic injury. For
1. Sagittal STIR, 512 matrix an understanding of the trauma mechanism and what
2. Sagittal Tl, 512 matrix to look for, any traumatic condition is better consid-
3. Oblique axial Tl, 512 matrix ered a primary soft tissue injury which may involve
4. Oblique or double oblique axial proton fat saturat- bone. In chronic injury and overuse syndromes, it is
ed, 512 matrix less appropriate to make a distinction between soft
5. Sagittal contrast-enhanced Tl fat saturated tissue and bone injury, considering the anatomy and
6. Oblique or double oblique axial contrast -enhanced the most common sites of injury at tendon attach-
Tl fat saturated ments to bone and cartilage (Fig. 13.34).
13.9.1.1
Imaging of the Knee with Acute Patellar
Dislocation
13.9.3
Overuse Syndromes
The term"jumper's knee" has been used to describe a Fig.13.51. Lateral radiographie view of avulsion from the tibial
common painful condition in athletes of the quadri- tuberosity (arrow) in a 10-year-old girl with anterior knee pain
244 N. Egund and L. Ryd
Fig. 13.53. Sagittal MR STIR sequence of the same knee as in Fig. 13.55. Jumper's knee. Sagittal MR STIR sequence dem-
Figs. 13.51 and 13.52. The defect of the patellar ligament at onstrating a bulging infiammatory lesion of the fat pad of
the site of the tibial tuberosity is visualized but the sequence Hoffa (arrow) and minor lesions of the patellar ligament at its
does not contribute any additional information compared attachment to the patella
with Figs. 13.51 and 13.52
Patellar and Quadrieeps Mechanism 245
ligament abnormalities (KARANTANAs et al. 2001). of the patella: the recurrence and site of osteoarthritis in 114
Both MR imaging and bone scintigraphy (KAHN and operated knees followed for 14 years. Int Orthop (in press)
Atkin DM, Fithian DC,Marangi KS, Stone ML, Dobson BE, Men-
WILSON 1987) may demonstrate bony involvement
delsohn C (2000) Characteristies of patients with primary
of the patella. acute lateral patellar dislocation and their recovery within
the first 6 months of injury. Am J Sports Med 28:472-479
Bashir A, Gray ML, Hartke J, Burstein D (1999) Nondestructive
imaging ofhuman cartilage glycosaminoglycan concentra-
13.10 tion by MRI. Magn Reson Med 41:857-865
Beaconsfield T, Pintore E, Maffulli N, Petri GJ (1994) Radiolog-
Conclusion ieal measurements in patellofemoral disorders. A review.
Clin Orthop 308:18-28
For many years the understanding and treatment of Blackburne JS, Peel TE (1977) A new method of measuring
patellofemoral dis orders has been influenced by con- patellar height. J Bone Joint Surg [Br] 59:241-242
ventional radiography and advanced imaging as well Blankevoort L, Huiskes R, de Lange A (1988) The envelope of
passive knee joint motion. J Biomech 21:705-720
as clinical examination obtained in the supine posi- Boegard T, Rudling 0, Petersson IF, Jonsson K (1998a) Correla-
tion. We feel that the weight-bearing examination of tion between radiographieally diagnosed osteophytes and
the patellofemoral joint may contribute not only to magnetie resonance detected cartilage defects in the patel-
more accurate diagnostics but also to the understand - lofemoral joint. Ann Rheum Dis 57:395-400
ing of why anterior knee pain occurs, as requested Boegard T, Rudling 0, Petersson IF, Jonsson K (1998b) Joint-
space width in the weight-bearing radiogram of the tibio-
by INsALL (1995). The weight-bearing techniques of femoral joint. Should the patient stand on one leg or two?
imaging do not immediately open up a new world Acta RadioI39:32-35
of treatments for patellofemoral instability, but they Boegard T, Rudling 0, Petersson IF, Sanfridsson J, Saxne T,
may prevent patients from undergoing the tradition- Svensson B, Jonsson K (1998c) Joint -space width in the axial
al surgical treatment, which from a "standing" radi- view of the patello-femoral joint. Definitions and compari-
son with MR imaging. Acta RadioI39:24-31
ologist's point of view commonly appears contrain- Bradley WG, Ominsky SH (1981) Mountain view of the patella.
dicated. Hopefully colleagues worldwide will soon AJR Am J Roentgenol136:53-58
regard the standing examination techniques with the Brandt KD, Fife RS, Braunstein EM, Katz B (1991) Radiograph-
same enthusiasm as has been, and still is, shown for ie grading of the severity of knee osteoarthritis: relation of
examination in the supine position. the Kellgren and Lawrence grade to a grade based on joint
space narrowing, and correlation with arthroscopie evi-
The role of seronegative arthritis and spondylar- dence of artieular cartilage degeneration. Arthritis Rheum
thropathy with possible involvement of entheses in 34: 1381-1386
anterior knee pain has not been explored, but should Brattström H (1964) Shape of the intercondylar groove nor-
always be kept in mind when imaging these patients. mally and in recurrent dislocation of patella. A clinieal and
x-ray-anatomieal investigation. Acta Orthop Scand Suppl
68:1-148
Brossmann J, Muhle C, Schroder C, Melchert UH, Bull CC, Spiel-
mann RP, Heller M (1993) Patellar tracking patterns during
References active and passive knee extension: evaluation with motion-
triggered eine MR imaging. Radiology 187:205-212
Ahl T, Dalen N, Selvik G (1988) Mobilization after operation Brossmann J, Muhle C, Bull CC, Schroder C, Melchert UH, Zie-
of ankle fractures. Good results of early motion and weight plies J, Spielmann RP, Heller M (1994) Evaluation of patel-
bearing. Acta Orthop Scand 59:302-306 lar tracking in patients with suspected patellar malalign-
Ahlbäck S (1968) Osteoarthrosis of the knee. A radiographie ment: eine MR imaging vs arthroscopy. AJR Am J Roent-
investigation.Acta Radiol Diagn (Stockh) [Suppl] 72:1-72 genol 162:361-367
Aleman 0 (1928) Chondromalaeia post-traumatiea patellae. Buckland-WrightJC (1994) Quantitative radiography of osteo-
Acta Chir Scand 63:194 arthritis.Ann Rheum Dis 53:268-275
Ando T, Hirose H, Inoue M, Shino K, Doi T (1993) A new Burstein D, Velyvis J, Scott KT, Stock KW, Kim YJ, Jaramillo
method using computed tomographie scan to measure the D, Boutin RD, Gray ML (2001) Protocol issues for delayed
rectus femoris-patellar tendon Q-angle comparison with Gd(DTPA)(2- )-enhanced MRI (dGEMRIC) for clinieal eval-
conventional method. Clin Orthop 289:213-219 uation of artieular cartilage. Magn Reson Med 45:36-41
Apostolaki E, Cassar-Pullieino VN, Tyrrell PN, McCall IW Caton J, Deschamps G, Chambat P, Lerat JL, Dejour H (1982)
(1999) MRI appearances ofthe infrapatellar fat pad in occult Patella infera. Apropos of 128 cases. Rev Chir Orthop Repa-
traumatie patellar dislocation. Clin RadioI54:743-747 ratriee Appar Mot 68:317-325
Arnbjörnsson A, Egund N, Rydling 0, Stockerup R, Ryd L Cook JL, Khan KM, Harcourt PR, Grant M, Young DA, Bonar SF
(1992) The natural his tory of recurrent dislocation of the (1997) A cross sectional study of 100 athletes with jumper's
patella. Long-term results of conservative and operative knee managed conservatively and surgieally. The Vietorian
treatment. J Bone Joint Surg [Br] 74:140-142 Institute of Sport Tendon Study Group. Br J Sports Med
Arnbjörnsson A, Egund N, Ryd L (2002) Recurrent dislocation 31:332-336
246 N. Egund and 1. Ryd
Cook JL, Khan KM, Kiss ZS, Purdam CR, Griffiths L (2000) Grelsamer RP, Meadows S (I992) The modified Insall-Salvati
Prospective imaging study of asymptomatie patellar tendi- ratio for assessment of patellar height. Clin Orthop
nopathy in elite junior basketball players. J Ultrasound Med 282:170-176
19:473-479 Hampson WG,Hill P (1975) Late results oftransfer ofthe tibial
Delgado-Martins H (1979) A study of the position of the patel- tubercle for recurrent dislocation of the patella. J Bone Joint
la using computerised tomography. J Bone Joint Surg [Br] Surg [Br] 57:209-213
61:443-444 Harrison MM, Cooke TD, Fisher SB, Griffin MP (1994) Patterns
Dupuy DE, Hangen DH, Zachazewski JE, Boland AL, Palmer W of knee arthrosis and patellar subluxation. Clin Orthop
(I997) Kinematie CT of the patellofemoral joint. AJR Am J 309:56-63
RoentgenoI169:211-215 Hawkins RJ, Bell RH, Anisette G (I986) Acute patellar disloca-
Edwards JCV (1998) Synovial physiology in the context of osteo- tions. The natural history. Am J Sports Med 14: 117-120
arthritis. In: Brandt KD, Doherty M, Lohmander LS (eds) Henche HR, Kunzi HU, Morscher E (1981) The areas of contact
Osteoarthritis. Oxford University Press, Oxford, pp 167-175 pressure in the patello-femoral joint. Int Orthop 4:279-281
Egund N (1986) The axial view of the patello-femoral joint. Hungerford DS, Barry M (1979) Biomechanies of the patel-
Description of a new radiographie method for routine use. lofemoral joint. Clin Orthop 144:9-15
Acta Radiol Diagn (Stockh) 27:101-104 Insali JN (I995) Foreword. In: Scuderi GR (ed) The patella.
Egund N (2001) The radiographie axial view of the patello- Springer, Berlin Heidelberg New York
femoral joint: examination in the supine or standing posi- Insali JN, Salvarti E (I971) Patella position in the normal knee.
tion? (abstract). Eur Radiolll:130 Radiology 101:101-104
Egund N, Friden T (1988) Lesion of the anterior cruciate liga- Insali JN, Falvo KA, Wise DW (1976) Chondromalacia patellae.
ment and sagittal disalignment of the knee in weight-bear- A prospective study. J Bone Joint Surg [Am] 58:1-8
ing. Acta RadioI29:559-563 Jaroschy W (1924) Die diagnostiehe Verwertbarkeit der Patel-
Egund N, Lundin A, Wallengren NO (I988) The vertieal posi- laraufnahmen. Fortsehr Roentgenstr 31:781-782
tion of the patella. A new radiographie method for routine Jones RB, Barlett EC, Vainright JR, Carroll RG (1995) CT deter-
use. Acta RadioI29:555-558 mination of tibial tubercle lateralization in patients present-
Egund N, Friden T, Hjarbaek J, Lindstrand A, Stockerup R ing with anterior knee pain. Skeletal RadioI24:505-509
(I993) Radiographie assessment of sagittal knee instability Juliusson R, Markhede G (1984) A modified Hauser procedure
in weight bearing. A study on anterior cruciate-deficient for recurrent dislocation of the patella. A long-term fol-
knees. Skeletal RadioI22:177-181 low-up study with special reference to osteoarthritis. Arch
Egund N, Jacobsen BW, J0rgensen A, Kristiansen SE (2000) Orthop Trauma Surg 103:42-46
Angles and shape of the artieular surfaces of the tibial con- Kahn D, Wilson MA (1987) Bone scintigraphie findings in
dyles: measurement by MR imaging in the sagittal plane. patellar tendonitis. J Nucl Med 28: 1768-1770
Presented at the European Society of Musculoskeletal Radi- Karantanas AH, Zibis AH, Papanikolaou N (2001) Increased
ology, Leiden signal intensity on fat -suppressed three-dimensional Tl-
Egund N, Hol PK, Johnston V, Tillung T, Fosse E (2001) weighted pulse sequences in patellar tendon: magie angle
Normal femorotibial rotation and patellar tracking in the effect? Skeletal Radio130:67 -71
weight-bearing standing and the non-loaded supine posi- Kärrholm J, Hansson LI, Svensson K (1983) Predietion of
tion: assessment and comparison by kinematic MR imag- growth pattern after ankle fractures in children. J Pediatr
ing (abstract). Radiology (Suppl) 609 Orthop 3:319-325
Felson DT (I998) Epidemiology of osteoarthritis. In: Brandt Kärrholm J, Selvik G, Elmqvist LG, Hansson LI (1988) Active
KD,Doherty M,Lohmander LS (eds) Osteoarthritis. Oxford knee motion after cruciate ligament rupture. Stereoradiog-
University Press, Oxford, pp 13-22 raphy. Acta Orthop Scand 59: 158-164
Fieat R, Hungerford DS (1977) Disorders of the patello-femoral Kärrholm J, Borssen B, Lowenhielm G, Snorrason F (1994a)
joint. Williams & Wilkins, Baltimore Does early mieromotion of femoral stern prostheses matter?
Friden T, Sommerlath K, Egund N, Gillquist J, Ryd L, Lind- 4-7-year stereoradiographie follow-up of 84 cemented
strand A (1992) Instability after anterior cruciate ligament prostheses. J Bone Joint Surg [Br] 76:912-917
rupture. Measurements of sagittal laxity compared in 11 Kärrholm J, Jonsson H, Nilsson KG, Soderqvist I (1994b) Kine-
cases. Acta Orthop Scand 63:593-598 maties of successful knee prostheses during weight-bear-
Friden T, Egund N, LindstrandA (1993) Comparison of symp- ing: three-dimensional movements and positions of screw
tomatic versus nonsymptomatie patients with chronie ante- axes in the Trieon-M and Miller-Galante designs. Knee Surg
rior cruciate ligament insufficiency. Radiographie sagittal Sports Traumatol Arthrosc 2:50-59
displacement during weightbearing. Am J Sports Med Keats TE (1996) Atlas of normal roentgen variants that may
21:389-393 simulate disease. Mosby, St Louis
Fulkerson JP (I 983) The etiology of patellofemoral pain in Kellgren JH, Lawrence JS (1957) Radiologie assessment of
young, active patients: a prospective study. Clin Orthop osteoarthritis.Ann Rheum Dis 16:494-501
179:129-133 Khan KM, Bonar F, Desmond PM, Cook JL, Young DA, Visentini
Fulkerson JP (I997) Disorders of the patellofemoral joint. Wil- PJ, Fehrmann MW, Kiss ZS, O'Brien PA, Harcourt PR, Dowl-
liams and Wilkins, Baltimore ing RJ, O'Sullivan RM, Criehton KJ, Tress BM, Wark JD (1996)
Fulkerson JP, Shea KP (I 990) Disorders of patellofemoral align- Patellar tendinosis (jumper's knee): findings at histopatho-
ment. J Bone Joint Surg [Am] 72:1424-1429 logie examination, US, and MR imaging. Vietorian Institute
Goodfellow J, Hungerford DS, Woods C (1976) Patello-femoral of Sport Tendon Study Group. Radiology 200:821-827
joint mechanics and pathology. 2. Chondromalacia patel- Khan KM, Cook JL, Kiss ZS, Visentini PJ, Fehrmann MW, Har-
lae. J Bone Joint Surg [Br] 58:291-299 court PR, Tress BW, Wark JD (1997) Patellar tendon ultra-
Patellar and Quadrieeps Mechanism 247
sonography and jumper's knee in female basketball players: tis by MRI: correlation between dynamie and statie gado-
a longitudinal study. Clin J Sport Med 7:199-206 linium-enhanced magnetie resonance imaging and miero-
Knutsson F (1941) Über die Röntgenologie des Femoropatel- scopie and macroscopie signs of synovial inflammation.
largelenks sowie eine gute Projektion für das Kniegelenk. Magn Reson Imaging 16:743-754
Acta Radiol Diagn (Stockh) 22:371-376 Outerbridge R (1961) The etiology of chondromalacia patellae.
Kuhl CK, Schild HH (2000) Dynamie image interpretation of J Bone Joint Surg [Br] 43:752-757
MRI of the breast. J Magn Reson Imaging 12:965-974 Post WR (1997) History and physieal examination. In: Fulker-
Langer M (1929) Über die Entwieklung des Kniegelenkes. Z son JP (ed) Disorders of the patellofemoral joint. Williams
Gesamte Anat 89:83-10 1 and Wilkins, Baltimore, pp 39-71
Laurin CA, Levesque HP, Dussault R, Labelle H, Peides JP Radin EL (1979) A rational approach to the treatment ofpatel-
(1978) The abnormal lateral patellofemoral angle: a diag- lofemoral pain. Clin Orthop 144: 107-109
nostie roentgenographie sign of recurrent patellar sublux- Radin EL, Rose RM (1986) Role of subchondral bone in the
ation. J Bone Joint Surg [Am] 60:55-60 initiation and progression of cartilage damage. Clin Orthop
Laurin CA, Dussault R, Levesque HP (1979) The tangential x-ray 213:34-40
investigation of the patellofemoral joint: x-ray technique, Ragnarsson JI, Eliasson P, Karrholm J, Lundstrom B (1992) The
diagnostie criteria and their interpretation. Clin Orthop accuracy of measurements of femoral neck fractures. Con-
144:16-26 ventional radiography versus roentgen stereophotogram-
Ledingham J,Regan M, Jones A, Doherty M (1993) Radiograph- metrie analysis. Acta Orthop Scand 63:152-156
ie patterns and associations of osteoarthritis of the knee in Resniek D, Niwayama G (1995) Degenerative diseases. In: Resn-
patients referred to hospital. Ann Rheum Dis 52:520-526 iek D (ed) Diagnosis of bone and joint disorders. Saunders,
Macnab J (2001) Recurrent dislocation of the patella. J Bone Philadelphia, pp 1263-137l
Joint Surg [Am] 34:957-967 Rünow A (1983) The dislocating patella. Etiology and progno-
Mäenpaa H, Lehto MU (1997a) Patellar dislocation. The long- sis in relation to generalized joint laxity and anatomy of the
term results of nonoperative management in 100 patients. patellar artieulation. Acta Orthop Scand SuppI201:1-53
Am J Sports Med 25:213-217 Ryd L, Lindstrand A, Rosenquist R, Selvik G (1986) Tibial
Mäenpaa H, Lehto MU (1997b) Patellofemoral osteoarthritis component fixation in knee arthroplasty. Clin Orthop
after patellar dislocation. Clin Orthop 339: 156-162 213:141-149
Martinez S, Korobkin M, Fondren FB, Hedlund LW, Goldner JL Ryd L, Albrektsson BE, Carlsson L, Dansgard F, Herberts P,
(1983a) Computed tomography of the normal patellofemo- Lindstrand A, Regner L, Toksvig-Larsen S (1995) Roentgen
ral joint. Invest RadioI18:249-253 stereophotogrammetrie analysis as apredietor of mechani-
Martinez S, Korobkin M, Fondren FB, Hedlund LW, Goldner JL calloosening of knee prostheses. J Bone Joint Surg [Br]
(1983b) Diagnosis of patellofemoral malalignment by com- 77:377-383
puted tomography. J Comput Assist Tomogr 7:1050-1053 Sallay PI, Poggi J, Speer KP, Garrett WE (1996) Acute disloca-
McNally EG, Ostlere SJ, Pal C, Phillips A, Reid H, Dodd C (2000) tion of the patella. A correlative pathoanatomie study. Am
Assessment of patellar maltracking using combined statie J Sports Med 24:52-60
and dynamie MRI. Eur RadioI1O:1051-1055 Sanfridsson J, Arnbjornsson A, Friden T, Ryd L, Svahn G, Jons-
Merchant AC (1988) Classification of patellofemoral disorders. son K (2001a) Femorotibial rotation and the Q-angle relat-
Arthroscopy 4:235-240 ed to the dislocating patella. Acta Radiol 42:218-224
Merchant AC, Mercer RL,Jacobsen RH, Cool CR (1974) Roent- Sanfridsson J, Ryd L, Svahn G, Friden T, Jonsson K (2001b)
genographie analysis of patellofemoral congruence. J Bone Radiographie measurement of femorotibial rotation in
JointSurg [Am] 56:1391-1396 weight-bearing. The influence of flexion and extension in
Moore TM, Harvey JP (1974) Roentgenographie measurement the knee on the extensor mechanism and angles of the lower
of tibial-plateau depression due to fracture. J Bone Joint extremity in a healthy population. Acta Radio142:207 - 217
Surg [Am] 56:155-160 Schutzer SF, Ramsby GR, Fulkerson JP (1986) Computed
Muneta T, Yamamoto H, Ishibashi T,Asahina S, Furuya K (1994) tomographie classification of patellofemoral pain patients.
Computerized tomographic analysis of tibial tubercle posi- Orthop Clin North Am 17:235-248
tion in the painful female patellofemoral joint. Am J Sports Scuderi GR (1995) The patella. Springer, Berlin Heidelberg
Med 22:67-71 NewYork
Nagamine R, Miura H, Inoue Y, Tanaka K, Urabe K, Okamoto Seil R, Muller B, Georg T, Kohn D, Rupp S (2000) Reliability
Y, Nishizawa M, Iwamoto Y (1997) Malposition of the tibial and interobserver variability in radiologieal patellar height
tubercle during flexion in knees with patellofemoral arthri- ratios. Knee Surg Sports Traumatol Arthrosc 8:231-236
tis. Skeletal Radio126:597 -601 Selvik G (1974) A roentgen stereophotogrammetrie method
Nagamine R, Miura H, Urabe K, Matsuda S, Chen WJ, Matsu- for the study of the kinematies of the skeletal system. Lund
nobu T, Iwamoto Y (1999) Radiologieal assessment of the University, Sweden
position of the tibial tuberosity by means of a marking Settegast J (1921) Typische Röntgenbilder von normalen Men-
wire in knees with patellofemoral arthritis. Skeletal Radiol schen. Lehmanns Med Atlanten 5:211
28:27-32 Shellock FG, Mink JH, Deutsch AL, Fox JM (1989) Patellar
Önsten I, Bengner U, Besjakov J (1993) Socket migration after tracking abnormalities: clinieal experience with kinematie
Charnley arthroplasty in rheumatoid arthritis and osteo- MR imaging in 130 patients. Radiology 172:799-804
arthritis. A roentgen stereophotogrammetrie study. J Bone Siu D, Cooke TD, Broekhoven LD, Lam M, Fisher B, Saunders
Joint Surg [Br] 75:677-680 G, Challis TW (1991) A standardized technique for lower
0stergaard M, Stoltenberg M, Lovgreen-Nielsen P, Volck B, limb radiography. Practiee, applications, and error analysis.
Sonne-Holm S, Lorenzen I (1998) Quantification of synovi- Invest Radiol 26:7l-77
248 N. Egund and L. Ryd
Sonin AH, Fitzgerald SW, Bresler ME, Kirsch MD, Hoff FL, video digital-fiuoroscopie study in healthy volunteers. AJR
Friedman H (1995) MR imaging appearance of the exten- Am J Roentgenol161 :617-620
sor mechanism of the knee: functional anatomy and injury Thomee R, Augustsson J, Karlsson J (1999) Patellofemoral
patterns. Radiographies 15:367-382 pain syndrome: a review of current issues. Sports Med
Spector TD, Hochberg MC (1994) Methodologieal problems 28:245-262
in the epidemiologieal study of osteoarthritis. Ann Rheum Uvehammer J, Karrholm J, Brandsson S (2000) In vivo kine-
Dis 53:143-146 maties of total knee arthroplasty. Concave versus poste-
Stanitski CL (1995) Articular hypermobility and chondral rior-stabilised tibial joint surface. J Bone Joint Surg [Br1
injury in patients with acute patellar dislocation. Am J 82:499-505
Sports Med 23:146-150 van den Berg WB, van der Kraan PM, van Beuningen HM
Stanitski CL, Paletta GA (1998) Artieular cartilage injury (1998) Synovial mediators of cartilage damage and repair in
with acute patellar dislocation in adolescents. OA. In: Brandt KD, Doherty M, Lohmander LS (eds) Osteo-
Arthroscopic and radiographie correlation. Am J Sports arthritis. Oxford University Press, Oxford, pp 157-167
Med 26:52-55 Wiberg G (1941) Roentgenographie and anatomie studies on
Starok M, Lenchik L, Trudell D, Resniek D (1997) Normal patel- the femoro-patellar joint. Acta Orthop Scand 12:319-410
lar retinaculum: MR and sonographie imaging with cadav- Woodland LH, Frands RS (1992) Parameters and comparisons
erie correlation. AJR Am J Roentgenol168: 1493-1499 of the quadrieeps angle of college-aged men and women
Stein LA, Endieott AN, Sampalis JS, Kaplow MA, Patel MD, in the supine and standing positions. Am J Sports Med
Mitchell NS (1993) Motion of the patella during walking: a 20:208-211
14 Infection
MOHAMED E. ABD EL BAGI, MONA S. AL SHAHED, BASSAM M. SAMMAK
biological aspiration cultures has been reported to methicillin-resistant Staph. aureus (MRSA), cell wall
be as low as 42% but was found to increase to 84% protein is resistant to all b-lactam drugs and glyco-
when culture tests were combined with histological peptides such as vancomycin are required (BOUZA
core biopsies (WHITE et al. 1995). Certain laboratory and MUNOZ 1999). Coagulase-negative staphyloc-
tests such as white blood count (WBC), erythrocyte coci commonly affect prosthetic joints. Anaerobes
sedimentation rate (ESR) and C-reactive protein (CRP) are common in vasculopathies and diabetes. Multi-
are helpful diagnostic and baseline tests. In one study drug-resistant tuberculosis is an emergent disease
a WBC alone was high in only 25% of children with in acquired immunodeficiency syndrome (AIDS).
osteomyelitis while the differential count was abnor- In tropical countries and predisposed patients, rare
mal in 65% (DIRSCH and KINDERS 1993). Serial CRP bone infections commonly involve the knee, e.g.
measurements are a more accurate means for moni- tuberculosis, brucellosis, salmonella and fungi (ABD
toring of response to treatment and predicting out- EL BAGI et al. 1999).
come than ESR or radiographs (ROINE et al. 1995). The radiological evaluation of knee infection is
extremely useful for treatment plans. Differentiation
between osteomyelitis, cellulitis and abscess forma-
14.1.2 tion is crucial in selecting treatment options such
Classification of Knee Infections as operative intervention, percutaneous drainage or
medical treatment (BOUTIN et al. 1998).
Knee infection may be classified into several types
based on the route of infection, the anatomical site,
the patient's age, the aetiology and the type of onset.
Predisposition to knee infections may result from 14.2
trauma, general illness, malnutrition, extremes of Infection Imaging Techniques
age, diabetes, neuropathies, sickle cell disease, sur-
gery, interventions, steroid treatment and immuno- 14.2.1
compromise. Understanding the pathological anato- Plain Radiographs
my of knee infections is aprerequisite for correct
diagnosis and treatment plans. Knee infection may Conventional radiographs in the frontal and lateral
arise in the bone (osteomyelitis), synovium (syno- projections are the first modality for the work-up of
vitis), articular structures (septic arthritis), synovial patients suspected to have knee infections. They are
cysts (bursitis), muscle (myositis) and subcutaneous simple, readily available and cheap (Fig. 14.1). The
tissues (cellulitis or fasciitis). earliest change appreciable in knee infections is peri-
Knee infections may be acute, subacute or chronic. articular soft tissue swelling or displacement of the
The chronic forms may be active or inactive. Infec- normal fat planes. This may be apparent as early as 2
tion may be due to (1) direct extension from a local days. Development of bone hyperaemia manifests as
focus, (2) haematogenous spread, (3) direct implan- osteopenia after 7-10 days. The earliest bone change
tation of a pathogen or (4) surgery. In considering the is loss of the trabecular pattern. Periosteal re action
choice of imaging technique, it is important to know may appear next. The location of the soft tissue swell-
whether infection is developing in a "violated knee", ing may indicate the infection harbour, e.g. swelling in
i.e. violated by surgery or injuries, or in a "non-vio- the suprapatellar recess in septic arthritis or behind
lated knee". Postoperative infections are dealt with in the joint in infected Baker's cyst. In acute osteomyeli-
Chap. 11. In this chapter we will concentrate on infec- tis,loss of bone density must reach 30%-50% before
tions in the non-violated knee. radiographs become abnormal. Overt bone erosions
Haematogenous spread ofbone infection can arise take 2-6 weeks to show on plain films. In chronic
from urinary tract infection, pneumonia, skin boils, osteomyelitis, diffuse sclerosis, sequestra, involucrum
otitis or endocarditis, but the source may remain formation and intramedullary abscess are evident on
unidentified (DIRSCHL and KINDERS 1993). The plain radiographs (DAVID et al. 1987).
aforementioned are said to induce bacteraemia, but Plain radiographs often help in the interpretation
transient bacteraemia is a daily event, e.g. after tooth of the plentiful findings of sophisticated tests in com-
brushing. Staphylococci are the commonest cause of plicated cases. An important but under-emphasized
all joint and bone infections in all groups. More than role for plain radiographs lies in indicating other
98% of these organisms are now resistant to peni- pathology or mimics. Sclerotic bone malignancies
cillins and are hence treated by cephalosporins. In may be confused with chronic osteomyelitis. Trau-
Infection 251
Fig. 14.1. Plain radiograph signs of knee infection: (1) soft Fig. 14.2. Sinogram demarcating the extent and direction of
tissue swelling (short white arrows); (2) periostitis (open white the small sinus opening in the skin
arrows); (3) bone erosions (short black arrowheads); (4) artic-
ular erosions (Zarge black arrowhead); (5) osteopenia and rar-
efaction (short black arrows)
matie lesions and osteochondritis can mimic infec- infection. If such uptake is combined with increased
tion. A Charcot's neuropathic joint may show charac- tracer accumulation in the delayed statie phase at 3 h,
teristie sclerosis, disorganization or periostitis. osteomyelitis should be diagnosed (Fig. 14.3). How-
ever, if there are sizable subperiosteal pustular pock-
Sinography. When a chronie discharging sinus is ets, large joint effusions or rapidly destructive pro-
present, the injection of water-soluble contrast agent cesses, cold spots will appear.
into the cutaneous opening may provide an excellent The sensitivity of 99ffiTc_MDP scintigraphy varies
demarcation of its extent and direction (Fig. 14.2). from 32% to 100% (DAVID et al. 1987). The highest
sensitivity is in the non-violated tissues, which have
not been subjected to trauma or orthopaedie hard-
14.2.2 ware. Scintigraphy is highly sensitive but non-specif-
Nuclear Medicine Scanning ic. Correlation with clinical and radiographie find-
ings is essential. Sensitivity is lower in children,
14.2.2.1 osteoporosis, vasculopathies and metabolic bone dis-
Three-Phase Bone Scintiscan ease. Specificity is much reduced in violated knees
because the physiological effects of trauma and infec-
Three-phase bone scintigraphy may become positive tion are similar. A single-photon emission computed
a few hours after the onset of focal infections, but tomography facility increases test sensitivity for small
in practiee is typically not performed until 24-48 h lesions owing to superior resolution. Other lesions,
from the onset of clinieal symptoms. Technetium- such as arthritis, healing fractures or previously treat-
99m (99ffiTc)-methylene diphosphonate (MDP) is the ed osteomyelitis show little abnormality in the first
principal radioisotope in most institut ions. Less com- two phases but may show markedly increased uptake
monly, hydroxymethylene diphosphonate (HMDP) is on delayed bone images. Furthermore, a four-phase
used. Increased uptake in the early first-pass perfu- scintiscan incorporating a 24-h delayed image will
sion phase and the blood pool phase at about 10 min help to clarify the significance of any doubtful activ-
post injection indicates soft tissue inflammation or ity. On a four-phase scan at 24 h, activity of normal
252 M. E. Abd EI Bagi et al.
98% (PALESTRO and TORRES 1997). lllIn-WBC scan is specificity of 93% and a sensitivity of 100% (RODDIE
less sensitive for chronie osteomyelitis (60%) although et al. 1998). 99mTc-HMPAO is very reliable and is
highly specific (96%). This is due to the lack of granu- probably the best leucocyte-Iabelling agent for the
locytes in chronic inflammatorytissues (SCHAUWECK- evaluation of chronie osteomyelitis (McAFEE 1990).
ER et al. 1984). Bone marrow activity is more prominent on 99mTc_
11lIn-Iabelled compounds can, however, give false- HMPAO-WBC scan than on lllIn-WBC scan; this can
positive results in trauma, arthritis, neuropathie joints give rise to false-positive results.
and tumours. Another disadvantage is the high radi-
ation dose. Furthermore, 11lIn labelling is a tedious
procedure as 40-60 cc of whole blood must be 14.2.3
withdrawn, labelled and re-injected. This is a prob- Magnetic Resonance Imaging
lem particularly in children. 11lIn-WBC test is more
expensive and not readily available, requiring special For the detection of osteomyelitis, magnetie reso-
orders. Despite these limitations, lllIn-labelled com- nance (MR) imaging is superior to 99mTc_MDP bone
pounds are probably the tracers of choiee for diagno- scintigraphy in terms of both sensitivity and spec-
sis and localization of osteomyelitis. Combined use of ificity (MAZUR et al. 1995). Thus several studies
lllIn-WBCs and 99mTc was found to have a high sen- have reported sensitivities and specificities for MRI
sitivity of 86% and a specificity of 94% when infec- of 92%-100% and 89%-100%, respectively (SCHAU-
tion occurred in the presence of pre-existing fr ac- WECKER et al. 1990). Similar to scintigraphy, the
ture, arthropathies, previous surgery or skin lesions MR appearance of osteomyelitis can be non-specific
(KOLINDOU et al. 1996). because infiltrating tumours and stress fractures can
give the same appearance (BOUTIN et al. 1998).
14.2.2.4 The excellent soft tissue and bone marrow con-
Indium-111 Labelled Polyclonallmmunoglobulin trast resolution and the multiplanar capability of MR
imaging offer greater detail than computed tomogra-
Clinieal studies indieate that this new agent is as effi- phy (CT) or conventional radiographs. Early changes
cacious as 11lIn-WBCs or 67Ga citrate for the evalua- in bone marrow signal on MR are a more sensitive
tion of focal infection. Its preparation is simpler than indicator of disease than the late-developing destruc-
that of llIIn-WBCs, evading phlebotomy and labo- tive bone lesions, whieh need further time to show
rious labelling methods. Unlike with 67Ga, there is radiographieally. MR imaging becomes positive after
no gastrointestinal or bone marrow uptake. A lower a few hours but a radio graph takes more than a week.
radiation dose is delivered, but lesion to background Active infections replace marrow fat with water sec-
activity ratios are relatively low. The agent is pre- ondary to oedema, exudate, hyperaemia and isch-
pared from pooled human serum gamma-globulin aemia. This leads to a high signal on T2-weighted
andlabelled with lllIn viaDTPA chelation ellln-IgG). images and a low to intermediate signal on Tl-
In the first 24 h the non-specific polyclonal IgG shows weighted images. Short tau inversion recovery (STIR)
lesion uptake equal to that of specific monoclonal and fat suppression before and immediately after con-
IgG uptake. Thereafter, the lesion uptake of specific trast enhancement improve the detection of osteo-
monoclonal IgG exceeds that of control non-specific myelitis and soft tissue infections.
IgG (NIJHOF et al. 1997). The tracer does not differ- The bone marrow signal becomes heterogeneous
entiate between infection and sterile inflammation. with the advent of chronicity. Granulation tissue
Indium is rarely taken up by neoplasms. False-posi- is hyperintense on T2" STIR and gradient echo
tive results have been reported in arthritis and Char- sequences (Fig. 14.4). Diffuse sclerosis can lower
cot's joint. the signal. When the primary signs of infection are
equivocal (low to intermediate signal on Tl weight-
14.2.2.5 ing and high signal on T2 weighting), secondary
Technetium-99m d,l-HMPAO Labelled Leucocytes signs of osteomyelitis will aid in the differential diag-
nosis from tumours, arthritis or neuropathie con-
Advantages of 99mTc-Iabelled over lllIn-labelled leu- ditions (MORISSON et al. 1998). Secondary signs of
cocytes include cost, availability, dosimetry, shorter infection include ulcer, cellulitis, abscess formation,
acquisition time and improved image quality. WBC cortical sequestration and subperiosteal abscess for-
scan using the lipophilic complex d,l-hexamethylpro- mation. Recently a penumbra sign on Tl-weighted
pylene amine oxime (HMPAO) was reported to have a images was reported to be helpful in differentiating
254 M. E. Abd EI Bagi et al.
subacute osteomyelitis from bone tumours, show- on X-rays. Changes of chronic osteomyelitis are weIl
ing a sensitivity of 75% and a specificity of 99% demonstrated by CT as there is no overlap, which is
(GREY et al. 1998). Contrast-enhanced MR imaging a problem on radiographs. The improved soft tissue
is a highly sensitive technique to diagnose musculo- contrast and the use of contrast enhancement are
skeletal infection and differentiate abscess from sur- useful for the diagnosis of bone and soft tissue infec-
rounding myositis or cellulitis (HOPKINS et al. 1995) tion, and particularly abscess formation. CT is espe-
(Fig. 14.5). Lack of contrast enhancement rules out cially important in patients in whom MR imaging is
infection with a high degree of certainty but pres- contra-indicated. It is also very useful when infec-
ence of enhancement cannot be used to exclude non- tion is superposed on comminuted fractures (LEDER-
infective inflammatory conditions entirely. Although MANN et al. 2000) (Fig. 14.6). CT can differentiate
MR imaging is increasingly useful in diagnosis and whether increased isotope accumulation is due to
the planning of surgical management, it is of lim- infection or new bone formation (Fig.14.7).A further
ited use for multifocal infection. This emphasizes great advantage of CT is its ability to accurately guide
the value of scintigraphy, which allows whole-body bone biopsies (Fig. 14.8a).
scanning; this capability is of particular use in chil-
dren, 7% of whom have multifocal disease.
14.2.5
Ultrasound
14.2.4
Computed Tomography Ultrasound is a quick, simple and inexpensive tech-
nique. Because it does not involve ionizing radiation
Computed tomography is superior to MR imaging for and does not require sedation, it is becoming more
visualization of bone destruction and gas formation popular in the imaging of children. The intro duc-
(RAM et al. 1981) (Fig. 14.6). CT demonstrates bone tion of high-resolution prob es in the 5- to 12-MHz
abnormalities much earlier than do radiographs. The range has rendered ultrasound a useful tool for
earliest sign on CT is increased attenuation values the evaluation of joint and superficial peri-articular
of the medullary cavity, an abnormality undetectable pathology.
a b
Fig. 14.6. a Confusing mixed appearance of intense enhancement and destruction on post-contrast MR imaging in a violated
knee. b CT dearly demonstrated gas formation, soft tissue swelling and periosteal elevation in a patient with non-united fracture
complicated by chronic osteomyelitis
Musculoskeletal infections may show ultrasound yses, infection can rapidly spread from metaphyse al
signs at as early as 1 or 2 days. Signs of knee infec- marrow with destruction of the growth plate and like-
tion include soft tissue abnormalities, cortical irregu- lihood of septic arthritis. Joint effusion is present in
larities and subperiosteal fluid collection. In addition, 70% of cases and focal growth retardation may occur.
ultrasound can reveal joint effusions and peri-artic- Infantile disease is areal challenge as it lacks systemic
ular abscesses or bursitis (FESSELL et al. 2000). Early and local clinical signs (DAVID et al. 1987). Between
diagnosis of septic arthritis requires analysis of joint the first year and puberty, in the juvenile form, vascu-
fluid (BUREAU et al. 1999). In this context, ultrasound- lar penetration of the growth plate no longer exists
guided fluid aspiration is a quick, simple and accurate and the infection localizes in the metaphyseal sinu-
procedure. Moreover, when there is not much fluid, soids. Raised pressure forces the exudate to track lat-
synovial biopsy can be performed under ultrasound erally through the channels of Volkman and Havers,
guidance (BUREAU et al. 1999) (Fig.14.8b). The lack of spreading to perforate the cortex and elevate the thick
widespread use of ultrasound for diagnosis of infec- but loosely attached periosteum (Fig. 14.9), forming
tions is due to lack of interest on the part of radiol- the characteristic subperiosteal collection. Spread to
ogists rather than limitations of the technique. The the joint space is unlikely in the juvenile form, except
field of view with high-resolution ultrasound is limit- in joints where the metaphyses are essentially intra-
ed but suffident for most superfidal infections. Ultra- articular, e.g. hips, elbow and shoulder. After growth
sound has limitations in deep-seated lesions, the vio- plate closure in adults and absorption of the epiphy-
lated knee, purely intra-osseous processes and detec- seal cartilage, infection can spread from metaphyseal
tion of occult or multiple sites. Anormal ultrasound marrow up to the subarticular region. Bone infection
scan does not exclude the presence of infection. is classified according to exact anatomical site, treat-
ment options and prognostic factors. According to
MADER, stage 1 is the early intramedullary stage of
a haematogenous spread. In stage 2 there is superfi-
14.3 dal osteomyelitis due to an infected wound. Stage 3
Common Knee Infections is characterized by full thickness cortical sequestra-
tion. Finally, stage 4 entails through and through bone
14.3.1 involvement which may require resection with stabili-
Osteomyelitis
zation (MADER et al. 1999). According to the modified cal malignancies who are on cytotoxic treatment, have
Roberts' criteria, osteomyelitis is classified according been proven to have gram-negative bacilli,Aspergillus
to the exact anatomieal site (ROBERTS et al. 1982). and Nocardia. Those with defects in humoral immu-
Type 1 refers to involvement of the central metaphy- nity, as in myeloma and leukaemia, tend to have encap-
seal region, type 2 to eccentric metaphyseal cortieal sulated bacteria and Strept. pneumoniae. Patients with
erosions, type 3 to eccentric diaphyseal erosion, type ceH-mediated immunodeficiency like AIDS tend to
4 to diffuse periostitis without erosions and type 5 acquire intraceHular pathogens like mycobacteria and
to primary epiphyseal plate lesions. We consider the Nocardia. Drug abusers have a combined deficiency of
lesion to be of type 6 when it involves both epiphyses neutrophil action and ceH-mediated immunity, and are
and metaphyses (Fig.14.1O). likely to acquire Pseudomonas and Candida (DIRSCHL
and KINDERS 1993). Although pyogenic infections
continue to cause the majority of bone infections,
non-pyogenie infections are caused by opportunistie
infections in the immunocompromised.An example is
Mycobacterium species and fungal colonies.
Osteomyelitis is considered acute if it is ofless than
1 month's duration, subacute if it has been present for
1-3 months and chronic if it has lasted for more than
3 months (GREY et al. 1998). Chronie osteomyelitis is
common around the knee, and predominantly occurs
in middle-aged males. A discharging sinus may be
the first presentation in chronic osteomyelitis. Radio-
graphs show a thiek periosteal and sclerotie bone
re action. FoHowing thrombosis of metaphyseal ves-
sels, cortical necrosis occurs, resulting in a devitalized
segment that forms asequestrum of detached, dense,
necrotic bone. Perioste al elevation leads to deposition
of new bone, forming an involucrum envelope of peri-
osteal bone around the dead sequestrum (Fig. 14.11).
The involucrum may have gaps, "cloacae", through
Fig. 14.10. T2-weighted sagittal image of a child's knee, show- which pus escapes. In adults, the periosteum is weH
ing a type 6 lesion involving both sides of the physes with attached and is less elevated but sinus tracts are
hyperintense lesions in the metaphyses (Zarge arrowhead) and common. Pathological fractures may be the present-
epiphyses (small arrowhead)
14.3.3
Pyogenic Myositis
14.3.4
Fig. 14.15. a High-resolution ultrasound showing extensive
Septic Bursitis
oedema and effusion in the medial belly of the gastrocne-
mius muscle due to infective myositis. b Normal appearance
Bursae are synovial cysts to facilitate gliding oflocomo- of the arrangement of the right gastrocnemius muscle fibres
tor system components. Sepsis is responsible for 30% on ultrasound
260 M. E. Abd EI Bagi et al.
Fig. 14.16. Coronal T2-weighted MR image showing diffuse Fig. 14.17. A ehild with painful knee swelling. Sagittal T2-
oedema of the right ealf muscles due to myositis weighted sequenee shows an oval distended hyperintense
fluid-filled strueture behind the knee, surrounded by soft
tissue oedema. There is a traee of effusion in the suprapatellar
reeess. Appearanees are of popliteal bursitis
14.3.5
Infective Cellulitis
14.4
Rare Bone Infections
14.4.1
Tuberculosis
~=iiii_ b,C
Fig. 14.19. a Young adult who presented with knee pain which was initially considered
degenerative. The patient responded weil to analgesics and anti-inflammatory drugs. Note
rarefaction at the medial femoral condyle (arrowheads). b The patient's symptoms recurred
and he underwent MR imaging. The cause of the symptoms was considered to be osteo-
chondritis at that time. c The patient deteriorated rapidly after 6 months. CT showed mili-
a ary mottling consistent with tuberculosis. (Courtesy of Dr. Munir Madkour)
262 M. E. Abd EI Bagi et al.
14.4.2 14.4.4
Brucellosis Mycetoma
Peripheral bone infection is rare in brucellosis but Mycetoma commonly affects the foot in sub-Saharan
the knee is one of the common extraspinal sites (AL Africa, India and parts of the Arabian peninsula. It
SHAHED et al. 1994). Synovial involvement is reported has distinct radiographie and MR findings. Radio-
in 81 % of all peripherallesions. This is an important logieal signs were classified by DAVIES (1958) and
differentiating point from tuberculosis, whieh tends MR imaging features were described by SHARIF et
to spare the joint untillate. Both diseases are prev- al. (1991). Infection proliferates beneath the skin to
alent in endemie areas and may co-exist. ELISA is form deep abscesses with sinus tracks and tumefac-
more sensitive than other serologieal tests. Advances tion leading to bizarre lesions (Fig. 14.23). The pres-
in molecular techniques, such as polymerase chain ence of coloured grains helps in the diagnosis and
re action (PCR), will soon become the method of choiee classification of the disease; these grains appear as
for diagnosis of brucellosis and other slow-growing signal-void spots on MR imaging.
Infection 263
nostie dilemma. Initial clinical symptoms are simi- (LEE and SARTORIS 1994). This can take the form
lar. Periosteal elevation and subperiosteal pustular of cellulitis, subcutaneous abscess, pyomyositis or
pockets are signs of infection (Fig. 14.9). The pres- osteomyelitis. Infection can be haematogenous even
ence of sizable soft tissue swelling and intense con- in adults. Local infections are commoner in drug
trast enhancement and breakthrough of the cortex abusers. HIV is blamed for the resurgence of tuber-
are features of osteomyelitis not seen in infarcts culosis, and the incidence of pyomyositis is rising
(Fig. 14.24). On plain films, infarcts are usually scle- in these patients. The commonest pathogen is Staph.
rotic whereas infections are usually lytic. Infection aureus, followed by Strep. pneumoniae. Because
pro duces a thicker periosteal re action and cortical unusual organisms can be encountered, aspirations
abnormalities. Both infections and infarcts affect the and biopsies are essential for accurate diagnosis.
medullary space of long bones and the short tubular Examples are Histoplasma capsulatum and bacillary
bones of the hands and feet. A pseudobone within angiomatosis, whieh is aggressively osteolytie. Drug
bone appearance is commoner in infection than in abusers ean aequire clostridial and non-clostridial
infarcts. No radiographie feature can give absolute myoneerosis or gas gangrene.
differentiation. By contrast, MR imaging is particu-
larly important for differentiation (Fig. 14.25).
On a bone scan, increased uptake may be due to an
infection or healing avascular necrosis, partieularly if 14.6
the scan is performed a week or more after the symp- Summary
toms. Addition of 67Ga can enhance the specificity:
patients with infarct will have decreased or normal The knee is a very common site for most bone
67Ga uptake. Furthermore,addition of a 99ffiTc-sulphur and joint infeetions. Rapid and aeeurate diagnosis is
colloid scan will show photon deficiency in infarcts. important to avoid eomplieations in this weight-bear-
Radiocolloid photon deficiency is not a sign of osteo- ing joint. The knee is also a eommon site for most
myelitis. tumours, peripheral degenerative disease, trauma
and infarets. Differentiation of these poses a diag-
nostie ehallenge. There is no gold standard for the
14.5.3 diagnosis of knee infeetion as no single modality is
The Immunocompromised Patient ideal all of the time. We reeommend an algorithm
for investigations of patients suspeeted to have knee
Musculoskeletal infections associated with immuno- infeetion (Fig. 14.26).
deficiency states are still rare (ESPINOZA and BERMAN
1999). Septie arthritis is commoner than osteomyeli-
tis. Both opportunistic and non-opportunistic infec-
tions can occur in the immunocompromised patient
a b
b c
Fig. 14.25. a A patient known to have sickle cell disease who presented with knee pain. Increased uptake is seen in the lower
femur and upper tibia on bone scan. This could have been due to infection or infarct. b Predominately sclerotic lower femur,
suggestive of osteonecrosis. c Geographical pattern ofbone marrow on MR image without periosteal or soft tissue involvement,
compatible with infarct
II. Signs of
I
2. Violated Knee
I
3. Normal
I
4. Signs ofEffusion
1
s. Signs ofnon-infective
1 1
osteomyelitis pathology
cr us
I
I
Normal Periosteal Soft tissue Abscess or Hot Normal Temperature
Elevation Oedema Bursitis
I I
Tc 99 m Isotope Bone Scan
•
Positive
t
Doubtful •
Negative
Treat •
Treat
- Ga 67. In WBC
- HMPA O, Co/loid Scan
•Re-evaluate Aspirate MRI
- ±,Aspiration I Biopsy
References
Abd EI Bagi ME, Sammak BM, Al Shahed MS et al (1999) limitations of magnetic resonance imaging in chronic post
Rare bone infections "excluding the spine". Eur Radiol traumatic osteomyelitis. Eur Radioll0:1815-1823
9:1078-1087 Lee DJ,Sartoris DJ (1994) Musculoskeletal manifestations of the
Al Shahed MS, Sharif HS, Haddad MC et al (1994) Imaging human immunodeficiency virus infection: review of imag-
features of musculoskeletal brucellosis. Radiographics ing characteristics. Radiol Clin North Am 32(2):399-411
14:333-348 Mader JT, Shirtliff M, Calhoun JH (1999) The host and the
Boutin RD, Brossmann J, Sartoris DJ et al (1998) Update on skeletal infection: classification and pathogenesis of acute
imaging of orthopaedic infections. Orthop Clin North Am bacterial bone and joint sepsis. Baillieres Clin Rheumatol
29(1):41-65 13(1):1-20
Bouza E,Munoz P (1999) Microorganisms responsible for osteo- Mazur JM, Ross G, Cummings RJ et al (1995) Usefulness of
articular infections. Baillieres Clin Rheumatol13(1):21-35 magnetic resonance imaging for the diagnosis of acute
Brook I (1999) Soft tissue and muscular infections (including musculoskeletal infections in children. J Paediatr Orthop
group A streptococci). Baillieres Clin Rheumatol 15:144-147
13(1 ):95-114 McAfee JG (1990) What is the best method for imaging focal
Bureau NJ, Chhem RK, Cardinal E (1999) Musculoskeletal infec- bone infections? J Nucl Med 31(4):413-416
tions: US manifestations. Radiographics 19:1585-1592 Merkel KD, Brown ML, Dewanjee MK et al (1985) Compari-
Datz FL, Thorne DA (1987) Cause and significance of cold son of indium labeled leucocyte imaging with sequential
bone defects on In III labeled leucocyte imaging. J Nucl Med technetium gallium scanning in the diagnosis of low grade
28:820-823 musculoskeletal sepsis. J Bone Joint Am Surg 67:465
David R, Barron BJ, Madewell JE (1987) Osteomyelitis acute Merkle E, Schulte M, Vogel J (1997) Musculoskeletal involve-
and chronic. Radiol Clin North Am 25(6):1171-1201 ment in cystic echinococcosis - areport of eight cases and
Davidson PT, Horowitx J (1970) Skeletal tuberculosis: a review review ofthe literature.AJR 168:1531-1534
with patients: presentation and discussion. Am J Med Morrison WB, Schweitzer ME, Granville BW et al (1998) Osteo-
48:77-84 myelitis of the foot: relative importance of primary and
Davies AGM (1958) The bone changes of Madura foot. Obser- secondary MR signs. Radiology 207:625-632
vations in Uganda Africans. Radiology 70:841-847 Nijhof MW, Oyen WSG, Van Kemper A et al (1997) Evaluation of
Dirschi DR, Kinders A (1993) Osteomyelitis: common causes infections of the locomotor system with indium lll labeled
and treatment recommendations. Drugs 45(1):29-43 human IgG scintigraphy. J Nucl Med 38: 1300-1305
Donohoe KJ (1998) Selected topics in orthopaedic nuclear Palestro q, Torres MA (1997) Radionuclide imaging in ortho-
medicine. Orthop Clin North Am 29(1):85-99 paedic infection. Semin Nucl Med 27(4):334-345
Eigazzar AH, Abd EI Dayem HM, Clark JD et al (1995) Pandey J, Talib VH (1993) Laboratory diagnosis of tubercu-
Multimodality imaging of osteomyelitis. Eur J Nucl Med losis, use of ELISA and PCR. Indian J Pathol Microbiol
22(9):1033-1063 35:512-518
Espinoza LR, Berman A (1999) Soft tissue and osteoarticular Peltola H, Vahvanen J (1984) A comparative study of osteomy-
infections in HIV infected patients and other immunodefi- elitis and purulent arthritis with special reference to aetiol-
cient states. Baillieres Clin Rhematol13(1):115-128 ogy and recovery. Infection 12:75
Fessell DP, Jacobson JA, Craig J et al (2000) Using sonography Perez LC (1999) Septic arthritis. Baillieres Clin Rheumatol
to reveal and aspirate joint effusions. AJR 174:1353-1362 13(1):37-58
Frink SJ, SnearlyW, Parsons TW III (1998) Malignant tumours Perlman MH, Patzakis MJ, Kumar PJ et al (2000) The incidence
about the knee.Am J Knee Surg 11(4):257-266 of joint involvement with adjacent osteomyelitis in paedi-
Grey AC, Davies AM, Mangham DC et al (1998) The 'penum- atric patients. J Paediatr Orthop 20(1 ):40-43
bra sign' on Tl' weighted MR Imaging in subacute osteo- Phemister DB, Hatcher CM (1933) Correlation of pathological
myelitis, frequency cause and significance. Clin Rad 53(8) and roentgenological findings in the diagnosis of tubercu-
587-592 losis. Arthritis Am J RoentgenoI29:736-7 52
Gupta NC, Prezio JA (1988) Radionuclide imaging in osteomy- Ram P, Martinez S, Korobkin M et al (1981) CT detection
elitis. Semin Nucl Med 19(4):287-299 of intraosseous gas. A new sign of osteomyelitis. AJR
Hobo TZ (1921) Zur der Vitolfarbungslehere Pathogenese de 137:721-723
Akuten Haematogenen Osteomyelitis. Mit Beruckishtigun Rasool MN, Govender S, Naidoo KS (1994) Cystic tuberculosis
Univ Imp Kioto 4:1-29 ofbone in children. J Bone Joint Surg [Br] 76B:I13-117
Hopkins KL, Li KC, Bergman G (1995) Gadolinium DTPA Roberts JM, Drummond DS, Bread AL et al (1982) Subacute
enhanced MRI of musculoskeletal infectious processes. hematogenous osteomyelitis in children, a retrospective
Skeletal RadioI24(5):325-330 study. J Pediatr Orthop 2:249-254
Hugosson L, Nyman R, Ny Brismar R (1996) Peripheralosteoar- Roddie ME,Peter AM,Osman S et al (1988) Osteomyelitis.Nucl
ticular and soft tissue tuberculosis. Acta RadioI37:512-516 Med Commun 9:713-717
Jaramillo D, Treves St, Kaser JR et al (1995) Osteomyelitis and Roine I, Faingezicht I, Arguedas A (1995) Serial serum C-reac-
septic arthritis in children; appropriate use of imaging to tive pro tein to monitor recovery from acute osteomyelitis
guide treatment. AJR 165(2):399-403 in children. Paediatr Infect Dis J 14(1):40-44
Kolindou A, Li Y, Ozker et al (1996) In lll WBC imaging of Rooney J Jr, Crocco JA, Kramer S et al (1976) Further dissec-
osteomyelitis in patients with underlying bone scan abnor- tions in tuberculin reaction in active tuberculosis. Am J
malities. Clin Nucl Med 21(3):183-191 Med 60:517-522
Ledermann HP, Kaim A, Bongartz G et al (2000) Pitfalls and Sammak BM, Abd EI Bagi ME, Al Shahed MS et al (1999) Osteo-
Infection 267
myelitis: a review of currently used imaging techniques. Tui H, Chink LK (1997) Osteomyelitis of the spine due to sal-
Eur Radiol 9:894-900 monella infection. Conservative treatment with quinolone:
Schauwecker DS, Park H, Mock BH et al (1984) Evaluation of a case report. Can J Surg 40:48-50
complicating osteomyelitis with Tc99m MDP, indium 111 Wane R (1997) Salmonella infection in siekle cell disease. A
and Ga67 citrate. J Nucl Med 25:849-958 clear and present danger. J Paediatr 130:350-351
Schauwecker DS,BraunsteinEM,WheatLJ (1990) Diagnosticimag- White LM, Schweitzer ME, Deely DM et al (1995) Study of
ing of osteomyelitis. Infect Dis Clin North Am 4(3):441-461 osteomyelitis: utility of combined histologie and mierobio-
SharifH, Clark D,Aabed M et al (1991) Mycetoma comparison logie evaluation of percutaneous biopsy samples. Radiol-
of MR imaging with CT. Radiology 178:865-870 ogy 197(3):840-842
Truetta J (1957) The three types of acute haematogenous osteo- Young EJ (2001) Immunology of brucellosis in Madkour's
myelitis. A clinieal and vascular study. J Bone Joint Surg [Br1 brucellosis. Springer, Berlin Heidelberg New York, pp
41:671-680 44-45
15 Arthritis
CHARLES WAKELEY and IAIN WATT
CONTENTS
is not as marked as in a metaearpophalangeal joint.
15.1 Introduetion 269
15.2 How Does One Investigate the Knee However, synovium is abundant, for example, in the
Radiologieally? 270 eapacious suprapatellar poueh. Henee, synovitis may
15.2.1 Basie Principles 270 be marked, whilst aetual bone erosion is seanty. Fur-
15.2.2 What Do the Various Imaging Teehniques ther, the knee has two large fibroeartilages. This mix-
Teaeh Us? 270 ture ofhyaline eartilage and fibroeartilage seems eru-
15.3 What Are the Arthritides That Involve
the Knee? 272 cial to the development of ehondroealcinosis. Again,
15.3.1 Osteoarthritis (OA) 272 the knee is the eommonest site for episodes of erystal
15.3.2 Inflammatory Arthritides 274 shedding - pseudogout.
15.4 What Other Conditions in the Knee May Present 2. The knee is a major weight-bearing joint. It is
as Arthritis? 276 at risk of injury especially from many and various
15.4.1 Tumours 277
15.4.2 Hypertrophie Osteoarthropathy 278
sporting aetivities. Further, internal derangement
15.4.3 Foreign Body Reaetion 278 enhanees the risk of "seeondary" OA. The relation-
15.4.4 Bleeding Disorders 278 ship between meniseal and hyaline eartilage pathol-
15.4.5 Neuropathie Involvement 279 ogies has only reeently begun to reeeive attention.
15.4.6 Infeetions 279 Certainly, the failure of normal meniseal movement
15.4.7 Metabolie Conditions 280
15.4.8 Systemie Metabolie Disorders 281 seems related to hyaline eartilage disease and, in
15.5 Summary 281 turn, the menisci are extruded from the joint line by
Further Reading 282 osteophytosis. Abnormalloading on the knee is eru-
cial to joint damage. In the Western world at least,
obesity is beeoming more prevalent, and with it an
inereasing incidenee of tibiofemoral OA.
3. The knee is not one, but at least two joints. A
15.1 tendeney exists to lump both together, but epidemio-
Introduction logical studies suggest that patellofemoral and tibio-
femoral OA are different disease states.
Every joint is anatomically different, and eaeh ana-
tomieal site earries individual risks and suseeptibil- When faeed with a patient whose eomplaint would
ity. In the ease of a knee, three important eonsider- seem to be ofknee arthritis, three eonsiderations arise:
ations influenee the manifestations of arthritis: 1. Is this loeal or systemic disease? If the former, then
1. The knee joint has the largest hyaline eartilage greater eonsideration will be given to infeetion,
volume of all joints. Henee, it is the "target" tissue in trauma or benign synovial tumour. If systemie,
the various knee arthritides, most noticeably osteo- have images of other important joint sites been
arthritis (OA). Even in rheumatoid disease the pre- obtained? Clearly, the most relevant are the hands
dominant effeet is diffuse hyaline eartilage thinning, and feet, sinee here almost all the major arthritides
rather than erosion of bone. Why? Beeause the "bare have manifestations, often at an earlier stage than
area" (where synovium eontaets bone direetly) is pro- in the knee. Thus, the diagnosis may be easier with
portionately small. Thus, bone erosion by synovium this knowledge.
2. Has there been a history of trauma, even remote?
C. WAKELEY, MD; I. WATT, MD Both the immediate effeets of meniseal or liga-
Department of Clinieal Radiology, Bristol Royal Infirmary, mentous trauma ean be masked as "joint pain" or
Bristol, BS2 8HW, UK "arthritis", but also the late sequelae include OA.
graded screens. The length of individual bones, limb tional radiographs, and also axial eT images acquired
shortening, joint angulation and the "clinical load through the hips, knees and ankles facilitate calcula-
line" can be measured. These are of particular impor- tion of the torsional profile of individual bones, again
tance in pre-operative surgical planning. for pre-operative surgical planning.
15.2.2.2 15.2.2.3
Computed Tomography Radionuclide Radiology
In general, MRI has become the cross-sectional imag- Skeletal scintigraphy using technetium-99m labelled
ing modality of choice owing to the ability to demon- diphosphonate bone scanning agents is useful for
strate all the appropriate tissue planes. However, scout identifying active sites of disease, and also their dis-
films (scanograms) can replace full-Iength conven- tribution. Various subsets of scintigraphic patterns
have been associated with the evolution and progno-
sis of OA (see below).
15.2.2.4
Ultrasound
15.2.2.5
Magnetic Resonance Imaging
has been described in patients with OA that not only local reduction in bone density and diffuse hyaline
permits classification of various subsets of knee OA cartilage thinning. In all, increased joint fluid will be
but also relates these changes to likely disease pro- seen on ultrasound with varying degrees of increased
gression and prognosis. blood flow on Doppler. MRI is also non-specific
OA secondary to other causes ofjoint damage is the unless, or until, other features emerge, such as ero-
result of a combination of secondary reparative pro- sion of bone.
cesses superimposed upon pre-existing joint pathol-
ogy. The initial joint insult may be due to congenital 15.3.2.1
or developmental deformity, trauma, previous infec- Rheumatoid Arthritis
tion or pre-existing inflammatory arthritis (e.g. rheu-
matoid arthritis). The compartmental distribution of Rheumatoid arthritis (RA) is the most common
the OA will depend on the initial joint deformity or specific inflammatory arthritis to involve the knee.
joint insult. Clues to the secondary nature of the OA Others include the various spondyloarthritides (espe-
include: cially ankylosing spondylitis, psoriasis and Reiter's
1. Moulding abnormalities. Is the joint squared off or syndrome).
are the former epiphyseal areas larger than usual, In RA the disease process is essentially an inflam-
suggesting a childhood overgrowth? matory synovitis with secondary bone and articular
2. Symmetrical involvement. Are all three compart- cartilage erosion that progresses from the marginal
ments involved uniformly? If so, is there evidence "bare areas" of the joint inwards. The so-called bare
ofbone softening as in rheumatoid disease? Anoth- areas are the intracapsular portions of the bone at the
er clue is the relative absence of osteophyte in the edge of joints that are not covered by articular carti-
face of extensive hyaline cartilage loss. lage. Characteristic features of RA include:
1. Soft tissue swelling and joint effusion
Sudden worsening of OA may be associated with 2. Marginal erosions
either further rapid focal hyaline cartilage loss or 3. Diffuse joint space narrowing
increased stress. The latter includes proximal tibial 4. Subarticular geodes
and femoral condylar insufficiency fractures and 5. Demineralisation of the skeleton
spontaneous osteonecrosis of the knee. Character-
istically subchondral insufficiency fractures involve Unlike OA, the changes in RA tend to involve all
predominantly the medial femoral condyle in elderly three compartments of the joint diffusely, with a lack of
osteoporotic women or middle-aged men. The sub- both subchondral sclerosis and osteophyte formation.
sequent subchondral collapse causes rapid advance-
ment of OA in that compartment. Insufficiency frac- Tissue Swelling and Joint Effusion. Not only may the
tures may result in increased eccentric loading of joint effusion be pronounced, but also, underlying
one of the compartments and resultant angulation. synovial hypertrophy and inflammation results in an
This feature has been previously ascribed to avas- ill-defined outline on plain films. Thus, it is difficult
cular necrosis (idiopathic medial femoral condylar to distinguish between the synovial hypertrophy and
necrosis, or "SONC" - sudden onset necrosis of the the joint effusion. Ultrasound and MRI can reliably
condyle). Whether or not true spontaneous avascular distinguish between these features. The latter may be
necrosis occurs in the knee is debatable. Both medul- used to assess the full extent of joint involvement and
lary and cortical avascularity do occur with system- to document therapy effects. Dynamic MRI using
ic disease as with steroid therapy or alcohol abuse. intravenous contrast agents (Fig. 15.6) and the cre-
Indeed, the femoral condyles and lower femoral med- ation of enhancement profiles for synovium is a
ullary canal are typical sites (see Chap. z). potent means of assessing the value of disease-mod-
ifying drugs.
a b
Fig.15.6a, b. Dynamic MRI of the knee. Sagittal Tl-weighted images; pre-contrast (a) and post-contrast (b). Note the low signal
joint effusion in the suprapatellar pouch with a high-signal peripheral band of synovial enhancement on the post-contrast
image (b)
Such rupture is associated with acute calf pain, rais- neck and metacarpal heads. Typically, the lesion is
ing the differential diagnosis of deep venous throm- found centrally in the upper tibia, subadjacent to the
bosis. Ultrasound is a safe and secure means of diag- tibial spines. The aetiology is obscure. Smaller, "cystic"
nosing the unruptured cyst, and may be useful in the lesions usually arise as part of secondary OA.
case ofleakage. Often the diagnosis of venous throm-
bosis in the calf will need venography, however. Diffuse Regional Osteopenia. In active RA, diffuse
regional osteopenia may be pronounced in the typical
Marginal Erosions. Marginal erosions in RA of the subchondral, transmetaphyseal regions associated with
knee are not as common as in the joints of the hand. algodystrophy. Later, RA is associated with bone atro-
They occur at the bare areas. These are proportion-
ately smaller in the knee. Consequently, bony ero-
sion at the knee is considerably less common than in
metacarpophalangeal joints. However, as in the latter,
ill-defined margins to the erosions suggest that they
are more acute, whereas corticated margins suggest
biological inactivity. No new bone formation is asso-
ciated with the rheumatoid erosions, whereas new
bone formation does accompany erosions associated
with spondyloarthritides.
15.3.2.2
Spondyloarthritis
a b
ciated with bone involvement, usually irregular bone tive former epiphyseal regions. An associated non-
destruction at capsular attachments. Obviously, such specific synovitis and soft tissue swelling is an addi-
rare tumours must be considered in the differential tional feature (Fig. 15.10).
diagnosis of a younger patient with knee pain and
a popliteal swelling. The use of ultrasound to distin-
guish between the cystic Baker's cyst and the echo- 15.4.3
genic synovioma is vital. Foreign Body Reaction
15.4.2
Hypertrophie Osteoarthropathy
the known history of haemophilia, Christmas disease in OA. Characteristieally joint failure is associated
or von Willebrand's disease. Significant involvement with gross osteochondral fragmentation of the joint
in childhood results in epiphyse al hypertrophy and surfaces, filling the joint cavity with debris and result-
premature growth plate fusion from the episodes of ing in end-stage collapse of the joint (Fig. 15.12).
hyperaemia. Excess iron deposition in the various
bleeding dis orders may result in increased radio-
opacity of synovium on plain film (Fig. 15.11) and 15.4.6
reduced synovial signal on all pulse sequences on Infections
MRI. Recurrent haemarthrosis after joint replace-
ment may be due to arteriovenous malformation in An infectious arthritis is relatively common in the
synovium. It is not known whether this phenome- knee, particularly due to atypieal organisms includ-
non arises de novo, or is an effect of surgery. The ing gonococcus and tuberculosis. Risk populations
major end-result is secondary OA. The typieal fea- include intravenous drug users and those with sex-
tures include diffuse hyaline cartilage loss, subchon- ually transmitted disorders. Further, the knee is a
dral radiolucencies, scanty osteophyte and opaque common site of presentation for tiek-related disor-
synovium. Childhood onset is associated with the ders, for example Lyme disease. Initial joint aspira-
secondary growth disturbances described above. tion may not be diagnostie, but persistent synovitis
results in significant hyaline cartilage destruction.
Similarly, the knee may be the presenting feature of
15.4.5 local bone infection, especially Brodie's abscess.Acute
Neuropathie Involvement septie arthritis (see also Chap. 14) in its early phase
may be indistinguishable from other acute infiam-
Neuropathie involvement of the knee is uncommon matory arthropathies. Septic arthritis is important,
and usually presents as the hypertrophie variant of a since it is treatable and, if undiagnosed and untreat-
Charcot joint. The atrophie neuropathie joint is rare ed, results in joint destruction. Thus, it is important
in the knee. A non-specific synovitis and joint effu- always to consider infection as a cause of an acute
sion may be the first signs. This may be clinieally arthritis and always to consider aspiration of the
silent, although it is worth remembering that Char- joint for bacterial mieroscopy as well as crystal analy-
cot's original description included painful knees. Sub- sis. In the early phase of septie arthritis, there is syno-
sequent hypertrophie changes with subchondral scle- vial thickening,joint effusion and soft tissue swelling.
rosis and osteophyte formation are seen, as described On conventional radiographs, the distortion of asso-
Fig.15.11. Haemophilia. Lateral conventional radiograph of the Fig. 15.12. Neuropathie joint. AP conventional radiograph of
knee demonstrating the high-density radiopaque synovium, the knee showing gross hypertrophie destructive change and
best appreciated in the suprapatellar pouch disorganisation
280 C. Wakeley and I. Watt
ciated fat planes may be difficult to visualise owing asymptomatie 80 year olds exhibit chondrocalcinosis
to the soft tissue oedema obscuring the normal out- on knee X-rays. Patients with hyperparathyroidism or
lines of fat planes. An associated hyperaemia usually haemochromatosis have a higher incidence of chon-
results in peri-artieular osteopenia. If left untreated, drocalcinosis than the normal population. The knee
septic arthritis will lead to cartilage destruction fol- is the classieal site for CPPD, as typically deposition
lowed by subchondral bone destruction and even- occurs in those joints with both fibro- and hyaline car-
tually collapse of the entire joint. The end-result is tilage within them. CPPD deposition results in linear
either fibrous or bony ankylosis. or granular calcification of fibro- and hyaline carti-
Skeletal scintigraphy is often helpful in distin- lage (Fig. 15.13). In gross cases, it will also be seen in
guishing between soft tissue inflammation and bone the capsule and at entheses. Apart from being a clas-
or joint involvement. MRI not only demonstrates the sieal site of acute crystal shedding ("pseudogout"),
synovial component, but also assesses the hyaline as superficiallayers of hyaline cartilage are shed into
cartilage and underlying bone marrow. the joint, the knee may exhibit GA, often with a
florid hypertrophie response (Fig. 15.14). It is doubt-
ful whether any real difference exists between GA with
15.4.7 and OA without CPPD crystals. Although PFJ OA has
Metabolie Conditions been emphasised in association with CPPD, this may
be aseparate subset of OA as such, and the crystal-
Two main subsets of disease affect the knee: related aetiology is again dubious.
- Those associated with the deposition of crystals in
or around the joint BCP. Calcium hydroxyapatite is most frequentlydepos-
- Those due to systemie metabolie disease ited in peri-artieular locations (such as the supraspi-
natus tendons of the shoulder). However, it is also
15.4.7.1 obtained from joint aspirations in patients with a
Crystal Deposition form of atrophie destructive OA of the knee and
shoulder. It is important to emphasise that aspiration
Three main crystals are associated with a major of fluid from normal knee joints will yield BCP crys-
inflammatory synovitis in humans. They comprise tals. They are a component of normal joint fluid. They
calcium pyrophosphate dihydrate (CPPD), calcium are in gross excess in atrophie destructive OA, how-
hydroxyapatite (also known as basie calcium phos- ever. The main radiographie features of the latter are
phate: BCP) and monosodium urate monohydrate rapidly progressive bone destruction in the absence
(MSUM). The role of crystals in the pathogenesis of of a secondary bone response. Large (often haemor-
arthritis is debated. The current view is that crystals rhagic) joint effusions containing abundant BCP crys-
are really a disease marker and not simply part of tals are obtained. As with CPPD, the role of these crys-
cause and effect. An association between CPPD and a
hypertrophie bone response and between BCP and an
atrophie bone response is recognised. These related
processes are thought to be caused by unknown stern
factors resulting in either a hypertrophie or an atro-
phie response to joint insult or failures. However,
these processes are processes and not diseases, and
not mutually exclusive. For example, the hypertrophie
form of disease may progress to an atrophie form. The
presumed systemic controllers of the various bone
responses to injury are not known. To add to the
confusion over the cause-effect relationship between
crystals and disease processes, there is also a degree
of association between these diseases. For example,
CPPD shares a strong association with primary hyper-
parathyroidism, gout and haemochromatosis.
Fig. 15.13. Chondrocalcinosis. AP conventional radiograph of
CPPD. Radiologieal chondrocalcinosis is a normal,age- the knee demonstrating chondrocalcinosis of the menisci and
related phenomenon. Approximately 40% of normal, hyaline cartilage
Arthritis 281
become more obese. Further, our ageing, "greying" all things, the radiologist is there to serve the patient
population rightfully demands to remain active. Yet, and steer clinical management in their best interest.
one potent risk to that mobility is knee arthritis. Radi- With a large joint like the knee, that role is crucial.
ology has a pivotal role in demonstrating the morpho-
logical changes in arthritis and in distinguishing them
from the normal ageing processes. This chapter has
summarised the major arthritides and the pitfalls in Further Reading
diagnosis that may be made. Most frequently, a radio-
logical report is made to a family physician, who will Resnick D (ed) (1995) Diagnosis of bone and joint dis orders,
never see the film and reHes solelyon the report. As in 3rd edn. Saunders, Philadelphia
16 Assessment of Knee Cartilage
SOUHIL ZAIM, ALl GUERMAZI, JOHN A. LYNCH, CHARLES G. PETERFY, HARRY K. GENANT
16.2 16.3
Anatomy, Function, and Structure MR Appearance
Artieular cartilage consists of sparsely scattered chon- The appearance of cartilage on MR imaging reflects
drocytes embedded in an extracellular matrix com- its structure and composition. Contrast results mainly
posed predominantly of collagen fibers, aggregated from the water content, i.e. the proton density. The
proteoglycans, and water. The collagen fibers impart high water content of artieular cartilage forms the
tensile strength to the matrix and res ist the swelling basis for the MR signal. However, this fundamental
of the proteoglycan gel. The proteoglycans are com- MR imaging signal is modulated by a number of pro-
posed of core proteins to whieh are attached numer- cesses, including TI relaxation, T2 relaxation, magne-
ous glycosaminoglycans, such as keratan sulfate and tization transfer and water diffusion. The thickness of
chondroitin sulfate, which have pendant sulfate and cartilage also poses achalienge as cartilage measures
carboxylate groups that are ionized under physiolog- onlya few millimeters at the most, the patellar carti-
ie conditions. These negatively charged moieties repel lage being the thickest. Not surprisingly, the patellar
each other and attract sodium ions that in turn draw cartilage has been the focus of most in vivo studies.
water osmotieally into the cartilage. The combination
of electrostatic repulsion and osmotie pressure pro-
duces a swelling pressure that keeps the artieular car- 16.3.1
tilage inflated and the collagen fibers under tension Normal Appearance and Artifacts
(AKESON et al. 1995).
The structure of cartilage is highly organized and The thiekness of artieular cartilage varies in the knee,
composed of several contiguous layers (Fig.16.1). The with the thiekest cartilage found at the patella and the
superficial or tangential zone is composed of tightly thinnest at the tibial plateau. At 1.5 Tesla, the Tl of car-
packed collagen fibers that are oriented parallel to the tilage is about 700 ms and the T2 varies between 30 ms
artieular surface, and the transitional or intermediate at the radial zone and 70 ms in the transitional zone
zone, just below the superficial zone, consists of col- (DARDZINSKI et al.1997). This pattern ofT2 variation,
lagen fibers with oblique orientations. The radial or also known as T2 anisotropy, is the result of the orien-
deep zone is characterized by collagen fibers that are tation of organized collagen fibers relative to the static
oriented perpendicular to the subchondral bone plate magnetic field (Ba)' or the "magie-angle" effect (ERICK-
and are anchored in the calcified zone of cartilage. SON et al. 1996). The magie-angle effect explains the
observed layering of cartilage related to its ultrastruc-
Superficial tural disposition when high-resolution imaging, espe-
~ne cially in vitro imaging with special attention to the
~nsitional
zone
technique, is performed (RUBENSTEIN et al. 1993; XIA
et al.1997).Additionally, the magie-angle phenomenon
Upper radial
is responsible for areas of increased signal observed
~ne at curved areas of cartilage, typically at the femoral
trochlea (Fig.16.2) and femoral condyles.
Deep radial
~ne
Calcified
~ne
16.3.3
Fat-Suppressed and Fast Spin Echo Imaging
16.3.4
Fat-Suppressed Spoiled Gradient Echo Imaging
a _ _ _
b
Fig. 16.4a-c. Sagittal conventional spin echo images with fat suppression
and TE values of 30 ms (a), 60 ms (b) and 90 ms (c) of the same anatomical
section. Cartilage loses signal gradually with apparent thickness decrease
c as T2 weighting increases (images are with same windowinglleveling)
Fig. 16.5. Focal defect of cartilage on fat -suppressed fast spin Fig.16.6. Edema and chondromalacia on a fat-suppressed fast
echo T2-weighted MR image. A large focal defect (arrowhead) spin echo T2-weighted MR image. A large area of chondroma-
is nicely within the medial posterior femoral cartilage and lacia of the weight -bearing femoral cartilage is associated with
highlighted by the adjacent synovial fluid edema of the subchondral bone marrow (arrowheads)
acquisition of thin slices in a three-dimensional mode ness (PETERFY et al. 1994a; ECKSTEIN et al.1996). This
with high contrast of cartilage (Fig. 16.7). It is widely technique requires, however, long imaging times of
available (3D SPGR, 3D FLASH) and has become the approximateiy 12 min for the entire knee and is prone
most sensitive clinical technique for delineating artic- to susceptibility artifacts and heterogeneity of satura-
ular cartilage lesions (DISLER et al. 1996; RECHT et tion of signal from fat. It does not allow analysis of
al. 1996) and measuring cartilage volume and thick- intrasubstance signal evaluation of the cartilage.
Assessment of Knee Cartilage 287
16.4.4 16.5.1
Diffusion Imaging Cartilage Volume and Thickness Measurement
Water diffusion in cartilage also contributes to signal Measurements of cartilage thickness and volume are
loss on T2-weighted MR images. Because water mole- typically taken from fat-suppressed 3D spoiled GRE
cules that have changed positions during a portion of images (Fig. 16.8). The earliest methods (PETERFY et
the MR imaging acquisition can no longer contribute al. 1994a, b; RECHT et al. 1996; ECKSTEIN et al. 1994;
maximally to the net signal, a resulting loss of phase MARSHALL et al. 1995; PILCH et al. 1994) involved
coherence is proportional to the distance traveled by region-growing techniques on a slice-by-slice basis,
the diffusing water protons and is, therefore, worse with manual tracing of subtle boundaries, where car-
on long-TE images. The presence of proteoglycans, tilage surfaces were in contact or where overlying tis-
particularly chondroitin sulfate, in normal cartilage sues with similar intensities occurred. These meth-
inhibits water diffusion and keeps this effect relative- ods were tedious and time consuming, requiring
ly small, although with very strong gradients and many subjective decisions from experienced human
specialized phase-sensitive pulse sequences, water observers, especially in regions where partial volume
diffusion can be demonstrated and quantified in averaging had occurred, but can be relatively reliable
normal cartilage (BURST EIN et al. 1993). With car- for measuring total cartilage volumes (ECKSTEIN et
tilage degeneration and proteoglycan loss, however, al. 1994, 1998a, b; HARDY et al. 2001; TIESCHKY et al.
water diffusion has been shown to increase consid- 1997). Such techniques have been used to examine
erably. Accordingly, diffusion may playa more sig- progression in osteoarthritis (PETERFY et al. 1998),
nificant role in cartilage signal modulation in knee the effects of exercise on cartilage (ECKSTEIN et al.
osteoarthritis. 1998c), and the distribution of cartilage within the
knee (ECKSTEIN et al. 1998d).
Other researchers (COHEN et al. 1999; LYNCH et al.
16.4.5 2000; SOLLOWAY et al. 1997; STAMMBERGER et al. 1999,
Sodium Magnetic Resonance Imaging 2000) have developed more automated techniques that
use information on edge strengths within the image.
Another imaging technique that has shown pro mise A technique which uses 2D deformable active shape
with respect to evaluation of the biochemical status models has been described (COHEN et al. 1999), and
of cartilage is sodium MR imaging (INSKO et al. preliminary results of the use of a 3D template-based
1997). Similar to the use of ionic gadolinium com- segmentation technique have also been described. A
pounds, sodium imaging also allows the assessment semiautomated technique using B-spline snakes has
of the proteoglycan content of cartilage. Also based been used to delineate cartilage boundaries on a slice-
on the charge-related loss paralleling proteoglycan by-slice basis (LYNCH et al. 2000), and a 3D model-
depletion, sodium imaging can depict the same based approach has been used to segment many struc-
regions. Unfortunately, this technique requires the tures within the knee, including cartilage. A compari-
use of sophisticated MR imaging equipment and son of a manual and a semiautomated segmentation
high-resolution imaging. In addition, it has very low technique has also been performed (STAMMBERGER
inherent signal-to-noise ratio, making it of question- et al. 1999). All these automated or semiautomated
able clinical value in its current form. techniques have the potential to reduce the amount of
Assessment of Knee Cartilage 289
Fig. 16.8. Quantification of cartilage volume using a 3D SPGR sequence and segmentation software. The outline of cartilage is
obtained on a slice-by-slice basis, allowing calculation of total volume and data that can be expressed as thickness mapping
References
Akeson WH, Amiel DA, Gershuni DH (1995) Articular cartilage relationships of normal cartilage volumes of the human
physiology and metabolism. In: Resnick D (ed) Diagnosis knee joint: assessment by MRI. Anat Embryol (Berl)
of bone and joint disorders, 3rd edn. Saunders, Philadel- 197:383-390
phia, pp 769-790 Erickson SJ, Waldschmidt JG, Czervionke LF, Prost RW (1996)
Bashir A, Gray ML, Boutin RD, Burstein D (1997) Glycos- Hyaline cartilage: truncation artifact as a cause of trilaminar
aminoglycan in articular cartilage: in vivo assessment appearance with fat-suppressed three-dimensional spoiled
with delayed Gd DTPA2- -enhanced MR imaging. Radiology gradient-recalled sequences. Radiology 201:260-264
205:551-558 Gray ML, Burstein D, Lesperance LM, Gehrke L (1995) Magne-
Bredella MA, Tirman PF, Peterfy CG, Zarlingo M, FeIler JF, Bost tization transfer in cartilage and its constituent macromol-
FW, Belzer JP, Wischer TK, Genant HK (1999) Accuracy ecules. Magn Reson Med 34:319-325
of T2-weighted fast spin-echo MR imaging with fat satu- Hardy PA, Nammalwar P, Kuo S (2001) Measuring thickness of
ration in detecting cartilage defects in the knee: compar- articular cartilage from MR images. J Magn Reson Imaging
ison with arthroscopy in 130 patients. Am J Roentgenol 13:120-126
172:1073-1080 Hargreaves BA, Gold GE, Lang PK, Conolly SM, Pauly JM, Berg-
Burstein D, Gray ML, Hartman AL, Gipe R, Foy BD (1993) Dif- man G, Vandevenne J, Nishimura DG (1999) MR imaging of
fusion of small solutes in cartilage as measured by nuclear articular cartilage using driven equilibrium. Magn Reson
magnetic resonance (NMR) spectroscopy and imaging. J Med 42:695-703
Orthop Res 11:465-478 Insko EK, Reddy R, Leigh JS (1997) High resolution, short echo
Chan WP, Lang P, Stevens MP, Sack K, Majumdar S, Stoller time sodium imaging of articular cartilage. J Magn Reson
DW, Basch C, Genant HK (1991) Osteoarthritis of the knee: Imaging 7: 1056-1059
comparison of radiography, CT, and MR imaging to assess Lynch JA, Zaim S, Zhao J, Peterfy CG, Genant HK (2000) Car-
extent and severity. Am J RoentgenoI157:799-806 tilage segmentation of 3D MRI scans of the osteoarthritic
Chandnani VP, Ho C, Chu P, D Trudell D, Resnick D (1991) Knee knee combining user knowledge and active contours. Proc
hyaline cartilage evaluated with MR imaging: a cadaveric SPIE 3979:925-935
study involving multiple imaging sequences and intraar- Lynch JA, Zaim S, Zhao J, Peterfy CG, Genant HK (2001) Auto-
ticular injection of gadolinium and saline solution. Radiol- mating measurement of subtle changes in articular carti-
ogy 178:557-561 lage from MRI of the knee by combining 3D image registra-
Cohen ZA, McCarthy DM, Kwak SD, Legrand P, Fogarasi F, Ciac- tion and segmentation. Proc SPIE 4322:431-439
cio EJ,Ateshian GA (1999) Knee cartilage topography, thick- Marshali KW, Mikulis DJ, Guthrie BM (1995) Quantification of
ness and contact areas from MRI: in vitro calibration, and articular cartilage using magnetic resonance imaging and
in vivo measurements. Osteoarthritis Cartilage 7:95-109 three-dimensional reconstruction. J Orthop Res 13:814-823
Dardzinski BJ, Mosher TJ, Li S, Van Slyke MA, Smith MB Peterfy CG, Majumdar S, Lang P, van Dijke CF, Sack K, Genant
(1997) Spatial variation of T2 in human cartilage. Radiol- HK (1994a) MR imaging of the arthritic knee: improved
ogy 205:546-550 discrimination of cartilage, synovium, and effusion with
Disler DG, McCauley TR, Kelman CG, Fuchs MD, Ratner pulsed saturation transfer and fat-suppressed Tl-weighted
LM, Wirth CR, Hospodar PP (1996) Fat-suppressed three- sequences. Radiology 191:413-419
dimensional spoiled gradient -echo MR imaging of hyaline Peterfy CG, van Dijke CF, Janzen DL, Gluer C, Namba R,
cartilage defects in the knee: comparison with standard MR Majumdar S, Lang P, Genant HK (1994b) Quantification of
imaging and arthroscopy.Am J RoentgenoI167:127-132 articular cartilage in the knee by pulsed saturation trans-
Eckstein F, Sittek H, Milz S, Putz R, Reiser M (1994) The mor- fer and fat-suppressed MRI: optimization and validation.
phology of articular cartilage assessed by magnetic reso- Radiology 192:485-491
nance. Surg Radiol Anat 16:429-438. Peterfy CG, White DL, Zhao J, van Dijke CF, Genant HK
Eckstein F, Gavazzeni A, Sittek H, Haubner M, Losch A, Milz (1998) Longitudinal measurement of knee articular carti-
S, Englmeier KH, Schulte E, Putz R, Reiser M (1996) Deter- lage volume in osteoarthritis. Arthritis Rheum 41: 1970
mination of knee joint cartilage thickness using three- Pilch L, Stewart C, Gordon D, Inman R, Parsons K, Pataki I, Ste-
dimensional magnetic resonance chondro-crassometry (3D vens J (1994) Assessment of cartilage volume in the femoro-
MR-CCM). Magn Reson Med 36:256-265 tibial joint with magnetic resonance imaging and computer
Eckstein F, Schnier M, Haubner M, Priebsch J, Glaser C, Engl- reconstruction. J RheumatoI21:2307-2319
meier KH, Reiser M (1998a) Accuracy of 3D knee joint car- Potter HG, Linklater JM,Allen AA, Hannafin JA, Haas SB (1998)
tilage volume and thickness measurements with MRI. Clin Magnetic resonance imaging of articular cartilage in the
Orthop Rel Res 52:137-148 knee. An evaluation with use of fast-spin-echo imaging. J
Eckstein F, Westhoff J, Sittek H, Maag KP, Haubner M, Faber Bone Joint Surg Am 80:1276-1284
S, Engelmeier KH, Reiser M (1998b) In vivo reproducibility Recht MP, Piraino DW, GA Paletta, JP Schils, Belhobek
of three-dimensional cartilage volume and thickness mea- GH (1996) Accuracy of fat-suppressed three-dimensional
surements with MRI. Am J RoentgenoI170:593-597 spoiled gradient-echo FLASH MR imaging in the detection
Eckstein F, Tieschky M, Faber SC, Haubner M, Kolem H, Engl- of patellofemoral articular cartilage abnormalities. Radiol-
meier KH (1998c) Effect of physical exercise on cartilage ogy 198:209-212
volume and thickness in vivo: an MRI study. Radiology Rubenstein JD, Kim JK, Morova-Protzner I, Stanchev PL, Hen-
207:243-248 kelman RM (1993) Effects of collagen orientation on MR
Eckstein F, Winzheimer M, Westhoff J, Schnier M, Haubner imaging characteristics of bovine articular cartilage. Radi-
M, Englmeier KH, Reiser M, Putz R (1998d) Quantitative ology 188:219-226
Assessment of Knee Cartilage 291
Rubin DA, Harner CD, Costello JM (2000) Treatable chondral suppressed MRI sequence with short acquisition time and
injuries in the knee: frequency of assodated focal subchon- 3D data processing. J Orthop Res 15:808-813
dral edema.Am J RoentgenoI174:1099-1106 Trattnig S, Mlynarik V, Breitenseher M, Huber M, Zembsch A,
Solloway S, Hutchison CE, Waterton JC, Taylor CJ (1997) The Rand T, Imhof H (1999) MRI visualization of proteoglycan
use of active shape models for making thickness measure- depletion in articular cartilage via intravenous administra-
ments of articular cartilage from MR images. Magn Reson tion of Gd-DTPA. Magn Reson Imaging 17:577-583
Med 36:943-952 Vahlensieck M, Peterfy CG, Wischer T, Sommer T, Lang P,
Stammberger T, Eckstein F, Michaelis M, Englmeier KH, Reiser Schlippert U, Genant HK, Schild HH (1996) Indirect MR
M (1999) Interobserver reprodudbility of quantitative car- arthrography: optimization and clinical applications. Radi-
tilage measurements: comparison of B-spline snakes and ology 200:249-254
manual segmentation. Magn Reson Med 17:1033-1042 Xia Y, Farquhar T, Burton-Wurster N, Lust G (1997) Origin of
Stammberger T, Hohe J, Engelmeier KH, Reiser M, Eckstein F cartilage laminae in MRI. J Magn Reson Imaging 7:887-894
(2000) Elastic registration of 3D cartilage surfaces from MR Yao L, Gentili A, Thomas A (1996) Inddental magnetization
image data for detecting local changes in cartilage thick- transfer contrast in fast spin -echo imaging of cartilage. J
ness. Magn Reson Med 44:592-601 Magn Reson Imaging 6:180-184
Tieschky M, Faber S, Haubner M, Kolem H, Schulte E, Englmei- Yelin E, Callahan LF (1995) The economic cost and sodal and
er KH, Reiser M, Eckstein F (1997) Repeatability of patel- psychological impact of musculoskeletal conditions. National
lar cartilage thickness patterns in the living, using a fat- Arthritis Data Work Groups. Arthritis Rheum 38: 1351-1362
17 Osteochondritis and Osteonecrosis
ARTHUR A. DE SMET
proposed etiologies for OCD inc1ude trauma, periph- mented, «It is difficult to conceive how avascular necro-
eral avascular necrosis, and congenital abnormalities sis ofbone could produce speeimens without attached
of ossification of the epiphysis (CLANTON and DELEE necrotic subchondral bone" (MILGRAM 1978).
1982; GARRETT 1991). In addition to trauma and avascular necrosis, some
authorities feel that OCD results from an abnor-
mality of endochondral ossification (SCHENCK and
17.2.1 GOODNIGHT 1996).A predisposition to OCD has been
Etiology described in several families but a familial assoeiation
is rare (PAES 1989; KOZLOWSKI and MIDDLETON 1985;
Although many etiologies have been proposed, most CLANTON and DELEE 1982; SCHENCK and GOOD-
investigators believe that direct and indirect trauma NIGHT 1996; MUBARAK and CARROLL 1979,1981).
plays an important role in the formation of osteo-
chondritic lesions (SCHENCK and GOODNIGHT 1996;
PAPPAS 1981;FAIRBANK 1933;MILGRAM 1978;CAHILL 17.2.2
et al. 1989). There is a history of prior knee trauma Clinical Presentation
in 40%-60% of patients who have OCD of the knee
(CLANTON and DELEE 1982). OCD is more common- The most common presenting symptom of patients
ly seen in athletes than in those who do not partiei- with OCD is vague joint pain. If the fragment is par-
pate in organized sports (SCHENCK and GOODNIGHT tially or completely detached, there may be symp-
1996). OCD also occurs two to three times more com- toms of catching, locking or swelling with use of the
monly in males than females (CLANTON and DELEE affected knee (CLANTON and DELEE 1982; SCHENCK
1982; SCHENCK and GOODNIGHT 1996; LINDEN 1976). and GOODNIGHT 1996). OCD of the knee presents
This male predominance has been used as evidence more commonly in childhood and early adolescence
of the traumatic etiology in view of the greater physi- than does OCD at other sites in the body. Femoral
cal activity levels of boys than girls, as was the pat- condylar osteochondritic lesions are two to three
tern until recent years. In addition, the location of times as likely to be unilateral as bilateral (CLANTON
these lesions on convex surfaces suggests a shearing and DELEE 1982; SCHENCK and GOODNIGHT 1996).
or impaction injury. The most common site for OCD Osteochondritis dissecans of the knee is divided
in the body is on the convex joint surfaces of the femo- into juvenile and adult forms based on patient age at
ral condyles (CLANTON and DELEE 1982). presentation (CLANTON and DELEE 1982; SCHENCK
Previous studies of the histological features of OCD and GOODNIGHT 1996). In juvenile OCD, the lesion
in the knee have not definitively identified an etiol- first becomes symptomatic prior to the c10sure of the
ogy. The fragments may consist of articular cartilage adjacent growth plate (Fig. 17.1). In the adult form,
alone or with variable amounts of attached viable or the adjacent growth plate is fused at the time of clini-
necrotic subchondral bone. Some biopsy studies of cal presentation (Fig.17.2). This distinction is impor-
the underlying bone found normal viable bone with tant for patient management. Juvenile lesions will
reparative granulation tissue, so at least in these cases, often heal with conservative treatment while patients
the osteochondritic lesions were not due to peripheral with adult femoral condylar OCD usually have per-
avascular necrosis (CHIROFF and COOKE 1975). How- sistent knee dis ability unless their lesions are treated
ever, others have noted ischemic changes in the base surgically.
of the lesions, causing them to believe that OCD rep- Although far less common than femoral condylar
resents «••• pathological fractures through subchondral lesions, patellar OCD is not a rare entity. A survey
necrotic bone.. :' (CAMPBELL and RANAWAT 1966). of the literature in 1977 found 40 recorded cases
Based on his extensive research, MILGRAM (1978) (EDWARDS and BENTLEY 1977). Since that time there
conc1uded that OCD is not a form of peripheral osteo- have been two series with 13 cases in one study and
necrosis. He analyzed 29 partially attached and 21 12 in the other (DESAI et al. 1987; RENU et al. 1994).
detached osteochondritic fragments, and found that Although speeific episodes of prior trauma have not
the foei of increased density seen on radiographs been assoeiated with patellar OCD, all 13 patients
within an OCD can be due to: (1) attached subchon- in one series had a history of repeated trauma in
dral bone, (2) caleification of degenerating cartilage, sports (DESAI et al. 1987). The most common pre-
(3) new bone formation, or (4) caleification of new senting symptom in these patients is the gradual
surface layers of cartilage and bone. Because of the onset of vague anterior knee pain. Patients with
absence of attached bone in 24 fragments, he com- patellar OCD are more commonly male than female
a
Fig. 17.2a-c. Unstable adult ocn of the medial femoral condyle in a 35-year-
old man. a Anteroposterior radiograph shows a crescentic lucent lesion only
partially filled with a central calcified density, which begins in the notch and
extends medially onto the weight-bearing surface. b Coronal spin echo T2-
weighted MR image shows high signal interface (black arrowheads) between
the ocn and the femur, indicating an unstable lesion. c Sagittal spin echo
T2-weighted MR image shows an anterior cyst (curved black arrow), high
signal intensity at the margin of the ocn, and a cartilage fracture posteriody
(straight white arrow), which are all signs of an unstable lesion
296 A.A.De Smet
and are usually in the second and third decades of osteochondritie lesions at sites other than the femo-
life (DESAI et al. 1987). ral condyles usually require debridement rather
than fixation. More recently, osteochondral allografts
have been used for the treatment of femoral OCD
17.2.3 when there is a large defect in the articular surface
Natural History and Treatment (SCHENCK and GOODNIGHT 1996; GARRETT 1991).
Good patient outcomes are also common in patients
In general, the younger the patient is at presentation with patellar OCD who are initially treated conserva-
of the OCD, the better the long-term prognosis for tively and then with debridement and drilling if the
good joint function. Juvenile OCD of the femoral lesion fails to heal under conservative management
condyles has a better prognosis than the adult form (DESAI et al. 1987; RENU et al. 1994; EDWARDS and
(LINDEN and TELHAG 1977; TwYMAN et al. 1991). BENTLEY 1977).
Most cases of juvenile OCD heal without sequelae if
the fragment does not displace (PAPPAS 1981; GAR-
RETT 1991). Limitation of activity is the recommend-
ed treatment when the fragment is not displaced 17.3
(CLANTON and DELEE 1982; CAHILL 1985; SCHENCK Imaging of Osteochondritis Dissecans
and GOODNIGHT 1996). Because of the muscle atro-
phy that occurs with long-term casting, protected 17.3.1
weight bearing with a brace or crutches is now Radiography of OCD of the Femoral Condyles
suggested as treatment (SCHENCK and GOODNIGHT
1996). In contrast to the good results for conservative The recommended radiographic filming for assess-
treatment in adolescent OCD, it is rare for a symp- ment of a patient who may have an OCD is a four-view
tomatic adult with OCD to become asymptomatie series consisting of anteroposterior, lateral, tunnel, and
without surgieal intervention (DE SMET et al.1997). sunrise views (MCGILL et al. 1995). The prototypieal
The size of the lesion is of considerable impor- OCD of the knee is alesion on the medial femoral con-
tance in predieting stability and the ability to subse- dyle, which begins near or in the intercondylar notch
quently heal. Multiple authors have shown that the and extends medially onto the weight-bearing surface
larger the lesion, the less likely that the lesion will of the condyle (Figs. 17.1,17.2). Lesions of the medial
heal under conservative management (MESGARZA- femoral condyle are three to four times more common
DEH et al. 1987; DE SMET et al.1997; HUGHSTON et al. than the lateral lesions (SCHENCK and GOODNIGHT
1984; CAHILL and BERG 1983). The area of an OCD is 1996; LINDEN 1976). Only 12% of medial condylar
determined by multiplying the length of the lesion lesions are on the central surface of the condyle
by its width. Stable lesions were found to have an (AICHROTH 1971). In contrast,laterallesions are almost
average area of 194-424 mm 2 while unstable lesions always centered on the central weight-bearing surface
had an average of 436-815 mm2 (MESGARZADEH (Fig.17.3) (AICHROTH 1971; GARRETT et al.1992).
et al. 1987; DE SMET et al. 1997; HUGHSTON et al. Medialiesions also tend to be more anterior than
1984; CAHILL and BERG 1983). Consistent with these lateral lesions. If a line is drawn on a lateral radio-
results, medial osteochondritie lesions are generally graph down the posterior femoral cortex of the
sm aller and thus more likely to heal than the lateral femur, it will usually intersect a medial femoral con-
lesions (GARRETT 1991). dyle lesion (Fig. 17.4a). The laterallesions are usually
Numerous surgical options are available for the posterior in a location, where there is weight bear-
treatment of OCD. When possible, preservation of ing with the knee in flexion (GARRETT et al. 1992).
the articular surface is the preferred management In these cases, the osteochondritic lesion is projected
(SCHENCK and GOODNIGHT 1996). Surgieal treatment posterior to aHne drawn down the posterior femoral
usually includes drilling through the osteochondral cortex on a lateral radiograph (Fig.17.4b). Because of
fragment to induce healing at the base. The frag- this more posterior location, lateralIesions are often
ment may be internally fixed with bone pegs, Kirsch- best seen on a tunnel view of the knee (Fig. 17.3c).
ner wires or screws recessed beneath the articular The typical radiographie appearance of a femoral
cartilage (SCHENCK and GOODNIGHT 1996; GARRETT condylar OCD is an elliptical radiolucentlesion extend-
1991). Occasionally, when the fragment is multipar- ing from the subchondral bone plate into the medul-
tite or too small to be stabilized, debridement of lary space with a proximal convex border with a well-
the lesion is required. Because of their smaller size, defined rim of sclerosis (CLANTON and DELEE 1982;
Osteochondritis and Osteonecrosis 297
a b
Fig. 17.3a-c. OCD of the lateral femoral condyle. a The lesion is hard to
appreciate on the anteroposterior radiograph but a subtle area of lucency and
sclerosis is seen laterally. b Lateral radiograph shows the posterior location
of the lesion (arrowheads). c Tunnel view radiograph best profiles the lesion,
c which is entirely radiolucent without central calcifications (black arrow)
a b
Fig. 17.4a, b. Line drawings illustrating typicallocations of OCD. a Medial femoral condyle lesion beginning in the notch and
extending onto the weight-bearing surface. On the lateral projection, it lies centrally on the weight-bearing surface, so that a
line drawn down the posterior femoral cortex intersects the lesion. b Lateral femoral condyle lesion that lies centrally in the
coronal plane but posteriorly in the sagittal plane, so that a line drawn down the posterior femoral shaft cortex passes anterior
to the lesion
298 A.A.De Smet
SCHENCK and GOODNIGHT 1996). The sclerosis can present on the inferior half of the patella, especially
vary in thickness from 1 to as much as 10 mm. The on the median ridge and medial facet, with a minor-
lesion itself may be entirely radiolucent (Fig. 17.3), ity on the lateral facet (DESAI et al. 1987; RENU et al.
may have one or more small fragments within the 1994; EDWARDS and BENTLEY 1977). No lesions have
lesion (Fig. 17.2), or may be almost completely filled by been noted on the odd facet or on the margins of the
a solitary bony fragment (Fig. 17.1). There is usually patella.
disruption of the subchondral bone plate at the mar- The anterior femoral lesions are almost always
gins of the lesion unless the lesion has healed. radiolucent with variable amounts of marginal sclero-
In children around 10 years of age, the major dif- sis ranging from minimal to extensive (Fig.17.6). The
ferential diagnosis is normal variation in ossifica- lesions may occur anywhere in the femoral sulcus,
tion of the epiphysis. In some children, the epiphysis with sites varying from the medial edge through the
will not ossify from a single enlarging ossification central groove onto the lateral edge of the intercondy-
center but will have islands of ossification. As the lar sulcus (CAYEA et al. 1981; KURZEWEIL et al.1988).
child develops, one of the small islands of ossification The tibial plateau and anterior tibial spine are
may persist at the articular surface and may mimic rare sites of OCD in the knee (BUI-MANSFIELD et al.
an OCD. The normal variation differs from an OCD 2000; SAGASTIBELZA et al. 1993; TOWBIN et al. 1982).
in that there is no marginal sclerosis at the base of Although not many lesions have been reported at
an ossification center, while an osteochondritic frag- these two sites, the radiographie appearance is usu-
ment usually has a rim of reactive sclerosis. ally a well-defined radiolucent area surrounded by
variable amounts of marginal sclerosis and often
containing a central ossified fragment.
17.3.2
Radiography of Less Common Sites of OCD
17.3.3
Patellofemoral osteochondritic lesions are consider- Radionuclide Bone Scintigraphy
ably less common than those in the femoral condyles.
However, they have distinctive findings, which may Radionuclide bone scintigraphy is highly sensitive
be easily overlooked without proper filming. Because in the detection of femoral osteochondritic lesions
they occur on the sloping surfaces of the femoral (LOOMER et al. 1993; Ly and FALLAT 1993; CAHILL
sulcus or the patellar facets, they are best visualized and BERG 1983). The typical appearance is an intense
on either the lateral view of the knee or tangential focus of radionuclide uptake in a femoral condyle
views of the patellofemoral joint, such as the Laurin that extends down to the subchondral bone plate.
view. MESGARZADEH et al. (l987) found that the presence
The patellar lesions appear as radiolucent defects of hyperemia on the blood pool phase was 100% sen-
adjacent to the articular surface with thin well- sitive and 83% specific in identifying unstable OCD
defined borders, which may or may not have central lesions and that intense uptake on the delayed images
bony fragments (Fig.17.5). They are most commonly was 90% sensitive and 100% specific. CAHILL and
a b
Fig. 17.5a, b. Unstable lateral patellar OCD. a Axial radiograph of the patellofemoral joint shows a subtle radiolucency (black
arrowhead) lateral to the median ridge of the patella. b Axial STIR MR image shows high signal intensity fluid filling the defect
on the lateral patellar facet, indicating end-stage OCD in which the fragment has become a loose body
Osteochondritis and Osteonecrosis 299
,
proton density-weighted MR image shows a long lesion with
disruption of the low signal intensity subchondral bone plate
anteriorly. c Fat-saturated sagittal fast spin echo T2-weighted
MR image shows high signal intensity within the lesion but
none of the signs of instability
c
b
BERG (1983) found that radionuclide bone scintigra- tation (HOWIE 1985; HOLLAND and DAVIES 1994;
phy performed every 6 weeks was useful in predicting LOOMER et al. 1993; Ly and FALLAT 1993). In my own
the potential for healing of OCD. If serial bone scin- practice, CT is often used to assess the size of osse-
tigrams show continued or increasing high levels of ous fragments within the lesion (Fig. 17.7). Determi-
uptake, then the site is undergoing constant remodel- nation of osseous fragment size allows surgical plan-
ing and the prognosis for healing with conservative ning for the method of stabilization of the fragment.
management is good (CAHILL and BERG 1983). Serial Larger fragments are amenable to screw fixation while
radionuclide scanning has not received widespread smaller fragments cannot be internally fixed.
application because of the radiation exposure to the
patient. Most physicians prescribe reduced weight
bearing in patients with juvenile OCD of the knee 17.3.5
and use serial radiographs and change in symptoms Arthrography
to assess healing.
Early reports suggested that arthrography would be
a useful technique to assess the stability of the lesion
17.3.4 and the integrity of the overlying articular cartilage
Computed Tomography (ALMGARD and WIKSTAD 1964). In theory, contrast
material injected into the joint should pass beneath
Computed tomography (CT) will occasionally detect the fragment. Visualization of radio-opaque contrast
alesion which is occult on radiographs and can be beneath the lesion is proof of an articular cartilage
used for preoperative staging by determining the defect and confirms an unstable lesion. However,
exact size of the lesion and the degree of fragmen- because these lesions commonly have granulation
300 A. A. De Smet
17.3.6
MRlmaging
a
Magnetic resonance imaging is the best non-invasive
way to evaluate the stability and extent of an OCD
lesion. Multiple reports have confirmed MR imaging
is an accurate method for determining whether a
lesion is stable or unstable (PFEIFFER et al. 1991). An
OCD lesion is considered unstable if the fragment is
partially or completed detached or if the lesion is bal-
lottable with a probe at arthroscopy.
A variety of MR imaging techniques have been
used for assessment of OCD. These have included T 1-
weighted, proton density-weighted, and T2-weighted
spin echo imaging, and gradient recalled echo imag-
ing (DIPAOLA et al. 1991; ADAM et al. 1991; ENGEL
et al. 1990). Because of the magnetic susceptibility
artifact that occurs within cancellous bone, non-con-
trast-enhanced gradient recalled echo imaging may
Fig. 17.7a, b. OCD of the medial femoral condyle. Coronal (a) make identification of instability difficult. T2-weight-
and sagittal (b) reformatted cr images show an entirely radio- ed imaging has been the most useful in assessment
lucent lesion without central calcified fragments that might be of instability of an osteochondritic lesion as detec-
amenable to internal fixation tion of granulation tissue or fluid beneath the lesion
is a reliable indicator of instability. In addition, the
arthrographic effect produced by the high signal
tissue beneath them, the amount of contrast that intensity joint fluid allows assessment of defects in
passes beneath the lesion is often small and hard to the articular cartilage.
detect by radiographs. For this reason, conventional In a study performed using an animal model for
arthrography is not very useful for assessing lesion OCD, Tl-weighted imaging after intravenous injection
stability. of gadolinium was as helpful as T2-weighted imaging
Computed arthrotomography is a useful modifi- in identifying whether a fragment was stable or unsta-
cation of conventional arthrography. With its high ble (ADAM et al. 1991). This same study provided his-
resolution, cross-sectional imaging and sensitivity tological confirrnation that the interface between an
to tissue contrast differences, CT after intra-articu- unstable lesion and the underlying femur consisted of
lar contrast injection allows excellent definition of vascularized granulation tissue (ADAM et al. 1991).
the articular cartilage (HOLLAND and DAVIES 1994; My current protocol for assessment of femoral
PAILLE et al. 1988). However, although CT arthrog- OCD is fast spin echo T2-weighted imaging with fat
raphy should be an accurate way to determine the saturation in the sagittal and coronal planes using a
integrity of the articular cartilage over an OCD, few field ofview of 14 cm, matrix of 256+ 192, two excita-
studies have compared its accuracy with operative tions, an echo train length of 8, a slice thickness of
findings of stability or instability. 3 mm, a TR of 3,000 and a TE of 80eff. This is com-
A more recent modification of knee arthrography bined with routine coronal Tl-weighted and sagittal
has been the use of magnetic resonance (MR) imag- proton density-weighted images to assess the menis-
Osteochondritis and Osteonecrosis 301
ci as patients with OCD are usually athletically active ed conservatively. Arecent study reported the long-
and their knee pain can be due to a meniscal tear term follow-up of 14 patients treated conservatively
rather than an unstable OCD lesion. for 1.5-8.5 years. Both patients with stable lesions as
The criteria for diagnosing instability of an OCD diagnosed by MR imaging had a good outcome while
lesion have been well described (DE SMET et al. 1990, 10 of the 12 patients whose lesions were unstable by
1996; MESGARZADEH et al. 1987). Alesion is unstable MR imaging had poor outcomes with continued knee
by MR criteria if: (l) there is a high signal intensity disability or required surgery (DE SMET et al. 1997).
line representing granulation tissue or fluid beneath Magnetic resonance imaging also appears to be
the deep portion of the lesion, (2) if there is a cyst of effective in confirming the healing of surgically treat-
more than 5 mm in width beneath the lesion, (3) if ed OCD. Fifty-six patients with knee OCD treated
there is a focal articular defect, or (4) if fluid is seen surgically by retrograde drilling were followed post-
passing through an articular cartilage fracture (DE operatively by physical examination, radiographs,
SMET et al.1996). If one or more of these is present, the and MR imaging (SCHNEIDER et al. 1998). MR imag-
lesion can be diagnosed as being unstable (Figs. 17.1, ing was superior to radiographs for predicting heal-
17.2,17.5). Using these criteria,MR imagingwas found ing of the lesions (SCHNEIDER et al. 1998). One case
to be 97% sensitive and 100% specific for diagnosing report noted that MR imaging could be used to assess
the instability of OCD lesions in 31 patients with femo- healing of OCD even when absorbable pins had been
ral condylar lesions and nine patients with talar dome used for fixation of the osteochondritic fragment
lesions (DE SMET et al. 1996). In this same study, the (SMITH et al. 1994).
most common sign was a high signal intensity line
beneath the lesion, which was seen in 72% of the unsta-
ble lesions. Fifty-six percent of the 40 cases had only
one of the aforementioned four signs on MR imaging 17.4
(DE SMET et al.1996). Osteonecrosis of the Knee
If the lesion itself has signal within it, but none of
the four signs are present, then the lesion should be Osteonecrosis of the knee is similar to OCD in that its
diagnosed by MR imaging as being stable (Fig. 17.6) etiology is unknown when it is a primary condition.
There have been only two reported cases in which a However, unlike OCD, where there is variable histol-
high signal intensity line was present beneath lesions ogy, osteonecrosis by definition always shows necrosis
that were surgically proven to be stable (MESGAR- on histological examination. In its earliest stages, the
ZADEH et al. 1987; DE SMET et al. 1990). Fast spin necrosis involves primarily bone marrow, as this is the
echo T2-weighted imaging with fat saturation should tissue most sensitive to ischemia (MUNK and VELLET
be an even more sensitive method than convention- 1993). With continued ischemia, there is death ofbone
al spin echo T2-weighted imaging in determining with necrosis of osteocytes. Being relatively resistant
lesion instability, but there have been no reported to ischemia, the fat within the medullary cavity is the
studies to confirm its accuracy. last tissue to undergo necrosis. If a predisposing cause
Although MR imaging is accurate in the assess- is known, the treating physician will often suspect
ment of whether or not alesion is unstable, it is not osteonecrosis when the patient presents with knee
definitive for assessment of the integrity of the artic- pain even before any imaging is performed.
ular cartilage. Alesion may be unstable with motion
detected by ballottement of its surface at surgery but
still have an intact overlying cartilage (DE SMET et al. 17.4.1
1990). Although instability may in itself be an indi- Etiology
cation for surgery in a child whose lesion has not
responded to conservative treatment, some surgeons Osteonecrosis may be considered to be of two types,
wish to prove that the cartilage is fractured before idiopathic and secondary. Although vascular injury
proceeding to surgery. In this scenario, MR arthrog- and trauma have both been proposed as causes for
raphy provides the best assessment of articular car- idiopathic osteonecrosis of the knee, the exact etiol-
tilage. One study reported an accuracy of 100% in ogy is unknown (LOTKE and ECKER 1988). Those who
assessing the status of the articular cartilage over feel that it is caused by vascular injury believe that
OCD in 25 knees (KRAMER et al. 1992). there is occlusion of the microcirculation, while those
Magnetic resonance imaging can also be useful who favor a traumatic etiology speculate that sub-
in predicting patient outcome when an OCD is treat- chondral microfractures in osteopenic bone cause
302 A.A.De Smet
bleeding with increased marrow pressure and then ical activity such as walking or climbing stairs. The
ischemia (LOTKE and ECKER 1988). pain can be quite disabling and collapse of the articu-
Secondary osteonecrosis has many possible causes lar surface is an unfortunate common sequela. Spon-
and is the most common type of osteonecrosis of taneous osteonecrosis is three times more common in
the knee. The list of secondary causes of osteone- women than in men (LOTKE and ECKER 1988).
crosis is extensive and includes corticosteroid usage, An association has been noted between sponta-
marrow packing dis orders such as sickle ceH anemia neous osteonecrosis of the knee and meniscal tears
and Gaucher's disease, alcoholism, gout, systemic (NORMAN and BAKER 1978).Since manypatients with
lupus erythematosus, dysbaric osteonecrosis due to spontaneous osteonecrosis have collapse of the artic-
undersea diving and working in caissons, femoral ular surface, it is not certain whether the meniscal
neck fracture and femoral head dislocation (CRUESS tear has preceded the osteonecrosis or has occurred
1986). Corticosteroids are used to treat many condi- secondary to the articular incongruity.
tions such as autoimmune disease and marrow-based The most common presentation for both idiopath-
malignancies like lymphoma. Unfortunately, osteo- ic and secondary osteonecrosis is pain. The pain is
necrosis is not infrequently a complication of either likely due to the increased pressure from a compart-
acute or chronic moderate to high dosage corti- ment syndrome, which occurs when edema and hem-
costeroid treatment. In my own practice, the most orrhage develop within the closed space of the medul-
common cause of osteonecrosis is the use of cortico- lary canal (CRUESS 1986). During the early stages, the
steroids to suppress transplant rejection in patients radiographs are normal and diagnosis of the cause
who have had organ transplants (Fig. 17.8). Recently of the patient's pain requires either radionuclide scin-
osteonecrosis of the femoral condyles and tibial pla- tigraphy or MR imaging (CRUESS 1986; LOTKE and
teau has been identified after the use of a laser for ECKER 1988). Although the pain can be severe, there is
meniscectomy (JANZEN et al. 1997). no mechanical abnormality of the knee. If the balance
between resorption of the necrotic bone and bone
repair shifts in favor of bone resorption, a subchon-
dral fracture and coHapse of the weight-bearing por-
tion of the bone may occur (Fig. 17.9). At this point,
there is incongruity of the articular surface and lim-
ited range of motion of the knee.
Fortunately, there is often sufficient repair of the
osteonecrosis that the bone may repair completely
without functional deficit for the patient (CRUESS
1986). Complete clearing of osteonecrosis has not
been documented for osteonecrosis of the knee but
has been confirmed in the femoral heads of patients
who have received corticosteroids after renal trans-
plantation. Using routine screening MR imaging of
the hips in 104 patients after renal transplantation,
Fig. 17.8. A 22-year-old woman on corticosteroids for a liver KOPECKY et al. (1991) found osteonecrosis in 25 hips
transplant. Anteroposterior radiograph shows collapse of the yet only seven of these hips became symptomatic.
articular surface of the lateral femoral eondyle from osteone- The MR examinations returned to normal in 6 of the
erosis
18 asymptomatic hips.
17.4.2
Clinical Presentation and Natural History
17.5
The most common form of idiopathic osteonecrosis Imaging of Osteonecrosis
of the knee is spontaneous osteonecrosis (LOTKE and
ECKER 1988). This is seen in the elderly and is much 17.5.1
more common in the medial femoral condyle than Radiography
in the lateral femoral condyle or the tibial condyles
(LOTKE and ECKER 1988). Patients usually present with In contrast to OCD of the knee, in which symptom-
sudden onset of severe pain while doing a minor phys- atic lesions are always seen on radiographs, the radio-
Osteochondritis and Osteonecrosis 303
a b
Fig.17.9a, b. A 55-year-old man with sudden onset ofknee pain due to spontaneous osteonecrosis of the medial femoral condyle.
a Coronal fat-saturated fast spin echo proton density-weighted image shows a crescentic low signal intensity area beneath the
subchondral bone plate of the medial femoral condyle and diffuse edema in the condyle. Radiographs at this time were normal.
b Anteroposterior radiograph obtained 2 months later shows a subchondral fracture with surrounding radiolucency
If a three-phase bone scintigram is performed, weighted images (Fig. 17.9). This is not a specific
there is increased blood flow at the site of acute appearance and can also be seen in osteomyelitis and
osteonecrosis with intense uptake within the lesion bone bruises from trauma. The clinical presentation
on the blood pool phase (GREYSON et al. 1982).As the will usually allow easily discrimination with fever
lesion matures, the hyperemia and increased radio- and systemic symptoms in osteomyelitis and a his-
nuclide activity on the blood pool phase are no longer tory of re cent trauma for bone bruises. In these cases,
seen (GREYSON et al. 1982). it is also important to compare the radiographs with
When the radionuclide uptake occurs in the tibial the MR images. Elderly patients with tibial plateau
condyles, an additional differential diagnosis in the insufficiency fractures can present clinically with
elderly would be a tibial insufficiency fracture. Insuf- spontaneous medial pain that may resemble the pain
ficiency fractures usually have a linear pattern of in spontaneous osteonecrosis. However, MR imaging
radionuclide uptake, which allows differentiation allows definite diagnosis of an insufficiency fracture,
from the globular uptake pattern of osteonecrosis. which shows a linear pattern of abnormal signal
Although the initial radiographs may be negative in intensity that is different from the globular pattern of
the early stages of a tibial insufficiency fracture, serial osteonecrosis (LE GARS et al. 1996).
films will show a characteristic linear pattern of scle- However, most cases of avascular necrosis have a
rosis in the tibial condyle paralleling the adjacent characteristic appearance once they reach the chronic
tibial plateau. phase. The resistance of the fatty marrow to ischemia
Computed tomography has litde value in the initial results in lesions that on Tl-weighted MR images
diagnosis of osteonecrosis of the knee. If the lesion is have the central high signal intensity of fat and a
not visible by radiographs, it is only occasionally vis- thin well-defined low signal intensity border. This
ible by CT. CT may be useful in determining the exact low signal intensity border may remain of low signal
amount of the bone loss for surgical planning prior to intensity on T2-weighted images, reflecting sclerosis
total knee arthroplasty, as has been shown in assess- or fibrosis, or may have a high signal intensity due
ment of osteonecrosis of the femoral heads. to granulation tissue. An active lesion mayaIso have
a large surrounding zone of edema, which is of low
signal intensity on Tl-weighted images and high
17.5.3 signal intensity on T2-weighted images.
MRlmaging
Aiehroth P (1971) Osteochondritis dissecans of the knee: a Garrett JC, Kress KJ, Mudano M (1992) Osteochondritis disse-
clinieal survey. J Bone Joint Surg 53B(3):440-447 cans of the lateral femoral condyle in the adult. Arthroscopy
Almgard LE, Wikstad I (1964) Late results of surgery for osteo- 8(4):474-481
chondritis dissecans of the knee joint. Acta Chir Scand Greyson ND, Lotem MM, Gross AE et al (1982) Radionuclide
127:588 evaluation of spontaneous femoral osteonecrosis. Radiol-
Bui-Mansfie!d LT, Kline M, Chew FS et al (2000) Osteochondri- ogy 142:729-735
tis dissecans of the tibial plafond: imaging characteristies Holland P, Davies AM (1994) Computed tomographie arthrog-
and a review of the literature. AJR 175:1305-1308 raphy in the assessment of osteochondritis dissecans of the
Cahill BR (1985) Treatment of juvenile osteochondritis disse- elbow. Clin RadioI49:231-235
cans and osteochondritis dissecans of the knee. Clin Sports Howie JL (1985) Computed tomography in osteochondritis
Med4:67 dissecans of the patella. J Can Assoc Radiol 36: 197-199
Cahill BR, Berg BC (1983) 99 m -Technetium phosphate com- Hughston JC, Hergenroeder PT, Courtenay BG et al (1984)
pound joint scintigraphy in the management of juvenile Osteochondritis dissecans of the femoral condyles. J Bone
osteochondritis dissecans of the femoral condyles. Am J Joint Surg 66A(9):1340-1348
Sports Med 11(5):329-335 Janzen DL, Kosarek FJ,Helms CA et al (1997) Osteonecrosis after
Cahill BR, Phillips MR, Navarro R et al (1989) The results of contact neodymium: yttrium aluminum gamet arthroscop-
conservative management of juvenile osteochondritis dis- ie laser meniscectomy. AJR 169:855-858
secans using joint scintigraphy: a prospective study. Am J Kopecky KK, Braunstein EM, Brandt KD et al (1991) Apparent
Sports Med 17:601 avascular necrosis of the hip: appearance and spontaneous
Campbell CJ, Ranawat CS (1966) Osteochondritis dissecans: resolution of MR findings in ranal allograft recipients.
the question of etiology. J Trauma 6(2):201-221 Radiology 179:523-527
Cayea PD, Pavlov H, Sherman MF et al (1981) Lucent artieular Kozlowski K,Middleton R (1985) Familial osteochondritis dis-
lesion in the lateral femoral condyle: source of patellar fem- secans: a dysplasia of artieular cartilage? Skeletal Radiol
oral pain in the athletie adolescent.AJR 137:1145-1149 13:207-210
Chiroff RT, Cooke CP III (1975) Osteochondritis dissecans: Kramer J,Stiglbauer R,Enge!A et al (1992) MR contrast arthrog-
a histologie and mieroradiographie analysis of surgieally raphy (MRA) in osteochondrosis dissecans. J Comput Assist
excised lesions. J Trauma 15(8):689-696 Tomogr 16(2):254-260
Clanton TO, DeLee JC (1982) Osteochondritis dissecans: his- Kurzeweil PR, Zambetti GJ, Hamilton WG et al (1988) Osteo-
tory, pathophysiology and current treatment concepts. Clin chondritis dissecans in the lateral patellofemoral groove.
Orthop 167:50-64 Am J Sports Med 16(3):308-310
Conklin JJ, Alderson PO, Zizie TM et al (1983) Comparison Le Gars LL, Orce! P, Liote F et al (1996) Osteonecrosis-like syn-
of bone scan and radiograph sensitivity in the detection drome of the medial tibial plateau can be due to a stress
of steroid-induced ischemie necrosis of bone. Radiology fracture. Rev Rheum 66( 6):323-330
147:221-226 Linden B (1976) The incidence of osteochondritis dissecans in
Cruess RL (1986) Osteonecrosis of bone: current concepts as the condyles of the femur. Acta Orthop Scand 47:664-667
to etiology and pathogenesis. Clin Orthop 208:30-39 Linden B, Telhag H (1977) Osteochondritis dissecans: a his-
D'Ambrosia RD, Shoji H, Riggins RS et al (1978) Scintigraphy tologie and autoradiographic study in man. Acta Orthop
in the diagnosis of osteonecrosis. Clin Orthop 130:139-143 Scand 48:682-686
Desai SS, Pate! MR, Miehelli LJ et al (1987) Osteochondritis Loomer R, Fisher C, Lloyd-Smith R et al (1993) Osteochondral
dissecans of the patella. J Bone Joint Surg [Br] lesions of the talus. Am J Sport Med 21: 13
69B(2):320-325 Lotke PA, Ecker ML (1988) Current concepts review osteone-
De Smet AA, Fisher DR, Graf BK et al (1990) Osteochondritis crosis of the knee. J Bone Joint Surg 70A(3):470-474
dissecans of the knee: value of MR imaging in determin- Ly PN and Fallat LM (1993) Transchondral fractures of the
ing lesion stability and the presence of artieular cartilage talus: a review of 64 surgieal cases. J Foot Ankle Surg
defects. AJR 155:549-553 32:352
De Smet AA, Ilahi OA, Graf BK et al (1996) Reassessment of Martel W, Sitterley BH (1969) Roentgenologic manifestations
the MR criteria for stability of osteochondritis dissecans in of osteonecrosis. AJR 106(3):509-522
the knee and ankle. Skeletal RadioI25:159-163 McGill JJ, Demos TC, Lomasney LM et al (1995) Osteo-
De Smet AA, Ilahi OA, Graf BK et al (1997) Untreated osteo- chondritis dissecans: imaging modalities. Orthopedics
chondritis dissecans of the femoral condyles: predietion of 18(12):1180-1185
patient outcome using radiographie and MR findings. Skel- Mesgarzadeh M, Sapega AA, Bonakdarpour A et al (1987)
etal Radiol 26:463-467 Osteochondritis dissecans: analysis of mechanical stability
Dipaola JD, Nelson DW, Colville MR et al (1991) MR: character- with radiography, scintigraphy, and MR imaging. Radiol-
izing osteochondrallesions by magnetie resonance imag- ogy 165:775-780
ing. Arthroscopy 7:101 Milgram JW (1978) Radiologieal and pathologieal manifesta-
Edwards DH, and Bentley G (1977) Osteochondritis dissecans tions of osteochondritis dissecans of the distal femur. Radi-
patellae. J Bone Joint Surg [Br] 59B(1):58-63 ology 126:305-311
Engel A, Hajek P, Kramer J et al (1990) Magnetie resonance knee Motohashi M, Morii, T, Koshino T et al (1991) Clinical course
arthrography. Acta Orthop Scand 61 [SuppI240]:33-43 and roentgenographic changes of osteonecrosis in the
Fairbank HAT (1933) Osteochondritis dissecans. Br J Surg femoral condyle under conservative treatment. Pro gnosis
21:67-82 Medial Gonarthrosis 266:156-161
Garrett JC (1991) Osteochondritis dissecans. Clin Sports Med Mubarak SJ, Carroll NC (1979) Familial osteochondritis dis-
10(3):569-593 secans of the knee. Clin Orthop 140: 131-136
306 A.A.De Smet
Mubarak sJ, Carroll NC (1981) Juvenile osteochondritis dis- roid- or alcohol-related osteonecrosis of the femoral con-
secans of the knee: etiology. Clin Orthop 157:200-211 dyle. Acta Orthop Scand 69( 6):598-602
Mulliken BD, Renfrew DL, Brand RA et al (1994) Prevalence of Schenck RC, Goodnight JM (1996) Current concepts review
previously undetected osteonecrosis of the femoral head in osteochondritis dissecans. J Bone Joint Surg
renal transplant recipients. Radiology 192:831-834 78A(3):439-456
Munk PL, Vellet AD (1993) Lesions of cartilage and bone Schneider T, Fink B, Jerosch J et al (1998) The value of mag-
around the knee. Top Magn Reson Imaging 5(4):249-262 netic resonance imaging as postoperative control after
Norman A, Baker ND (1978) Spontaneous osteonecrosis of the arthroscopic treatment of osteochondritis dissecans. Arch
knee and medial meniscal tears. Radiology 129:653-656 Orthop Trauma Surg 117:235-239
Paes RA (1989) Familial osteochondritis dissecans. Clin Radiol Siddiqui AR, Kopecky KK, Wellman HN et al (1993) Pro-
40:501-504 spective study of magnetic resonance imaging and SPECT
Paille P, Quesnel C, Baunin C et al (1988) Computed arthrog- bone scans in renal allograft recipients: evidence for a
raphy: its role in the screening of joint diseases in pediatric self-limited subclinical abnormality of the hip. J Nucl Med
radiology. Pediatr RadioI18:386-390 34:381-386
Pappas AM (1981) Osteochondrosis dissecans. Clin Orthop Smith DS, Sharp DC, Resendes M et al (1994) MRI of healing
158:59-69 osteochondritis dissecans fragment with absorbable pins. J
Pfeiffer WH, Gross ML, Seeger LL et al (1991) Osteochondritis Comput Assist Tomogr 18(5):832-833
dissecans of the patella: MRI evaluation and a case report. Tervonen 0, Mueller DM, Matteson EL et al (1992) Clinically
Clin Orthop 271:207-211 occult avascular necrosis of the hip: prevalence in an
Renu JM, Bou CV, Portet RV et al (1994) Osteochondritis dis- asymptomatic population at risk. Radiology 182:845-847
secans of the patella: 12 cases followed for 4 years. Acta Towbin J, Towbin R, Crawford A et al (1982) Osteochondritis
Orthop Scand 65(1):77-79 dissecans of the tibial plateau. J Bone Joint Surg [Am]
Sagastibelza J, Zuniga JJ, Blasco JJ et al (1993) Osteochondritis 64A(5):783-784
dissecans of the anterior tibial spine. Arthroscopy Twyman RS,Desai K,Aichroth PM et al (1991) Osteochondritis
9(6):695-696 dissecans of the knee: a long-term study. J Bone Joint Surg
Sakai T, Sugano N, Ohzono K et al (1998) MRI evaluation of ste- [Br] 73(B):461
18 Tumours and Tumour-like Lesions
A. MARK DAVIES and DANIEL VANEL
CONTENTS the distal femur and proximal tibia are the sites of
predilection for many benign and malignant bone
18.1 Introduction 307
18.2 Detection 307
tumours (Table 18.1). The purpose of this chapter is
18.3 Diagnosis 308 to detail the role of imaging in the detection and diag-
18.3.1 Diagnosis of Bone Tumours 308 nosis of bone and soft tissue tumours in and around
18.3.2 Diagnosis of Soft Tissue Tumours 314 the knee joint, as well as in the surgieal staging
18.3.3 CT and MR Imaging in Diagnosis 314
and follow-up of patients after initial treatment. We
18.4 Surgical Staging 315
18.5 Imaging Follow-up 318 shall also highlight the numerous tumour-like lesions
18.6 Bone Tumours 320 whieh may confuse the unwary observer. Unless oth-
18.6.1 Benign Bone Tumours 320 erwise stated, incidence data quoted have been calcu-
18.6.2 Malignant Bone Tumours 323 lated by combining results from several authoritative
18.6.3 Patellar Tumours 325
texts on the subject (MuLDER et al. 1993; UNNI 1996;
18.7 Soft Tissue Tumours 326
18.7.1 Benign Soft Tissue Tumours 326
CAMPANACCI 1999).
18.7.2 Malignant Soft Tissue Tumours 326
Tah1e 18.1. Relative overall incidence of the four commonest
18.8 Joint Tumours 327
sarcomas of hone (column N) and incidence of each type
18.8.1 Benign Joint Tumours 327
around the knee (column B)
18.8.2 Malignant Joint Tumours 331
18.9 Tumour-like Lesions of Bone 332 A B
18.9.1 Avulsive Cortical Irregularity 332 Osteosarcoma 35% 56%
18.9.2 Stress Fractures 332 Chondrosarcoma 26% 17%
18.9.3 Inflammatory Conditions 333 Ewing's sarcoma 16% 14%
18.9.4 Brown Tumours 333 Fihrosarcoma/MFH 6% 38%
References 333
MFH, Malignant fihrous histiocytoma
18.1 18.2
Introduction Detection
Bone sarcomas are uncommon when compared with The majority of patients with a bone or soft tissue
other malignancies, accounting for only 0.2% of all tumour will present with pain and/or swelling. Alter-
tumours (DORFMAN and CZERNIAK 1995). Their natively, a pathologieal fracture may be the precipi-
incidence is approximately one-tenth that of soft tous presenting feature. Occasionally, a bone tumour,
tissue sarcomas (MAcK 1995) and one-sixtieth that typieally a benign lesion, can be an incidental radio-
of either lung or breast carcinoma. The annual inci- graphie finding. Despite newer imaging techniques,
den ce for bone sarcomas is approximatelyO.8/1 00,000 the radio graph is the preliminary and single most
(DORFMAN and CZERNIAK 1995). The subject is par- important imaging investigation. Frequently, the diag-
ticularly pertinent when dealing with the knee, as nosis may be obvious to the trained eye, and further
imaging, if required, is then directed towards staging
A.M. DAVIES, MD the lesion. Alternatively, if an abnormality is present
Consultant Radiologist, MRI Centre, Royal Orthopaedic Hos- on the film and the precise nature is not immediately
pital, Bristol Road, Birmingham, B31 2AP, UK
D. VANEL,MD
apparent, certain features will indieate a differential
Department of Radiology, Institut Gustav Roussy, 39 rue diagnosis and other forms of imaging can then be
Camille Desmoulins, Villejuif 94895, France employed to assist in establishing a more definitive
radiological diagnosis. If the initial radiograph is ing the smallest bone, the proximal fibula, would be
normal, however, with persisting and increasing radiographically more conspicuous than those in the
symptoms, a repeat radio graph may be indicated in femur or tibia. Beware, however, the patella. Although
due course. it is a rare site, tumours of the patella are easily over-
Early signs of a bone tumour, or for that matter looked as the bone is projected over the distal femur
infection, include subtle areas of ill-defined lysis or on the anteroposterior projection.
sclerosis, cortical destruction, periosteal new bone In the presence of anormal radiograph, referred
formation and soft tissue swelling (Fig.18.I) (ROSEN- pain needs to be considered. Hip joint pathology pre-
BERG et al. 1995). Not surprisingly, bone lesions are senting with referred pain to the knee is a well-recog-
frequently missed or overlooked on the initial radio- nised entity in the child. If referred pain is suspected,
graph. In an audit performed at one of the authors' then radiographs of the pelvis and proximal femur
institutions (A.M.D.), in approximately 20% of cases are indicated. On occasion, radiographically occult
neither the clinician nor the radiologist at the refer- lesions may be detected by bone scintigraphy and/or
ring centre detected the bone tumour on the initial magnetic resonance (MR) imaging. It is important to
radiographs, although evidence was present on ret- stress that, owing to its high sensitivity, MR imaging
rospective review of the films (GRIMER and SNEATH all too frequently reveals abnormalities of little or no
1990). This is a bigger problem with tumours of flat clinical significance. An increasing number of knee
bones, such as the pelvis, than around the knee joint. MR scans are performed each year for a variety of
The pathological process may be well established conditions. Incidental medullary abnormalities will
even in the presence of anormal radio graph. At be revealed in the distal femoral meta-diaphysis in
least 40%-50% of trabecular bone must be destroyed a small percentage of cases which alm ost invariably
before a discrete area of lucency can be discerned prove to be innocuous chondromas (Fig.18.2).
on the radiographs (ARDRAN 1951; EDELSTYN et al.
1967). Erosion or destruction of the cortex is more
readily apparent. It is self-evident that the smaller
the bone involved, with a greater proportion of corti- 18.3
cal to medullary bone, the easier it will be to detect Diagnosis
an abnormality on the radiographs. Therefore, were
three tumours of similar size to arise around the 18.3.1
knee joint, it is fair to assurne that the lesion affect- Diagnosis of Bone Tumours
a b
and myeloma should always be considered in a It is at this stage that attention should turn to
patient over 40 years of age (Fig. 18.3). Similarly, the imaging. The radiograph remains the most accu-
metastatie neuroblastoma should be in the differ- rate of all the imaging techniques currently available
ential diagnosis at 2 years of age or under. Con- in determining the differential diagnosis of a bone
versely, a tumour arising in adolescence or early lesion (KRICUN 1983). Although many lesions will
adult life is unlikely to be a metastasis. be instantly recognisable, it is prudent to analyse
2. Gender and ethnic origin. Many bone tumours the radiographie features present. The analysis can
occur more commonly in boys but this fact does not be performed by answering the following questions:
play a significant role when formulating the dif- Whieh bone and what part of the bone is involved?
ferential diagnosis. Ewing's sarcoma is unusual in
that it is prevalent in Caucasians but is rarely seen
in Afro-Caribbeans. A number of non-neoplastie
lesions that may on occasion simulate neoplasia
also showa racial disposition, e.g. siekle ceIl, Gau-
cher's disease and Paget's disease. The geographie
origin of the patient may also be significant in that
the incidence of bone and joint infection is much
more common in the underdeveloped countries.
3. Family history. There is little evidence of a famil-
ial predisposition to the formation of musculosk-
eletal neoplasms in most instances. The exception
is certain congenital bone conditions which may
undergo malignant transformation, e.g. diaphyse-
al aclasis (Fig. 18.4), Ollier's disease and Maffucci's
syndrome.
4. Multiplicity. It is critieal early in the management
of a patient to establish whether alesion is soli-
tary or multiple as this will influence the differen-
tial diagnosis. Frequently this question will not be Fig. 18.3. AP radiograph in a 60-year-old male with a solitary
definitively answered until the staging imaging is renal metastasis in the tibial metaphysis. The radiographie fea-
performed. tures mimie a GeT
310 A. M. Davies and D. Vanel
flBRIlUS DYSPlASIA
flBROSARCOMA
flBROXANTHOMA
IfiBRIlUS COl!llCAl DEfECT:
NON-QSSlfYING fIBROMA!.
OSTEOCHONDROIM
ENCHONDROMA }
-H----++
t----------------- ~~=,~T'"
CHONDROSARCOIM
GIANT Cut TUMOR
CHllD, METAPHYSE!
ADULT, ' ~ND Of BOI<
.~ CHONDROillASTOMA --cH---F"I
~ ARTICUlAR OSTEOCHONDROMA
] -~~~~~~II~~:~~-------
loma. Following skeletal fusion, subarticular lesions, marginated lesion (Fig. 18.8). The faster the growth,
analogous in the adult to the epiphyseallesions, the more aggressive the pattern of destruction and
include GCT, clear cell chondrosarcoma (rare) and the wider the zone of transition between tumour and
intra-osseous ganglion. Most cases of osteomyelitis normal bone (Fig. 18.1). Aggressivity per se does
will arise within the metaphysis of a long bone, typi- not conclusively indieate malignancy, but the malig-
cally the tibia and femur. nant tumours tend to be faster growing than their
It can also be helpful to identify the origin of the benign counterparts. Geographie bone destruction
tumour with respect to the transverse plane of the is the term applied to bone lesions that appear well
bone. Is the tumour central, eccentric or cortieally marginated with a thin zone of transition. The thick-
based? For example, a simple bone cyst, fibrous dys- er the sclerotic border, the longer the host bone has
plasia and Ewing's sarcoma will tend to be centrally had to respond to the lesion and, therefore, by impli-
located. Chondromyxoid fibroma and fibrous corti- cation, the slower the rate of growth of the lesion. The
cal defect/non-ossifying fibroma (Fig. 18.8) are typi- vast majority of bone tumours in children showing a
cally eccentric. Lesions that usually arise in an eccen- geographic pattern of destruction are benign, such as
tric position may appear central if the tumour is simple bone cyst (SBC), ABCs, fibrous dysplasia and
particularly large or the involved bone is of a small enchondroma.
calibre. Therefore, most tumours arising in the proxi-
mal fibula will appear "central". There are numerous Moth-eaten and permeative bone destruction are
surface lesions of bone which are related to the outer terms used to describe bone destruction in whieh
cortex (KENAN et al. 1993a; SEEGER et al. 1998). The there are multiple tiny corticallucencies with an ill-
majority of the malignant surface lesions of bone are defined zone of transition. These patterns indieate the
the rare forms of osteosarcoma, e.g. periosteal, high- aggressive nature of these lesions in contrast to those
grade surface and parosteal osteosarcoma. Most of with a geographie pattern. The rapid growth of these
the cases of parosteal osteosarcoma arise from the lesions does not allow the host bone sufficient time to
posterior metaphysis of the distal femur. react and produce a response. Typically malignancies,
including osteosarcoma, Ewing's sarcoma and neu-
Pattern of Bone Destruction. Analysis of the interface roblastoma metastasis, exhibit a moth-eaten or per-
between tumour and host bone is a good indieator of meative pattern of bone destruction (Figs. 18.1, 18.6,
the rate of growth of the lesion. A sharply marginat- 18.9). Acute osteomyelitis is the "benign" condition
ed lesion usually denotes slower growth than a non- whieh mayaiso give a moth-eaten appearance.
312 A. M. Davies and D. Vanel
a b
a b
Periosteal Reaction. The periosteum is normallyradio- osteal reaction is frequently valuable in narrowing
lucent but will mineralise when stimulated by an adja- down the differential diagnosis of a bone tumour.
cent osseous or para-osseous process. The rate o( A "shell" is used to describe a lytie lesion with
mineralisation is partly dependent on the age of the bone expansion. The shell is the periosteal new bone
patient. The younger the patient, the more rapid the laid in response to the growing tumour. The thieker
appearance of radiographie change and vice versa. the shell, the slower growing the lesion and vice
Periosteal reaction, otherwise known as periosteal versa. Shells are typically found in benign lesions
new bone formation, may occur in any condition such as SBC, ABC, fibrous dysplasia and chondro-
whieh elevates the periosteum, whether it be blood, myxoid fibroma (Fig. 18.10). They mayaIso be seen
pus or tumour. The appearance and nature of a peri- with a telangiectatie osteosarcoma, which frequently
Tumours and Tumour-like Lesions 313
a Fig. 18.10. AP radiograph (a) and CT (h) of an ABC of the proximal fibula.
There is a typical expanded shell with fluid-fluid levels on the CT
mimics an ABC. In the older age group, shells are popcorn in appearance (Figs. 18.2a, 18.11). Identi-
found in expansile metastases from renal and thyroid fying the pattern of matrix calcification will signif-
primaries and plasmacytoma. icantly reduce the differential diagnosis but matrix
A lamellar periosteal reaction is seen in many per se has no influence as to whether the lesion is
traumatic and inflammatory conditions. The lamel- benign or malignant. The distribution can be help-
lated periosteal reaction, otherwise known as onion ful. For example, both enchondroma and medullary
skin, is seen in Ewing's sarcoma, osteosarcoma, eosin- infarction, which frequently arise in the distal femur,
ophilic granuloma and acute osteomyelitis (Figs.18.6,
18.9). A spiculated periosteal re action occurs when
the mineralisation is oriented perpendicular to the
cortex and denotes a more rapidly evolving process.
It is typical of malignant tumours such as osteosarco-
ma and Ewing's sarcoma (Fig.18.9). It may be seen in
benign tumours such as haemangioma of bone and
non-neoplastic conditions such as thalassaemia and
thyroid acropachy, but not in relation to the knee.
may show caleification of a similar nature. However, ganglia, from solid tumours. Doppler ultrasound can
the distribution is typically central in enchondroma, be employed to assess the vascularity of alesion and
while it is peripheral in medullary infarction. ultrasound is ideally suited for image-guided biopsy.
18.3.2 18.3.3
Diagnosis of Soft Tissue Tumours CT and MR Imaging in Diagnosis
The lack of contrast resolution is a well-recognised The prineipal role of computed tomography (CT)
limitation of radiography, but the value of the exami- and MR imaging in the management of the patient
nation should not be underestimated in the evalua- with a suspected musculoskeletal tumour is in stag-
tion of soft tissue masses. It will not identify the pre- ing (see Sect. 18.4). In selected cases both techniques
eise diagnosis in the majority of cases, but it can still can be useful in establishing a differential diagnosis.
provide valuable information, e.g. on the presence of The CT features that should be assessed are similar
caleification and bone involvement. The absence of to those described above when evaluating the radio-
any bony abnormality in the presence of a clinically graphs. This reflects that fact that both are radio-
palpable mass immediately indieates that the pathol- graphie techniques relying on the attenuation of an
ogy is of soft tissue origin, albeit with a large differ- X-ray source. Cortieal breaching, soft tissue exten-
ential diagnosis. sion and faint mineralisation are all more readily
The radiodensity of most soft tissue masses appreeiated on CT scans than on radiographs.Assess-
approximates to that of water and is similar to that ment of CT attenuation values will allow distinction
of muscle; such masses are, therefore, only revealed between fat-containing and fluid-containing masses.
by virtue of their mass effect. In a minority of cases, Although the physical basis of MR imaging is very
part or all of the tumour may exhibit a radiodensity different, similar morphological information can be
suffieiently different from that of water for it to be easily identified. The exception is the signal voids of
visualised directly on radiographs. Lipomas, the com- fine mineralisation, whieh can be easily missed on
monest of all soft tissue tumours, produce a low MR imaging. Potentially misleading MR features that
radiodensity between that of muscle and air. For this might suggest a bone sarcoma are prominent marrow
reason, lipomas are typieally well demarcated from oedema and soft tissue oedema (HAYES et al. 1992).
the surrounding soft tissues and, if of suffieient size, These are, however, common with osteoid osteoma,
can be diagnosed on radiographs with moderate con- osteoblastoma, chondroblastoma, stress fracture and
fidence. It should be noted that low-grade liposarco- infection. Many soft tissue sarcomas will appear well
mas may contain variable amounts of fat that will defined on MR imaging owing to the presence of
also appear relatively radiolucent on radiography. A a pseudocapsule, whereas inflammatory processes,
low-kilovoltage technique will accentuate the differ- such as abscesses, will appear poorly defined owing
ences between fat and muscle. to the surrounding inflammatory exudate.
Increased radiodensity may be seen in the tissues The majority of tumours will have prolonged Tl
due to haemosiderin, calcification or ossification. and T2 relaxation times, thereby showing low to
Haemosiderin deposition typically occurs in synovi- intermediate signal on Tl-weighted and high signal
al tissues exposed to repeated haemorrhage such as is on T2-weighted sequences. Tl shortening, with a
seen in pigmented villonodular synovitis. Caleifica- high signal intensity, will be seen in fat-containing
tion or ossification in the soft tissues is a feature of a tumours, subacute haemorrhage and gadolinium
large spectrum of pathologies, including congenital, chelate enhancement. A low signal intensity on T2-
metabolic, endocrine, traumatic and parasitie infec- weighted images is seen with dense mineralisation,
tions. Primary soft tissue tumours are one of the less hypocellular/fibrous tumours, signal voids from flow-
common causes of caleification that the general radi- ing blood,haemosiderin deposition,surgiealimplants
ologist can expect to come across in his or her routine and bone cement. Fluid-fluid levels are well dem-
practiee. onstrated on both CT and MR imaging in a large
Ultrasound is an important technique in the ini- number of different musculoskeletal conditions. In
tial assessment of a suspected soft tissue mass. First, the immature skeleton with the appropriate radio-
it can confidently confirm or exclude the presence of graphie appearances fluid-fluid levels are most com-
a mass. Second, it can to a degree characterise the monly seen in ABCs (Fig. 18.lOb) (DAVIES and CAS-
lesion by distinguishing purely cystic lesions, such as SAR-PULLICINO 1992).
Tumours and Tumour-like Lesions 315
Dynamic contrast-enhanced MR imaging has been MR imaging for the purposes of staging a suspected
used to differentiate benign from malignant bone bone sarcoma are as follows:
lesions using the slope of the derived time-intensity
curves (VERSTRAETE et al. 1994). Benign bone lesions Extent in Bane? To assess the extent of bone involve-
tend to show a low slope as compared with the high ment by tumour, a Tl-weighted sequence should be
or steep slope of malignant lesions; however, there performed oriented along the long axis of the bone
is considerable overlap such that this technique is of involved (Figs. 18.2b, 18.12a, b). This sequence is par-
limited value in routine practice. For example, highly ticularly sensitive to marrow changes. It is necessary
vascularised or perfused lesions such as ABC, eosino- to measure the tumour extent from a recognised ana-
philic granuloma, osteoid osteoma and acute osteo- tomical reference point, which, for the purposes of
myelitis may all show slope values in the malignant a bone sarcoma arising around the knee, can be the
tumour range. Similarly, in the soft tissues, early myo- articular cortices of the femur or tibia. A gadolini-
sitis ossificans will show a steep slope mimicking um-chelate should not be used at this stage as uptake
malignancy. of the contrast medium may well render the tumour
iso-intense with marrow fat. This problem can be
overcome by utilising a contrast-enhanced fat-sup-
pressed Tl-weighted sequence but this is an expen-
18.4 sive way of achieving the same result. Many benign
Surgical Staging as well as malignant bone tumours show a variable
degree of peritumoral oedema. These include osteoid
Accurate surgical staging is a fundamental requisite osteoma, chondroblastoma, GCT and osteosarcoma
of all oncological imaging. The staging system reg- (Fig. 18.13b). The oedema appears as a zone of inter-
ularly used for bone and soft tissue sarcomas is mediate signal intensity merging imperceptibly with
that adopted by the Musculoskeletal Tumor Society the main tumour. With sarcomas it can be difficult
(ENNEKING et al. 1980). This assigns one of three to distinguish where tumour ends and oedema com-
grades according to the local extent of the tumour, mences. Arguably, it is prudent to include all reduced
the presence or absence of metastases and the his- marrow signal within the measurements of the
tological grade. Clarification of the first two features tumour extent as malignant cells may contaminate
of the staging system relies entirely on imaging. The the oedematous area beyond the immediate confines
value of a straightforward staging system, such as of the main tumour. Some researchers have suggest-
this, is that it is easily applied, correlates well with ed that it is possible on MR imaging to distinguish
prognosis and allows valid comparison of studies between tumour tissue and peritumoral oedema
of differing treatments and treatment centres. An using a dynamic contrast-enhanced sequence, but it
alternative staging system for bone sarcomas is the is difficult to believe this technique would pick up
American Joint Committee on Cancer, which iden- isolated nests of malignant cells. It is interesting to
tifies tumour extension, i.e. whether it is confined speculate whether in the future diffusion-weighted
to bone (Tl) or extends beyond bone (T2), grade MR imaging might help distinguish tumour from
(Gl-4), nodal involvement (N) and distant tumour oedema. Fortunately, this is not a significant manage-
spread (M). ment problem in the majority of patients with a sar-
Determination of local tumour extent around the coma arising in the distal femur or proximal tibia.
knee usually relies on MR imaging. One study has Increasing the length of a custom -made prosthesis by
shown CT to be as good as MR imaging in staging several centimetres to accommodate the oedematous
(PANICEK et al. 1997a) although there has been some zone is unlikely to affect the functional outcome.
doubt expressed as to whether the technique and
quality of technology used in that multicentre study Extent in Soft Tissue? If the tumour is confined to
were stricdy comparable (STEINBACH 1998). Howev- bone the cortex will remain intact. Cortical bone
er, where access to MR imaging remains limited, CT is appears black on all MR sequences as it does not pro-
an adequate alternative, albeit with a significant radi- duce a signal. Cortical destruction with loss of the
ation burden. The MR scan should preferably be per- black line is a frequent and characteristic finding of
formed before the biopsy as the trauma of the proce- bone malignancy (Fig. 18.13). Not infrequently, how-
dure may result in haemorrhage and oedema which ever, highly malignant sarcomas such as osteosarco-
can exaggerate the true extent of the tumour. The ma and Ewing's sarcoma can penetrate the cortex
tumour characteristics that should be assessed on without frank destruction. In this situation, best
316 A. M. Davies and D. Vanel
demonstrated on axial images, the dark contour of relatively high signal of fat, which may limit the con-
the cortex will persist with a permeative appearance trast with tumour. This problem can be overcome
analogous to the permeative or moth-eaten pattern with the use of fat suppression. A STIR sequence is
on radiographs. It is convention to describe any an alternative, albeit with a poorer signal to noise
tumour tissue identified outside the cortex as extra- ratio (Fig. 18.13b). The STIR sequence will also tend
osseous or soft tissue extension. Strictly speaking this to overstage the extent of the tumour owing to its
is often incorrect as the tumour can remain con- increased sensitivity to raised water content in a tissue.
fined by a largely intact periosteum. Nevertheless, As in bone, the distinction of soft tissue tumour from
this convention persists and is usually only a source of perineoplastic oedema can be problematic (SHUMAN
problems when resolving the findings of MR imaging et al. 1991). Knowledge of the compartmental anato-
versus the examination of the pathological specimen. my around the knee is essential when determining the
The relatively high water content of most tumours, stage of the tumour (ANDERSON et al. 1999).
both bone and soft tissue sarcomas, renders them
iso-intense and therefore indistinguishable from sur- Joint Involvement? It is important to identify knee
rounding muscles on Tl-weighted images. It is for joint invasion by a sarcoma because, although the
this reason that to assess soft tissue extension a joint will usually be excised at the time of definitive
T2-weighted sequence, with good contrast between surgery, prior knowledge will prevent the surgeons
tumour and muscle, is required. A disadvantage of from opening the joint and thereby potentially con-
the widely used fast spin echo (turbo) T2-weighted taminating the surgical field with tumour cells. MR
sequence is slightly reduced spatial resolution and imaging is highly sensitive for detecting joint inva-
Tumours and Tumour-like Lesions 317
sion but false positives due to subsynovial rather than to grow in the line of least resistance. Around the knee
true intra-articular spread can lead to overstaging this is typically into the popliteal fossa. MR imaging
(SCHIMA et al. 1994). This is problematic in the knee, can demonstrate whether a tumour is close to or in
where anterior extra-osseous spread of a distal femo- contact with a neurovascular structure in the poplite-
ral sarcoma will appear to invade the suprapatellar al fossa, but usually cannot distinguish mere contact,
pouch while in reality it is frequently displacing it. adherence or early invasion (PANICEK et al. 1997b).
Of significance is the fact that the absence of a joint Fortunately, the prevalence of neurovascular involve-
effusion has a high negative predictive value for joint ment in bone sarcomas is less than 4%, such that,
invasion (SCHIMA et al. 1994). The articular cartilage although the positive predictive value of MR imaging
is a relative barrier to tumour growth and is usually for involvement is poor, the negative predictive value
only involved in very large or late presenting tumours. is over 90% (PANICEK et al. 1997b). MR angiography
Therefore, typical sites of joint invasion in the knee can be used to delineate the relationship of the tumour
are the meniscocapsular reflections and the intercon- to vessels (LANG et al. 1995; SWAN et al. 1995).
dylar notch along the cruciate ligaments (Fig.18.13c).
Transarticular spread is rare. Identification of tumour Skip Metastases and Lymph Node Involvement? Small
on both sides of the knee should suggest that it arose synchronous foci of tumour, usually osteosarcoma,
de novo in the joint rather than the bone. that are present within the same bone as the pri-
mary tumour, or within a bone on the other side of
Neurovascular Involvement? Once a sarcoma has an unaffected joint, are called skip metastases. Skip
extended beyond the confines of the bone it will tend metastases in osteosarcoma have been reported in
318 A. M. Davies and D. Vanel
up to 25% of eases (ENNEKING and KAGAN 1975) Bone scintigraphy is used to exclude skeletal metas-
although in the author's experienee the true inci- tases. However, over 95% of scintigraphie abnormali-
denee is less than 5%. The majority will be deteeted ties oeeurring at the time of presentation of osteosar-
by bone scintigraphy (Fig. 18.12e) but scintigraph- eoma at loeations distant from the primary tumour
ieally negative skip metastases have been reported do not represent metastatie disease (KELLER et al.
(BHAGIA et al. 1997). When staging a sareoma around 1984). It is, therefore, important to eorrelate scinti-
the knee, the best resolution images will undoubt- graphie abnormalities with radiographs of the rele-
edly be obtained utilising the knee eoil (Fig. 18.l2a). vant area.
It is prudent however, to include a single large At the same time as the staging imaging studies
field of view Tl-weighted sequenee along the line are performed, it is the usual praetiee in the author's
of the femur or tibia to exclude a skip metastasis unit to obtain measurement radiographs of both the
(Fig. 18.12b). In the author's experienee, for some affeeted and the eontralateral lower limbs to aid
unknown reason most transartieular skip metastases the manufaeture of a eustom-made prosthesis. Also,
are identified in the proximal tibia in patients pre- the bone age of the skeletally immature patients is
senting with a distal femoral osteosareoma. Lymph estimated as eertain designs of prosthesis allow for
node spread in bone sareomas is uneommon and growth, i.e. are extendable.
usually a late manifestation of extensive disease.
As at other sites, imaging has diffieulty distinguish-
ing metastatie infiltration from reaetive hyperplasia
(BEARCROFT and DAVIES 1999). The exeeption is 18.5
those eases of osteosareoma with mineralisation, Imaging Follow-up
indieating metastatie involvement, whieh ean be
easily deteeted on radiographs or CT and will show The imaging follow-up for a patient with a proven
inereased aetivity on bone scintigraphy. A false diag- sareoma arising around the knee ean be divided into
nosis of a skip metastasis may oeeur on scintig- short-term (i.e. pre-definitive surgery) andlong-term
raphy when inereased aetivity within an involved (i.e. post-definitive surgery). In the short-term many
lymph node is projeeted over the distal femur when patients with a sarcoma will be ente red into one of
only anterior or posterior projeetions are obtained the international adjuvant ehemotherapy trials. After
(BEARCROFT and DAVIES 1999). Rarely, in patients a predetermined number of eycles of ehemotherapy
with longstanding prostheses, regional lymphade- and immediately before surgery, the patient is re-
nopathy may oeeur owing to a foreign body reae- staged with an MR sean of the primary tumour and a
tion in response to the lymphatie uptake of metal CT sean of the ehest. This is to ensure that the stage of
debris (DAVIES et al. 2001). In patients treated by the tumour has not altered and that the planned sur-
amputation, post-traumatie neuromas may mimie gery is still appropriate. Also, this is an opportunity
lymphadenopathy. to use imaging to assess the response of the tumour
Gd-DTPA has little value in the initial staging of to the ehemotherapy. Histologieal response to ehe-
bone sareomas (SEEGER et al. 1991). It may help motherapy expressed as pereentage neerosis is one
distinguish subsynovial spread from true joint inva- of the most important prognostie indieators in both
sion (SCHIMA et al. 1994), and adynamie eontrast- osteosareoma and Ewing's sareoma. Over the years
enhaneed sean ean be obtained at this stage as a base- all types of imaging have been used to estimate the
line study for the subsequent assessment of tumour response to ehemotherapy.
response to ehemotherapy (see Seet.18.5). Post -ehemotherapeutie radiographie and CT find-
The principles of staging soft tissue sarcomas aris- ings do not consistently differentiate the good from
ing around the knee are very similar to those detailed the poor responder (SHAPEERO and VANEL 2000).
above for bone sareomas. For both bone and soft For example, an inerease in tumour volume may sug-
tissue sareomas the exclusion or eonfirmation of pul- gest a poor response but mayaiso represent haemor-
monary metastases requires ehest CT. The sensitivity rhage seeondary to neerosis in a responsive tumour
of pulmonary CT has been improved with the intro- (VAN DER WOUDE et al. 1998). Conventional angi-
duetion of spiral CT (GINSBERG and PANICEK 2000). ography is eonsidered too invasive a proeedure
Overstaging with spiral CT is a potential hazard as up for monitoring tumour response to ehemotherapy.
to 70% of solitary nodules less than 5 mm in diameter Although it ean identify over 90% of responders, it
at initial presentation in ehildren with solid extratho- will miss 50% of the poor responders (CARRASCO et
racie tumours may be benign (GRAMPP et al. 2000). al. 1989). It remains to be seen whether MR angiog-
Tumours and Tumour-like Lesions 319
raphy can fulfil a useful role in this respect (LANG et with or without bone destruction. Locally recurrent
al. 1995). bone sarcoma will usually occur within the soft tis-
If there is a significant extra-osseous component sues at the site of the initial surgery as the host bone
to the tumour, Doppler ultrasound can be used to will have been excised and replaced with a prosthe-
monitor response (VAN DER WOUDE et al. 1995). sis. Detection on radiographs is easier if there is evi-
The technique is operator dependent, which may dence of matrix mineralisation. Recurrent tumours
affect reproducibility of results on sequential scan- with the propensity to mineralise (i.e. osteosarcoma)
ning. Scintigraphy using technetium-99m methylene will usually exhibit focal increased activity on scin-
diphosphonate, thallium-201 and gallium-6? and tigraphy but it is rarely used for this purpose.
fluorine-18 fluorodeoxyglucose positron emission MR imaging is the technique of choice in the
tomography (PET) scanning have been advocated for detection of early recurrence when local control may
the estimation of tumour response (SHAPEERO and still be surgically achievable. While ultrasound does
VANEL 2000). Inherent to all of these methods is the have some attractions (CHOI et al. 1991), MR imag-
limited anatomical resolution and, with PET scan- ing will still be required for preoperative evaluation
ning, limited availability. To date these techniques are if a recurrence is identified. Depending on the pres-
largely reserved for research purposes. ence or absence of mineralisation, most recurrences
Unenhanced MR imaging has a limited role. will show a high signal intensity mass on T2-weight-
Increased or unchanged tumourvolume and increased ed or STIR images (VANEL et al. 1994). Diffuse high
peritumoral oedema after chemotherapy suggest signal intensity is frequently seen shortly after sur-
a poor histological response in osteosarcoma and gery or can be prolonged following radiation therapy
Ewing's sarcoma. Virtual obliteration of the extra- (RICHARDSON et al. 1996). Contrast medium may be
osseous component combined with a hypo-intense required to distinguish enhancing recurrent tumour
rim in Ewing's sarcoma usually indicates a good from seromas, haematomas etc. Dynamic contrast-
response. It is, however, impossible to exclude small enhanced MR imaging can be helpful in differen-
foci of viable tumour without contrast medium. Stan- tiating small recurrences from other postoperative
dard contrast-enhanced MR imaging is also of lim- changes.
ited value as viable tumour, revascularised necrotic It is generally accepted that it is usually the meta-
tissue, reactive hyperaemia etc. may all enhance. It static disease that will eventually kill the patient, and
is for this reason that much of the work on imaging not the primary tumour itself. It is for this reason that
assessment of the response of sarcoma to chemo- follow-up imaging is concentrated on the site where
therapy over the past decade has concentrated on metastases are likely to occur, namely the lungs. Chest
dynamic contrast-enhanced MR imaging.A number radiographs are usually considered adequate. Serial
of different techniques have been described but all ehest CT scans are of doubtful value in view of the
rely on the underlying principle that viable tumour considerable radiation dose involved. The natural
enhances rapidly (i.e. within seconds of the contrast history of osteosarcoma has been modified by che-
medium arriving in the adjacent artery) whereas all motherapy in that up to 20% of those who develop
other enhancing tissues take much longer. It is pos- metastases will first do so in bone prior to there
sible on the console of most modern scanners to being any evidence of pulmonary metastases. The
plot a time/intensity curve showing the uptake of the prognosis for a patient with osseous metastases is so
contrast medium. By comparing the curve obtained poor that serial follow-up scintigraphy is unlikely to
before commencement of chemotherapy with that modify the outcome. Scintigraphy is indicated should
obtained afterwards, the tumour response can be a patient on follow-up develop bone pain.
estimated. It should be noted that this is a time-con- It should be recognised that the prolonged medical
suming and costly exercise, with numerous variables and surgical management of a patient with a sarcoma
that directly influence patient management in very is not without risk of complications. Prostheses may
few cases. become loose or infected or require replacement if
In the long-term the patients are closely moni- a child has outgrown the extended length of a grow-
tored for evidence of local recurrence (DAVIES and ing prosthesis (KASTE et al. 2001). Allografts may also
VANEL 1998), metastatic disease (BEARCROFT and become infected and are prone to fracture. In the
DAVIES 1999) and complications of treatment. Local long-term follow-up of patients who have received
recurrence of a sarcoma is almost inevitable if the radiotherapy, pain or functional impairment within
original resection margin was not wide. Recurrence the radiation field should lead to consideration or
may be detected on radiographs as a soft tissue mass bone necrosis or radiation-induced sarcoma.
320 A. M. Davies and D. Vanel
18.6.1.3
Fibrous Tumours
in the distal femur than the other bon es. The radio- pathognomonic, feature of ABCs seen on CT and MR
graphie appearances are a well-defined lytie lesion, imaging (Figs. 18.10b, 18.15b ).It is necessary for the
cent rally located within the metaphysis and migrat- scan to be perpendieular to the fluid-fluid levels for
ing with time into the diaphysis. A typieal, but not them to be visible. The differential diagnosis for an
pathognomonic sign, is the so-called fallen fragment. ABC includes SBC, chondromyxoid fibroma, non-
This represents a fragment of fractured cortex that ossifying fibroma, GCT (in older patients) and, most
descends to the dependent portion of the cyst. important of all, telangiectatie osteosarcoma. To
The differential diagnosis around the knee includes the unwary, telangiectatie osteosarcoma may resem-
aneurysmal bone cyst (ABC), fibrous dysplasia and ble an ABC both on imaging (including fluid-fluid
non-ossifying fibroma; the last-mentioned is usually levels) and histology (Fig. 18.16). Misdiagnosis can
eccentrically located. have potentially disastrous consequences as the man-
Aneurysmal bone cyst is a non-neoplastie lesion agement of an ABC is curettage while that of a tel-
consisting of multiple blood-filled spaces with vary- angiectatie osteosarcoma is chemotherapy and wide
ing amounts of fibrous, riehly vascular connective surgical excision.
tissue. Three-quarters of cases present in patients Giant cell tumour (GCT) is a locally aggressive
under 20 years of age. Approximately 26% of cases tumour representing approximately 5% of all prima-
arise in the bones around the knee,ABC being slight- ry bone tumours and 22% of benign bone tumours.
ly more common in the proximal tibia than in the Approximately 50% arise around the knee, more
other bones. The radiographie features are a well- commonly in the distal femur. Patients are usually
defined multiloculated expansile lesion arising in between 20 and 40 years of age. The tumour is consid-
the metaphysis (Fig. 18.10). Occasionally ABCs may ered benign although occasionally it may be multi-
arise in a subperiosteallocation but these are usually focal, metastasise to the lungs or undergo malignant
in the diaphysis. If the ABC is growing particularly transformation. The characteristie radiographie fea-
fast, the expanded shell may be thinned or absent tures are an expansile, eccentric, lytic, subarticular
(Fig. 18.15). ABC may be a secondary phenomenon lesion which is weIl defined without marginal scle-
occurring in pre-existing bone lesions, in whieh case rosis (Fig. 18.17). On the AP projection of the knee,
there may be imaging evidence of the underlying distal femoral GCTs may appear remote from the
abnormality. Fluid-fluid levels are a typical, but not articular margin but inspection of the lateral will
a b
Fig. 18.15. a Lateral radiograph of an ABC of the femur. Because of the rapidity of growth only some peripheral mineralisation
("shell") is visible distally. b Multiple fluid-fluid levels are visible on the sagittal STIR image
Tumours and Tumour-like Lesions 323
a a
b b
Fig. 18.16a, b. MR imaging of a telangiectatic osteosarcoma of Fig. 18.17. Typical giant ceIl tumours of the proximal tibia (a)
the distal femur. The sagittal Tl-weighted image (a) shows a and proximal fibula (b)
multiloculated lesion with high signal intensity cysts due to
subacute haemorrhage. The fluid-fluid levels are more conspic-
uous on the sagittal STIR image (b)
WYNNE-DAVIES 1983). Clinical features that suggest plasia and non-ossifying fibroma. As with chondro-
malignant change include pain and increasing size sarcoma, they tend to present in a slightly older age
following skeletal fusion. Measurement of the thick- group than most other primary sarcomas of bone,
ness of the cartilage cap using ultrasound, CT or MR with a peak in the fourth decade. Typical radio-
imaging can be helpful. A cartilage cap of less than 2 graphie appearances are geographie bone destruc-
cm is likely to be benign, whereas as the cap exceeds tion with a wide zone of transition in an eccentric
2 cm in thickness the likelihood of chondrosarcoma metaphyseal or meta-diaphyseallocation. They tend
increases. Complications of osteochondromas such not to extend to the articular margin. Cortical
as overlying bursitis and pseudo-aneurysm forma- destruction with soft tissue extension is common but
tion can mimic malignant change. periosteal new bone formation is unusual and there
is no matrix mineralisation. In the older patient the
18.6.2.3 appearances can be indistinguishable from metasta-
Fibrous Tumours sis or lymphoma of bone.
18.6.3
Patellar Tumours
18.7.2
Malignant Soft Tissue Tumours
18.8.1.2
Pigmented Villonodular Synovitis
Fig.18.25. Lateral radiograph (a) and sagittal T2* MR image (b) of synovial chondromatosis. The thin fod of cartilaginous min-
eralisation on the radiograph are seen as multiple small signal voids within synovial masses on the sagittal T2* MR image
MR imaging will reveallobulated homogeneous intra- the degree of mineralisation. In severe cases, bone ero-
articular soft tissue masses of intermediate signal sion will be present. In longer-standing cases, discrete
intensity on Tl-weighted images and high signal inten- intra-articular bodies can be identified. This condi-
sity on T2-weighted images (Fig. 18.25b) (KRAMER et tion of primary synovial osteochondromatosis should
al. 1993). The extent of focal areas of signal void, best be distinguished from the much commoner secondary
seen on T2-weighted or STIR images, will depend on form. In the latter there are osteocartilaginous loose
330 A. M. Davies and D. Vanel
bodies of varying size associated with more advanced PVNS. Lipoma arborescens can be confidently diag-
osteoarthritis. The loose bodies in the secondary form nosed on MR imaging, and this has probably led to
are assumed to originate from the debris shed by the recent increase in the number of reported cases.
the damaged articular cartilage. There is a rare asso- To give some idea of the relative incidence of this
ciation between synovial osteochondromatosis and condition, the diagnosis is made in approximately 1
synovial chondrosarcoma (see Sect.18.8.2.2). in 1,400 knee MR scans performed in the unit of one
of the authors (A.M.D.).
18.8.1.4
Lipoma Arborescens 18.8.1.5
Synovia I Haemangioma
Lipoma arborescens is a cause of slowly increasing
monarticular pain and swelling in the absence of a Soft tissue haemangioma arising in the synovium
history of trauma. The aetiology is unknown; how- accounts for less than 1% of all haemangiomas
ever, it has been described in patients with rheu- (RESNICK and OLIPHANT 1975). Over 65% of these
matoid arthritis, psoriasis, degenerative joint disease involve the knee joint, typically the suprapatellar
and diabetes mellitus (KLOEN et al. 1998), suggesting pouch. Approximately one-third will also have extra-
a reactive rather than a neoplastic process. It occurs articular extension. On MR imaging a lobulated intra-
most commonly in the knee joint, particularly the synovial mass with or without an extra-articular
suprapatellar pouch (FELLER et al. 1994; MARTIN et component will be seen containing prominent ser-
al. 1998), but the shoulder, hip, wrist, elbow and ankle piginous vessels (Fig. 18.27). Fatty elements, identi-
are also recognised sites. Although usually monartic- fied as high signal intensity on Tl-weighted images,
ular, cases of bilateral involvement of the knees may be seen, similar to an intramuscular haemangi-
and hips have been reported (MARTIN et al. 1998; oma. Repetitive intra-articular bleeding will result in
SUMEN et al.1998). It occurs more commonly in men haemosiderin deposition in the synovium similar in
than women, usually in the fourth and fifth decades. appearance to PVNS and haemophiliac arthropathy.
Patients as young as 9 years (DONNELY et al. 1994) In the author's experience the degree of haemosid-
and as old as 90 years (LAoRR et al. 1995) have been erin deposition in synovial haemangioma tends to
reported. The appearance on MR imaging is typically be less florid than in cases of either PVNS or hae-
that of a large joint effusion with a frond-like mass mophiliac arthropathy. Localised pressure erosion of
arising from the synovium, which has fat density on the adjacent subcortical bone is common, particular-
all sequences (Fig. 18.26). The mass will therefore ly affecting the distal femoral metaphysis (Fig. 18.27)
appear dark on fat-suppressed sequences. Changes of
degenerative joint disease are seen in cases of long-
standing disease; erosions are less prominent than in
18.8.1.6
Intracapsular/Para-articular Chondromas
18.8.2 this arose in the ankle joint (KAUL and UNNI 1998).
Malignant Joint Tumours The commonest primary joint malignancies are syno-
vial sarcoma and synovial chondrosarcoma.
Malignancies arising within the knee joint, or any
synovial joint for that matter, are extremely rare. 18.8.2.1
Because of their relatively common occurrence in Synovia I Sarcoma
general, on occasion one might expect to see metasta-
ses or lymphoma arising within a joint. A single case Less than 10% of synovial sarcomas arise in an intra-
of malignant change in PVNS has been reported but articular location (Fig.18.29) (McKINNEY et al. 1992).
Fig. 18.29. Axial T2-weighted MR images of an intra-articular synovial sarcoma. The tumour mass can be identified invading
the lateral patellar retinaculum on the axial T2-weighted MR images
332 A. M. Davies and D. Vanel
More commonly the tumour arises within the soft (BUFKIN 1971; BARNES and GWINN 1974). The radio-
tissues dose to the knee, with or without secondary graphie appearances are erosion of the outer cortex
joint invasion. The imaging features of the intra- with small bony spicula which can simulate an
articular form mimic other synovial proliferative or aggressive or malignant lesion. It is seen in adoles-
infiammatory disorders such as primary synovial cents, particularly males, and is often bilateral. There
chondromatosis and gout. The bone destruction, is some debate as to whether it is a normal variant
however, tends to be more infiltrative than erosive or represents a traction enthesopathy at the sites of
and the disease shows rapid progression. attachment of the tendinous fibres of the adductor
magnus or medial head of gastrocnemius musdes.
18.8.2.2 Bone scintigraphy is typically normal, which would
Synovia I Chondrosarcoma be unusual for a traumatic lesion (BURROWS et al.
1982; VELCHIK et al. 1984). MR imaging, however,
Synovial chondrosarcoma is rare, with the knee joint can show some oedematous change in relation to the
the commonest site. Although it may arise de novo, cortical irregularity (HYMAN et al. 1995; YAMAZAKI
most cases described in the literature are due to et al. 1995) which would not be expected with a
malignant degeneration of synovial chondromatosis normal variant.
(KENAN et al. 1993b ). There can be considerable diffi-
culty in differentiating between synovial chondrosar-
coma and synovial chondromatosis, both on imaging 18.9.2
and on histology.As with intra-articular synovial sar- Stress Fractures
coma, the bone destruction in synovial chondrosar-
coma is more infiltrative than erosive. Extra-articular The fatigue form of stress fracture can cause diag-
extension, rapid local recurrence following surgery nostic difficulties in that the healing fracture may fre-
and the development of metastatic disease dearly quently be mistaken for a malignant lesion (LEVIN et
favour the diagnosis of a sarcoma. al. 1967; SOLOMAN 1974). The tibia, particularly the
proximal third, is the commonest site for fatigue frac-
tures in the adolescent and young adult and also the
commonest site to be misinterpreted radiologically
18.9 as a sarcoma of bone (DAVIES et al. 1988). A less
Tumour-like Lesions of Bone common site where similar problems may occur is
the distal femoral diaphysis (DAVIES et al. 1989).
There are a large number of disparate bone condi- Some indication of the in eiden ce of misdiagnosis is
tions which can have similar imaging appearances to indicated by two studies conducted in an orthopae-
tumours. What constitutes a tumour mimic depends die oncology unit. In the first, fatigue fractures of the
very much on the expertise of the individual review- proximal tibia accounted for 11 % of referrals with
ing the imaging. The majority can be dassified as tibial lesions and was the second commonest final
normal variants or post -traumatic, infiammatory and diagnosis after osteosarcoma (DAVIES et al. 1988).
metabolie conditions. The prudent radiologist will In the second study on femoral diaphyseallesions,
always have dose at hand, when reporting, one of the fatigue fractures were the second commonest diag-
standard reference texts on normal variants which nosis after Ewing's sarcoma but the commonest diag-
illustrates the wide spectrum of developmentallucen- nosis if only skeletally immature patients were con-
eies and irregularities of ossification that can be seen sidered (DAVIES et al. 1989).
on radiographs of the growing knee. Computed tomography will show the periosteal
and endosteal new bone formation, the fracture
line (if viewed on appropriate bone windows) and,
18.9.1 most important of all from the point of view of
Avulsive Corticallrregularity exduding malignancy, absence of a soft tissue com-
ponent (SOMER and MEURMAN 1982; YOUSEM et al.
Avulsive cortical irregularity (also known as peri- 1986; DAVIES et al. 1989). The increased attenuation
osteal desmoid, cortical desmoid and distal femoral within the medulla around the fracture site due to
metaphyseal defect) is a tumour-like fibrous pro- oedema and/or haemorrhage should not be mistak-
liferation of the periosteum which affects the pos- en for intra-osseous tumour tissue (ARRIVE et al.
teromedial cortex of the distal femoral metaphysis 1988; DAVIES et al. 1988). The value ofMR imaging in
Tumours and Tumour-like Lesions 333
Edelstyn GA, Gillespie PJ, Grebbel FS (1967) The radiologieal cance of bone scan abnormalities in patients with primary
demonstration of osseous metastases: experimental obser- osteogenie sarcoma. J Surg Onco126:122-129
vations. Clin RadiolI8:158-162 Kenan S, Abdelwahab IF, Klein MJ, Herman G, Lewis MM
Enneking WF, Kagan A (1975) The implications for skip metas- (1993a) Lesions of juxtacortieal origin (surface lesions of
tases in osteosarcoma. Clin Orthop 111:33-41 bone). Skeletal Radio122:337-357
Enneking WF, Spanier SS, Goodman MA (1980) A system Kenan S,Abdelwahab IF, Klein MJ et al (1993b) Synovial chon-
for the surgieal staging of musculoskeletal sarcoma. Clin drosarcoma secondary to synovial chondromatosis. Skel-
Orthop 153:106-120 etal Radio122:623-626
Feller JF, Rishi M, Hughes EC (1994) Lipoma arborescens of the Kim MG, Kim BH, Choi JA et al (2001) Intra-artieular ganglion
knee: MR demonstration. Am J Roentgen 163:162-164 cysts of the knee: clinieal and MR imaging features. Eur
Flemming DJ, Murphey MD (2000) Enchondroma and chon- Radiol11:834-840
drosarcoma. Semin Musculoskeletal Radio14:59-71 Kloen P, Keel SB, Chandler HP, Geiger RH, Zarins B, Rosenberg
Geirnaerdt MJA, Hermans J, Bloem JL, Kroon HM, Pope TL, AE (1998) Lipoma arborescens of the knee. J Bone Joint
Taminiau AHM, Hogendoorn PCW (1997) Usefulness of Surg 80B:298-301
radiography in differentiating enchondroma from central Kramer J, Recht M, Deely DM, Schweitzer M, Pathria M, Gentili
grade I chondrosarcoma.Am J Roentgen 169:1097-1104 E, Greenway G, Resniek D (1993) MR appearance of idio-
Geirnaerdt MJA, Hogendoorn PCW, Bloem JL, Taminiau AHM, pathie synovial chondromatosis. J Comput Assist Tomogr
van der Woude HJ (2000) Cartilaginous tumors: fast con- 17:772-776
trast-enhanced MR imaging. Radiology 214:539-546 Kransdorf M, Moser R, Vinh T et al (1989) Primary tumours of
Ginsberg MS, Panieek DM (2000) Subcentimetre pulmonary the patella. Skeletal RadiolI8:365-371
nodules detected in patients with sarcoma. Sarcoma Kransdorf MJ, Murphey MD (1997) Soft tissue tumors in a
4:63-66 large referral population: prevalence and distribution of
Grampp S, Bankier AA, Zoubek A, Wiesbauer P, Schroth B, diagnoses by age, sex and location, In: Kransdorf MJ, Mur-
Henk CB, Grois N, Mostbeck GH (2000) Spiral CT of the phey MD (eds) Imaging of soft tissue tumors. Saunders,
lung in children with malignant extra-thoracic tumors: dis- Philadelphia, pp 3-35
tribution of benign vs malignant pulmonary nodules. Eur Krieun ME (1983) Radiographie evaluation of solitary bone
RadiollO:1318-1322 lesions. Orthop Clin North Am 14:39-64
Greenspan A, Azouz EM, Matheus IJ, Decarie JC (1995) Synovi- Lang P, Grampp S, Vahlensieck M, Johnston JO, Honda G,
al hemangioma: imaging features in 8 histologieally proven Rosenau W, MaUhay KK, Peterfy C, Higgins CB, Genant HK
cases, review of the literature and differential diagnosis. (1995) Primary bone tumors: value of MR angiography for
Skeletal Radio124:583-590 preoperative planning and monitoring response to chemo-
Grey AC, Davies AM, Mangham DC, Ritchie DA (1998) The therapy. Am J RoentgenolI65:135-142
"penumbra sign" on Tl-weighted MR imaging in subacute Laorr A, Peterfy CG, Tirman PF, Rabassa AE (1995) Lipoma
osteomyelitis: frequency, cause and significance. Clin Radiol arborescens of the shoulder: MR imaging findings. J Can
53:587-592 Assoc Radio146:311-313
Grimer RJ, Sneath RS (1990) Diagnosing malignant bone Lee JK, Yao L (1988) Stress fractures: MR imaging. Radiology
tumours: editorial. J. Bone Joint Surg [Br] 72B:754-756 196:217-220
Hayes CW, Conway WF, Sundaram M (1992) Misleading aggres- Levin DC, Blazina ME, Levine E (1967) Fatigue fractures of
sive MR imaging appearance of some benign musculoskel- the shaft of the femur: simulation of a malignant tumor.
etalIesions. RadioGraphies 12:1119-1134 Radiology 89:883-885
Hughes TH, Sartoris DJ, Schweitzer ME, Resniek DL (1995) Pig- Liu J, Hudkins PG, Swee RD et al (1987) Bone sarcomas associ-
mented villonodular synovitis: MRI characteristies. Skel- ated with Ollier's disease. Cancer 59:1376-1385
etal Radio124:7 -12 Llauger J, Monill JM, Palmer J, Clotet M (1995) Synovial hem-
Hyman AA, Heiser WJ, Kim SE et al (1995) An excavation ofthe angioma of the knee: MR imaging findings in two cases.
distal femoral metaphysis: a magnetie resonance imaging Skeletal Radio124:579-581
study: case report. J Bone Joint Surg 77A:1897-1901 Mack TM (1995) Sarcomas and other malignancies of soft
Janzen DL, Peterfy CG, Forbes JR, Tirman PFJ, Genant HK tissues, retroperitoneum, pleura, heart, mediastinum and
(1994) Cystic lesions around the knee joint: MR imaging spleen. Cancer 75 [Suppl]:211-244
findings. Am J RoentgenolI63:155-161 Madewell JE, Ragsdale BD, Sweet DE (1981) Radiologie and
Jelinek JS, Kransdorf MJ, Utz JA, Berrey BH, Thomson JD, pathologie analysis of solitary bone lesions, part 1. Internal
Heekin RD, Radowieh MS (1989) Imaging of pigmented margins. Radiol Clin North Am 19:715-748
villonodular synovitis with emphasis on MR imaging. Am Martin S, Hernandez L, Romero L, Lafuente J, Poza AI, Ruiz
J RoentgenolI52:337-342 P, Jimeno M (1998) Diagnostie imaging of lipoma arbore-
Jones BC, Sundaram M, Kransdorf MJ (1993) Synovial sarco- scens. Skeletal Radio127:325-329
ma: MR imaging findings in 34 patients. Am J Roentgenol McKinney CD, Mills SE, Fechner RE (1992) Intraartieular syno-
161:827-830 vial sarcoma. Am J Surg Patholl6: 10 17-1020
Kalil RK, Unni KK (1998) Malignancy in pigmented villonodu- Milgram JW (1977) Synovial osteochondromatosis. A his-
lar synovitis. Skeletal Radio127:392-395 topathologieal study of thirty cases. J. Bone Joint Surg
Kaste SE, Neel MN, Rao BN, Thompson VF, PraU CB (2001) 59A:792-801
Complications oflimb-sparing procedures using endopros- Morrison JL, Kaplan PA (2000) Water on the knee: cysts, bursae
thesis replacements about the knee for paediatrie skeletal and recesses. MRI Clin North Am 8:349-370
sarcomas. Paediatr Radio131:62-71 Moser RP, Madewell JE (1987) An approach to primary bone
Keller SM, Rosenbaum RC, Rosenberg SA (1984) The signifi- tumors. Radiol Clin North Am 25:1049-1093
Tumours and Tumour-like Lesions 335
Mulder JD, Schutte HE, Kroon HM, Taconis WK (1993) Radio- Steinbach LS, Neumann CH, Stoller DW, Mills CM, Crues JV,
logie atlas of bone tumors. Elsevier, Amsterdam Lipman JK, Helms CA, Genant HK (1989) MRI of the knee
Murphey MD, Flemming DJ, Boyea SR et al (1998) Enchon- in diffuse pigmented villonodular synovitis. Clin Imaging
droma versus chondrosarcoma in the appendieular skel- 13:305-316
eton: differentiating features. RadioGraphies 18: 1213-1237 Steiner GC, Meushar N, Norman A, Present D (1994) Intra-
Murphey MD, Choi JJ, KransdorfMJ, Flemming DJ, Gannon FH capsular and paraartieular chondromas. Clin Orthop
(2000) Imaging of osteochondroma: variants and complica- 303:231-236
tions with radiologie-pathologie correlation. Radiograph- Suh JS,Hwang G,Hahn SB (1994) Soft tissue hemangiomas: MR
ies 20:1407-1434 manifestations in 26 patients. Skeletal Radiol 26:621-626
O'Mara JW, Keeling J, Montgomery EA, Aaron AD (2000) Pri- Sumen Y, Ochi M, Soda Y, Ikuta Y (1998) Lipoma arborescens
mary lesions of the patella. Orthopedies 23:328, 348, 370, in bilateral knee joints. Arch Orthop Trauma Surg
376,377 117:105-107
Oxtoby J, Davies AM (1996) The MR imaging appearances of Swan JS, Grist TM, Sproat JA, Heiner JP, Wiersma SR, Heisey
chondroblastoma. Clin RadioI51:22-26 DM (1995) Musculoskeletal neoplasms: preoperative evalu-
Palumbo RC, Matthews LS, Reuben JM (1994) Localized pig- ation with MR angiography. Radiology 194:519-524
mented villonodular synovitis of the patellar fat pad: a Tyrrell PNM, Davies AM (1994) MR imaging appearances of
report of two cases. Arthroscopy 10:400-403 fatigue fractures of the long bones of the lower limb. Br J
Panieek DM, Gatsonis CG, Rosenthai DI, Seeger LL, Huvos AG, RadioI67:332-338
Moore SG, Caudry DJ, Palmer WE, McNeii BJ (1997a) CT Tyrrell PNM, Cassar-Pullicino VN, McCall IW (2000) Intra-
and MR imaging in the local staging of primary malignant artieular ganglion cysts of the cruciate ligaments. Eur
musculoskeletal neoplasms: report of the radiology diag- RadiollO: 1233-1238
nostie oncology group. Radiology 202:237-246 Unni KK (1996) Dahlin's bone tumors: general aspects and
Panieek DM, Hilton S, Schwartz LH (1997b) Assessment of data on 11,087 cases, 5th edn. Lippincott-Raven, Philadel-
neurovascular involvement by malignant musculoskeletal phia
tumors. Sarcoma 1:61-63 van der Woude HJ, Bloem JL, Oostayen JA, Nooy MA, Tamin-
Resniek D, Oliphant M (1975) Hemophilia-like arthropathy of iau AH, Hermans J, Rejnierse M (1995) Treatment of
the knee associated with cutaneous and synovial heman- high-grade bone sarcomas with neoadjuvant chemother-
giomas. Radiology 114:323-326 apy: the utility of sequential color Doppler sonography
Riehardson ML, Zink-Brody GC, Patten RM, Koh WJ, Conrad in predieting histopathologie response. Am J Roentgenol
EU (1996) MR characterization of post-irradiation soft 165:125-133
tissue edema. Skeletal RadioI25:537-543 van der Woude HJ, Bloem JL, Hogendoorn PCW (1998) Preop-
Rosenberg ZS, Lev S, Schmahmann S, Steiner GC, Beltran J, erative evaluation and monitoring chemotherapyin patients
Present D (1995) Osteosarcoma: subtle, rare and misleading with high-grade osteogenie and Ewing's sarcoma, review of
plain film features. Am J RoentgenoI165:1209-1214 current imaging modalities. Skeletal RadioI27:57-71
Schima W, Amann G, Stiglbauer R, Windhager R, Kramer J, Vanel D, Shapeero LG, de Baere T, Gilles R, Tardivon A, Genin
Nieolakis M, Farres M, Imhof H (1994) Preoperative staging J, Guinebretiere JM (1994) MR imaging in the follow-up of
of osteosarcoma: efficacy of MR imaging in detecting joint malignant and aggressive soft tissue tumors: results of 511
involvement. Am J Roentgenol163: 1171-1175 examinations. Radiology 190:263-268
Seeger L, Widoff BE, Bassett LW, Rosen G, Eckardt JJ (1991) Varma DGK (1999) Optimal radiologie imaging of soft tissue
Preoperative evaluation of osteosarcoma: value of gadolin- sarcomas. Semin Surg OncoI17:2-1O
ium dimeglumine-enhanced MR imaging. Am J Roentgenol Velchik MG, Heyman S, Makler PT Jr, Goldstein HA, Alavi A
157:347-351 (1984) Bone scintigraphy: differentiating benign cortieal
Seeger LL, Yao L, Eckardt JJ (1998) Surface lesions of bone. irregularity of the distal femur from malignancy. J Nucl
Radiology 206:17-33 Med 25:72-74
Shapeero LG,Vanel D (2000) Imaging evaluation of the response Verstraete KL, De Beene Y, Roels H, Dieriek A, Uyttendale D,
ofhigh-grade osteosarcoma and Ewing sarcoma to chemo- Kunnen M (1994) Benign and malignant musculoskeletal
therapy with emphasis on dynamie contrast-enhanced MR lesions: dynamie contrast-enhanced MR imaging. Radiol-
imaging. Semin Musculoskeletal RadioI4:137-146 ogy 192:835-843
Shuman WP, Patten RM, Baron RL, LiddelI RM, Conrad EU, Voutsinas S, Wynne-Davies R (1983) The frequency of malig-
Riehardson ML (1991) Comparison of STIR and spin-echo nant change in diaphyse al aclasis and neurofibromatosis. J
MR imaging at 1.5 T in 45 suspected extremity tumors; Med Genet 20:345-349
lesion conspieuity and extent. Radiology 179:247-252 Willis RB, Rozencwaig R (1996) Pediatric osteomyelitis mas-
Solomon L (1974) Stress fractures of the femur and tibia querading as skeletal neoplasia. Orthop Clin North Am
simulating malignant bone tumours. South Afriean J Surg 27:624-634
12:19-25 Yamazaki T, Maruoka S, Takahashi S, Saito H, Takase K, Naka-
Somer K, Meurman KOA (1982) Computed tomography of mura M, Sakamato K (1995) MR findings of avulsive cortieal
stress fractures. J Comput Assist Tomogr 6: 109-115 irregularity of the distal femur. Skeletal RadioI24:43-46
Steinbach L (1998) CT and MR imaging in the local staging of Yousem D, Magid D, Fishman EK, Kuhajda F, Siegelmann SS
primary malignant musculoskeletal neoplasms: comment. (1986) Computed tomography of stress fractures. J Comput
Sarcoma 2:57-58 Assist Tomogr 10:92-95
Subject Index
illiotibial band 71
Insall-Salvati index 16 N
neuroblastoma metastasis 309,311
neurofibromatosis 81,85
L neuropathie arthropathy 272,276,279
lateral collateralligament 71,121,160-162 non -accidental injury 97-106
lead poisoning 9 - metaphyseal fractures 98,99,103
lipohaemarthrosis 5,9,68 non-ossifying fibroma see fibrous cortieal defect
lipoma 314,326
liposarcoma 314,327
lipoma arborescens 35,68,277,330 o
loose bodies 49,50,66,74,273 Ollier's disease 308,320,324
Osgood-Schlatter disease 8,9,70,119,121,218
osteoarthritis 3,8,10-14,35,83,188,189,269-274
M - grading 12,236
Maffucci's syndrome 308 osteoblastoma 314,320
magnetie resonance 41-63 osteochondroma 65,321,324,325
- chemieal shift artefact 58 osteochondritis dissecans 8,9,293-302
- colls 42,129 osteochondromatosis 8,35,36
- contrast to noise 41,44 osteogenesis imperfecta 80,81,86,87,104,105
- diffusion weighted imaging 55,315 osteoid osteoma 32,38,39,278,310,314,315,320,321
- echo plan ar 47 osteoma 320
- fat suppression 47,48,51,54,56,57,130,156,174,235,254, osteomyelitis 8,32,79,106,250-265,310,311,313,315,333
260,285,287,316 - sequestrum 32,250,257
- field of view 45, 56 osteonecrosis 56,301-304
- field strength 41,42,48,133,139 osteopoikilosis 320
- gadolinium 48,49,54,56,57,130,132,138,175,187,188, osteosarcoma 3,278,307,308,310-312,315-319,322-325,
235,254,287 332
- - dynamie enhancement 192,274,275,315,318,319,324 - extraskeletal 326
- gradient echo 46,47,51,53,54,58-60,129,130 - parosteal 311,324
- magie angle phenomenon 54,58,59,131,132 - skip metastases 316-318
- matrix 42-46,56,57 osteotomy 212-214
- MR arthrography 19,47-50,54,56,57,132,148,175,184,
188,287,300
- magnetisation transfer 48,55, 130,285 P
- metal artefact 178, 193 PACS (pieture archiving communieation system) 4
- signal to noise 41-45,48,51,55,129 Paget's disease 29,308
- slice thiekness 43,56,57 - sarcoma 324,325
Subjeet Index 339
Q
T
quadrieeps
thalassaemia 86, 313
- muscle 67
tibial bowinglpseudarthrosis 81
- tendon 57,66,67,129,217,218
transverse ligament 135, 136
- Questor Precision Radiographie System (QPR) 14
tubereulosis 261,262
ultrasound 65-74, 254-256
R - Doppler 66,270,273,319
radiography
- eomputed 3,4
- eonventional analogue 3,4,15 U
- digital 3,4 unieameral bone eyst see simple bone eyst
List of Contributors
J. BELTRAN, MD A. A. DE SMET, MD
Chairman and Clinical Professor of Radiology Professor, Department of Radiology - E3/311
Department of Radiology University ofWisconsin Hospital and Clinics
Maimonides Medical Center 600 Highland Avenue
4802 Tenth Avenue Madison, WI 53792
Brooklyn, NY 11219 USA
USA
N.EGUND,MD
Professor, Department of Radiology
T. H. BERQUIST, MD, FACR
Aarhus University Hospital
Diagnostic Radiology
Noerrebrogade 44
Mayo Clinic
8000 Aarhus C
4500 San Pablo Road
Denmark
Jacksonville, FL 32224
USA
J. GARCIA, MD
Professor, Division of Radiodiagnosis
S. BIANCHI, MD
and Interventional Radiology
Division of Radiodiagnosis and Interventional Radiology
Hopital Cantonal Universitaire
Hopital Cantonal Universitaire
24 rue Micheli-du-Crest
24 rue Micheli-du-Crest
1211 Geneva 14
1211 Geneva 14
Switzerland
Switzerland
H. K. GENANT, MD
T. BOERGARD, MD, PhD Professor, Osteoporosis and Arthritis Research Group
Department of Radiology Department of Radiology
University Hospital University of California San Francisco
22185 Lund 350 Parnassus Avenue, Suite 150
Sweden San Francisco, CA 94143-1349
USA
V. N. CASSAR-PULLICINO, MD
Consultant Radiologist A. GUERMAZI, MD
Department of Diagnostic Imaging Osteoporosis and Arthritis Research Group
The Robert Jones and Agnes Hunt Department of Radiology
Orthopaedic and District Hospital University of California San Francisco
Oswestry 350 Parnassus Avenue, Suite 150
Shropshire, SYI0 7AG San Francisco, CA 94143-1349
UK USA
342 List of Contributors
C.HERON,MD L. RYD,MD
Consultant Radiologist Associate Professor, Department of Orthopedics
St. George's Hospital University Hospital of Lund
Blackshaw Road 22185 Lund
London SW17 OQT Sweden
UK
B. M. SAMMAK, MD, FRCR
A.HINE,MD Consultant Radiologist
Consultant Radiologist Department of Radiology
Central Middlesex Hospital Riyadh Armed Forces Hospital
Acton Lane PO Box 7897
London, NWlO 5NS Riyadh 11159
UK Kingdom of Saudi Arabia
C.P. Ho, MD, PhD
S. SHANKMAN, MD
National Orthopaedic Imaging Associates
Vice Chairman and Pro gram Director
Sand HilI Imaging Centre
Department of Radiology
2882 Sand HilI Road, Suite 118
Maimonides Medical Center
Menlo Park, CA 94025
4802 Tenth Avenue
USA
Brooklyn, NY 11219
USA
K. JOHNSON, MD
Consultant Paediatric Radiologist
P. N. M. TYRRELL, MD
Princess of Wales Birmingham Children's Hospital
Consultant Radiologist
Steelhouse Lane
Department of Diagnostic Imaging
Birmingham
The Robert Jones and Agnes Hunt
West Midlands, B4 6NH
Orthopaedic and District Hospital
UK
Oswestry
K. JONSSON, MD, PhD Shropshire, SYI0 7AG
Professor, Department of Radiology UK
University Hospital
D.VANEL,MD
22185 Lund
Department of Radiology
Sweden
Institut Gustav Roussy
J.A. LYNCH, PhD 39 rue Camille Desmoulins
Osteoporosis and Arthritis Research Group Villejuif 94895
Department of Radiology France
University of California San Francisco
C. WAKELEY, MD
350 Parnassus Avenue, Suite 150
Department of Clinical Radiology
San Francisco, CA 94143-1349
Bristol Royal Infirmary
USA
Bristol, BS2 8HW
C. MARTINOLI, MD UK
Istituto di Radiologia
I.WATT,MND
Universita di Genova
Department of Clinical Radiology
Largo Rosanna Benzi 1
Bristol Royal Infirmary
16100 Genoa
Bristol, BS2 8HW
Italy
UK
c.G. PETERFY, MD, PhD
Osteoporosis and Arthritis Research Group R.W. WHITEHOUSE, MD
Department of Radiology Department of Clinical Radiology
University of California San Francisco Manchester Royal Infirmary
350 Parnassus Avenue, Suite 150 Oxford Road
San Francisco, CA 94143-1349 Manchester, MB 9WL
USA UK