PERPUSTAKAAN

PRIBADI
AN-NUR
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

M. E. Abd El Bagi· M. S. Al Shahed . J. Beltran . T. H. Berquist . S. Bianchi . T. Boegard

V. N. Cassar-Pullidno . S. Chapman . A. M. Davies . A. A. De Smet . N. Egund . J. F Garda
H. K. Genant . A. Guermazi . C. Heron . A. Hine· C. P. Ho . K. Johnson . K. Jonsson
J. A. Lynch . C. Martinoli . C. G. Peterfy· S. N. J. Roberts . L. Ryd . B. M. Sammak
S. Shankman· P. N. M. Tyrrell· D. Vanel· C. Wakeley· I. Watt· R. W. Whitehouse· S. Zăim

Foreword by
A.L.Baert

With 345 Figures in 624 Separate Illustrations, Some in Color

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

MEDICAL RADIOLOGY . Diagnostic Imaging and Radiation Oncology

Series Editors: A. 1. Baert . L. W. Brady . H.-P. Heilmann . F. MolIs ' K. Sartor
Continuation of
Handbuch der medizinischen Radiologie
Encydopedia of Medical Radiology

ISBN 978-3-540-00250-5
Library of Congress Cataloging-in-Publicalion Dala

Jmaging of the knee: Itdtniques and applications / A. M. Davin, V. N. Cassar·PuIlidno

(eds.) ; with oonlributions by L Beltran; foreword by A. L Baert.
p.; em. - (Medical radiology)
Includes bibliographical rtferences and index.
ISBN 978-]-540-0025()"5 ISBN 978-]-642-55912-9 (eBook)
DOI 10.1007/978-3-642-55912-9
1. Knee __ Jmaging. 2. Diagnostic imaging. 3. Knee-Disea~s--Diagnosis. 1. Davies. A
M. (Arthur Mark), 1954- TI. Gassar-Pullicino, V. N. (Victor N.), 1954- ITI. Behran,
Javier. IV. Se rie •.
[DNLM: 1. Knee-·anatomy & histology. 2. Diagnostic Imaging-mdhods. 3. Kn ee
Injurin--diagnosis. WN 870 131 2002J
RC95].J4(i 2002
617.5'820754 --dcl l 2001049635
This work is subiect 10 copyright AU righu are reserved, whdhu the whole or part of the material is concerned, sp«ifi-
cally the rights of translation, reprinling, reu~ of ilIustralions, recitalions, broadtuting, rtproduclion on microfilm or
in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the
provisions of the German Copyright Law nf SeplemNr 9, 1965, in its current ~rsion,and permission for u~ must always
be obtained from Springer-Verlag. Violalions are liable fo r proseculion under the German Copyright Law.

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

In addition to radiography, sonography, computed tomography and magnetie resonance

are now used routinely to depiet the complex normal anatomy and the pathologie al find-
ings of one of the most important artieulations of the human body: the knee. Impor-
tant technieal progress has been achieved in recent years in each of these cross-sectional
imaging modalities. It is important for all professionals involved in the management of
patients with a knee condition to be weH informed about the specific advantages as weH
as the limitations of the various imaging techniques, including arthrography, that are
now available to study the knee.
This book sets out to provide a sorely needed update of our knowledge in this
radiologie field and forms a welcome addition to our series "Medieal Radiology",
whieh aims to cover aH important aspects of modern diagnostie radiology. It will be of
great interest for general and musculoskeletal radiologists and orthopedie surgeons. It
comprehensively covers both common and less common congenital and developmental
anomalies, non-accidental injury and trauma as weH as infectious, infiammatory and
degenerative conditions of the knee. As operative procedures of the knee are becoming
more and more sophistieated, ample attention has been given to the clinieal problems
and the radiologie depietion of the postoperative knee.
The editors, Dr. Davies and Dr. Cassar-Pullicino, both from the UK, are inter-
nationaHy renowned experts who have, over the years, accumulated unique experience
in the radiologie al management of the knee. They have been very successful in recruiting
several other leading authorities from Europe and the USA to deal with specific prob-
lems of the knee.
I would like to congratulate the editors and all contributors most sincerely for their
outstanding work on this volume: the content is comprehensive, the illustrations superb.
As always I would be pleased to receive any constructive criticism or comments that
readers may care to express.

Leuven ALBERT L. BAERT

Preface

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

3 Computed Tomography (CT) and CT Arthrography

RICHARD WILLIAM WHITEHOUSE ........................................... 23

4 Magnetic Resonance Imaging

CHRISTINE HERoN and ANDREW HINE ...................................... 41

5 Ultrasound
STEFANO BIANCHI, CARLO MARTINOLl, and JEAN GARCIA . . . . . . . . . . . . . . . . . . . . .. 65

Clinical Applications ........................................................ 75

6 Congenital and Developmental Abnormalities of the Knee

KARL JOHNSON and A. MARK DAVIES ....................................... 77

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

10 The Cruciate and Collateral Ligaments

CHARLES P. Ho ........................................................... 153

11 The Postoperative Knee 1: Menisci and Cruciate Ligaments, Cartilage

VICTOR N. CASSAR-PULLICINO and S. N. ROBERTS ............................. 165

12 The Postoperative Knee 2: Arthroplasty, Arthrodesis and Osteotomy

THOMAS H. BERQUIST ..................................................... 199

13 Patellar and Quadriceps Mechanism

NIELS EGUND and LEIF RYD ................................................ 217
x Contents

14 Infection
MOHAMED E. ABD EL BAGI, MONA S. AL SHAHED, and BASSAM M. SAMMAK ....... 249

15 Arthritis
CHARLES WAKELEY and IAIN WATT .......................................... 269

16 Assessment ofKnee Cartilage

SOUHIL ZAIM, ALl GUERMAZI, ]OHN LYNCH, CHARLES PETERFY,
and HARRY K. GENANT .................................................... 283

17 Osteochondritis and Osteonecrosis

ARTHUR A. DE SMET ...................................................... 293

18 Tumours and Tumour-like Lesions

A. MARK DAVIES and DANIEL VANEL ........................................ 307

Subject Index ............................................................... 337

List of Contributors .......................................................... 341

Imaging Techniques
1 Radiography
KJELL JONSSON and TORSTEN BOEGARD

CONTENTS nostie work-up, and in the vast majority of cases with

knee problems, radiography is sufficient for diagno-
1.1 Introduction 3
1.2 Radiographie Techniques 3 sis and as a guide to management.
1.3 Radiographie Projections 6
1.3.1 Standard Series 7
1.3.2 Trauma Series 9
1.3.3 Degenerative Disease 10
1.4 Radiographie Measurement Techniques 13
1.2
1.4.1 Assessment of Alignment 13 Radiographie Teehniques
1.4.2 QPR or SKI 14
1.4.3 Roentgen Stereophotogrammetrie Analysis 14 Three different techniques are used to obtain a radio-
1.4.4 Femoral Condyle Configuration 15 graph: the conventional analogue technique, comput-
1.4.5 Patellar Position 16
References 16
ed radiography and digital radiography.
A cassette with film and intensifying screens is
used for the analogue technique. This technique is
well known and used all over the world.
A cassette with an image storage phosphor plate
1.1 (image plate, IP), is used for computed radiography.
Introduction This cassette is of the same size as a conventional cas-
sette with film and is used in the same way during the
The knee joint is vulnerable in many respects. It is examination. Energy proportional to the amount of
prone to trauma, both in osteoporotie patients radiation is stored in the IP when electrons excited by
and in the younger population who partake in the X-ray photons become trapped at higher energy
modern life-style sports activities. The cartilage may levels in the phosphor crystals. The stored image is
develop degenerative changes (osteoarthritis, OA) read out in an IP reader where the IP is scanned
with increasing age, but in the younger population with a laser beam, whieh allows the trapped elec-
degenerative change is often the sequel of local trons to return to a lower energy level. The light
trauma to the cartilage or secondary to instability emitted during this transition is converted into a dig-
after cruciate or collateralligament damage. In this ital image, which is then processed according to the
group of patients, premature OA often occurs many examination selected. The image processing is pro-
years prior to the natural onset. The knee joint is grammable for each examination and must be done
often affected in arthritis, both septie and non-septie, with care in order to avoid image artefacts and to
and osteochondroses. Rare osteosarcomas in preteen optimise visualisation of the pathology. Image arte-
children and teenagers are most common around the facts can appear as a radiolucent zone at high-con-
knee. trast steps, for instance at the interface between
In all these conditions, radiographie examination metal, cement and bone in a joint prosthesis. The
of the knee joint should be the first step in the diag- image processing is automatieally adjusted to give
the desired image density regardless of the radiation
dose within a large range.
K. JONSSON, MD, PhD The advantage of computed radiography is the
Professor, Department of Radiology, University Hospital,
broad exposure range and the free choiee of data pro-
22185 Lund, Sweden
T. BOEGARD, MD, PhD cessing. The dynamie range is more than 10,000:1,
Department of Radiology, University Hospital, 22185 Lund, which is more than 100 times that of analogue film-
Sweden screen systems. This reduces the need for retakes,
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
 4 K. Jonsson and T. Boegard.
reducing the dose burden to the patient due to faulty
exposure factors. The radiation dose to the patient can
be decreased without affecting the diagnostic accura-
cy, especially with measurements of lines and angles
(J6NSSON et al. 1996; SANFRIDSSON et al. 1998). With
the analogue technique it is possible to overexpose a
film, rendering the pathology less obvious. The broad
exposure range with CR may, however, be a disad-
vantage. A heavily overexposed picture does not show
on the image obtained, and appears as a normally
exposed film. This is a threat, because the patient may
receive a heavy radiation dose, which is not easily
detected. This can be checked by looking at the sensi-
tivity value (S-value) that is available for every image.
The S-value can be used for selection of the correct
exposure settings on the X-ray generator. A high S-val-
ue corresponds to a low dose and a low S-value indi-
cates a high exposure. The S-value is related to the
speed of a screen-film system and it indicates which
speed a screen-film system should have to produce the
correct density on the film. Images with a low S-value
have low noise while those with a high S-value are
grainier due to high noise.
The image is either transferred to a laser image
printer for film output or to a PACS system (picture
archiving and communication system), where the
image is reviewed on a workstation and stored in a
digital archive. It is possible to perform a complete
reprocessing of the images if the IP reader transfers
the image raw data to a dedicated workstation. The
reprocessing capabilities can also be incorporated in
the IP reader in some models. A PACS workstation
a b
allows some image processing (windowing, zoom,
image rotation and so on) depending on the software
in the workstation. The free choice of data processing
means that from a single exposure, it is possible to
obtain the image that is best suited for excluding and
displaying the pathology looked for in each individu-
al case. In knee examinations this is advantageous for
the diagnosis of suprapatellar pouch distension with
blood (Fig.l.l) and ßake fragments (Fig.1.2) and for
the disclosure of small intra-articular calcifications.
Image output on a film makes it possible to show
two images on one film, where the images are pro-
cessed independently of each other. One image is
normally processed to give the impression of a con-
ventional image, and the other image is usually pro-
cessed with a higher degree of edge enhancement to
increase the contrast for the desired spatial frequen-
cy. Computed radiography has a lower spatial resolu-
tion than conventional radiography, but this has not
proven to be a limiting factor with modern systems
(J6NSSON et al. 1995; SCOTT et al. 1993).
A ßat panel detector is used for digital radiogra-
phy. The size of the detector is approximately that of
a large cassette, up to 43x43 cm. The detector con-
verts the X-ray image into a digital signal, which is
transferred to a computer for image processing, dis-
play and transfer to a PACS system or a laser printer.
The detector is usually mounted on a stand and some
stands allow the detector to be moved for different
X-ray projections. The conversion of X-ray photons
into an electrical signal can be done with two differ-
ent processes, either an indirect or a direct process.
Fig.1.la, b. The supra-
patellar pouch. aNormal;
CR technique with edge
enhancement. b Disten-
sion with blood. Conven-
tional film-screen combi-
nation
Radiography 5

Fig. 1.2. a Horizontal lateral view of a knee with a

single fluid-fluid level (arrow). No obvious fracture
is seen. b Oblique lateral view of the same knee. A
flake fragment is seen (arrowhead). The origin of
this fragment is not clear. c Postoperative examina-
tion of the same knee. The flake fragment came
from the articular surface of the patella

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
examined part of the body remains stationary. The
motion of the X-ray tube and the film may be either
unidirectional (lineartomography) or pluridirection-
al (circular, elliptieal, spiral or hypocycloidal tomog-
raphy). The more complex the movement, the better
the quality of the image with less longitudinal streak-
ing, which may be seen in unidirectional tomogra-
phy. The disadvantage with tomography is the long
examination time and the high radiation exposure to
the patient. Computed tomography (CT) has super-
seded tomography with multiplanar reconstruction.
If CT is not available, conventional tomography can
be used to demonstrate the degree of depression of
a tibial plateau fracture (Fig. 1.3), to assess prema-
ture growth plate fusion, to evaluate healing of bony
pseudarthrosis and to identify sequestra in chronie
osteomyelitis.
choiee depends on the available equipment. The min-
imum radiographie projections are AP (anteropos-
terior) and lateral views. The knee joint should be
straight in the AP view and slightly flexed in the lat-
eral projection. These radiographs are taken with the
patient in the supine position.
The tunnel view ("einblick") is taken with knee
flexion of 40-50° and with the X-ray beam angulated
to the same degree to visualise the intercondylar
notch and tibial spine; ideally the beam will be
tangential to the tibial plateau (Fig. 1.4). Other meth-
ods employed to obtain the tunnel view have been
described (BALLINGER 1991; BONTRAGER 1997).
In the evaluation of the joint space, the radio-
graphs should be taken with the patient weight bear-

1.3
Radiographie Projections

The technique and radiographie projections in the

examination of the knee depend on the clinieal indi-
cations. A large number of textbooks on radio graph -
ie positioning have been published over the years,
forinstance BALLINGER (1991), BERNAU (1995), BON-
Fig.l.4. Position of the patient for tunnel view of the knee.
TRAGER (1997) and PAVLOV et al. (1999).
The knee is flexed 40-50° with the cassette under the flexed
It is important to use procedures that are easily knee. The X-ray beam is angulated to the same degree from
reproducible. Different techniques are possible and below to obtain a view of the intercondylar notch

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
ing and with the knee in slight flexion. At the Univer-
sity Hospital, Lund, the weight load is placed almost
entirely on the leg being examined. The lateral view
is obtained after fluoroscopic monitoring, so that the
dorsal aspects of the femoral condyles are superim-
posed (Figs. 1.5, 1.12b). Using fluoroscopic controt, an
AP view is obtained with the X-ray beam centred over
the joint space and tangential to the tibia plateau,
with the tibial spines centred in the femoral notch.
BOEGARD et al. (1997) described a similar technique
with the PA (postero-anterior) projection, where the
patella and the big toe touch the upright examination
table. To obtain reproducible projections, the medial
border of the foot is parallel with the X-ray beam.
Radiographs obtained with weight bearing on both
legs have, however, proved to be advantageous and
are increasingly used.
a ~-----
b
Fig.1.5. a Position of the patient for lateral weight-bearing
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
examined leg; there is knee flexion of 20-30°. The projection
is checked with fluoroscopy
7
Other protocols for reproducible examination of
the weight-bearing knee have been reported.
BUCKLAND-WRIGHT (1995) designed a protocol
where the patient flexes the knee until the tibial pla-
teau is horizontal and parallel to the central X-ray
beam. This requires aknee flexion of 11-20°,depend-
ing on the inclination of the tibial plateau. The posi-
tion is checked with fluoroscopy. The foot is rotated
internally or externally until the tibial spines appear
centrally placed relative to the femoral notch. The
position of the foot is drawn and recorded on a piece
of paper placed in a defined position on the floor.
Reproducible views at follow-up examinations are
then possible using the same position of the foot.
In another protocol for flexed PA knee examina-
tion, the patient is positioned with the patella and the
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
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
The minimum projections are AP, lateral and tunnel
views. The last-mentioned projection is often very
helpful when looking for loose bodies in the joint.
Only radio-opaque loose bodies are seen, while carti-
lage loose bodies are not visualised. The tunnel view
8 K. Jonsson and T. Boegärd.
Fig.1.6. Position of the patient for axial weight-bearing view
of the patella
a b
m----------_/I ~D--------- /
~~~
c d
Il'
e f
Fig. 1.7a-f. Different radiographie techniques used to obtain
axial non-weight-bearing views of the patella. [From MER-
CHANT et al. 1994. Reproduced with permission from the Jour-
nal 0/ Bone and Joint Surgery (Ameriean edition)]
is of value to delineate osteochondritis dissecans,
which in most cases is located on the inner aspect of
the medial femoral condyle (Fig.1.8).
In standard series it is important not to over-
or underexposed analogue films, and to check the
computed radiography picture on the monitor with
Fig. 1.8. Tunnel view of the knee. Osteochondritis dissecans
lesion of the inner aspect of the medial femoral condyle
(arrow)
different settings for window width and level. If
this is not done, subtle changes such as small cal-
cified loose bodies in osteochondromatosis may be
missed. In tumours and osteomyelitis the first find-
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-
phytes are present on the femoral or tibial condyles,
there are always local degenerative changes of the
cartilage, seen with magnetic resonance imaging
(MRI) (BOEGARD et al. 1998a) or arthroscopy. How-
ever, in some instances of degenerative changes
there are no accompanying osteophytes. Osteo-
phytes at the tibial spine and at the intercondylar
fossa are unreliable signs of knee OA (BOEGARD et
al. 1998a).
In children with pain over the tibial tub ero sity,
Osgood-Schlatter disease is clinically suspected. A
coned down view with soft tissue exposure should
be obtained to reveal bone fragmentation (Fig. 1.9).
Fragments mayaiso be seen with a skyline view of the
patella, where the fragments overlie the patellofemo-
ral joint. A similar type of osteochondrosis is Sind-
 Radiography 9

a b

Fig. 1.9. a Lateral view of the knee in a

12-year-old girl with pain over the tibial
tuberosity. b Magnification and opti-
mal setting of window width and level
to reveal minor fragmentation of the
tuberosity in Osgood-Schlatter disease

ing-Larsen disease of patella, with fragmentation of 1.3.2

the distal pole of the patella at the insertion of the Trauma Series
patellar tendon.
A wide spectrum of normal variants of the knee The minimum radiographie projections are AP and
can be seen on radiography and have been described lateral views in the supine position. The knee should
in standard texts like KEATS (1996) and BROSSMANN be as straight as possible. Option al views for further
et al. (2001). Benign cortical defects are common information may be obtained, for instance medial
around the knee joint in children, in both the distal oblique to show the head of the fibula and the prox-
femur and the proximal tibia. When these defects imal tibio-fibular joint. Some trauma patients may
heal they fill out with sclerosis, whieh will eventually have difficulty in lying on the injured side for a lateral
resolve. Irregularities of the distal femoral epiphysis view, and the lateral view is taken with a horizontal
are common with spieula-like ossification of the car- X-ray beam (Fig. 1.2).
tilage.A not uncommon normal variation of ossifica- A fracture is not always shown on the views, but the
tion is a defect of the distal surface of the femur that suprapatellar pouch is often distended due to intra-
looks like osteochondritis dissecans. These defects artieular bleeding. This can be seen both with the con-
are seen prior to closure of the growth plate and will ventionallateral view and with the horizontal X-ray
eventually fitl out with normal ossification. A round- beam technique (Fig. 1.1). A fluid-fluid level is only
ed defect of the dorsal, artieular surface of the patel- seen on the latter view (Fig.1.2). This is commonly
la may look like osteochondritis dissecans, but is a described as "lipohaemarthrosis" because in cases of
normal variation of ossification. Sometimes accesso- intra-articular fracture, marrow fat leaks into the joint.
ry ossification centres occur around the knee, giving LUGO-OUVIERI et al. (l996), however, found that a
rise to accessory bones, as in bipartite patella or an single fluid-fluid level is rarely due to fat in the joint,
accessory ossicle at the superior end of the proximal but just represents separation of cellular elements
fibula. from supernatant serum. Double fluid-fluid levels are
Relatively thin radiodense bands are sometimes a more specific finding, indieating fat in the joint. As
seen in the proximal metaphysis of the tibia. These intra-articular fractures may not always be apparent
bands are often called arrest lines and are consid- on standard films, extra projections are often neces-
ered to be sequelae of previous fractures or attacks sary. These fractures often represent subtle osteochon-
of arthritis or infection. These benign bands must dral injury where the bony part is small in comparison
be differentiated from sclerotie bands that are seen to the cartilage flake. These osteochondral fragments
after heavy metal poisoning, such as lead poisoning. may be easily missed on overexposed films. It is often
In lead poisoning the sclerotic bands are thiek and difficult to exactly localise the origin of such a flake
occur in the metaphyses of the femur, tibia and fibula fragment on standard radiographs (Fig.1.2), which
adjacent to the growth plate. can migrate away from the site of injury.
10
Most fractures of the knee are apparent on the stan-
dard trauma series. Sometimes a finding may be subtle
or equivocal, particularly in the case of tibial condylar
fractures. There may be a subtle and localised depres-
sion of the joint surface, but the degree and extension
are not clear on plain films. In such cases the stan-
dard examination should be completed with conven-
tional tomography or CT with reconstruction in mul-
tiple planes for evaluation (Fig.1.3).
Stress views of the knee may be valuable for indi-
reet demonstration of ligamentous disruption. Stress
may be exerted manually during fluoroscopy or by
using a special device to produce standardised stress
(STEDTFELD and STROBEL 1983). If stress radio-
graphs are done in the acute stage after trauma, the
test should be done under local or general anaesthe-
sia to avoid muscle spasm. Stress radiographs seem
to be of limited value and should, according to STRO-
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
to obtain this view in a trauma patient (Fig. l.7).After
patellar dislocation, a small fragment is often avulsed
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.
Fig. 1.1 O. Axial view of the patella after patellar disloeation. A
small fragment is avulsed from the medial aspeet of the patella
(arrow). In this ease a fragment has also been avulsed from
the lateral femoral troehlea (arrowhead)
1.3.3
Degenerative Disease
Osteoarthritis is a disorder that affects both the hya-
line cartilage and the subchondral bone. With plain
radiography,OA of the tibiofemoral joint may be diag-
K. Jonsson and T. Boegard.
nosed by the presence of marginal osteophytes, while
the degree of the degeneration is indicated by the
severity of joint space narrowing. AHLBÄCK (l968) and
LEACH et al. (l970) found that weight-bearing exam-
ination was superior to that obtained in the supine
position for demonstration of joint space narrowing
in tibiofemoral OA. With general degeneration of the
joint cartilage, joint space narrowing is easily found
with examination under weight bearing, but with local
chondrallesions this may be difficult. It is important
to obtain tangential views of the joint cartilage in
the optimal position (Fig. 1.l1a). A number of factors
influence the possibility of achieving reproducible and
optimal views of the joint space. One factor is that the
slope of the tibia plateau differs between the medial
and lateral condyles. The shape of the tibial condyles
differs in that the medial condyle is slightly concave
and a tangential view shows the superimposition of
~~~-'-------ID
a d
I____. -~r-=----­
t-_ \
b
L_((
r----~
~
1) .._
~
c
Fig. l.lla-e. Line drawing of examination teehnique in OA.
a Normal joint. The eentral beam from the X-ray tube (R) is
tangential to the subehondral bone. The width of the eartilage
(L) is eorreetly outlined on the film (F). b Normal joint. The
X-ray beam is not tangential and the eartilage width is falsely
depieted, being too narrow. c Cartilage defeet on both sides
of the joint. The X-ray beam is eorreetly tangential to the sub-
ehondral bone. Joint spaee narrowing is eorreetly visualised. d
Cartilage defeet on one side of the joint is rotated away from
a tangential position. Only the depth of the lower defeet is
depieted. e The upper eartilage defeet has rotated away from
the tangential position and the lower joint surfaee has slid pos-
teriorly. The joint spaee width appears normal on radiography.
This pieture ean illustrate the situation of the knee with loeal
eartilage degeneration posteriorly upon radiographie exam-
ination with a straight or overextended joint (see Fig. 1.12)
(Published with permission from Nordisk Lärobok I Radiologi,
Studentlitteratur, Sweden 1993)
Radiography 11

the anterior and posterior margins over the true joint

surface (Fig.1.l3). The true joint surface was defined
by BUCKLAND-WRIGHT (1995) as "the line between the
tibial spine and the medial or outer margin, across the
centre of the floor of the articular fossa in the midcor-
onal plane of the joint. This line is defined by the supe-
rior margin of the bright radiodense band of the sub-
chondral cortex and appears below the anterior and
posterior articular margins of the tibial plateau". The
joint space is measured between this line and the distal
convex margin of the femoral condyle in the medial
compartment. The lateral tibial plateau is slightly less
concave as compared to the medial tibial condyle,
and sometimes convex. BUCKLAND-WRIGHT (1995)
defined the measuring line of the lateral tibial con-
dyle as "the proximal margin of the articular surface,
defined bythe superior margin of the bright radiodense
band of the subchondral cortex extending from near a
the tibial spines to the lateral or outer margin". Thus
there is a difference between the medial and lateral
compartments in that it is easy to define the joint sur-
face of the lateral tibial plateau when the X-ray beam is
tangential to the surface. In the medial compartment
the anterior and posterior margins of the joint sur-
face superimpose the true joint surface (Fig.1.l3). The
slope of the tibial plateau is 5-20°. This means that a
horizontal X-ray beam will not run tangential to the
plateau and the inclination of the X-ray beam has in
most cases to be adjusted by fluoroscopy.
Previous studies have shown that degeneration
of the joint cartilage of the femoral condyles starts
at a position posterior to the apex of the condyle
(BOEGARD et al. 1997; BOEGARD 1998). Therefore a
patient with non-specific knee pain should be exam-
ined with weight bearing and 20-30° flexion of the
tibiofemoral joint, especially if he or she is over the
age of 30, in order to reveal joint space narrowing.
b
In patients who are not able to stand, a tunnel view
of the knee may be valuable. RESNICK and VINT
(1980) described six patients in whom the tunnel
view showed the most striking evidence of cartilage
damage when compared with the weight-bearing AP
view (Fig. 1.12).

Fig.1.12a-c. Weight-bearing knee. a AP view; b lateral view;

c tunnel view. The knee is examined with slight flexion. The
femoral condyles are superimposed in the lateral view. Osteo-
phytes are present laterally on the femur and tibia (arrows).
The joint space width appears normal medially and laterally.
With the tunnel view the joint space is markedly reduced lat-
erally because of degeneration of joint cartilage posteriorly
on both the femur and the tibia. The posterior horn of the
c
lateral meniscus has been resected. These findings were dem-
onstrated with MRI (not shown here)
12 K. Jonsson and T. Boegärd.

In some Swedish hospitals the standard technique

Fig.1.13. The principle ofHKA measurement in two examples.
Lines are drawn from the centre of the femoral head to the
middle of the knee (represented by the tibial spines) and from
the middle of the ankle to the middle of the joint, to evaluate
the degree of varus deformity
employed is with the patient standing, bearing almost
the entire weight on the examined leg, with slight
knee flexion. In most centres, however, the patient
stands on both legs, with equalload on both knees.
BOEGARD et al. (1998b) compared the two techniques,
i.e. standing on one or both legs. They found that
there was a reduction of the minimal joint space
width in the medial compartment when the patient
stood with the weight equally distributed on both
legs. With weight load on one leg, the joint space
reduction was most prominent in the lateral com-
partment. Because OA is 10 times more common in
the medial compartment than in the lateral compart-
ment, examination with equal weight on both legs is
preferred. One must be aware that the slope of the
tibial plateau differs between the medial and the lat-
eral side, and if an examination is aimed at the medial
compartment there may be difficulties in assessing
the lateral joint space.
Different grading systems for knee OA have been
proposed. Those most commonly employed are the
systems suggested by KELLGREN and LAWRENCE
(1957) and AHLBÄCK (1968). The two grading sys-
tems are compared in Table 1.1. AHLBÄCK (1968)
proposed that narrowing, as a sign of cartilage loss,
exists if the minimal joint space width is less than
3 mm, and most authors have accepted this limit. This
limit has been verified by BOEGARD et al. (1997).

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

Grade 1 Minute osteophytes,

"doubtful" doubtful significance
Grade 2 Definite osteophytes,
"minimal" unimpaired joint space
Grade 1 Joint space narrowing Grade 3 Moderate diminution
(Joint space <3 mm) "moderate" of joint space
Grade 2 Joint space obliteration Grade 4 Joint space greatly of
"severe" impaired with sclerosis
subchondral bone
Grade 3 Minor bone attrition Grade 4 As above
(0-5 mm)
Grade 4 Moderate bone attrition Grade 4 As above
(5-10 rnrn)

Grade 5 Severe bone attrition Grade 4 As above

(>lOmm)
Radiography 13

As mentioned previously, the presence of osteo-

phytes is an important indicator of OA in the
early stage of the disease. In the advanced stage,
osteophytes are often extensive. Several atlases of
radiographs separately define and re cord individu-
al radiographie features of knee OA, such as joint
space narrowing, osteophytes and subchondral scle-
rosis (ALTMAN et al. 1995; SPECTOR et al. 1992).
These atlases have been created for semiquantitative
assessment of the degree of disease in an attempt to
reduce intra-observer and inter-observer variance
in the evaluation of knee joints for research pur-
poses.
In patients with suspected degenerative disease,
an axial view of the patella should be included to
evaluate the patellofemoral joint. The examination
should be performed with weight bearing,knee flex-
ion and a vertical X-ray beam (Fig. 1.6). No com-
parative studies between weight bearing and non-
weight bearing have been published so far, but it is
our feeling that the only way to assess the function-
al relationships of the patella is with weight bear-
ing. In the measurement of the patellofemoral joint
space, BUCKLAND-WRIGHT (1995) defined the mea-
suring surfaces as "the articular surface of the cortex
at the medial and lateral surfaces" for the femur
and "the bright radiodense band of the subchondral
cortex at the medial and lateral articular surfaces
lying deep, or anterior, to the profile of the inferior
articular margin" for the patella. With the examina-
tion technique used, BOEGARD et al. (1998c) found
the critical joint space width to be 5 mm. If the joint
space is below 5 mm, there is a low sensitivity (50%)
and a high specificity (94%) for MR-detected carti-
lage defects. Osteophytes at the patellofemoral joint
are associated with MR-detected cartilage defects
in the same joint. The relationship was strong for
osteophytes at the lateral femoral trochlea and in
joints with narrowing «5 mm) but weak in joints
without narrowing (>5 mm) (BOEGARD et al. 1998d).
Extensive review articles on plain radiography in
OA are available (BOEGARD and ]ONSSON 1999;
MAZZUCA and BRANDT 1999). Fig.1.14. The examination table for QPR. (From SANFRIDSSON
et al. 1996, with permission from Acta Radiologica)
In patients undergoing implantation of a knee
prosthesis, the same technique is used as in primary
OA, i.e. with weight bearing and flexion of the knee.
In these patients it is important to have the central
X-ray beam parallel to the tibia plateau in the AP or
PA view in order to have the same projection at fol- thesis should be superimposed. The aim of pros-
low-up examinations (Fig. 1.14). If the X-ray beam thesis examination is to obtain standardised projec-
is not parallel to the tibial plateau, a thin zone of tions in order to diagnose possible complications,
lucency may be hidden. In the lateral view the poste- such as loosening, infection or wear of the plastic
rior elements of the femoral component of the pros- co at.
14
1.4
Radiographie Measurement Techniques
1.4.1
Assessment of Alignment
Accurate measurements of lower extremity align-
ment and configuration are aprerequisite for
research and the clinical handling of patients with
OA. Such measurements are mandatory in the pre-
operative planning ofhigh tibial osteotomy and knee
arthroplasty as weIl as for postoperative outcome
assessment. The mechanical axis of the lower extrem-
ity, defined as the hip-knee-ankle (HKA) angle dem-
onstrated on long films exposing the hip, knee and
ankle, has been used (ODENBRING et al. 1993). Ideal-
ly, measurement of the HKA angle is performed with
the patient bearing weight on one leg, just as with
examination for evaluation of the joint space. First,
the knee joint is monitored in the lateral view, so that
the femoral condyles are superimposed. Then the AP
view is obtained 90° to the lateral view. In the AP
viewa long radiograph including the knee and ankle
and a radio graph of the hip are obtained with the
film cassettes in a calibration frame. After the radio-
graphs have been developed, they are mounted on a
similar frame on a light box. In this way the radio-
graphs are centred correctly in relation to each other
and lines drawn on the films allow evaluation of the
HKA angle. The lines are drawn from the cent re of
the femoral head to the cent re of the knee and from
the middle of the ankle joint to the middle of the
knee (Fig. 1.l3). In this way a varus or valgus defor-
mity of the knee can be outlined and graded. Anoth-
er important angle is the HKS angle (hip-knee-
shaft), which is the angle between the line from the
centre of the femoral head to the middle of the knee
and the line from the middle of the femoral shaft at
the level of the lesser trochanter to the centre of the
knee. This angle is important for correct positioning
of surgical equipment at osteotomy or placement of
prosthesis.
In moderate or early knee OA, patients may be
operated on with high tibial osteotomy to correct the
alignment of the knee and redistribute the load to the
healthy compartment, i.e. to the lateral compartment
in medial arthrosis (INSALL et al. 1974). Traditional-
ly an osteotomy included rem oval of a bone wedge
to correct the alignment, but more recently callus
distraction or hemicallotasis osteotomy has become
popular in some centres (DE PABLOS et al. 1995). In
order to perform a correct osteotomy and estimate
the size of wedge or callus distraction, HKA measure-
K. Jonsson and T. Boegard.
ment is necessary both preoperatively and postoper-
atively to check the result.
Preoperative measurement in gonarthrosis can
also be evaluated in the supine non-weight-bearing
position, but with the application of stress views
(EDHoLM et al. 1976).
1.4.2
QPR or SKI
To improve the measuring routines of the knee, other
methods have been described, such as QPR (Questor
Precision Radiographie system), which is a stan-
dardised examination method for the lower limb in
the weight-bearing position. This method has now
been renamed "SKI" (standardised knee imaging). The
method was developed for conventional radiography
bythe Clinical Mechanics Group, Division of Orthopae-
die Surgery and Department of Mechanical Engineer-
ing at Queens University, Kingston, Ontario, Canada
(SIU et al. 1991). The patient stands on a turntable
inside a frame that has protractor scales for recording
foot rotation (Fig.1.14). The ankles are positioned
against adjustable blocks to represent the location of
each mid-malleolar point and the hips are supported
by adjustable hip pads. A lateral view of the knee is
obtained and then the turntable can be rotated 90°
without the patient changing position for an orthogo-
nal AP view of the knee and hip. The QPR frame
contains two Plexiglas panels spaced 10 cm apart with
embedded radio-opaque markers and reference lines
for correction of parallax and magnification. The
panels are located in front of the patient between the
X-ray source and the film. The bony landmarks, ref-
erence lines and markers are located and digitised on
a digitising tablet. A software program in a personal
computer processes the information and generates a
display of nine angles and ten distances. Our routine
method for measurement of the HKA and HKS angles
was compared with QPR, and the study showed a good
correlation between the two methods (SANFRIDSSON
et al. 1996). More recently, the QPR system has been
adapted to computed radiography and PACS (SAN-
FRIDSSON et al.1997), with good reproducibility. With
the CR technique it is possible to reduce the radiation
dose considerably in this environment (SANFRIDSSON
et al. 2000). The same group has also demonstrated
that with the QPR technique it is possible to show the
rotation of the tibia in relation to the femur and the
migration of the patella when changing the position of
the knee from slight flexion to the fully extended posi-
tion (SANFRIDSSON et al. 2001).
 Radiography 15

It has to be pointed out that this technique is not
generally available, because it requires special equip-
ment and computer pro grams for evaluation. The
method is used at some centres for routine evaluation
ofknee alignment (COOK et al. 1999).
1.4.3
Roentgen Stereophotogrammetric Analysis
Stereophotogrammetry (RSA) was described by
SELVIK (1989). The idea behind the method is that
mobility between two structures can be defined and
analysed by a stereophotogrammetrie technique.
The structures are defined by small metallic beads
of known size and density. The beads are placed
in two structures that may move in relationship to
each other, for instance the tibial component of a
prosthesis in relation to the proximal tibia. Abso-
lutely spherieal 0.8-mm beads made of tantalum are
implanted into the bone and/or the plastie coating
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
X-ray tubes in stereoscopie convergent-ray mode or
at 90° to each other. Before these radiographs are
taken, special calibration procedures must be per-
formed. The films are mounted in a special analy-
ser that, due to the stereoscopie radiographie tech-
nique, can locate the bodies in space and define the
position of each body in relation to the other by
using a special software program. At repeat exami-
nations, differences in position between the bodies
can be revealed.
The tantalum beads can be easily introduced per-
cutaneously under local anaesthesia. The knee is
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
conventional radiography is in the order of 2-3 mm,
but with RSA the accuracy is 0.2-0.3 mm. Also, rota-
tion can be estimated with the same degree of accu-
racy (RYD 1992; FR IDEN et al.1992).
There is no other method available that can com-
pete with RSA in terms of accuracy of measurements,
but the method requires considerable time and effort.
The markers must be implanted with great care, and
the radiographs must be taken with patients cor-
rectly positioned. Digitisation and analysis are time
consuming and demand thorough knowledge of the
technique. RSA is best suited for smaller well-defined
investigations that address kinematie problems. So
far it is not a method for general use.
1.4.4
Femoral Condyle Configuration
The configuration of the femoral condyles may influ-
ence knee stability. In a study by FRIDEN et al. (1993),
100 consecutive patients with anterior cruciate liga-
ment (ACL) rupture were studied prospectively for
5 years. During this time 16 patients developed dis-
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
Line A
Fig.1.l6. Line drawing to show the principle of evaluation
of the configuration of the femoral condyles. The measure-
ments are performed on a non-weight-bearing standard lat-
eral radiograph. Une A is drawn from the most distal point
of the angle of Blumenstaat's line (PI) to the centre of the
femoral shaft 10 cm above the joint line (P2). Perpendieular to
this axialline, two additionallines are drawn to the most pos-
terior point on each femoral condyle. The midpoint between
these lines on line A is called P3 and the mean length of the
lines is called x, representing the sagittal depth of the condyles.
The tangents of the most distal parts of each condyle perpen-
dieular to line Aare drawn and the midpoint between these
lines on line A is called P4. The distance P3 to P4 represents
the height of the condyles, whieh is called y. A quotient was
determined by dividing the condyle height (y) by the sagittal
depth (x). This quotient describes the relation between height
and depth
remaining 84 patients did not develop any function-
allimitations. The configuration of the femoral con-
dyles in all patients was measured (Fig.1.l6). The
measurements showed that, compared with the non-
operated patients, the 16 patients who needed sur-
gery had femoral condyles that were more spherical.
This indicates that the articular geometry of the fem-
oral condyles is of importance for function after ACL
tear. One may speculate as to whether the configura-
tion of the femoral condyles is of importance in the
development of OA.
1.4.5
Patellar Position
Patellar problems have previously been mentioned in
trauma and degenerative disease. The position of the
K. Jonsson and T. Boegard.
patella is considered to be a significant aetiological
factor in retropatellar chondromalacia. A distinction
is drawn between high-riding patella - patella alta
- and low-riding patella - patella baja. The most
common method employed to determine the height
of the patella is the Insall-Salvati index. In this
method the greatest diagonal length of the patella
(LP) is divided by the length of the patellar tendon
(LT). An Insall-Salvati index (LP/LT) of less than 0.8
indicates patella alta, which results in stresses on the
central and distal portions of the retropatellar carti-
lage. If the Insall-Salvati index is greater than 1.2 it
indicates patella baja, with resultant loading on the
proximal portion of the retropatellar cartilage. Sev-
eral other methods have been described. The reliabil-
ity and inter-observer variability of the Insall-Salvati
index and several other indices have been compared
(SEIL et al. 2000). There was great variability between
different indices and, depending on the index used,
the same knee could show patella alta, patella baja
or normality. In SEIL et al.'s study, the method
described by BLACKBURNE and PEEL (1977) showed
the lowest inter-observer variability and discrimi-
nated best among the groups "alta", "normal" and
"baja". In this method the ratio of the articular sur-
face length of the patella to the height of the lower
pole of the articular surface above a tibial plateau line
is measured.
EGUND et al. (1988) described a method in which
the relation of the patella and the tibia is measured
in weight bearing and 30-40° of knee flexion. This
seems to be the most physiological method, but it was
not studied by SEIL et al. (2000).
References
Ahlbäck 5 (1968) Osteoarthrosis of the knee: a radiographie
investigation. Acta Radiol 5uppI277:7-72
Altman RD, Hochberg M, Murphy WA Jr et al (1995) Atlas of
individual radiographie features in osteoarthritis. Osteoar-
thr Cartil3 [5uppl A]:3-70
Ballinger PW (1991) Merrill's atlas of radiographie positions and
radiologie procedures, 7th edn. Mosby Year Book, 5t Louis
Bernau A (1995) Orthopädische Röntgendiagnostik. Einstell-
technik. Urban and 5chwarzenberg, Munieh
Blackburne J5, Peel TE (1977) A new method of measuring
patellar height. J Bone Joint 5urg [Br] 59:241-242
Boegard T (1998) Radiography and bone scintigraphy in osteo-
arthritis of the knee-comparison with MR imaging. Acta
Radiol39 [5uppl418]:3-37
Boegard T, Jonsson K (1999) Radiography in osteoarthritis of
the knee. 5keletal Radio128:605-615
Boegard T, Rudling 0, Petersson IF et al (1997) Postero-anterior
radiogram of the knee in weight bearing and semiflexion:
comparison with MR imaging. Acta Radiol 38: 1063-1 070
Radiography
Boegard T, Rudling 0, Petersson lF et al (1998a) Correlation
between radiographieally diagnosed osteophytes and mag-
netie resonance detected cartilage defects in the tibiofemo-
ral joint.Ann Rheum Dis 57:401-407
Boegard T, Rudling 0, Petersson lF et al (1998b) Joint-space
width in the weight-bearing radiogram of the tibio-femo-
ral joint. Should the patient stand on one leg or two? Acta
RadioI39:32-35
Boegard T, Rudling 0, Petersson lF et al (1998c) Joint -space
width in the axial view of the patello-femoral joint. Def-
initions and comparison with MR imaging. Acta Radiol
39:24-31
Boegard T, Rudling 0, Petersson lF et al (1998d) Correlation
between radiographieally diagnosed osteophytes and mag-
netie resonance detected cartilage defects in the patello-
femoral joint. Ann Rheum Dis 57:395-400
Bontrager KL (1997) Textbook of radiographie positioning
and related anatomy, 4th edn. Mosby, St Louis
Brossmann J, Freyschmidt J, Czerny C (2001) Grenzen des
Normalen und Anfänge des Pathologischen in der Radi-
ologie des kindlichen und erwachsenen Skeletts, 14th edn.
Thieme, Stuttgart
Buckland-Wright C (1995) Protocols for precise radio-ana-
tomieal positioning of the tibiofemoral and patellofemoral
compartments of the knee. Osteoarthritis Cartilage 3
[Suppl A]:71-80
Cook TDV, Kelly BP, Harrison L, Mohamed G, Khan B (1999)
Radiographie grading for knee osteoarthritis. A revised
scheme that relates to alignment and deformity. J Rheuma-
toI16:641-644
De Pablos J, Azcarate J, Barrios C (1995) Progressive open-
ing-wedge osteotomy for angular long-bone deformities in
adolescents. J Bone Joint Surg [Br] 77:387-391
Egund N, Lundin A, Wallengren NO (1988) The vertieal posi-
tion of the patella. A new radiographie method for routine
use. Acta RadioI29:555-558
Edholm P, Lindahl 0, Lindholm B et al (1976) Knee instability.
An orthoradiographie study. Acta Orthop Scand
47:658-663
Friden T, Ryd L, Lindstrand A (1992) Laxity and graft fixation
after reconstruction of the anterior cruciate ligament. A
roentgen stereophotogrammetrie analysis of 11 patients.
Acta Orthop Scand 63:80-84
Friden T, Jonsson A, Erlandsson T et al (1993) Effect offemoral
condy1e configuration on disability after an anterior cruci-
ate ligament rupture. 100 patients followed for 5 years. Acta
Orthop Scand 64:571-574
lnsall J, Shoji H, Mayer V (1974) High tibial osteotomy. J Bone
Joint Surg [A]56:1397-1405
J6nsson A, Hannesson P, Herrlin K et al (1995) Computed
vs film-screen magnification radiography of fingers
in hyperparathyroidism. An ROC analysis. Acta Radio1
36:71-75
J6nsson A, Herrlin K, Jonsson K et al (1996) Radiation dose
reduction in computed skeletal radiography: effect on
image quality. Acta RadioI37:128-133
Keats TE (1996) Atlas of normal roentgen variants that may
simulate disease, 6th edn. Mosby-Year Book, st. Louis
Kellgren JH, Lawrence JS (1957) Radiologic assessment of
osteoarthritis. Ann Rheum Dis 16:494-501
Leach RE, Gregg T, Silber FJ (1970) Weight bearing radiogra-
phy in osteoarthritis of the knee. Radiology 97:265-268
Lugo-Olivieri CH, Scott WW Jr, Zerhouni EA (1996) Fluid-fluid
17
levels in injured knees: do they always represent lipohem-
arthrosis? Radiology 198:499-502
Mazzuca ST, Brandt KD (1999) Plain radiography as an out-
come measure in clinieal trials involving patients with knee
osteoarthritis. Rheum Dis Clin North Am 25:467-480
Merchant AC, Mercer RL,Jacobsen RH et al (1974) Roentgeno-
graphie analysis of patellofemoral congruence. J Bone Joint
Surg [Am] 56A:1391-1396
Odenbring S, Berggren AM, Peil L (1993) Roentgenographie
assessment of the hip-knee-ankle axis in medial gonarthro-
sis. A study of reproducibility. Clin Orthop 289: 195-196
Pavlov H, Burke M, Giesa M, Seager KR, White E (1999) Ortho-
paedist's guide to plain film imaging. Thieme, New York
Piperno M, Conrozier T, Bochu M, Carillon Y, Fantino 0,
Vignon E (1998) Re-evaluation of conventional radiogra-
phy in arthritis (in French). Rev Prat 48 SuppI17:S5-S8
Ravaud P,Aule1ey GR, Chastang C, Rousselin B Paolozzi L,Arnor
B, Dougados M (1996) Knee joint space width measurement:
an experimental study of the influence of radiographie pro-
cedure and joint positioning. Br J RheumatoI35:761-766
Resniek D, Vint V (1980) The "tunnel" view in assessment of carti-
lage loss in osteoarthritis of the knee. Radiology 137:547-548
Ryd L (1992) The role of roentgen stereophotogrammetrie
analysis (RSA) in knee surgery. Am J Knee Surg 5:44-54
Sanfridsson J, Ryd L, Eklund K et al (1996) Angular configu-
ration of the knee. Comparison of conventional measure-
ments and the QUESTOR Precision Radiography system.
Acta RadioI37:633-638
Sanfridsson J, Svahn G, Jonsson K et al (1997) Computed radi-
ography for characterisation of the weight-bearing knee.
Acta RadioI38:514-519
Sanfridsson J, Svahn G, Ryd L et al (1998) Assessment ofimage
post-processing and of measuring assistance tools in com-
puted radiography. 39:642-648
Sanfridsson J, Holje G, Svahn G et al (2000) Radiation dose and
image information in computed radiography. A phantom
study of angle measurements in the weight-bearing knee.
Acta RadioI41:310-316
Sanfridsson J, Ryd L, Svahn G, Friden T, Jonsson K (2001) Radio-
graphie measurement of femorotibial rotation in weight
bearing. The influence of flexion and extension in the knee
on the extensor mechanism and angles of the lower extrem-
ity in a healthy populaton. Acta Radio142:207 - 217
Scott WW Jr, Rosenbaum JE, Ackerman SJ et al (1993) Subtle
orthopaedie fractures: teleradiology workstation versus
film interpretation. Radiology 187:811-815
Seil R, Müller B, Georg T et al (2000) Reliability and interob-
server variability in radiologie al patellar height ratios. Knee
Surg Sports Traumatol Arthrosc 8:231-236
Se1vik GA (1989) A roentgen stereophotogrammetrie system
for the study of the kinematies of the skeletal systems. Uni-
versity of Lund, Sweden, Thesis 1974 (reprint). Acta Orthop
Scand 60 [SuppI232]:1-51
Siu D, Cook DV, Broekhoven LD et al (1991) A standardized
technique for lower limb radiography. luvest Radio126:71
Spector TD, Co oper C, Cushnaghan Jet al (1992) A radiograph-
ie atlas of knee osteoarthritis. Springer, Berlin Heide1berg
NewYork
Stedtfeld HW, Strobel M (1983) A new holding deviee for the
preparation of stress roentgenograms of the knee joint.
Unfallheilkunde 86:230-235
Strobel M, Stedtfeld HW (1990) Diagnostie evaluation of the
knee. Springer, Berlin Heidelberg New York
2 Arthrography
A. MARK DAVIES and VICTOR CASSAR-PULLICINO
CONTENTS
2.1 Introduction 19
2.2 Indications 19
2.3 Contraindications 19
2.4 Technique 19
2.5 Complications 20
2.6 Interpretation 20
2.7 Accuracy 21
2.8 Conclusion 21
References 21
2.1
Introduction
Arthrography is the imaging technique by which the
intra-articular structures of a joint are rendered vis-
ible or more conspicuous by the introduction of a
contrast medium.1t was first described only 10 years
after the discovery ofX-rays using a negative contrast
medium (air or oxygen), but it was not until the
1930s that the use of positive contrast media (iodine
based) was introduced in the knee. As early as 1934
the value of the double-contrast technique was recog-
nised (BIRCHER and OBERHOLZER 1934). Over the
next 30 years, further refinements were introduced
(FREIBERGER et al. 1966) such that it was the imag-
ing technique of choice for internal derangements of
the knee until the mid-1980s (BUTT and McINTYRE
1969; STOKER 1980) roday it has been largely ren-
dered obsolete in the knee by the introduction of
magnetic resonance (MR) imaging,computed tomog-
raphy (Cr) and ultrasound. It remains, however, a
useful adjunct to both MR imaging and cr with so-
called MR arthrography and cr arthrography (see
Chaps. 3 and 4).
A.M. DAVIES, MD
Consultant Radiologist, MRI Centre, Royal Orthopaedic Hos-
pital, Bristol Road, Birmingham, B31 2AP, UK
V. CASSAR-PULLICINO, MD
Consultant Radiologist, Department of Diagnostic Imaging,
The Robert Jones & Agnes Hunt Orthopaedic and District
Hospital, Oswestry, Shropshire, SYlO 7AG, UK
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
2.2
Indications
Historically the main indication for knee arthrog-
raphy was soft tissue injury, be it meniscal and/or
ligamentous. Although in the developed world knee
arthrography has been superseded by other tech-
niques, it may be of value if the cost of the procedure
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.
2.3
Contraindications
While there are no absolute contra-indications to
knee arthrography, local sepsis and a history of iodine
sensitivity are a cause for caution.
2.4
Technique
There are two elements to the successful performance
of a knee arthrogram: first, the satisfactory intro duc-
tion of the contrast agents into the joint; second, the
prompt production of high -quality radiographs.
The technique of ne edle puncture is identical irre-
spective of the nature of the subsequent imaging, be
it conventional arthrography, cr or MR arthrogra-
phy. rhe authors' preferred technique is the lateral
approach under aseptic conditions (GOLDMAN 1984).
The patient is supine with the knee mildly flexed over
a small pillow or foam pad in order to relax the patel-
lar mechanism. A sandbag or similar device is placed
against the ankle joint to prevent external rotation of
the lower limb. This ensures that the patellofemoral
joint is parallel to the table top. Using the non-dom-
inant hand, the patella is deviated laterally with the
fingers while the thumb palpates the space between
20 A. M. Davies and V. Cassar-Pullicino

the midpole of the patella and the underlying femoral 2.6

eondyle. Using the dominant hand, the skin and sub- Interpretation
eutaneous tissues are infiItrated with loeal anaesthet-
ic. A 21-gauge needle is horizontally advaneed into Exquisite images of both normal and abnormal
the patellofemoral joint, upon whieh a "popping sen- menisci ean be obtained (Fig. 2.1). The aesthetie qual-
sation" is usually feit. If the needle is intra-artieular, ity of the images is all the more appreciated beeause
no resistanee will be feit during the injeetion of of the radiographie skill and dexterity required to
air. Typieally 2-5 ml of non-ionie eontrast medium aehieve an adequate examination. It is beyond the
(300 mg I/ml) and 20 ml air will be injeeted for the seope of this ehapter to detail the normal and abnor-
double-eontrast method. The addition of adrenaline mal appearanees on double-eontrast knee arthrog-
(0.2 ml of 1:1,000 solution) will improve the visualisa- raphy. The interested reader is referred to the texts
tion of the intra -artieular structures by eausing vaso- cited above. In the assessment of leaking popliteal
eonstrietion of the synovial vessels, thereby deereas- eysts, arthrography has been supplanted by uItra-
ing the rate of absorption of the contrast medium sound, and for eruciate ligament injuries, by MR
and the amount of synovial fluid formed. Where imaging. Arthrography is of little value in the assess-
possible, any pre-existing joint fluid should be aspi- ment of medial eollateralligament tears as the defeet
rated before injeetion of the contrast medium to pre- in the synovium tends to seal off within 1 or 2 days
vent unneeessary dilution (nE CARVALHO and JURIK of injury.
1985). The knee is briefly exereised to ensure that all
the surfaees are adequately eoated with the contrast
medium. With the double-contrast method, over-vig-
orous exercising may eause frothing, thereby redue-
ing image quality.
Metieulous attention to the radiographie eompo-
nent of the examination is essential to ensure good-
quality images showing the menisci in their entirety.
A 10w-kilovoItage teehnique (50-55 kVp), a short
exposure time to minimise movement blur, and a
small foeal spot to reduee geometrie unsharpness are
all essential. Detailed deseriptions of the fluoroscopie
teehniques ean be found in the following texts: BUTT
and McINTYRE (1969), RICKLIN et al. (1979), FREI-
BERGER (1979), DALINKA (1980), STOKER (1980) and
GOLDMAN (1984). While stress is applied to open up
eaeh eompartment in turn, aseries of well-eollimated
a
spot films are obtained in varying degrees of rotation
to demonstrate the menisci. Subsequently, reeumbent
and sitting lateral radiographs of the knee are per-
formed to assess the cruciate ligaments (PAVLOV et al.
1983).

2.5
Complications

Minor swelling and diseomfort persisting for several

days after the proeedure are usually the worst that
most patients experienee. Some advoeate aspiration
of the air at the end of the examination to minimise
symptoms. Signifieant eomplieations are rare, with a
cited incidenee of septie arthritis of 1/25,000 (WATT
1987).
b
Fig. 2.1a, b. Double-contrast knee arthrography. aNormal pos-
terior horn medial meniscus. b Vertical tear, posterior horn
medial meniscus (arrows)
Arthrography
2.7
Accuracy
In experienced hands, accuracy is reported to exceed
90% with respect to meniscal tears (FREIBERGER et
al. 1966; KISS and MOIR 1968; BUTT and McINTYRE
1969; BRowN et al. 1978) and to be approximately
90% for anterior cruciate ligament tears (PAVLOV and
TORG 1978; PAVLOV et al.1983).
2.8
Conclusion
There is little doubt that knee arthrography is some-
what passe and has litde role in modern imaging
other than as an adjunct to examining the knee with
CT or MR imaging. It is worth noting, however, that its
diagnostic accuracy of over 90% compares remark-
ably weH with today's gold-standard, MR imaging.
References
Bircher E, Oberholzer J (1934) Die Kniegelenkkapsel im Pneu-
moradiographiebild.Acta RadioI15:452-459
Brown DW, Allman FL Jr, Eaton SB (1978) Knee arthrography,
21
a comparison of radiographie and surgieal findings. Am J
Sports Med 6: 165-172
Butt WP, Mclntyre JL (1969) Double contrast arthrography of
the knee. Radiology 92:487 -499
de Carvalho A, Jurik AG (1985) Joint fluid after aspiration: a
disturbing factor in knee arthrography. Acta Radiol Diagn
26:715-719
Dalinka MK (1980) Arthrography. Springer, Berlin Heidelberg
NewYork
Freiberger RH, Killoran PJ, Cardona G (1966) Arthrography
of the knee by double contrast method. Am J Roentgenol
97:736-747
Freiberger RH (1979) Techniques of knee arthrography. In:
Freiberger RH, Kaye JJ (eds) Arthrography. Appleton-Cen-
tury-Crofts, New York, pp 5-30
Goldman AB (1984) Knee arthrography. In: Goldman AB (ed)
Procedures in skeletal radiology. Grune and Stratton, Orlan-
do,pp 85-164
Kiss J, Moir JD (1968) Experience with arthrographie examina-
tion of the knee joint. J Can Assoc RadioI19:187-191
Pavlov H, Torg JS (1978) Double contrast arthrographie
evaluation of the anterior cruciate ligament. Radiology
126:661-665
Pavlov H, Warren RF, Sherman MF, Cayea PD (1983) The accu-
racy of the double contrast arthrographie evaluation of
the anterior cruciate ligament. A retrospective review of
163 surgieally confirmed cases. J Bone Joint Surg [Am]
65:175-183
Rieklin P, Ruttiman A, DeI Buono MS (1979) Meniscus lesions.
Grune and Stratton, New York
Stoker DJ (1980) Knee arthrography. Chapman and Hall,
London
Watt I (1987) Contrast medium investigations. In: Park WM,
Hughes SPF (eds) Orthopaedie radiology. Blackwell Scien-
tific Publieations, Oxford pp 45-104
3 Computed Tomography (CT) and CT Arthrography
RICHARD WILLIAM WHITEHOUSE

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.

Over the last two decades, sectional imaging has

developed rapidly. Computed tomography (CT) and
magnetic resonance (MR) imaging are now estab-
lished methods of investigation of the knee and both
methods continue to develop such that descriptions
of the current state of the art remain valid for only
months. The dinical value of comparisons between
techniques such as CT and MR imaging depends
R.W. WHITEHOUSE, MD
Department of Clinical Radiology, Manchester Royal Infir-
mary, Oxford Road, Manchester, Ml3 9WL, UK
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
3.2
Developments in Cl
A CT image is a Cartesian co-ordinate map of nor-
malised X-ray attenuation coefficients, generated by
electronicaHy filtered computerised back projection
of X-ray transmission measurements in multiple
directions through a section of the object in ques-
tion. This description is as true today as when Houn-
sfield first described the technique. Those areas where
24
re cent developments have been made include helical
scanning, multislice acquisition and real-time CT
"fluoroscopy". These developments have been made
on the back of improving technology which includes
slip rings for power and data transmission to and
from the gantry, higher heat loading X-ray tubes,
high-efficiency solid state X-ray detectors, and faster
data transmission and processing abilities of the elec-
tronics.
3.2.1
Slip Rings
In order to acquire X-ray transmission data in all
directions across a slice of the patient, the X-ray tube
has to travel around the entire circumference of a
circle around the slice. If the tube is supplied with
power by cables, then these have to wrap around
the circle as the tube moves. In order to unwrap the
cables, the next slice is performed by rotating the
tube in the opposite direction. This design requires
more than 360° of tube rotation as initial acceleration
and final deceleration distances are also required.
Powerful motors and brakes are required to cope
with the inertia of this system (which may include
the X-ray detectors and counterweights to balance
the gantry) and a significant time delay is necessary
between each slice acquisition to allow for these
acceleration and deceleration phases. Replacing the
cables with slip rings (large-circumference electrical-
ly conducting rings) which encircle the X-ray tube
path, and transferring power from the rings to the
X-ray tube via conducting brushes on the X-ray tube
gantry, allows the gantry to be continuously rotated
in one direction. This has several advantages - rapid
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
time delay between slices need be no longer than that
required for table movement in conventional acquisi-
tion mode and the potential for acquiring continu-
ously updated X-ray transmission data paves the way
for both helical scanning and CT fluoroscopy.
3.2.2
X-rayTubes
The development of slip rings resulted in a require-
ment for X-ray tubes to have both a higher heat
capacity and a higher maximum tube current, as the
mAs required for a single slice remained much the
R. W. Whitehouse
same but the time in which the slice was acquired was
reduced. In addition, for helical scanning continu-
ous X-ray output for up to 60 s may be required. The
dis advantage of these X-ray tubes is the increased
ease with which high radiation doses can be given to
patients during CT investigations. This disadvantage
may be mitigated by the developments of solid state
and multislice detectors.
3.2.3
X-ray Detectors
Xenon gas detectors, used in CT scanners for many
years, have a conversion efficiency (X-rays to signal
strength) of around 60%, which can diminish further
if the detectors are not maintained. Solid state crystal
detectors may have conversion efficiencies of nearly
100%, resulting in a 40% reduction in patient radia-
tion dose for the equivalent scan appearances. The
tendency for solid state detectors to continue emit-
ting light after the X-rays had terminated (afterglow)
and other technical problems with respect to the size
of the front face of the individual detectors and the
interspace material between adjacent detectors have
been largely overcome.
Although one of the earliest EMI CT head scan-
ners acquired two slices at the same time with a
pencil beam of X-rays passing to two adjacent detec-
tors, the ease with which solid state detectors can
be stacked in parallel adjacent channels has facilitat-
ed the re-development of multislice scanners. These
scanners can acquire several sections simultaneously,
which can be separately proeessed to give large num-
bers of thin seetions, or recombined to give fewer
thicker seetions with lower noise.
3.2.4
Helical eT (Spiral or Volume Scanning)
The requirement for a break in the X-ray emission
whilst the table is moved to the next slice position
was overeome by the development of helical sean-
ning. Helical scanning is performed by moving the
table continuously during the exposure, from the first
slice loeation to the last. Thus a helix of X-ray trans-
mission data through the sean volume is acquired.
To generate a CT image the data from adjaeent turns
of the helix are interpolated to produee transmission
data which are effeetively from a single slice loeation.
This process ean be performed at any loeation within
the helix (except the first and last 1800 's, where there
Computed Tomography (CT) and CT Arthrography
is no adjacent helix of data for interpolation). In
this way overlapping slices can be produced without
overlapping irradiation of the patient. The relation-
ship between the X-ray fan-beam collimation and the
table movement per rotation of the gantry is called
the pitch ratio. Extended or stretched pitch scans
are performed with pitch ratios greater than 1. Such
extended pitches can be used to trade off between
greater scan volumes, shorter sc an acquisition times
and lower sc an radiation doses. Stretching the pitch
ratio to 1.25 has little effect on the image appearanc-
es, but pitch ratios greater than 1.5 produce images
with an effective slice thiekness significantly greater
than the nominal fan-beam collimation thiekness. By
increasing the number of detector arrays ("multislice
scanner"), several interlaced helices can be acquired
simultaneously (Fig. 3.1), with the table in creme nt
per gantry rotation increased proportionately.
Fig.3.1. Pictorial representation of the path followed by a
single column of detectors for a four-beam multislice he1ical
CT scan. The patient positioning illustrates a way of scanning
a single knee at a time
The combination of multislice and helieal scan-
ning results in volume scan acquisition times whieh
are 4 times faster than a single-slice helical scanner
with the same gantry rotation speed, and an order
of magnitude faster than a non-spiral scanner. Mul-
tislice scanning reduces X-ray tube loading require-
ments as it acquires several slices simultaneously
with the same tube loading as a single slice would
require on a conventional scanner. The patient radia-
tion dose, however, is not directly reduced and may
be increased if greater volumes are scanned.
25
3.2.5
CT "Fluoroscopy"
In conventional CT transmission data from a 3600
gantry rotation are required to generate an image.
This is because two opposing beam paths then exist
for each ray across the imaging volume. This pro duc-
es improved signal to noise, corrections for the effects
of divergent X-ray beams along each ray and beam-
hardening effects. Images can also be produced using
270 0 or even 1800 gantry rotation datasets. Such "par-
tial scan" images have acquisition tim es proportion-
ately shorter than full rotation scans. This can be
useful for reducing movement artefacts in selected
patients. For a 0.5 s per rotation scanner, the effective
scan acquisition time will be one quarter of a second
(250 ms). If the gantry continues to rotate and
acquire data without table movement, continuously
updated transmission data will be collected from
whieh revised images can be generated. At any one
time, image data of between 0 and 250 ms old will be
available, i.e. data with an average age of 125 ms.
With extremely rapid processors and appropriate
reconstruction algorithms, further delay for image
reconstruction can be minimised and a continuous-
ly updated CT image displayed in "ne ar real time"
(HSIEH 1997). Such "CT fluoroscopy" imaging can
be used for CT-guided interventional procedures. As
with all fluoroscopic procedures, care should be taken
to reduce fluoroscopy time to the minimum neces-
sary and to avoid operator irradiation: instruments
designed to keep the operator's hands out of the CT
section (DALY et al.1998) and use of the lowest select-
able tube current (50 mA is sufficient; FROELICH et
al. 1999) are advocated. To assist in maintaining short
CT fluoroscopy exposure tim es, routine recording
and auditing of fluoroscopy exposure times are advo-
cated. An audible alarm after apreset exposure time
mayaiso assist in keeping exposures as short as pos-
sible. The use of a lead drape adjacent to the irradi-
ated volume has been demonstrated to reduce opera-
tor exposure (NAWFEL et al. 2000). High skin doses to
patients and operators will occur if care is not taken.
3.2.6
Data Manipulation
The vast mass of image data acquired from a mul-
tislice spiral scanner produces problems of data stor-
age and interpretation. It is no longer feasible to pro-
duce hard-copy images of every available section.
With isometrie voxels, reformatted images in any
 26
other image plane will have the same image quality
as the acquisition images, potentially requiring even
further hard copies.
Fast workstations, allowing rapid reformatting
and display of examinations in the most appropri-
ate plane for the pathology being demonstrated,
are therefore necessary, with hard copy restricted
to representative images. Other image reconstruc-
tion methods [curved planes, surface-rendered 3D
images, minimum or maximum intensity projec-
tions (Fig.3.2)] can produce a bewildering array of
visually stunning images, though demonstration of
the clinical utility of these methods is currently lim-
ited.
The current state of the art device has a mul-
tislice helical scanner with solid state detectors, sub-
second scan acquisition and image reconstruction
times, CT fluoroscopy capability and a link to a
powerful workstation with real-time image manip-
ulation software.
a b
c patient with nail-patella syndrome
R. W. Whitehouse
3.2.7
Reformatted Images
As spiral multislice scanning produces overlapping
sections and thinner slice collimation (less than
1 mm), in-plane and reformatted plane spatial resolu-
tions are now potentially similar, even for CT images
from small fields of view (Table 3.1). Volume acqui-
sitions obtained in any plane can therefore be refor-
matted into other planes without marked loss of
image quality. For scanners not capable of such fine
collimation, CT in the most appropriate plane for the
expected pathology is still preferable, if achievable.
As the best effective z-axis resolution for a spiral
scan acquisition is approximately half the X-ray
beam collimation thickness, isometric voxels can be
achieved by selecting a combination of scan param-
eters such as those identified in Table 3.1. For non-
helical scanners, overlapping transverse sections will
provide better z-axis resolution (e.g. 25-cm field of
Fig.3.2a-c. Examples of workstation manipulation of CT data
sets. a Coronal reformat through a comminuted distal femoral
fracture. b Maximum intensity projection (MIP) reconstruc-
tion through the same femoral fracture. (a and b kindly pro-
vided by Philips Medical Systems from images acquired with a
Tomoscan AVEl and processed on the EasyVision workstation
by Dr. R.C. Berlin, St. Tohn's Hospital, Tackson, Wyo., USA.) c
3D surface rendering of a hypoplastic bipartite patella in a
Computed Tomography (CT) and CT Arthrography 27

Table 3.1. Effect of field of view, collimation and reconstruction 50 1

interval on pixel and voxel sizes for CT images _ _ !arge ph8nlOm (185Osq cm)

_ _ smaI phantom (960 sq cm)

Field of Pixel diameter on Collimation/reconstruction
view 512X512 matrix interval required for iso-
metrie voxels (spiral mode)
1 4O
~
12.5 cm 0.25mm 0.5 mm/0.25 mm 130
'ö
25cm 0.5mm 1.0 mm/O.5 mm
50 cm 1.0mm 2.0 mm/1.0 mm ~GI
.~ 20
'c

t ....... ....
view, I-mm slice collimation, D.S-mm table increment
between sections) but at twice the radiation dose to
the patient.
- 10
.... ----- ..
O+-----~----~----~--~-----,

o 200 400 600 600 1000

3.3
Scan Image Quality
The amount of noise, beam-hardening and streak
artefacts in a CT image are dependent upon the fol-
lowing factors:
- Collimation slice thickness
- Partial or full rotation dataset
- Mass and distribution of tissue in the scan plane
- Scan time/movement
- High-density extraneous material (the contralat-
erallimb, contrast medium spills, surgical metal-
work)
- kVp andmAs
- Field of view
- Matrix size
- Reconstruction algorithm
- Post-processing image sharpening or softening fil-
ters
- Viewing window width and level settings
Most of these factors are amenable to selection or
modification by the scanner operator and can mark-
edly affect the quality of the final image. As Fig.3.3
demonstrates, the relationships between image noise,
mAs and patient size are non-linear, with a halving of
patient size resulting in a quarte ring of image noise,
whilst a fourfold increase in mAs is needed to halve
the image noise. The figure also demonstrates that for
small patients, image noise is low at all mAs settings
and the absolute reduction in image noise achieved
by quadrupling the mAs is small. For these reasons,
only one knee should be scanned wherever possible,
rather than both together. This reduces image noise,
streak and beam-hardening arte facts caused by the
contralaterallimb. A lower mAs setting can then be
used with consequent reduced patient radiation dose,
mAs
Fig.3.3. The influence of size and mAs on image noise for a
CT scanner. Sections were performed through water density
phantoms with lO-mm sliee collimation
no tube loading limitations (cooling time, limited
spirallength) and potentially increased tube life. If
the pathology being imaged is osseous, the width of
the usual viewing window renders noise impercepti-
ble, and even when soft tissue lesions are viewed on a
narrow window, the noise is not intrusive. In practice,
unfortunately, not all patients can achieve or main-
tain the position necessary to scan one knee (Fig. 3.1)
for long enough for a satisfactory scan. I would, how-
ever, most strongly advocate this positioning in chil-
dren, where the position may be easier to achieve and
the importance and benefits of dose reduction are so
much greater.
Streak arte facts can be generated by high-density
material within the scan plane hut outside the field
of view of the scanner. Tabletops which contain
edge grooves, tracks for the fixing of attachments or
detachable mattresses can act as traps for spilt con-
trast media. Contrast drop lets on the gantry window
will also cause image artefacts. Scrupulous care to
keep the tabletop and gantry clean is needed to
remove these sources of artefact.
3.3.1
Internal Metalwork from Fixation Devices
The streak artefact generated from in situ intramed-
ullary rods is rarely excessive and does not prevent
adequate assessment of the bone cortex, making CT
of value in assessing fracture union in selected cases.
28
More intrusive streak artefact is seen when the CT
plane is through locking screws in intramedullary
rods, bone surface plates or fixation screws. Care in
patient positioning (including decubitus positions
where necessary), combined with gantry angulation
in order to align the scan plane with the long axis
of any screws present, will reduce the number of sec-
tions degraded by streak artefact from the screws to a
minimum. In scanners with operator-selectable kVp,
the use of the highest kVp setting will reduce streak
artefact, as will the selection of a higher mAs (though
the combination of increased kVp and mAs results
in considerably greater tube loading and patient irra-
diation). Streak artefact also may appear visually less
intrusive on volume-rendered (3D) images (PRETO-
RIUS and FISHMAN 1999).
3.3.2
Cl Number, Hounsfield Units, Window Width
and Levels
The scale of numbers used to define the grey scale in
CT images is artificially limited by data storage con-
straints. In the earliest days of CT, the number scale
ran from -500 for air, through zero for water and up
to +500 at the top of the scale. This allowed numeri-
cal values from -512 to +512 to be stored as a lO-bit
binary number. This "EMI number" scale was soon
replaced by the CT number scale, still used now,
where air has a value of -1,000 and water has a value
of zero. The top end of the scale is usually con-
strained to fit into a 12-bit binary number (allowing
number values from -1,024 to +3,072 to be stored).
The Hounsfield unit (HU) is the true value which the
CT number should represent. Scanner drift, calibra-
tion error, artefact or other limitation may render
this inaccurate, which is why measurements made
from scan images are best called CT numbers.
The Hounsfield unit value for any material is
defined by:
where
HUs = the Hounsfield unit value for substance S;
/ls = the linear attenuation coefficient for substance
s;and
Ilw =the linear attenuation coefficient for water.
This formula relates the HU value to the linear
attenuation coefficients of the material being mea-
sured and water. As the linear attenuation coefficients
R. W. Whitehouse
of all materials change with X-ray beam energy, there
are consequently only two fixed points on the Houn-
sfield scale. These are -1,000, which is the HU value
for no X-ray attenuation (i.e. a vacuum), and zero,
which corresponds to the HU value for water (at the
calibration pressure and temperature for the scan-
ner). The HU scale is, in fact, open ended, with high
atomic number, high -density materials having values
way in excess of the upper end of the usual scale (even
on "extended sc ale" scanners) (Table 3.2).
Table 3.2. Theoretical RU values for a variety of materials at
65keV
Material RUvalue
Adipose tissue -80
Water o
Collagen 250
Dense cortical bone 1,600
Aluminium 2,300
lron 34,000
Iodine 141,300
Lead 205,000
As implied above, EMI units are converted into CT
units by doubling them. In the eady days of CT, it was
usually stated that dense cortical bone had an EMI
number of around 500 (i.e. the top of the scale),
which became 1,000 when the scale changed from
EMI to CT numbers. However, the theoretical Houn-
sfield value for dense cortical bone calculated at an
effective beam energy of 65 ke V (equivalent to a
scanner operating at around 120 kVp) is in the region
of 1,600 (Fig. 3.4). At 10wer energies (e.g. 55 keV -
the approximate effective energy of a scanner oper-
ating at 80 kVp), the HU value for dense cortical
bone is over 2,000. Other high atomic number mate-
rials (contrast media, aluminium and metal fixation
devices) also show marked variation in HU value
with beam energy. By contrast, the HU values of soft
tissues, collagen and fat vary very little with effective
beam energy as the linear attenuation coefficients for
these materials closely follow those of water.
Consequently, in scanners which allow the operat-
ing voltage to be changed, the CT number for bone
can be increased byusing a low kVp (around 80 kVp).
This increases the dependence of the CT number on
the presence of bone or calcification and is particu-
lady used for quantitative measurement of mineral
density. A high k VP (usually around 140 k Vp) can be
selected to reduce the CT number ofbone and metal-
work, which has some effect in reducing streak arte-
facts.
Computed Tomography (CT) and CT Arthrography 29

3000 , - - - - - - - - - , . - - - - - - - - - - - - - , at an overview of trabecular bone which may be inap-

propriate for subtle cortical bone lesions (Fig.3.5).
The most appropriate window level for cortical bone
2000 will be influenced by the bone density and the effec-
.BE
-bone tive scan energy, whilst the window width may need
- -collagen
e
I-
• • • soft tissue to be quite narrow to demonstrate subtle intracorti-
o 1000 - • fat cal density changes.
Reviewing images on the console prior to printing
hard copies is recommended to obtain the best image
settings for individual patients and to avoid over-
----------
looking pathology not demonstrated at "standard"
~ ~ ~ 00 ~ ro ~ 00
settings.
Effective kV

Fig.3.4. The influence of scanner kVp on CT numbers for

bone, collagen, soft tissue and fat

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
and surrounding structures in all imaging planes is
necessary for adequate interpretation. Knowledge of
anatomieal structures not easily or consistently dem-
onstrated on eT is still needed to assess the likeli-
hood of their involvement by any pathology whieh is
demonstrated. The anatomy of the region has been
covered in other chapters. Selected eT images are
included here for comparative purposes (Fig. 3.6).
For the clearest depietion of artieular surfaces and
fractures, images perpendicular to the plane of the
artieulation or fracture are usually best (vide infra),
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).
Although eT of the knee is a rapid procedure, immo-
bilisation may be necessary to prevent movement
artefacts, partieularly in children. Sandbags, Velcro
straps and stieking tape will usually suffice. Even
better immobilisation is achieved routinely in trauma
patients by the pIaster of Paris cast or backslab they
are usually fitted with. Scanning through a pIaster
cast does not significantly interfere with image quali-
tywhilst the immobilisation achieved is usually excel-
lent, such that a temporary cast is also worth consid-
ering for occasional patients inadequately immobil-
ised by other methods.

/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
3.5
Indications
CT of the knee is particularly suited to the dem-
onstration of bony anatomy such as the evaluation
of bony morphological abnormalities, pateHofemo-
ral tracking and fractures (GRAY et al. 1997). Intra-
articular fractures in particular should be assessed
by CT. Intraosseous tumours are weH demonstrat-
ed; for example, the nidus of an osteoid osteoma,
which can be overlooked on MR imaging, is charac-
teristic and clearly demonstrated on CT. The pres-
ence of tumour matrix ossification or calcification is
also clear on CT. In osteomyelitis, the presence and
location of sequestra are revealed. With the addi-
tion of arthrography, osteochondrallesions are weH
demonstrated. Soft tissue pathology is less weH
demonstrated than with MR imaging, and intrave-
nous contrast medium injection provides less sat-
isfactory contrast enhancement than the equivalent
MR examination but valuable information on soft
tissue lesions is still obtainable from CT (for exam-
pIe: size, extent, tumour calcification, enhancement,
articular involvement). CT can be used to guide
biopsy and cyst aspiration procedures (ANTONACCI
et al. 1998) (Fig. 3.7).
The accurate three-dimensionallocalisation of the
bone anatomy with CT can be used to calculate the
mechanical axes of long bones and the relationships
of the joints. This information can then be used in
the preoperative planning of joint replacements, and
morphological differences have also been demon-
strated between patients with anterior knee pain and
Fig. 3.7. CT-guided aspiration of an anterior cruciate ligament
ganglion cyst. A posteromedial approach is used to avoid the
popliteal vessels
R. W. Whitehouse
patients with asymptomatic knees. CT has also been
used after joint replacement to identify the relation-
ship between knee morphology, alignment of the
prosthesis and outcome.
The CT scanogram, usuaHy used to identify the
start and finish points for a CT investigation, can also
be used for limb length measurements.
Use of CT to demonstrate meniscal tears was most
prominent in the literature in the late 1980s, but
results were variable and MR imaging has subsumed
this role. This application has not been widely re-
investigated with modern scanners but may warrant
consideration where MR imaging is contra-indicat-
ed. One recent study gave dual-detector spiral CT
arthrographya sensitivity of 98% and a specificity of
94% for meniscal abnormalities (VAN DE BERG et al.
2000). Where helical scanning is not available, sim-
ilar results can be obtained from CT arthrography
with 1-mm seetions performed at 0.5-mm table incre-
ments on conventional transverse scans (MUGHETTI
et al. 1998).
The limitations of CT are usuaHy described in rela-
tionship to MR imaging, and consequently the poorer
soft tissue co nt rast of CT is top of the list. Where MR
imaging is available and not itself contra-indicated,
it is the most appropriate modality for imaging soft
tissue lesions. The other limitations of CT in relation
to MR imaging are the direct multiplanar capability
of the latter and the use of ionising radiation with
CT.
3.5.1
Trauma
In the acutely traumatised patient, speed and patient
safety are important requirements for a satisfactory
examination. This gives limited scope for scan tech-
nique modifications. As the primary aim of the exam-
ination is to determine the size and disposition of
fracture fragments and joint alignments, the aim is
to ensure adequate coverage of the injured region in
a single helical acquisition with effective slice thick-
ness appropriate to the size of the fracture fragments.
For purely osseous detail, a low mAs is sufficient. An
AP scout view to determine appropriate start and
end points for the acquisition should be routine. The
smaHer the collimation thickness, pitch and recon-
struction interval, the better the quality of reform at-
ted planar and 3D images, giving overlapping helical
acquisition seetions a small advantage over conven-
tional contiguous transverse seetions for fracture
classification (Fig. 3.8).
Computed Tomography (CT) and CT Arthrography 33

3.5.2
Knee Morphology and Surgery

CT is ideal for localising anatomical reference points

in three dimensions. The physics of the process
involves elearly defined table positions and utilises
the straight-line paths ofX-rays from the known tube
locations to the known detector locations. Conse-
quently there should be no distortion (compared with
MR imaging, where magnetic field variations cause
Fig. 3.8. Coronal reformation of a depressed tibial plateau frac- image distortion) or magnification (unlike conven-
ture. Scan acquired helically with 3-mm collimation, a pitch tional radiography) of the images. This is true for CT
ratio of 1 and reconstructed at 2-mm increments
images acquired with no gantry angulation but not
otherwise. Scout views also suffer from magnifica-
tion perpendicular to the direction of table travel but
However, at a collimation thickness of 3 mm, no not along it. With these provisos, CT provides three-
statistically significant difference in measured artic- dimensionallocalisation byvirtue of the x and y coor-
ular surface depression or fragment displacement is dinates in the CT image and the z coordinate by the
evident between conventional and helical scan pro- table location. The scout view can be used to identify
tocols (LOMASNEY et al. 1999). If helical scanning is the structures to be localised, single transverse sec-
not available, a mixed protocol of thicker sections to tions performed at each required level, and x, y and
cover the extent of the fracture, with thinner sections z coordinates recorded. From these points, distances
through the region of the articular surface depres- and angles can be calculated or measured directly on
sion, can be used. Three-dimensional surface recon- appropriate transverse or reformatted images. Such
structions provide an easily interpreted overview of measurements have been used to define the "Q" angle
fracture fragment disposition, particularly useful in (alignment of rectus femoris to the patellar tendon)
badly comminuted injuries (PRETORIUS and FISH- (ANDO 1999), the angle of rotation of the tibia with
MAN 1999). respect to the femur (NAGAO et al. 1998), the morphol-
High-energy, badly comminuted proximal tibial ogy of the femoral trochlear groove (MARTINO et al.
fractures may be better treated by Ilizarov techniques 1998), the location of the tibial tuberosity in flexed
than by conventional open reduction and fixation. osteoarthritic knees (NAGAMINE et al. 1997) and knee
Such injuries require CT assessment of fragment dis- version in anterior knee pain (ECKHOFF et al. 1997).
position, particularly with reference to the size and In preoperative assessments, CT has been used to
position of the fragment containing the tibial tuber- develop individual templates for joint replacements or
ele, but also to assist in defining the extent of soft intraoperative spatiallocalisation of prosthetic liga-
tissue injury. Any necessary limited open reduction ment attachments (RADERMACHER et al. 1998; SATI et
can then be performed through appropriate full- al. 1997). Finally, CT has been used for postoperative
thickness flaps with less risk of devitalising tissue assessment of intercondylar notchplasty (no change
(WATSON and COUFAL 1998). in the appearance of the notch up to 1 year after sur-
Avulsion injuries are well recognised around the gery, MANN et al. 1999), to demonstrate malrotation
knee. Sites inelude the tibial tuberosity and inferior causing patellofemoral complications after total knee
pole of the patella (patellar tendon), the posterior arthroplasty (BERGER et al. 1998) and to assess the
tibial plateau (posterior cruciate ligament, rare), the location of the tibial tunnel after anterior cruciate lig-
tibial eminen ce (anterior cruciate ligament), the lat- ament reconstruction (MCGUIRE et al. 1997). A scan
eral tibial plateau ("Segond fracture", lateral capsu- technique appropriate to the location of the struc-
lar ligament) and the fibular head (conjoint tendon of tures to be measured is required. These applications
lateral collateralligament and biceps femoris). Most currently remain largely of research or limited elinical
avulsion injuries are usually evident on conventional use.
plain film examination but the tibial eminence and
posterior tibial plateau avulsion injuries in particular
are less elearly visualised on plain film. CT can more
elearly demonstrate these (STEVENS et al. 1999 ): a
thin-section protocol (3 mm or less) is suggested.
34
3.5.3
Patellofemoral Joint
Knee flexion to varying degrees has been advocated
for the assessment of the patellofemoral joint align-
ment. The plain film assessment by skyline views
in 45° of knee flexion (Merchant's views) has been
largely superseded by CT or MR evaluation with
the increasing realisation that, once flexed to this
degree, the patella is forced into the intercondylar
groove even when prone to subluxation or maltrack-
ing (WALKER et al. 1993). Imaging at lesser degrees
of flexion is not easily or consistently achieved with
plain film techniques; hence CT methods were devel-
oped. Single 10-mm sections through the mid patel-
la with the knee fully extended and then passively
placed at 10°, 20° and 30° of flexion (Fig.3.9) can
demonstrate malalignment, joint space narrowing
and/or tilting not visible on plain film examination
(SCHUTZER et al. 1986).
Fig.3.9. Static unloaded patellar tracking study in a patient
with nail-patella syndrome (hypoplastic patellae). Single
10-mm-thick low-mAs seetions were performed through the
patella at 10°, 20° and 30° of knee flexion
R. W. Whitehouse
The addition of loading to put tension on the
extensor mechanism further increases the sensitivity
of the examination for maltracking. With the advent
of spiral scanning, yet further sophistication of the
technique is possible, with continuous data acquisi-
tion during active flexion and extension of the knee
to give kinematic CT imaging. This can also be per-
formed with or without additionalloading (Dupuy
et al. 1997).
The images obtained from static, dynamic, loaded
or unloaded studies are assessed for lateral shift of
the patella (Fig. 3.10), the lateral patellofemoral angle
(Fig. 3.11), thinning of the joint cartilage and other
osteoarthritic changes.
3.5.3.1
Static Cl of the Patellofemoral Joint
With the patient lying supine on the scanner table, flex-
ion of the knee over aseries of padded wedges to pro-
duce 10°,20° and 30° of angulation and a single low
mAs lO-mm-thick section through the mid part of the
patella at each angulation are all that is needed for an
unloaded CT assessment of the patellofemoral joint. In
an average-sized adult, in whom both the leg and the
thigh are approximately 450 mm long, raising the back
of the knee by only 40, 80 and 115 mm from its extend-
ed position achieves these respective angulations. For
a static loaded examination, the knee needs to be held
against resistance applied to the anterior shin, ankle
or foot to put the extensor mechanism into tension
whilst similar sections as above are obtained. Asking
the patient to hold their foot up against arestraining
bar or strap over the ankle will achieve this. Greater
loading can be achieved with weighted boots worn
during the same manoeuvres. Abnormal patellofemo-
ral relationships were found to be greatest during max-
imal quadriceps loading with the knee straight, being
less marked at 30° of flexion and appearing normal at
60° flexion (BIEDERT and GRUHL 1997).
3.5.3.2
Dynamic Cl of the Patellofemoral Joint
To achieve a truly dynamic assessment of the patello-
femoral joint, continuous imaging during movement
of the knee is required. Originally described using
ultrafast CT (MUHLE et al. 1999), this is now feasible
with helical scanning performed continuously during
active flexion and extension of the knee. If the rate of
movement is kept relatively slow and well controlled,
movement artefact on the subsequent images is not
intrusive. This method has been described by DuPUY
Computed Tomography (CT) and CT Arthrography
E patellofemoral articulation and centred equidistant
Fig. 3.10. Assessment of lateral shift of the patella is made by
measuring the percentage of the patella (DB/CB+ 100) that lies
lateral to a line (EF) drawn perpendicular to the plane on the
anterior condylar margins (AA), centred on the lateral edge of
the lateral condyle (P). M, Medial; L, lateral
B thickness and volume in the knee (Fig. 3.12).
Fig. 3.11. The patellofemoral angle (a) is measured between a
line joining the anterior condylar margins (AA) and the lateral
patellar articular facet (BB). The angle is normally positive (as
illustrated)
et al. (1997) and reviewed by MUHLE et al. (1999). The
technique can be performed with or without addi-
tionalloading, where the wearing of a weighted boot
is used to achieve this. The distal thigh is placed in a
custom-made holder so that the rest of the lower limb
extends beyond the end of the table. A lateral scout
view with the knee extended is then obtained and
the gantry angled to be perpendicular to the mid
35
between the physeal closure plate and Blumensatt's
line. Continuous helical scanning is then performed
(for 10 s on als per rotation scanner) without table
movement, during which time the patient flexes and
extends the knee from fully extended to 45° flexion
and back to fully extended. The data are recon-
structed into ten images. As only one slice location
is imaged with this technique, the patella moves
through the section and the patellofemoral relation-
ships are only truly demonstrated in a few of the
images. The preliminary study by DuPUY et al. (1997)
showed a tendency for patellofemoral relationships
to improve during dynamic studies compared with
static unloaded examinations, suggesting a lower sen-
sitivity but higher specificity in dynamic studies.
3.5.4
Articular Cartilage
Thin-section CT (particularlymultislice helical) com-
bined with arthrography (see below) has been used
as the gold standard for measuring articular cartilage
These measurements are of increasing clinical
importance as targeted treatments for knee osteoar-
thritis are being developed (HANGODY et al. 1998).
Many of these studies are aimed at validating MR
methods of cartilage measurement rather than advo-
cating the use of CT arthrography (ECKSTEIN et al.
1997, 1998; HAUBNER et al. 1997). Nevertheless, CT
arthrography currently remains more sensitive than
non-arthrographic MR imaging for subtle cartilage
defects (DAENEN et al. 1998) and the minimally inva-
sive technique allows accurate measurement of cap-
sular volume, fluid aspiration for laboratory studies
and injection of therapeutic agents as required (BER-
QUIST 1997).
3.5.5
Soft Tissues
The imaging examination of soft tissue masses and
synovial diseases of the knee is best undertaken by
MR imaging (with or without Gd-DTPA enhance-
ment), augmented by radiographs and possibly spe-
cialist ultrasound examination. CT has a limited role
where these methods are contra -indicated or unavail-
able, but some pathologies [e.g. fatty tumours such
as lipoma arborescens, calcified lesions - synovial
osteochondromatosis (Fig.3.13), gouty tophi and
36
dense lesions such as pigmented villonodular syno-
vitis 1 may have characteristic appearances on cr
(CHEN et al. 1999; LIN et al. 1999). cr is unreliable for
follow-up scanning of the resection site of soft tissue
sarcoma (HUDSON et al. 1985).
Contrast enhancement can be used with volume
rendering to demonstrate arte rial and graft stenosis
and obstructions after vascular surgery down to the
Fig. 3.12. Patellar cartilage thickness demonstrated by double-
contrast CT arthrography. (Courtesy of Dr. S. Bianchi, Höpital
Cantonal, Geneva, Switzerland)
Fig.3.13. Synovial osteochondromato-
sis. Multiple peripherally calcified lesions
are present within a joint effusion
R. W. Whitehouse
popliteal vessels (ISHIKAWA et al.1999) and othervas-
cular lesions (Fig. 3.14). Similarly, softtissue enhance-
ment in masses or synovium can be demonstrated
but timing is critical, with peak enhancement being
later, less marked and more variable in onset than in
the abdomen. cr scanner software which pre-scans
at low mA to detect the onset of enhancement and
triggers the study at that point may have a role.
Fig. 3.14. a Maximum intensity projection from scans acquired
during intravenous contrast infusion demonstrate the arte-
rial vascular anatomy of the knee and proximal leg. Contrast
extravasation from the distal muscular (sural) artery is noted.
b 3D surface-rendered image from the same data as a. (Images
provided courtesy of Philips Medical Systems and Dr. R.C.
Berlin, St. John's Hospital, Jackson, Wyo., USA)
Computed Tomography (CT) and CT Arthrography
3.6
Arthrography
3.6.1
Role and Indications
er arthrography ean be performed as an adjunet to
eonventional knee arthrography. With double-con-
trast arthrography, er ean be performed immedi-
ately after the eonventional examination as only a
small amount of iodinated eontrast medium is injeet-
ed (4 ml of 300 mg IIml). If single-eontrast arthrogra-
phy is performed with larger quantities of dense eon-
trast, an interval of 2-3 h between the eonventional
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)
Fig.3.16. The normal lateral and medial
menisci are demonstrated on sagittal ref-
ormations from CT arthrography using
the same technique as in Fig. 3.10. (Cour-
tesy of Dr. S. Bianchi, Höpital Cantonal,
Geneva, Switzerland)
37
teehnique and the er examination allows the eon-
trast medium density to dilute to a level appropriate
for er. If an immediate er arthrogram is planned,
appropriate reduetion in eontrast medium eoneen-
tration is needed, e.g. 150 mg I/mI non-ionie water-
soluble eontrast medium.
Indications for er arthrography are the dem-
onstration of synovial plieae, intra-articular loose
bodies, ehondrallesions and osteoehondral defeets.
Meniseal ossicles ean also be demonstrated (MAR-
TINOLl et al. 2000) (Fig.3.15). Demonstration of
meniseal tears requires multisliee helical seanning
with isometrie voxels of less than 0.5 mm in diameter
to aehieve aeeuraey similar to MR imaging (Fig. 3.16).
38 R. W. Whitehouse

An adequate examination can be achieved with non-

spiral scanners if 1-mm slice collimation is available
and overlapping seetions are obtained at O.5-mm
table increments (MUGHETTI et al. 1998), though
there is a radiation dose penalty to the patient with
this technique.

3.6.2
Technique

Fluoroscopy screening is rarely needed for needle

placement but confirms correct location during injec-
tion of contrast medium. Local anaesthetic should
not be necessary, though preparation of the skin with
topical anaesthetic cream may be useful in children.
Using an aseptic technique, a small-gauge ne edle is
introduced into the patellofemoral joint from either
the medial or the lateral side, holding the patella
and displacing it towards the side chosen for needle
introduction. The patellar displacement emphasises
the location of the articulation and assists in needle
placement. A double-contrast technique with 5 ml of
contrast and 50 ml can be performed if a conven-
tional arthrogram is also required. A single-contrast
technique with 20 ml of contrast is satisfactory for
a CT arthrogram and avoids the streak arte fact that
may occur at air/fluid interfaces. Screening during
injection confirms correct needle location as contrast
should flow rapidly away from the needle tip into the
joint. A misplaced ne edle results in focal accumula-
tion of contrast at the needle tip. After removing the
needle, gentle manipulation ensures that the contrast
extends throughout all the joint capsular recesses.
After a conventional arthrogram if required, trans- b
verse CT through the knee is performed. Delayed
post-arthrography scanning may be necessary if com-
munication between the knee joint and any nearby
cyst or ganglion is suspected as there may be a 1-
to 2-h delay before contrast appears within the cyst
(MALGHEM et al. 1998).

Fig.3.17. Post-contrast CT seetions demonstrate the location

of a arteries and b veins to assist in c CT-guided ablation of
an osteoid osteoma. Note the incidental bone island. (Images c
courtesy of Dr. P. Hughes, Derriford Hospital, Plymouth, UK)
Computed Tomography (CT) and CT Arthrography 39

3.7 a biopsy needle, have been used. In either case, to avoid

Cl-Guided Interventions complications the lesion to be treated should be more
than a centimetre from neurovascular or other critical
Computed tomography is being increasingly used to structures. A preliminary biopsy for histological con-
guide interventional procedures, recently encouraged firmation of the diagnosis is necessary as in one series
by the development of CT fluoroscopy, which enables 16% oflesions were not osteoid osteomas (SANS et al.
more rapid and accurate placement of needles and 1999). Osteoid osteomas can cause severe pain when
interventional devices (FROELICH et al. 1999; DE MEY biopsied, and although some series report the use of
et al. 2000). As described above (Sect. 3.2.5), care local anaesthesia, epidural or general anaesthesia may
needs to be taken to minimise operator and patient be necessary.
X-ray exposure during CT-guided biopsy. CT fluoros-
copy times of around 10 s should suffice for most
biopsy procedures (GOLDBERG et al. 2000). Limiting
the fluoroscopy to identification of the needle tip 3.8
rather than the entire ne edle will also reduce oper- Conclusion
ator and patient radiation dose (SILVERMAN et al.
1999). The CT section thickness should be appropri- With appropriate attention to technique, CT continues
ate to the size of the lesion, otherwise partial volume to have a role in the diagnosis and management of many
averaging may include both the needle tip and the conditions in and around the knee. Some pathologies,
lesion in the same section, erroneously suggesting an e.g. meniscal tears, will, however, only be adequately
accurate needle location. CT can be used to guide demonstrated using CT arthrography and a scanner
biopsy of bone and soft tissue lesions. Where pri- capable of sub-millimetre resolution in the y-axis.
mary malignancy is present then the course of the
biopsy track and the compartment(s) through which
it passes may need excision with the tumour at the References
time of definitive surgery. Biopsy of such lesions must Ando T (1999) Factors affecting the rectus femoris-patellar
therefore only be performed after consultation and tendon Q-angle, measured using a computed tomographie
agreement on the approach with the surgeon who will scan. J Orthop Sei 4:73-77
carry out the definitive treatment. Accuracy of CT- Antonacei VP, Foster T, Fenlon H, Harper K, Eustace S (1998)
Technieal report: CT-guided aspiration of anterior crueiate
guided biopsy is increased if specimens are obtained
ligament ganglion cysts. Clin Radio153:771-773
for both cytology and pathology: an overall accuracy Berger RA, Crossett LS, Jacobs JJ, Rubash HE (1998) Malrota-
of around 80% should be achieved (HODGE 1999). tion causing patellofemoral complications after total knee
Due to their deep location, ganglion cysts of the arthroplasty. Clin Orthop 356:144-153
anterior cruciate ligament may be difficult to identify Berquist TH (1997) Imaging of artieular pathology: MRI, CT,
arthrography. Clin Anat 10:1-13
on ultrasound imaging. CT-guided aspiration of these
Biedert M, Gruhl C (1997) Axial computed tomography of the
cysts with a wide-bore (14G) needle followed by injec- patellofemoral joint with and without quadrieeps contrac-
tion of 40 mg depomedrone and 1-2 ml 0.5% bupiva- tion. Arch Orthop Trauma Surg 116:77-82
caine (ANTONACCI et al. 1998) has been successful for Chen CK, Yeh LR, Pan HB, Yang CF, Lu YC, Wang JS, Resnick D
symptomatic relief. Using a posterior approach, care is (1999) Intra -artieular gouty tophi of the knee: CT and MR
imaging in 12 patients. Skeletal Radio128:75-80
taken to avoid the popliteal vessels (Fig. 3.7).
Daenen BR, Ferrara MA, Marcelis S, Dondelinger RF (1998)
Percutaneous treatment of osteoid oste oma can Evaluation of patellar cartilage surface lesions: compari-
also be performed with CT guidance. Again, a planned son of CT arthrography and fat-suppressed FLASH 3D MR
approach avoiding vascular structures is required, and imaging. Eur Radio18:981-985
a preliminary contrast-enhanced scan to clearly iden- Daly B, Templeton PA, Krebs TL, Carroll K, Wong You Cheong
tify the vessels is valuable (Fig. 3.17). It is possible to JJ (1998) Evaluation ofbiopsy needles and prototypie needle
guide deviees for percutaneous biopsy with CT fluoroscopie
treat osteoid osteomas by complete removal via CT- guidance in simulated organ tissue. Radiology 209:850-855
guided biopsy (VOTO et al.1990), but this may be diffi- De Mey J, Op de Beeck B, Meysman M, Noppen M, De Maese-
cult to achieve with biopsy needles unless a large-bore neer M, Vanhoey M, Vincken W, Osteaux M (2000) Real time
needle is used and several passes are made through CT-fluoroscopy: diagnostie and therapeutie applications.
Eur J Radio134:32-40
the lesion. More recently, techniques aimed at destroy-
Dupuy DE, Hangen DH, Zachazewski JE, Boland AL, Palmer
ing the tumour with heat, from either a radiofrequen- W (1997) Kinematie CT of the patellofemoral joint. AJR
cy ablation probe or a laser-heated probe, both of 169:211-215
which are available with fine probes for passage down Eckhoff DG, Brown AW, Kilcoyne RF, Stamm ER (1997) Knee
40
version associated with anterior knee pain. Clin Orthop
339:152-155
Eckstein F, Adam C, Sittek H, Becker C, Milz S, Schulte E, Reiser
M, Putz R (1997) Non-invasive determination of cartilage
thiekness throughout joint surfaces using magnetie reso-
nance imaging. J Biomech 30:285-289
Eckstein F, Schnier M, Haubner M, Priebsch J, Glaser C, Engl-
meier KH, Reiser M (1998) Accuracy of cartilage volume
and thiekness measurements with magnetie resonance
imaging. Clin Orthop 352:137-148
Froe!ieh JJ, Ishaque N, Saar B, Regn J, Walthers EM, Mauer-
mann F, Klose KJ (1999) Steuerung von perkutanen Biop-
sien mittels CT-Fluoroskopie. [Control of percutaneous
biopsywith CT fluoroscopy.] Rofo Fortsehr Geb Rontgenstr
Neuen Bildgeb Verfahr 170:191-197
Goldberg SN, Keogan MT, Raptopoulos V (2000) Percutaneous
CT-guided biopsy: improved confirmation of sampling site
and needle positioning using a multistep technique at CT
fluoroscopy. J Comput Assist Tomogr 24:264-266
Gray SD, Kaplan PA, Dussault RG (1997) Imaging of the knee.
Current status. Orthop Clin North Am 28:643-658
Hangody L, Kish G, Karpati Z, Udvarhe!yi I, Szigeti I, Be!y M
(1998) Mosaieplasty for the treatment of artieular carti-
lage defects: application in clinieal practiee. Orthopedies
21:751-756
Haubner M, Eckstein F, Schnier M, Losch A, Sittek H, Becker
C, Kolem H,Reiser M,Englmeier KH (1997) A non-invasive
technique for 3-dimensional assessment of artieular carti-
lage thiekness based on MRI, part 2. Validation using CT
arthrography. Magn Reson Imaging 15:805-813
Hodge JC (1999) Percutaneous biopsy of the musculoskeletal
system: a review of 77 cases. Can Assoc Radiol J 50:121-125
Hsieh J (1997) Analysis of the temporal response of computed
tomography fluoroscopy. Med Phys 24:665-675
Hudson TM, Schakel M, Springfie!d DS (1985) Limitations of
computed tomography following excisional biopsy of soft
tissue sarcomas. Skeletal Radiol13:49-54
Ishikawa M, Morimoto N, Sasajima T, Kubo Y (1999) Three-
dimensional computed tomographie angiography in lower
extremity revascularization. Surg Today 29:243-247
Lin J, Jacobson JA, Jamadar DA, Eilis JH (1999) Pigmented vil-
lonodular synovitis and related lesions: the spectrum of
imagingfindings.AJR 172:191-197
Lomasney LM, Griffiths AA, Tobin JP, Rabin SI, Cronin TR
(1999) Comparison of 4 computed tomographie imaging
protocols for lateral tibial plateau fractures. Can Assoc
RadiolJ 50: 11 0-116
Malghem J, Vande-berg BC, Lebon C, Lecouvet FE, Maldague
BE (1998) Ganglion cysts of the knee: artieular communiea-
tion revealed by delayed radiography and CT after arthrog-
raphy.AJR 170:1579-1583
Mann TA, Black KP, Zanotti DJ, Barr M, Teater T (1999) The
natural his tory of the intercondylar notch after notchplasty.
Am J Sports Med 27:181-188
Martino F, De Serio A, Macarini L, Rizzo A, Laforgia R, Roton-
do A,Ange!elli G (1998) Ultrasonographyversus computed
tomography in evaluation of the femoral trochlear groove
morphology: a pilot study on healthy, young volunteers. Eur
RadioI8:244-247
Martinoli C, Bianchi S, Spadola L, Garcia J (2000) Multimodal-
ity imaging assessment of meniscal ossicle. Skeletal Radiol
29:481-484
McGuire DA, Hendricks SD, Sanders HM (1997) The rela-
tionship between anterior cruciate ligament reconstruction
R. W. Whitehouse
tibial tunnellocation and the anterior aspect of the poste-
rior cruciate ligament insertion. Arthroscopy 13:465-473
Mughetti M, De-Masi M, Teodorani A, Gagliardi S (1998) Falsa
immagine con Tomografia Computerizzata di lesione de!
menisco mediale dei ginocchio. Studio retrospettivo in 400
casi. Radiol Med Torino 96:310-312
Muhle C, Brossmann J, Heller M (1999) Kinematie CT and MR
imaging of the patellofemoral joint. Eur RadioI9:508-518
Nagamine R, Miura H, Inoue Y, Tanaka K, Urabe K, Okamoto
Y, Nishizawa M, Iwamoto Y (1997) Malposition of the tibial
tubercle during flexion in knees with patellofemoral arthri-
tis. Ske!etal Radio126:597 -60 1
Nagao N, Tachibana T, Mizuno K (1998) The rotational angle
in osteoarthritie knees. Int Orthop 22:282-287
Nawfel RD, Judy PF, Silverman SG, Hooton S, Tuncali K,
Adams DF (2000) Patient and personne! exposure during
CT fluoroscopy-guided interventional procedures Radiol-
ogy 216:180-184 (and comments pages 9-10)
Pretorius ES, Fishman EK (1999) Volume-rendered three-
dimensional spiral CT: musculoske!etal applications. Radio-
graphies 19:1143-1160
Radermacher K, Portheine F, Anton M, Zimolong A, Kaspers
G, Rau G, Staudte HW (1998) Computer assisted orthopae-
die surgery with image based individual templates. Clin
Orthop 354:28-38
Sans N, Morera-Maupome H, Galy-Fourcade D, Jarlaud T, Chi-
avassa H, Bonnevialle P, Giron J, Railhac JJ (1999) Resec-
tion percutanee sous controle tomodensitometrique des
osteomes osteoides. Suivi a moyen terme de 38 cas. [Per-
cutaneous resection under computed tomography guid-
ance of osteoid osteoma. Mid-term follow-up of 38 cases.]
J RadioI80:457-465
Sati M, deGuise JA, Drouin G (1997) Computer assisted knee
surgery: diagnosties and planning ofknee surgery. Comput
Aided Surg 2:108-123
Schutzer SF, Ramsby GR, Fulkerston JP (1986) The evaluation
of patellofemoral pain using computerized tomography: a
preliminary study. Clin Orthop 204:286-293
SilvermanSG, Tuncali K, Adams DF, Nawfe! RD, Zou KH,
Judy PF (1999) CT fluoroscopy-guided abdominal inter-
ventions: techniques, results, and radiation exposure. Radi-
ology 212:673-681
Stevens MA, EI-Khoury GY, Kathol MH, Brandser EA, Chow S
(1999) Imaging features of avulsion injuries. Radiographies
19:655-672
Van de Berg BC, Lecouvet FE, Poilvache P, Dubuc JE, Bedat
B, Maldague B, Rombouts JJ, Malghem J (2000) Dual-detec-
tor spiral CT arthrography of the knee: accuracy for detec-
tion of meniscal abnormalities and unstable meniscal tears.
Radiology 216:851-857
Voto SJ, Cook AJ, Weiner DS, Ewing JW,Arrington LE (1990) Treat-
ment of osteoid osteoma by computed tomography guided
excision in the pediatrie patient. J Pediatr Orthop 10:510-513
Walker C, Cassar-Pullicino VN, Vaisha R, McCall IW (1993)
The patello-femoral joint - a critieal appraisal of its geo-
metrie assessment utilizing conventional axial radiography
and computed arthro-tomography. Br J RadioI66:755-761
Watson JT, Coufal C (1998) Treatment of complex lateral
plateau fractures using Ilizarov techniques. Clin Orthop
353:97-106
Whitehouse RW, Economou G, Adams JE (1993) The influence
of temperature on quantitative computed tomography:
implications for mineral densitometry. J Comput Assist
Tomogr 17:945-951
4 Magnetic Resonance Imaging
CHRISTINE HERON and ANDREW HINE

CONTENTS knee MR imaging. This particularly applies to the

question of scanning techniques. The optimal MR
4.1 Introduction 41
4.2 General Considerations 41 study obtains a maximum amount of information in
4.2.1 Signal to Noise Ratio 42 a given time period. Local practices and other factors
4.2.2 Contrast to Noise Ratio 44 dictate certain aspects of the examination. In every
4.2.3 Spatial Resolution 44 clinical setting there must always be a compromise
4.2.4 Scan Time 46
between a lengthy examination which may be unnec-
4.2.5 Image Contrast and Pulse Sequences 46
4.2.6 Image Display 47 essarily thorough and a cursory study which leaves
4.2.7 MR Arthrography 47 major questions unanswered. In practice, most insti-
4.2.8 Patient Positioning and Imaging Planes 48 tutions establish a routine protocol which provides
4.2.9 Menisci 48 adequate visualisation of the knee and answers the
4.2.10 Ligaments 51
4.2.11 Hyaline Cartilage 51
majority of clinical questions. Supplementary pro-
4.2.12 Bone 54 tocols may be employed to address specific clinical
4.3 MR Protocol for problems. This chapter will consider both of these
Routine Examination of the Knee 54 requirements. The demonstration of most pathology
4.4 MR Protocols for Specific Clinical Problems 55 within the knee requires pulse sequences that provide
4.4.1 Synovium and Soft Tissue Masses 55
optimal visualisation of mobile water-bound pro-
4.4.2 Femoral Trochlear Dysplasia and
Patellar Tracking Studies 55 tons. Increased water is seen in joint effusions, and
4.4.3 Tendons 55 in ligaments, tendons and muscles following injury.1t
4.4.4 Posterolateral and is also present in inflammatory change in soft tis-
Posteromedial Aspects of the Knee 55 sues and as bone marrow oedema following contu-
4.4.5 MR Imaging of the Postoperative Knee 56
sions. In the case of meniscal tears it is necessary to
4.5 Artefacts 56
References 59 visualise protons which become bound to macromol-
ecules along the edge of the tear. The highest achiev-
able spatial resolution is a further requirement for
knee imaging.

4.1
Introduction
Following the introduction of magnetic resonance
(MR) imaging of the knee into clinical practice there
was an exponential rise in publications documenting
its efficacy. Research into this area of musculoskele-
tal MR imaging has now plateaued and the present
emphasis is on the fine tuning of certain aspects of
C.HERON,MD
Consultant Radiologist, St. George's Hospital, Blackshaw Road,
London, SW17 OQT, UK
A.HINE,MD
Consultant Radiologist, Central Middlesex Hospital, Acton
Lane, London, NW10 5NS, UK
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
4.2
General Considerations
Many factors contribute to the production of opti-
mum images. In addition to the hardware employed,
these include the signal to noise ratio (SNR), con-
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
high. Low field strength magnets operate at 0.2 T and
below. Medium strengths are between 0.3 and 0.9 T
but are usually 0.5- T magnets. High fields are 1-1.5 T.
The field strength influences a number of factors that
are particularly important in relation to the SNR.
42 C. Heron and A. Hine

Higher field strength magnets have better SNR and

when lower field strengths are employed the number
of acquisitions must be increased to produce ade-
quate signal. As a consequence, the longer examina-
tion time may result in artefacts from patient move-
ment.
There has been recent interest in dedicated MR
systems operating at low magnetic field strengths
(0.1-0.2 T) for use in knee imaging. The advantages
of these systems are that they are easier to install and
cheaper to maintain than systems operating at higher
a
field strengths. As only the limb under investigation
enters the scanner, problems with claustrophobia do
not arise. Variable claims have been made regarding
the accuracy of these instruments in the detection of
internal derangement of the knee (AHN et al. 1998;
KOSKINEN et al. 1997; MASCIOCCHI et al. 2000; RAND
et al. 1999b ). The general consensus appears to be that
the lower field strength results in images of slightly
lower resolution. RAND et al. (1999b) observed that
although a low field strength (0.2 T) offered diagnos-
tic potential concerning the presence or absence of
a meniscal tear, the level of confidence in decision b
making was significantly superior with high field
strength (1.5 T) imaging. There is probably little dif-
Fig.4.1. a Quadrature knee coil. b Flexible knee coil
ference in accuracy between systems operating at
high field strength and mid field strengths.
Recent studies employing field strengths above
those in routine clinical use (3 T) to produce higher weighted images, variations in the TE influence the
resolution images have shown promising results T2 weighting and consequently the signal attenua-
(NIITSU et al. 2000; PETERSON et al. 1999). The other tion. Proton density-weighted images are only slight-
technical factors which influence the SNR are consid- ly attenuated by Tl or T2 relaxation and thus have a
ered below but the use of a dedicated surface coil is higher SNR than either Tl-weighted or T2-weighted
essential for optimal signal. A rigid send-receive coil images.
with either a quadrature or phased array design or a A consideration of operator-dependent factors
flexible coil may be employed (Figs. 4.1, 4.2). reveals that the field of view is one of the most impor-
tant determinants of the SNR. A voxel is a volume of
sampled tissue within the patient and the relation-
4.2.1 ship between voxel volume and field of view is indi-
Signal to Noise Ratio cated in the following equation:

The SNR is influenced by both operator-dependent
and non-operator-dependent factors. The non-oper-
ator-dependent factors include the field strength of
the magnet and the intrinsic molecular structure of
the tissue being examined. The molecular structure
of the tissue determines the density of protons and
the Tl, T2 and T2* relaxation times. Operator-depen-
dent factors include the type of coil, the field of
view, the number of acquisitions, sampling band-
width, matrix size, slice thickness and the TR, TE and
flip angle. Changes in TR influence the degree of Tl
weighting and hence the signal attenuation. In T2-
field of view dimensions x slice thickness
voxeI = - - - - - - - - - - - - - -
matrixsize
The matrix size is determined by the number of
frequency encodings and/or phase encodings. Low
frequency and phase encoding numbers result in a
coarse matrix with large voxels (assuming a fixed
square field of view). Large voxels have a higher SNR
than small voxels because they contain more pro-
tons to produce signals. The SNR is influenced by any
change in parameter which alters the voxel volume.
It can be appreciated, for instance, that doubling the
Magnetic Resonance Imaging 43

Fig.4.2a-c. The effect of different coils on

image quality is demonstrated. The field of
view, matrix size and scan time are the same.
In a the main body coil of the scanner has
been employed with no surface coil. In b
a non-phased array wrap around flexi-coil
(I6x13.5 cm) has been used and in c a dedicated
quadrature knee coil has been employed. Image
quality is considerably improved by the use of
surface coils in band c

field of view results in a doubling in the length of two
sides of the voxel and hence a fourfold increase in
the SNR, whereas doubling the slice thickness dou-
bles the length of only one side of the voxel and
hence only doubles the SNR. Both matrix size and
number of acquisitions affect the SNR. Doubling the
matrix size in both the phase encoding and frequen-
cy encoding directions for a fixed field of view results
in the SNR being reduced by a factor of 2 square
root of 2, but the scan time is doubled because there
are twice as many phase encoding steps to sampie
(Fig. 4.3). Doubling the number of acquisitions dou-
bles the scan time but only increases the SNR by a
maximum of square root of 2 (about 1.4).
In the frequency encoding axis the sampling band-
width determines the range of frequencies sampled.
Reducing the sampling bandwidth results in less high
frequency noise being sampled and consequently, as
the signal remains unaltered, there is an increase
in the SNR. However, it is necessary to increase the
sampling time as the sampling bandwidth decreases
to prevent any loss of resolution. This results in an
increase in the minimum TE that can be employed.
The choice of pulse sequence influences the SNR. In
spin echo sequences all of the longitudinal magneti-
sation is converted into transverse magnetisation, but
in gradient echo pulse sequences only a proportion
of the longitudinal magnetisation is converted into
transverse magnetisation as in this case the flip angle
is less than 90°. As a result of this, the signal is gener-
ally greater in spin echo sequences. In summary, SNR
can be increased by using a large field of view, a coarse
44 c. Heron and A. Hine

Fig.4.3a-d. The effect of

altering the matrix size
on image quality and scan
time is demonstrated. All
images were acquired in
a quadrature knee coll
with a field of view of
160x120 cm. a Matrix of
64x39, scan time of 1 min;
b matrix of 128x77, scan
time of 2 min; c matrix
of 256x153, scan time of
4 min; d matrix of
512x307, scan time of
8.5 min

matrix and a large slice thickness. Spin echo sequenc- 4.2.3

es, and in particular those with proton density weight-
ing, produce the highest SNR. In addition, the acquisi-
tion number should be as high as possible.
4.2.2
Contrast to Noise Ratio
The CNR is dependent on the difference in signal
strength between the two tissues under comparison.
Certain factors such as the administration of MR
contrast agents can influence this, but in the absence
of such factors the CNR is influenced by the same
parameters as the SNR.
Spatial Resolution
Resolution is the ability to distinguish between two
points and it is improved as voxel size is decreased.
Separate tissues within the same voxel are not sepa-
rately visualised on the MR image but separate tis-
sues in adjacent voxels are differentiated. When a
voxel contains more than one type of tissue, the signal
intensity of that voxel is the average of the signal
intensities of the different tissues and this results in
partial volume averaging. There is a direct relation-
ship between the size of the voxel and the resolution,
with a small voxel size resulting in better resolution.
As has already been shown in the discussion on SNR,
 Magnetic Resonance Imaging 45

the field of view has the largest effect on the voxel
volume. Halving the field of view results in a four-
fold decrease in the voxel volume, with correspond-
ing improvement in the resolution (Fig. 4.4). Decreas-
ing the slice thickness and therefore reducing the
voxel size also increases the spatial resolution. Whilst
pixels which are square result in the best spatial reso-
lution, it is possible for pixels to be rectangular. This
may arise when the field of view remains square but
an uneven matrix is chosen. The frequency number
of the matrix is usually the highest and the phase
number is entered to adjust the scan time and res-
olution. When the phase number is less than the
frequency number, the spatial resolution is reduced
along the phase axis. In some systems the pixels
remain square irrespective of the matrix selected.
Assuming that the frequency encoding is double the
phase encoding, the field of view in the phase direc-
tion will be half that in the frequency direction. The
pixels remain square and the spatial resolution is
maintained.
It may be necessary to increase the field of view in
the phase direction in order to obtain coverage. This
will increase scan time, and the SNR from the smaller
square pixels will be lower than that from the rectan-
gular pixels. Typically such systems permit the selec-
tion of rectangular pixels which keep the field of view
square and increase the SNR in the phase direction
with adequate coverage. The phase encoding number
and scan time are maintained and the pixels are
rectangular in the phase direction. As each pixel is
larger, the SNR increases and the spatial resolution is

a b

c d

Fig.4.4a-d. The effect

of altering the field of
view on image quaIity is
demonstrated. All images
were acquired in a
quadrature knee coil with
a matrix of 256x153 and
scan time of 4 min. A
rectangular field of view
has been employed. a
Field of view of 320x240;
b field of view of
220x165; c field of view
of 160x120; d field of
view of 120x90
46
reduced. If it is required to maintain spatial resolu-
tion with an uneven matrix in these systems, there
is generally an option of a rectangular field of view.
The dimension of the rectangular field of view in the
phase encoding direction is half that in the frequen-
cy encoding direction. Scan time is halved as a result
of the number of phase encodings and spatial reso-
lution is maintained. Rectangular fields of view may
therefore be used to decrease scan time at a given
resolution or to increase spatial resolution without
increasing scan time.
4.2.4
ScanTime
Ideally, the data acquisition should be completed
in the shortest possible time without compromising
image quality in order to reduce movement artefact.
A doubling of the TR results in a doubling of the scan
time, as does a doubling of the number of acquisi-
tions. Scan time also doubles as a result of a doubling
in phase encodings.
4.2.5
Image Contrast and Pulse Sequences
Two methods are commonly employed to generate
the echo following the initial excitation. The method
employed influences the characteristics of the signal
detected. In the spin echo technique, a second radio-
frequency (RF) pulse is used to reverse the magnet i-
sation and form an echo at a time equal to the inter-
val between the two RF pulses. In addition to reversing
the magnetisation, this second RF pulse also refocuss-
es the dephasing of magnetisation which has occurred
due to field inhomogeneities (T2* relaxation). The
signal intensity in this pulse sequence decays with a
time equal to the T2 relaxation time, and the T2* decay
has been eliminated by the refocussing process. As a
result of the 180° reversal pulse, the contrast in spin
echo images is related to the Tl and T2 relaxation
times of the tissues and to the chosen TR and TE.
Gradient echo images simply refocus the readout
gradient and form a gradient echo. This is measuring
dephasing of the magnetisation due to magnetic field
inhomogeneities (T2* relaxation). T2* relaxation is
shorter than T2 relaxation time. When short TEs are
employed, the signal attenuation from T2 and T2*
is similar but when longer TEs are employed (great-
er than 10 ms), the signal intensity is strongly influ-
enced byT2*.
C. Heron and A. Hine
4.2.5.1
Tl-Weighted Spin Echo Sequence
Tl-weighted images have a relatively short TR and
TE. As a result of the short TR, structures with a long
Tl relaxation time become progressively saturated
and demonstrate low signal intensity (e.g. muscles
and fluid-containing structures). Fat, however, dem-
onstrates high signal intensity.
4.2.5.2
Proton Density-Weighted Spin Echo Images
Proton density-weighted spin echo images have a
relatively long TR and short TE and provide signal
which re fleets the density of protons within the imag-
ing field. Because of this they have a high SNR, and
are useful for providing anatomical detail. The Tl
and T2 relaxation times of the tissues play only a
small part in the image contrast.
4.2.5.3
T2-Weighted Spin Echo Sequence
T2-weighted images with relatively long TRs and TEs
result in high signal intensity fluid owing to its T2
relaxation characteristics. Muscle, however, is of rela-
tively low signal intensity and fat is of intermediate
signal intensity. The SNR in these images is inferior
to the SNR in proton density-weighted images, but
the pathology, which is characterised by an increased
fluid content, is accentuated.
4.2.5.4
Gradient Echo Sequences
In gradient echo sequences the longitudinal magne-
tisation is flipped through an angle other than 90°.
Only part of the longitudinal magnetisation is there-
fore converted to transverse magnetisation. As soon
as the RF pulse terminates, the free induction decay is
produced owing to magnetic field inhomogeneities.
This is known as the T2* effect. A gradient pulse is
used to rephase this transverse component of magne-
tisation. The signal is then received and is known as
a gradient echo.
In the knee, gradient echo images may be used for
the evaluation of menisci. Some centres use gradient
echo images to assess hyaline cartilage, particularly
in conjunction with fat suppression. Gradient echo
images do not provide adequate visualisation ofbone
marrow oedema and they are not recommended for
the evaluation of bony pathology.
Magnetic Resonance Imaging
By adjusting the imaging parameters it is possible
to influence the weighting of a gradient echo
sequence. Tl-weighted gradient echo images are pro-
duced by employing a large flip angle, short TR and
short TE. T2* -weighted images result from a small
flip angle and a TR which is relatively long in order
to permit sufficient recovery of the longitudinal mag-
netisation of the protons. Gradient echo images have
scan times which are typically significantly shorter
than spin echo images. The main disadvantage is
their susceptibility to magnetic field inhomogene-
ities.
Gradient echo sequences are frequently obtained
in volume acquisitions. This is a particularly advan-
tageous technique in MR imaging of the knee. It can
be used for visualising very small structures and for
reformatting images of structures which do not lie
in a single anatomical plane, such as ligaments. An
entire volume of tissue is imaged and the chosen
slice thickness may be as low as 1 mm. Thinner slices
result in longer scan times, as illustrated in the equa-
tion below:
Scan time = TR x number of excitations x phase encoding
number x slice encoding number
Using this technique there is no inter-slice gap
and no cross-talk between images. In order to be
able to observe the images in multiple planes with-
out a reduction in image quality it is essential that
the voxels are nearly isotropie. Volume imaging gen-
erates large datasets with a requirement for consider-
able storage capacity. Obtaining hardcopy of all avail-
able images is impractical. Prompt review of images
on the console is therefore necessary.
4.2.5.5
Fast Spin Echo Sequences
Using the fast spin echo technique, proton density and
T2-weighted images can be obtained in a fraction of
the time required for conventional spin echo sequenc-
es. For each 90 0 pulse, multiple echoes are obtained
as opposed to one. This is achieved by rapidly apply-
ing multiple 1800 pulses with different phase encoding
gradients for a set TE. The different phase encodings
result in multiple echoes (up to 32), thereby potential-
ly reducing imaging time by 32-fold. Fast spin echo
images are susceptible to blurring and edge artefacts
but these can be minimised by using a short echo
train length and a long effective TE. As a result of the
frequent refocussing pulses applied in the fast spin
echo technique, fat appears of higher signal intensity,
47
while solid structures are of lower signal intensity
than in conventional spin echo images.
4.2.5.6
Echo Planar Imaging
Echo planar imaging (EPI) is a further fast imaging
technique in which multiple gradient echo es are
acquired for each excitation. In effect, EPI is the gra-
dient echo equivalent of fast spin echo imaging. The
shortest echo plan ar images can be obtained in less
than 100 ms using single-shot techniques. There is
only one excitation pulse and all of the echoes are
obtained from this. Movement artefacts do not arise
with these very fast imaging times but spatial reso-
lution and SNR can pose problems. These may be
improved by employing multi-shot techniques in
which, after each excitation pulse, several echo es are
obtained. Movement artefact is more problematic in
multi-shot techniques but can be reduced by shorten-
ing the echo train length. At present, EPI is not used
routinely in the knee. In the future it may have a role
in dynamic studies of the knee during movement and
in MR fluoroscopy for guiding interventional pro ce-
dures.
4.2.5.7
Fat Suppression Techniques
Three methods of fat suppression are described (DEL-
FAUT et al. 1999). These are fat saturation, inversion
recovery and opposed phase imaging. The last-men-
tioned technique is principally used for the detec-
tion of small amounts of fat in lesions, e.g. adrenal
gland tumours. It mayaiso be used to determine
whether bone marrow disease is likely to be neoplas-
tic (DISLER et al. 1997) but it will not be considered
further in this section.
4.2.5.7.1
Fat Saturation
In the fat saturation technique an RF pulse with the
same resonance frequency as fat is added to each
slice-selection RF pulse. This is followed by a homo-
geneity spoiling gradient pulse which dephases the
lipid signal.
The dephasing results in no signal emanating from
the fat in the MR images. The advantages of this tech-
nique are that fat only is suppressed and it may be
used with any imaging sequence. It is useful for post-
contrast scanning, e.g. following intravenous contrast
enhancement and MR arthrography. The fat satura-
48 c. Heron and A. Hine

tion technique has a number of disadvantages, how-

ever. Fat saturation may be unreliable owing to the
frequency of the fat saturation RF pulse not equalling
the resonance frequency of the lipid. This is mainly
the result of inhomogeneities in the static magnetic
field. These are small in modern systems and can be
further reduced by auto-shimming, centring on the
region of interest and decreasing the field of view.
The water component of fat and some fatty acids
are not suppressed and can result in inadequately
fat -suppressed images. Fat suppression is poor with
magnets of low field strength because the chemical
shift between lipid and water increases with the mag-
netic field strength. A further disadvantage is that
scanning time is increased. It is relatively short with
T2-weighted imaging but longer with fast gradient
echo imaging.

4.2.5.7.2
Inversion Recovery

The inversion recovery sequence is also known as

the short Tl inversion recovery (STIR) sequence. It
depends on the differences between the Tl of fat and
water (the Tl offat is less than the Tl ofwater).After
a 1800 inversion pulse the Tl (longitudinal) magne-
tisation of fat recovers faster than that of water. By
applying a 90 0 pulse at the null point of fat (the point
at which the longitudinal magnetisation equals zero),
the signal from fat is completely suppressed. One
advantage of this technique is that it suppresses
the signal from all of the fat, including the small
water component. Furthermore, the sequence can be
employed on low field strength magnets and it is not
affected by gradient field inhomogeneities. The T2
and Tl differences are combined, resulting in high
tissue contrast. Tissues with long Tl and T2 (e.g.
water) therefore appear very bright. Oedema and
increased fluid due to inflammation are easily visu-
alised (Fig. 4.5).
There are several disadvantages of the inversion
recovery technique. The images have a relatively low
SNR. Tissues with a similar Tl to fat are also sup-
pressed, for example mucoid tissue, haemorrhage,
proteinaceous fluid, melanin and gadolinium. For
this reason inversion recovery images are not suitable
for MR arthrography or scanning following intrave-
nous gadolinium. Tissues with a short Tl and long
Tl may have the same signal on inversion recovery
images. Spin echo and fast spin echo sequences are
the only sequences which are suitable for use with
inversion recovery techniques. Fat suppression tech-
niques are employed in the knee for the suppression
Fig.4.5. Inversion recovery (STIR) sagittal sequence showing
a high signal intensity focal region of subchondral oedema in
the medial femoral condyle. There is good suppression of the
normal fat in the femoral and tibial condyles but the SNR is
decreased, causing a "grainy" image
of normal fat in post-gadolinium imaging. This may
either be in MR arthrography or following intra-
venous gadolinium. The fat saturation technique is
most effective for post-gadolinium scanning. Fat sup-
pression techniques are also used to assist tissue
characterisation. This is especially true for the detec-
tion of bone marrow oedema and demonstration
of the extent of infiltration by tumour. Either fat
saturation or inversion recovery techniques may be
employed for these purposes.
4.2.5.8
Magnetisation Transfer
Contrast in magnetisation transfer images relies on
the fact that protons bound in macromolecules do
not participate directly in the production of the MR
signal whereas free protons in water do. Macromo-
lecular bound protons are capable of being excited
by a wide range of frequencies. Some of the energy
that they absorb when exposed to such frequencies is
transferred to the protons in surrounding water. The
Magnetic Resonance Imaging
water protons become partially saturated and they
therefore emit less signal. In this way tissues which
contain bound water demonstrate lower signal inten-
sity on magnetisation contrast images. The co nt rast
generally resembles that of T2-weighted images and
considerable signal is lost in solid tissues but this
does not occur in fluid or adipose tissue.
4.2.6
Image Display
At the present time MR images are generally report-
ed from hard copy. Some centres re-image the menis-
ci using greater magnification and special narrow
window settings to enhance meniscal pathology. This
practice is time consuming and costly and has not
been shown to offer significant advantages. Cost sav-
ings may be made by reporting investigations from a
work station. This also permits manipulation of the
images and reformatting of three-dimensional datas-
ets at the time of reporting. Limited hard copy of rele-
vant images can be produced for clinical use. Report-
ing times are generally longer using this technique.
Three-dimensional images have been obtained in one
plane and images reformatted in additional planes
prior to reporting of the scans (WIESLANDER et al.
1998). Whilst this practiee reduces imaging time, the
time required for reformatting and reporting in this
study was greater than the imaging time which it
saved.
4.2.7
MR Arthrography
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
gadolinium intravenously and scanning the knee
after an interval of approximately 40 min during
49
whieh some contrast diffuses into the joint. In order
to facilitate this process, the knee should be actively
exercised following the intravenous injection of con-
trast. Whilst certain success has been described with
this technique (SUH et al. 1996), it fails to distend the
joint with fluid, which is the most valuable aspect of
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-
forward procedure. It is possible for the experienced
radiologist to perform this in the MR room without
the need for radiographie screening but screening
guidance may be preferred. Either a medial or a lat-
eral approach is employed. The posterior border of
the patella is palpated and the ne edle inserted imme-
diately posterior to this at amid point between the
superior and inferior patellar poles. Free flow of fluid
confirms correct needle placement. The preferred
contrast agent differs between centres. The majority
use a dilute gadolinium solution (004 cc of gadolini-
um added to 100 cc of saline). If the procedure is to be
performed under radiographie screening guidance,
part of the saline solution may be replaced by iodin-
ated contrast.Alternatively, iodinated contrast may be
injected initially in order to check that the needle is
correct1y positioned and then the infusate is changed
for the gadolinium solution. Approximately 50 ml of
contrast agent is injected into the knee joint. As the
use of intra-articular gadolinium has not received
universal acceptance by pharmaceutical regulatory
bodies, some centres inject normal saline into the
joint. Following the injection of a gadolinium-based
contrast agent, Tl-weighted images with fat satura-
tion are employed. When normal saline is used as the
contrast agent, T2-weighted or gradient echo images
with T2* weighting are employed (Fig. 4.6).
Fig.4.6. MR arthrogram employing normal saline. Fat-satu-
rated T2-weighted axial image showing a detached portion of
hyaline cartilage on the medial facet of the patella
50
The invasive nature of MR arthrography should
not unduly limit the use of this technique, which
most patients find acceptable (ROBBINS et al. 2000).
The complication rate from MR arthrography should
be very low. The risk of the introduction of infection
should, however, be borne in mind. BROSSMANN et
al. (1996) compared MR imaging, MR arthrography
with both saline and gadolinium and CT and CT
arthrography in the detection of osseous and carti-
laginous loose bodies in the knees of cadavers. MR
arthrography with gadolinium was the most accu-
rate technique, followed by MR arthrography with
saline. The best sequences for the detection of loose
bodies on both unenhanced scans and images fol-
lowing saline arthrography were a T2-weighted spin
echo sequence and a spoiled GRASS sequence.
It is very likely that, as MR software continues to
improve, the indications for MR arthrography will
decrease.
4.2.8
Patient Positioning and Imaging Planes
Images should be obtained in the sagittal, coronal
and axial planes. The menisci are principally evalu-
ated on sagittal images, supplemented by images in
the coronal plane (Fig. 4.7). Sagittal images are used
to visualise the cruciate ligaments, which are further
assessed on coronal and sometimes axial images.
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
C. Heron and A. Hine
Coronal images are essential for the evaluation of
the collateralligaments and axial images supplement
these. The patellofemoral joint is best assessed on
axial images but can also be visualised in the sagittal
plane. Traditionally, MR imaging of the knee has been
performed with the knee fully extended and posi-
tioned in 10° to 15° of external rotation. In this posi-
tion the anterior cruciate ligament is aligned paral-
lel with sagittally orientated images. This permits the
fulilength of ligament to be assessed on one sagittal
section. There are, however, different ways of achiev-
ing this aim. Positioning of sagittal images may be
determined by visualising the alignment of the ante-
rior cruciate ligament on axial or coronal images
and planning the sagittal images in this orientation
(Fig. 4.8). Alternatively, images of the anterior cruci-
ate ligament can be reformatted from three-dimen-
sional datasets. In centres using thin slice thickness-
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
the neutral position. Obliquely orientated images are
therefore not necessary.
4.2.9
Menisci
The menisci are principally composed of fibro-carti-
lage. They have few mobile protons and are of low
Magnetic Resonance Imaging
Fig. 4.8. A coronal image is used for planning an oblique sagit-
tal imaging plane along the line of the anterior cruciate liga-
ment
signal intensity on all pulse sequences. In the pres-
ence of a meniscal tear, protons become adherent to
macromolecules at the edges of the tear. As a result
of this there is some degree of lengthening of the T2
relaxation time in the tissues at the site of the tear.
Assuming that no fluid has entered the tear and
that there are no unbound protons, when using spin
echo sequences, such tears are best demonstrated on
images with short TEs (Tl-weighted or proton den-
sity). Similar signal changes are seen in the degen-
erate meniscus. Although the T2 relaxation time has
lengthened, it is still not long enough to detect a
signal on images with true T2 weighting (long TEs).
When fluid has entered the tear, unbound protons are
present and the relaxation time is similar to that of
water. The te ar may then be seen on images with T2
weighting.
Early studies established the accuracy of spin echo
sequences in the evaluation of meniscal pathology
(CRUES et al. 1987; TYRRELL et al. 1988). Meniscal
tears and the various grades of meniscal degenera-
tion may be demonstrated using Tl-weighted and
proton density spin echo sequences but the latter
are gene rally preferred as they demonstrate a great-
er SNR. Exact imaging parameters are determined
by available hardware and software. Arecent study
51
has indicated the disadvantage of the current trend to
employ increasingly shorter TEs for Tl-weighted and
proton density-weighted sequences (PEH et al. 1999).
Using a porcine model, spurious signal hyperintensi-
ty was demonstrated within normal menisci at short
TE values.
The use of spin echo sequences was challenged by
the development of gradient echo sequences with T2*
weighting. Gradient echo sequences are very sensi-
tive to the presence of meniscal degeneration or a
meniscal tear (REEDER et al. 1989; TYRRELL et al.
1988). They are, however, less sensitive than spin echo
sequences in the evaluation of other areas of the knee
and particularly in the assessment of bone marrow
oedema. Additionally, gradient echo images are sus-
ceptible to magnetic field inhomogeneities, and arte-
fact can be particularly problematic in postoperative
patients. The use of a 3D gradient echo technique
permits the acquisition of thin slices (1 to 1.5 mm)
with no inter-slice gap. Images should have reason-
ably strong T2* weighting and therefore short TEs
are recommended (Fig. 4.9).
The ability to reformat three-dimensional images
may be advantageous in meniscal assessment. Some
centres consider radial images to be useful for visual-
ising the periphery of the meniscus and the menisco-
capsular junction. This is known to be a particularly
difficult region to evaluate. In practice, this technique
is probably best reserved for use as an additional
imaging option when doubt about these regions per-
sists following assessment of images obtained in the
standard planes. The use of intravenous gadolinium
with a view to assessing the vascularity of the periph-
ery of the meniscus and improve evaluation of this
region has not proved beneficial (HAUGER et al.
2000).
Fast spin echo techniques may also be employed
for meniscal assessment. Initial studies indicated that
because resolution was inferior to spin echo sequenc-
es and there could be some blurring of margins, fast
spin echo techniques were not appropriate (JARA-
MILLO et al. 1994; LISTERUD et al. 1992). Although
doubts continued to be expressed, recent studies have
provided results which indicate that state of the art
fast spin echo images are at least comparable in accu-
racy to conventional spin echo techniques (ANDER-
SON et al. 1995; CHEUNG et al. 1997; ESCOBEDO et al.
1996; EUSTACE et al. 1999; KOWALCHUK et al. 2000;
RUBIN et al. 1994). Proton density-weighted fast spin
echo images are preferred for meniscal evaluation
and they are frequently obtained as part of a sagittal
double-echo sequence. As the addition of fat sup-
pression facilitates visualisation of ligamentous inju-
52 c. Heron and A. Hine

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

ries and bone marrow oedema as weH as meniscal 4.2.10

pathology (Fig. 4.9), it should be routinely employed
(STABLER et al. 2000; WHITE et al. 1996). Although
magnetisation transfer imaging appears to show
some benefits in assessment of the menisci (ADLER
et al. 1996), it has not been widely adopted in clinical
practice.
Ligaments
The sagittal plane is preferred for cruciate ligament
evaluation and the coronal plane for evaluation of the
coHateralligaments, but valuable information can be
obtained by imaging the cruciate and coHateralliga-
 Magnetic Resonance Imaging 53

ments in all three planes (ROYCHOWDHURY et al. 1997).
The requirement to include the fulliength of the ante-
rior cruciate ligament on a single sagittal image has
already been indicated. Several studies have investi-
gated the value of scanning the anterior cruciate liga-
ment in some degree of flexion (NAKANISHI et al.1997;
NUTSU et al. 1996, 1998; PEREIRA et al. 1998). It has
been demonstrated that this generally aids assessment
of the integrity of the ligament. As with the menisci,
conventional spin echo sequences provide reasonably
accurate assessment. Tl-weighted spin echo images
visualise the anatomy of the non-injured ligament
well, but in the presence of a joint effusion the integrity
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
the ligament and joint fluid. T2*-weighted gradient
echo images are sensitive to signal intensity changes
within the ligament resulting from injury, and they
also accurately assess the integrity of the ligament. The
anterior cruciate ligament is well visualised on fast
spin echo images (HA et al. 1998; MUNK et al. 1997).
T2-weighted fast spin echo images, particularly with
fat suppression, provide excellent contrast between
the ligament and adjacent joint fluid and are especial-
ly useful for assessing the integrity of the ligament
(Fig. 4.10). They do not appear to be as sensitive as
T2* -weighted gradient echo images for the detection
of subtle signal change within the ligament (Fig. 4.10).
Proton density-weighted fast spin echo images with
fat suppression are also useful in the assessment of
both the cruciate and collateralligaments. In princi-
pIe, inversion recovery images provide excellent con-
trast between the injured ligament, adjacent joint fluid
and soft tissues. In practice, however, this sequence is
sometimes of limited value as motion artefact from
adjacent vascular structures may impair visualisation
of the ACL. Inversion recovery images in the coronal
plane, however, are of value in the assessment of the
collateralligaments.
4.2.11
Hyaline Cartilage
Numerous researchers have investigated the value
of MR imaging in the assessment of hyaline car-
tilage abnormalities using a wide variety of pulse
sequences. The search for a technique which accu-
rately assesses hyaline cartilage has been intensified
recently by the further refinement of treatments
for hyaline cartilage disease. Surgical interventions
include shaving and repair of the hyaline cartilage
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
change in articular cartilage. Probably of greater
importance is the knowledge that a significant
number of patients presenting with internal derange-
ment of the knee will be found to have abnormalities
of hyaline cartilage as the sole cause for their symp-
toms. In addition, it is known that the outcome of
treatment for other causes of internal derangement
such as meniscal tears is influenced by the state of
the articular cartilage. Accurate information relating
to the articular cartilage is therefore of considerable
importance to many clinical decisions.
The requirement in terms of MR imaging is for a
technique which provides high resolution and sharp
contrast between hyaline cartilage and the adjacent
bone, soft tissues and fluid which may be present
within the joint. Histologically, hyaline cartilage has
three layers (a superficial, a transitional and a deeper
radial zone). Although MR images demonstrating
layers have been reported by several researchers
(DISLER et al. 1996, 2000; ERICKSON et al. 1996; ERICK-
SON and PROST 1997; FRANK et al. 1997; KNEELAND
2000; MCCAULEY and DISLER 1998; RUBENSTEIN et
al. 1993, 1997; UHL et al. 1998; WALDSCHMIDT et al.
1997; XIA et al. 1997), they do not appear to corre-
spond exacdy to the histologicallayers. As yet it is
not clear to what extent truncation arte fact and the
"magie angle" phenomenon are responsible for the
appearance of layers in MR images of hyaline carti-
lage (ERICKSON et al. 1996; FRANK et al. 1997; RUBEN-
STEIN et al. 1997; WACKER et al. 1998).
Hyaline cartilage is of low signal intensity on Tl-
weighted spin echo sequences. It is therefore indis-
tinguishable from joint fluid and this sequence is
not of value in hyaline cartilage assessment. Articu-
lar cartilage also appears as low signal intensity on
conventional T2-weighted images but in this case
joint fluid is of high signal intensity and an arthro-
gram effect is produced. This sequence, however,
suffers from inadequate differentiation between
hyaline cartilage and cortical bone and a lack of
sensitivity to hyaline cartilage defects and areas of
inhomogeneity.
Gradient echo images offer considerable advantag-
es over conventional spin echo techniques. Using a fat-
saturated T2* -weighted three-dimensional spoiled
gradient -echo sequence, hyaline cartilage is of high
signal intensity and contrasts sharply with adjacent
structures of low signal intensity (Fig.4.11a). High
sensitivity and specificity rates for the detection of
hyaline cartilage defects have been reported using
this sequence (DISLER et al. 1996; RECHT et al. 1996b;
WANG et al. 1999). The dual echo in the steady state
(DESS) sequence produces one averaged image from
C. Heron and A. Hine
the two images acquired (one from the first echo and
the other from the second echo). By optimising the
imaging parameters it is possible to produce images
which have the highest contrast between intermedi-
ate signal hyaline cartilage (from the first echo) and
high signal fluid (from the second echo) (Figs.4.11b,
4.12). Additional contrast is obtained by filtering the
second echo image to eliminate noise before adding
it to the first echo image (HARDY et al. 1996; RUEHM
et al. 1998). The sequence is not,however,provided by
all MR manufacturers.
Fat-saturated proton density and T2-weighted fast
spin echo images have been used for hyaline carti-
lage imaging (KAWAHARA et al. 1998; BREDELLA
et al. 1999). The results using these sequences are
comparable to those employing gradient echo and
DESS sequences with intrasubstance cartilage damage
shown more clearly (Fig. 4.11). The abilityoffast spin
echo techniques to visualise hyaline cartilage better
than conventional spin echo techniques is thought to
be due to fast spin echo images providing magnetisa-
ti on transfer contrast in addition to T2 contrast (YAO
et al. 1996).
As has already been indicated, MR arthrography
enhances hyaline cartilage visualisation. Gadolini-
um-chelate contrast agents have been used intrave-
nously and injected direcdy into the knee to identify
regions of hyaline cartilage degeneration (BASHIR
et al. 1997). These agents diffuse into the surface of
cartilage in which the proteoglycan matrix is disrupt-
ed but the cartilage is grossly intact. This has been
reported to enhance visualisation of early degenera-
tive changes on MR images.
There have been several reports of the measure-
ment of hyaline cartilage volume using MR imaging
(Dupuy et al. 1996; ECKSTEIN et al. 1996, 1998,2000;
HYHLIK-DuRR et al. 2000; KSHIRSAGAR et al. 1998;
PIPLANI et al. 1996; SITTEK et al. 1996; STAMMBERGER
et al. 1999). The present techniques are labour inten-
sive as they require predominandy manual delinea-
tion of hyaline cartilage and it seems unlikely that
they will be widely adopted until the process can be
automated. In general, the measurements appear to
be accurate and reproducible.
Studies using short echo time projection recon-
struction MR imaging, and measurements of T2
relaxation time of articular cartilage, have demon-
strated T2 lengthening at sites of early cartilage
degeneration (BROSSMANN et al. 1997; GOLD et al.
1998; MOSHER et al. 2000). Their accuracy in hyaline
cartilage assessment has been shown to be superior
to that of fat-suppressed three-dimensional spoiled
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
shown promise in the evaluation of hyaline cartilage
(BURSTEIN et al. 1993; FRANK et al. 1999). As sodium
nuclei are abundant in articular cartilage, sodium-
based MR imaging has been employed. The sodium
nuclei attach to proteoglycan macromolecules. These
are reduced in degenerative disease, and sodium-
based MR studies have been used to detect early
degenerative change (REDDY et al. 1998). The poor
SNR of these images means that they are unlikely
to have clinical applications. Magnetisation transfer
techniques have been described above. It was thought
that the presence of increased unbound water pro-
tons in areas of degenerate hyaline cartilage would be
visualised using this technique. The present results
suggest that magnetisation transfer images are not
comparable to other sequences for the evaluation of
hyaline cartilage (ADLER et al. 1996).
56
Fig.4.12. A DESS sagittal image permits differentiation
between the intermediate signal intensity hyaline cartilage on
the medial femoral condyle and tibial condyle, and the high
signal intensity joint fluid
4.2.12
Bone
The spectrum ofbony abnormalities which affect the
knee ranges from common findings such as degener-
ative change and bone bruising due to trauma, to rare
primary bone tumours. The majority of bony prob-
lems generate an increased signal intensity on fat-
suppressed images, and it is therefore important to
include images with fat suppression in at least one
plane. Fast STIR and T2-weighted images with fat
suppression are comparable to conventional sequenc-
es (ARNDT et al. 1996). In fractures and avascular
necrosis the lesion can be further evaluated using Tl-
weighted spin echo sequences. If there is difficuhy
determining whether alesion is solid or cystic, Tl-
weighted spin echo sequences, supplemented if nec-
essary by Tl-weighted sequences following intrave-
nous gadolinium, are employed.
C. Heron and A. Hine
4.3
MR Protocol for
Routine Examination of the Knee
The following protocol permits adequate assessment
of the knee and may be supplemented by additional
sequences for specific clinical indications which are
outlined below. Imaging should be performed in all
three planes. The parameters are chosen to achieve
high-resolution images with a matrix size of, for
example, 512x256, a field of view of 14-16 cm and
slice thicknesses of between 1.5 and 3 mm. Since
the most important requirement is accurate assess-
ment of the menisci, images in the sagittal and coro-
nal planes, which maximise visualisation of menis-
cal tears, should be chosen. Possible sequences for
this purpose include a T2* gradient echo sequence,
proton density or preferably proton density with fat
saturation. The cruciate and collateralligaments can
also be adequately assessed on these sequences. A
STIR sequence or T2-weighted sequence with fat sat-
uration is routinely obtained if the technique chosen
for meniscal assessment has not included images with
fat suppression. In the sagittal and coronal planes,
images to visualise hyaline cartilage may be added
depending on local preferences. The axial sequence
should be chosen to provide optimal visualisation of
hyaline cartilage.
4.4
MR Protocols for
Specific Clinical Problems
4.4.1
Synovium and Soft Tissue Masses
Normal synovium is not visible on non-contrast-
enhanced MR images, while the thickened inflamed
synovium enhances following intravenous gadolini-
um DTPA. Thickened synovium can sometimes be
appreciated on non-enhanced images but is not reli-
ably demonstrated (RAND et al. 1999a). Synovial
thickening and synovial plicae mayaIso be more con-
spicuous following MR arthrography as the joint dis-
tension facilitates visualisation of synovial surfaces.
Soft tissue masses should be evaluated with a range of
Tl-weighted and T2-weighted images with the addi-
tion of fat suppression in order to permit char-
acterisation and determine their extent. Tl-weight-
ed images following intravenous contrast determine
enhancement characteristics.
Magnetic Resonance Imaging
4.4.2
Femoral Trochlear Dysplasia and
Patellar Tracking Studies
Trochlear dysplasia and patellar tracking abnormali-
ties are important causes of anterior knee pain. The
lack of ionising radiation, cross-sectional imaging
capability and potential for kinematic studies mean
that MR imaging is an ideal technique for assessing
these abnormalities. Dysplasia of the femoral troch-
lea may be assessed by using mid-sagittal and trans-
verse MR images 3 cm above the femorotibial joint
space (PF1RRMANN et al. 2000).
In assessing patellar tracking abnormalities the
requirement is to visualise the alignment of the patella
in the early stages of knee flexion (up to 30° or 45°).
Ideally, flexion should be observed and recorded in a
real time setting with the knee under normal load. MR
imagers that have vertically orientated coils ("dough-
nut" magnets) offer potential for this type of physiologi-
cal examination and for acquiring images during weight
bearing. Several MR studies employing ratchet systems
with various loading devices and increasingly fast pulse
sequences have approximated these conditions reason-
ably successfully (MUHLE et al. 1996; POWERS et al.
1998). Most recently, motion-triggered ultra-fast MR
imaging and CT scanning have been assessed (MUHLE
et al. 1999b). MR imaging has been advocated for the
demonstration of the iliotibial band syndrome (EKMAN
et al. 1994; N1SH1MURA et al. 1997; MUHLE et al. 1999a).
Imaging of the knee in some degree of flexion appears
to provide optimal evaluation.
4.4.3
Tendons
MR imaging may be undertaken to evaluate acute
injuries or chronic conditions arising in tendons
around the knee joint. The patellar tendon followed
bythe quadriceps insertion are most frequently inves-
tigated. If the symptoms are thought to be localised
to the tendon, initial imaging using ultrasound is rec-
ommended. MR imaging may be employed for fur-
ther evaluation or may detect abnormalities which
have not been suspected clinically. A 3-inch surface
coil positioned directly over the tendon and the use
of a high-field system provide optimal visualisation.
Images are obtained in the sagittal and axial planes. A
12-cm field of view is recommended. Preferred pro-
tocols include Tl-weighted spin echo sequences,
T2-weighted fast spin echo imaging with fat satura-
tion, T2* -weighted gradient echo imaging and STIR
57
sequences. Slice thickness and matrix size are opti-
mised to obtain high-resolution images.
4.4.4
Posterolateral and
Posteromedial Aspects of the Knee
Special imaging protocols have been recommended
for the detailed study of the anatomy and pathology
which arise in the posterolateral and posteromedial
aspects of the knee.
In the posterolateral aspect of the knee, the proxi-
mal popliteus tendon is orientated obliquely, and it
has been demonstrated that coronal oblique imaging
along the plane of the tendon provides superior visu-
alisation of the anatomical structures in this region
(Yu et al. 1996). Using this method the popliteus
tendon is visualised on sagittal oblique images (par-
allel with the anterior cruciate ligament), and the
coronal oblique images are positioned parallel with
the tendon. In the aforementioned study, proton den-
sity and T2-weighted coronal oblique images were
obtained in this plane.
Coronal oblique images have also been shown
to be of value in the assessment of ligamentous
structures in the posteromedial aspect of the knee
(LOREDO et al. 1999). The posterior oblique ligament
is an important stabilising structure, and it is best
appreciated on a combination ofaxial images and
coronal oblique images which have been angled at
25° to vertical.
4.4.5
MR Imaging of the Postoperative Knee
The most frequent indications for performing post-
operative MR imaging in the knee are to assess the
menisci following previous resection or repair and the
anterior cruciate ligament following reconstruction.
Whatever the nature of the previous surgery to the
knee, there is a likelihood of metal artefact on subse-
quent MR images. For this reason, spin echo and fast
spin echo images are generally preferred to gradient
echo images. In the assessment of the postoperative
meniscus, proton density-weighted and T2-weighted
fast spin echo images with fat saturation produce ade-
quate visualisation (L1M et al. 1999). The postoper-
ative meniscus may be difficult to assess as the crite-
ria for the diagnosis of a meniscal tear in an unoper-
ated meniscus no longer apply. When doubt exists, MR
arthrography may be of value (SCIULLI et al. 1999).
58
MR imaging of the anterior cruciate ligament fol-
lowing reconstruction is employed to determine the
integrity of the ligament (IRIZARRY and RECHT 1997;
SCHATZ et al.1997). It is also of value to assess wheth-
er the ligament is optimally positioned and whether
there is impingement on the ligament in addition to
the appearances of the intercondylar region follow-
ing notchplasty (MAY et al. 1997; RECHT et al. 1996a;
TOMCZAK et al. 1997). Furthermore, it may reveal
other complications such as a cyelops lesion (BRAD-
LEY et al. 2000) or fibrosis. Images in the sagittal,
coronal and axial planes (MURAKAMI et al. 1998)
employing Tl-weighted and fast spin echo proton
density- and T2-weighted images with fat saturation
can be used to address questions relating to the
position and integrity of the ligament. Tl-weighted
images with fat saturation following intravenous gad-
olinium are used to assess soft tissue problems such
as cyelops lesions and fibrosis, and to obtain infor-
mation about the degree of revascularisation of the
reconstructed ligament.
4.5
Artefacts
Artefacts may be broadly divided into two categories.
There are those resulting from the inherent nature of
the MR scanning technique and those which relate to
patient factors such as movement.
Chemieal shift artefact is one of the better known
artefacts resulting from the physical characteristics of
MR imaging and it is more conspicuous at high field
strengths (HOOD et al. 1999). In spin echo sequences,
chemical shift artefact arises because hydrogen pro-
tons within fat precess at a different frequency from
hydrogen protons within water. The hydrogen pro-
tons in fat are surrounded by electrons which cause
the proton to precess at a lower frequency than water-
based hydrogen protons. The precessional frequency
determines the position that the proton is assigned
along the frequency-encoding gradient. The assigned
position along the frequency-encoding gradient
therefore cannot be an accurate spatial representa-
tion at the site of fat/water interfaces. Chemical shift
artefact results in black and white lines appearing at
the interface.
Gradient echo images produce an alternative type
of chemical shift artefact. It also arises as a result of
the differences in precessional frequency for protons
embedded in fat compared with those in water. Imme-
diately after excitation the protons precess together
c. Heron and A. Hine
and will be pointing in the same direction. With the
increasing lapse of time following excitation, this dif-
ference in precession results in the two protons being
out of phase. They subsequently return to being in
phase and, depending upon the TE employed, imag-
ing may be undertaken when they are either in or out
ofphase.
In the case of a voxel which contains both fat and
water (at a fat/water interface), out of phase protons
will partially cancel each other and result in a low
signal intensity line. In-phase protons within a simi-
lar voxel result in a high signal intensity line. These
differences may be exploited to enhance the visual-
isation of bone marrow disease using out of phase
images. Chemical shift artefact can be minimised by
the application of fat suppression. Alternatively, selec-
tive water excitation techniques may be employed but
are not available on all systems.
The magie angle phenomenon arises in structures
which are composed of ordered parallel fibres. In the
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
the T2 dephasing. However, when the fibres are posi-
tioned at 55° to the main magnetic field, the acceler-
ated loss of T2 magnetisation is decreased and struc-
tures which are usually of low signal intensity dem-
onstrate an increase in signal (Fig.4.13) (ECHIGO et
al. 1999; OLEAGA and KRESSEL 1990; PETERFY et al.
1994). The effect is maximal on images with relatively
short TEs.
The truneation artefaet can arise at sites where
there is a significant contrast difference between two
adjacent structures. In the knee this applies at the
boundary between hyaline cartilage and the menisci
(tuRNER et al. 1991). As a result of Fourier transfor-
mation of acquired data, the boundary is not accu-
rately represented and aseries of alternating high
and low signal intensity lines are seen parallel to the
boundary. These artefacts are most marked when
a low phase encoding matrix is used in the supe-
ro-inferior direction. Increasing the phase encoding
matrix and applying it in the anteroposterior direc-
tion should resolve uncertainty.
A further artefact which can simulate a meniscal
tear or loose bodywithin a joint is the vaeuum phenom-
enon. When there is a small pocket of air between the
articular surfaces this is seen as a region of low signal
intensity. On gradient echo images the low signal focus
increases in size due to blooming but it appears smaller
on spin echo images. Further scanning following repo-
sitioning which redistributes the gas should remove
 Magnetie Resonance Imaging 59

Fig. 4.13a, b. Two sagittal

images demonstrating
the posterior cruciate lig-
ament. The magie angle
phenomenon is more
obvious on the short TE
gradient echo sequence
(a) than on the STIR
sequence (b)

.AIo_ _ _ _... b

the artefact. Inadvertent injection of air bubbles at MR
arthrography will generate similar artefacts.
In the past "wrap around" artefact could be prob-
lematic in knee MR imaging. This artefact arises
when signal from tissues outside of the imaging
region are assigned a point within it. Wrap around
arte fact has now been virtually eliminated by the
use of local surface coils and also by the use of
a technique known as "over sampling". The latter
involves the acquisition of data from a large area
but presentation of data from only the region of
interest with resultant elimination of wrap around
artefact.
Metal objects in the knee produce their own local
magnetic fields and may therefore markedly distort the
MR image. Even small metal fragments resulting from
bone drilling can cause significant artefact which is
more pronounced on gradient echo images (Fig. 4.14).
Artefacts from patient motion (Fig.4.15) may be
minimised by positioning the patient comfortably and
applying padding within rigid coils in order that the
potential for movement is reduced. Artefact from vas-
cular structures, for example in the popliteal fossa,
may be troublesome but altering the frequency and
phase encoding directions can ensure that regions of
particular interest are not obscured (Fig. 4.16).

Fig. 4.14. a A gradient

echo sequence showing
an anterior cruciate
ligament reconstruction
with artefact related to
metal at the proximal
and distal ends of the
reconstruction. b A spin
echo TI -weighted
sequence demonstrates
considerably less artefact
with clear delineation of
the reconstructed liga-
ment
 60 c. Heron and A. Hine

<l Fig.4.15. Gradient echo coronal image demonstrating alternating linear

regions of high and low signal intensity in the periphery of the medial
meniscus (arrow). These appearances are due to motion artefact. Their
artefactual nature can be appreciated because the low signal intensity lines
extend across the hyaline cartilage. Similar less marked changes are seen
in the region of the periphery of the lateral meniscus

Fig. 4.16. a Vascular artefact from popliteal vessels

partially obscures the patellar hyaline cartilage. b
In a different patient this is avoided by chan ging
the phase encoding direction such that the artefact
is horizontally rather than vertically positioned
v

References Anderson M, Raghaven N, Seidenwurm D et al (1995) Eval-

Adler RS, Swanson SD, Doi K et al (1996) Effect of magneti-
zation transfer in meniscal fibrocartilage. Magn Res Med
35:591-595
Ahn JM, Kwak SM, Kang HS et al (1998) Evaluation of patellar
cartilage in cadavers with a low-field-strength extremity-
only magnet: comparison of MR imaging sequences, with
macroscopic findings as the standard. Radiology 208:57-62
uation of meniscal tears; fast spin-echo versus conven-
tional spin-echo magnetic resonance imaging. Acad Radiol
2:209-214
Arndt WF III, Truax AL, Barnett FM et al (1996) MR diagnosis
of bone contusions of the knee: comparison of coronal T2-
weighted fast spin-echo with fat saturation and fast spin-
echo STIR images with conventional STIR images. AJR
166:119-124
Bashir A, Gray ML, Boutin RD et al (1997) Glycosaminoglycan
Magnetie Resonance lmaging
in artieular cartilage; in vivo assessment with delayed
Gd(DTPA)-enhanced imaging. Radiology 205:551-558
Bradley DM, Bergman AG, Dillingham MF (2000) MR imaging
of cyelops lesions.AJR 174:719-726
Bredella MA, Tirman PF, Peterfy CG et al (1999) Accuracy of
T2-weighted fast spin-echo MR imaging with fat saturation
in detecting cartilage defects in the knee: comparison with
arthroscopy in 130 patients. AJR 172: 1073-1 080
Brossmann J, Preidler K-W, Daenen B et al (1996) lmaging
of osseous and cartilaginous intraartieular bodies in the
knee: comparison of MR imaging and MR arthrography
with CT and CT arthrography in cadavers. Radiology
200:509-517
Brossmann J, Frank LR, Pauly JM et al (1997) Short echo time
projection reconstruction MR imaging of cartilage: com-
parison with fat -suppressed spoiled GRASS and magnetiza-
tion transfer contrast MR imaging. Radiology 203:501-507
Burstein D, Gray ML, Hartman AL et al (1993) Diffusion of
small solutes in cartilage as measured by nuelear magnetie
resonance (NMR) spectroscopy and imaging. J Orthop Res
11:465-478
Cheung LP, Li KC, Hollett MD et al (1997) Meniscal tears of the
knee: accuracy of detection with fast spin-echo MR imag-
ing and arthroscopie correlation in 293 patients. Radiology
203:508-512
Crues JV III, Mink J, Levy TL et al (1987) Meniscal tears of the
knee: accuracy of MR imaging. Radiology 164:445-448
Delfaut EM, Beltran J, Johnson G et al (1999) Fat suppression
in MR imaging: techniques and pitfalls. Radiographies
19:373-382
Disler DG, McCauley TR, Kelman CG et al (1996) Fat-sup-
pressed three-dimensional spoiled gradient -echo MR imag-
ing ofhyaline cartilage defects in the knee: comparison with
standard MR imaging and arthroscopy AJR 167:127-132
Disler DG, McCauley TR, Ratner LM et al (1997) In-phase
and out-of-phase MR imaging ofbone marrow: predietion
of neoplasia based on the detection of co existent fat and
water. AJR 169: 1439-1447
Disler DG, Recht MP, McCauley TR (2000) MR imaging of
artieular cartilage. Skeletal Radio129:367-377
Dupuy DE, Spillane RM, Rosol MS et al (1996) Quantification
of artieular cartilage in the knee with three-dimensional
MR imaging. Acad Radio13:919-924
Echigo J, Yoshioka H, Takahashi H et al (1999) Signal intensity
changes in anterior cruciate ligament autografts: relation
to magnetic field orientation. Acad Radio16:206-210
Eckstein F, Gavazzeni A, Sittek H et al (1996) Determination
of knee joint cartilage thickness using three-dimensional
magnetie resonance chondro-crassometry (3D MR-CCM).
Magn Res Med 36:256-265
Eckstein F, Westhoff J, Sittek H et al (1998) In vivo reproduc-
ibility of three-dimensional cartilage volume and thiekness
measurements with MR imaging. AJR 170:593-597
Eckstein F, Stammberger T, Priebsch J et al (2000) Effect of
gradient and section orientation on quantitative analysis of
knee joint cartilage. J Magn Reson lmaging 11:161-167
Ekman EF, Pope T, Martin DF et al (1994) Magnetie resonance
imaging of iliotibial band syndrome. Am J Sports Med
22:851-854
Eriekson SJ, Prost RW (1997) Laminar structures on MR
images of artieular cartilage: reply. Radiology 204:16-18
Eriekson SJ, Waldschmidt JG, Czervionke LF et al (1996) Hya-
line cartilage: truncation artifact as a cause of trilaminar
61
appearance with fat -suppressed three-dimensional spoiled
gradient-recalled sequences. Radiology 201:260-264
Escobedo EM, Hunter JC, Zink-Brody GC et al (1996) Useful-
ness of turbo spin-echo MR imaging in the evaluation of
meniscal tears: comparison with a conventional spin-echo
sequence. AJR 167:1223-1227
Eustace S, Hentzen P,Adams J et al (1999) Comparison of con-
ventional and turbo spin-echo Tl-weighted MR imaging in
acute knee trauma. AJR 172: 1393-1395
Frank LR, Brossmann J, Buxton RB et al (1997) MR imaging
truncation artifacts can create a false laminar appearance
in cartilage. AJR 168:547-554
Frank LR, Wong EC, Luh W-M et al (1999) Artieular cartilage
in the knee: mapping of the physiologie parameters at MR
imaging with a local gradient coil - preliminary results.
Radiology 210:241-246
Gold GE, Thedens DR, Pauly JM et al (1998) MR imaging of
artieular cartilage of the knee: new methods using ultra-
short TEs. AJR 170: 1223-1226
Grainger AJ, Elliott JM, Campbell RSD et al (2000) Direct
MR arthrography: a review of current use. Clin Radiol
55:163-176
Ha TPT,Li KCP, Beaulieu CF et al (1998) Anterior cruciate liga-
ment injury: fast spin-echo MR imaging with arthroscopie
correlation in 217 examinations.AJR 170:1215-1219
Hardy PA, Recht MP, Piraino D et al (1996) Optimisation of a dual
echo in the steady state (DESS) free-precession sequence for
imaging cartilage. J Magn Reson Imaging 6:329-335
Hauger 0, Frank LR, Boutin RD et al (2000) Characterization
of the "Red Zone" of knee meniscus: MR imaging and his-
tologie correlation. Radiology 217: 193-200
Hood MN, Ho VB, Smirniotopoulos JG et al (1999) Chemieal
shift: the artifact and elinieal tool revisited. RadioGraphs
19:357-371
Hyhlik-Durr A, Faber S, Burgkart R et al (2000) Precision of
tibial cartilage morphometry with a coronal water-excita-
tion MR sequence. Eur Radioll0:297-303
lrizarry JM, Recht MP (1997) MR imaging of the knee liga-
ments and the postoperative knee. Radiol Clin North Am
35:45-76
Jaramillo D, Laor T, Mulkern R (1994) Comparison between fast
spin-echo and conventional spin-echo imaging of normal
and abnormal musculoskeletal structures in children and
young adults. luvest Radio129:803-811
Kawahara Y, Uetani M, Nakahara N et al (1998) Fast spin-echo
MR of the articular cartilage in the osteoarthrotic knee.
Correlation of MR and arthroscopie findings. Acta Radiol
39:120-125
Kneeland JB (2000) Magnetie resonance imaging of articular
cartilage. Semin Roentgeno135:249-255
Koskinen SK, Yla-Outinen H, Aho HJ et al (1997) Magnetiza-
tion transfer and spin lock MR imaging of patellar cartilage
degeneration at 0.1 T.Acta Radio138:1071-1075
Kowalchuk RM, Kneeland JB, Dalinka MK et al (2000) MRI
of the knee: value of short echo time fast spin-echo using
high performance gradients versus conventional spin-echo
imaging for the detection of meniscal tears. Skeletal Radiol
29:520-524
Kshirsagar AA, Watson PJ, Tyler JA et al (1998) Measurement
of localized cartilage volume and thiekness of human knee
joints by computer analysis of three-dimensional magnetie
resonance images. lnvest Radio133:289-299
Lim PS, Schweitzer ME, Bhatia M et al (1999) Repeat tear of
62
postoperative meniscus: potential MR imaging signs. Radi-
ology 210:183-188
Listerud J, Einstein S, Outwater E et al (1992) First principles
of fast spin echo. Magn Reson Q 8:199-244
Loredo R, Hodler J, Pedowitz R et al (1999) Posteromedial
corner of the knee: MR imaging with gross anatomie cor-
relation. Skeletal RadioI28:305-311
Masciocchi C, Barile A, Satragno L (2000) Musculoskeletal
MRI: dedieated systems. Eur RadioI1O:250-255
May DA, Snearly WN, Bents R et al (1997) MR imaging findings
in anterior cruciate ligament reconstruction: evaluation of
notchplasty.AJR 169:217-222
McCauley TR, Disler DG (1998) MR imaging of artieular car-
tilage. Radiology 209:629-640
Mosher TJ, Dardzinski BJ, Smith MB (2000) Human artieular
cartilage: influence of aging and early symptomatie degen-
eration on the spatial variation of T2 - preliminary findings
at 3 T. Radiology 214:259-266
Muhle C, Brossmann J, Heller M (1996) Kinematie MRI of
the knee using a specially designed positioning deviee. J
Comput Assist Tomogr 20:522-525
Muhle C, Ahn JM, Yeh L et al (1999a) Iliotibial band frietion
syndrome: MR imaging findings in 16 patients and MR
arthrographie study of six cadaverie knees. Radiology
212:103-110
Muhle C, Brossmann J, Heller M (1999b) Kinematie CT and MR
imaging of the patellofemoral joint. Eur RadioI9:508-518
Munk PL, Hilborn MD, VelletAD et al (1997) Diagnostie equiv-
alence of conventional and fast spin echo magnetie res-
onance imaging of the anterior cruciate ligament of the
knee. Aust RadioI4l:238-242
Murakami Y, Sumen Y, Ochi M et al (1998) MR evaluation
of human anterior cruciate ligament autograft on oblique
axial imaging. J Comput Assist Tomogr 22:270-275
Nakanishi K, Horibe S, Shiozaki Y et al (1997) MRI of normal
anterior cruciate ligament (ACL) and reconstructed ACL:
comparison of when the knee is extended with when the
knee is flexed. Eur Radiol 7:1020-1024
Nishimura G, Yamato M, Tamai K et al (1997) MR findings in
iliotibial band syndrome. Skeletal RadioI26:533-537
Niitsu M, Ikeda K, Fukubayashi T et al (1996) Knee extension
and flexion: MR delineation of normal and torn anterior
cruciate ligaments. J Comput Assist Tomogr 20:322-327
Niitsu M, Ikeda K, Itai Y (1998) Slightly flexed knee position
within a standard knee coil: MR delineation of the anterior
cruciate ligament. Eur RadioI8:113-115
Niitsu M, Nakai T, Ikeda K et al (2000) High-resolution MR
imaging of the knee at 3 T. Acta Radiol4l :84-88
Oleaga L, Kressel H (1990) High signal intensity in the pos-
terior cruciate ligament on sagittal gradient -echo images:
normal variant. Radiology 177:295
Peh WC, Cassar-Pullicino VN (1999) Magnetie resonance
arthrography: current status. Clin RadioI54:575-587
Peh WC, Chan JH, Shek TW et al (1999) The effect of using
shorter echo times in MR imaging of knee menisci: a study
using a porcine model. AJR 172:485-488
Pereira ER, Ryu KN, Ahn JM et al (1998) Evaluation of
the anterior cruciate ligament of the knee: comparison
between partial flexion true sagittal and extension sag-
ittal oblique positions during MR imaging. Clin Radiol
53:574-578
Peterfy CG, Janzen DL, Tirman PFJ et al (1994) 'Magic angle'
phenomenon. A cause of increased signal in the normal
C. Heron and A. Hine
lateral meniscus on short -TE MR images of the knee. AJR
163:149-154
Peterson DM, Carruthers CE, Wolverton BL et al (1999) Appli-
cation of a birdcage coil at 3 Tesla to imaging of the human
knee using MRI. Magn Reson Med 42:215-221
Pfirrmann CW, Zanetti M, Romero J et al (2000) Femoral troch-
lear dysplasia: MR findings. Radiology 216:858-864
Piplani MA, Disler DG, McCauley TR et al (1996) Artieular
cartilage volume in the knee: semiautomated determina-
tion from three-dimensional reformations of MR images.
Radiology 198:855-859
Powers CM, Shellock FG, Pfaff M (1998) Quantification of
patellar tracking using kinematie MRI. J Magn Reson Imag-
ing 8:724-732
Rand T, Imhof H, Czerny C et al (1999a) Discrimination
between fluid, synovium, and cartilage in patients with
rheumatoid arthritis: contrast enhanced spin echo versus
non-contrast-enhanced fat-suppressed gradient echo MR
imaging. Clin Radiol 54: 107-11 0
Rand T, Imhof H, Turetschek K et al (1999b) Comparison of
low field (0.2 T) and high field (1.5 T) MR imaging in the
differentiation of torned from intact menisci. Eur J Radiol
30:22-27
Recht MP, Piraino DW, Applegate G et al (1996a) Complica-
tions after anterior cruciate ligament reconstruction: radio-
graphie and MR findings. AJR 167:705-710
Recht MP, Piraino DW, Paletta GA et al (1996b) Accuracy
of fat-suppressed three-dimensional spoiled gradient-echo
FLASH MR imaging in the detection of patellofemoral
artieular cartilage abnormalities. Radiology 198:209-212
Reddy R, Insko E, Noyszewski EA et al (1998) Sodium MRI
of human artieular cartilage in vivo. Magn Reson Med
39:697-701
Reeder JD, Matz SO, Becker L et al (1989) MR imaging of
the knee in the sagittal projection: comparison of three-
dimensional gradient-echo and spin-echo sequences. AJR
153:537-540
Robbins MI, Anzilotti KF Jr, Katz LD et al (2000) Patient
perception of magnetie resonance arthrography. Skeletal
RadioI29:265-269
Roychowdhury S, Fitzgerald SW, Sonin AH et al (1997) Using
MR imaging to diagnose partial tears of the anterior cruci-
ate ligament: value ofaxial images.AJR 168:1487-1491
Rubenstein JD, Kim JK, Morava-Protzner let al (1993) Effects
of collagen orientation on MR imaging characteristies of
bovine artieular cartilage. Radiology 188:219-226
Rubenstein J, Recht M, Disler DG et al (1997) Laminar struc-
tures on MR images of articular cartilage. Radiology
204:15-16
Rubin DA, Kneeland JB, Listerud J et al (1994) MR diagnosis of
meniscal tears of the knee: value offast spin-echo vs con-
ventional spin-echo pulse sequences. AJR 162:1131-1135
Ruehm S, Zanetti M, Romero J et al (1998) MRI of patellar artie-
ular cartilage: evaluation of an optimized gradient echo
sequence (3D-DESS). J Magn Reson Imaging 8:1246-1251
Schatz JA, Potter HG, Rodeo SA et al (1997) MR imaging of ante-
rior cruciate ligament reconstruction. AJR 169:223-228
Sciulli RL, Boutin RD, Brown RR et al (1999) Evaluation of
the postoperative meniscus of the knee: a study compar-
ing conventional arthrography, conventional MR imaging,
MR arthrography with iodinated contrast material, and
MR arthrography with gadolinium-based contrast mate-
rial. Skeletal RadioI28:508-514
Magnetic Resonance Imaging
Sittek H, Eckstein F, Gavazzeni A et al (1996) Assessment of
normal patellar cartilage volume and thiekness using MRI:
analysis of currently available pulse sequences. Skeletal
RadioI25:55-62
Stabler A, Glaser C, Reiser M (2000) Musculoskeletal MR: knee.
Eur Radioll0:230-241
Stammberger T, Eckstein F, Englmeier KH et al (1999) Deter-
mination of 3D cartilage thiekness data from MR imaging:
computational method and reproducibility in the living.
Magn Reson Med 41:529-536
Suh JS, Cho JH, Shin KH et al (1996) Chondromalacia of the
knee: evaluation with a fat-suppression three-dimensional
SPGR imaging after intravenous contrast injection. J Magn
Reson Imaging 6:884-888
Tomczak RJ, Hehl G, Mergo PJ et al (1997) Tunnel placement
in anterior cruciate ligament reconstruction: MRI analysis
as an important factor in the radiologieal report. Skeletal
RadioI26:409-413
Turner DA, Rapoport MI, Erwin WD, McGould M et al (1991)
Truncation artefact: a potential pitfall in MR imaging of the
menisci of the knee. Radiology 179:629-633
Tyrrell RL, Gluckert K, Pathria M et al (1988) Fast three-dimen-
sional MR imaging of the knee: comparison with arthros-
copy. Radiology 166:865-872
Uhl M, Ihling C, Allmann KH et al (1998) Human artieular
cartilage: in vitro correlation of MRI and histologie find-
ings. Eur RadioI8:1123-1129
63
Wacker FK, Bolze X, Felsenberg D et al (1998) Orientation-
dependent changes in MR signal intensity of artieular car-
tilage: manifestation of the "magie angle" effect. Skeletal
RadioI27:306-310
Waldschmidt JG, Rilling RJ, Kajdacsy-Balla AA (1997) In vitro
and in vivo MR imaging of hyaline cartilage: zonal anat-
omy, imaging pitfalls, and pathologie conditions. Radio-
graphies 17:1387-1402
Wang SF, Cheng HC, Chang CY (1999) Fat-suppressed three-
dimensional fast spoiled gradient-recalled echo imaging:
a modified FS 3D SPGR technique for assessment of patel-
lofemoral joint chondromalacia. Clin Imaging 23: 177 -180
White LM, Schweitzer ME, Johnson WJ et al (1996) Role of T2-
weighted fast-spin-echo imaging in the diagnosis of menis-
cal tears. J Magn Reson Imaging 6:874-877
Wieslander SB, Rappeport ED, Lausten GS et al (1998) Multi-
planar reconstruction in MR imaging of the knee. Compar-
ison with standard sagittal and coronal images. Acta Radiol
39:116-119
Xia Y,Farquhar T,Burton-Wurster N et al (1997) Origin of car-
tilage laminae in MRI. J Magn Reson Imaging 7:887-894
Yao L, Gentili A, Thomas A (1996) Incidental magnetization
transfer contrast in fast spin-echo imaging of cartilage. J
Magn Reson Imaging 6:180-184
Yu JS, Salonen DC, Hodler J et al (1996) Posterolateral aspect
of the knee: improved MR imaging with a coronal oblique
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
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
thigh and is formed by the vastus lateralis, the vastus
medialis (whieh are respectively located externally
and internaIly), the superfidally located rectus femo-
ris and the vastus intermedius (deeply located in
the midline of the thigh). The vastus lateralis and
medialis join in the midline to insert in a common
tendon. The rectus femoris inserts in a thin superfi-
dal tendon while the vastus intermedius inserts in
a deep thiek tendon. The quadriceps tendon is then
formed by the overlaying of three different tendons:
the superfidal tendon of the rectus femoris, the inter-
mediate common tendon of the vastus lateralis and
medialis, and the deep tendon from the vastus inter-
medius. This multilayered appearance is readily evi-
dent at US and can be weIl imaged on longitudinal
sonograms (BIANCHI et al. 1994). While the superfi-
dal tendon is always weIl seen, differentiation of the
intermediate and the deep tendon can be more dif-
ficult. The possibility of evaluating the internal struc-
ture of the quadriceps tendon to discriminate the
different components has practical value since this
allows differentiation between complete tears, which
are treated by surgery, and partial tears involving
only one or two tendons, whieh may be amenable
to conservative treatment. Transverse sonograms can
help in assessing the extension of the tear in the coro-
nal plane.
The suprapatellar synovial recess or suprapatellar
bursa lies posterior to the quadriceps tendon and
suprapatellar fat pad and anterior to the prefemoral
fat. In uterine life the synovial recess is completely
separated from the artieular cavity by a septum. A
perforation of the septum normally occurs at the end
of the 5th foetal month and allows free communiea-
tion between the two spaces. In a small percentage of
cases, incomplete resorption of the septum leads to
67
the presence of the so-called suprapatellar synovial
plica, whieh, if an effusion is present, can be imaged
at USo On longitudinal sonograms the plica appears as
an oblique curvilinear structure showing a posterior-
inferior concavity. The plica, which rarely has clinieal
relevance, must be differentiated from a postinflam-
matory fibrous septum. In normal conditions, on both
longitudinal and transverse sonograms the suprapa-
tellar recess appears as a hypoechoie structure owing
to the overlap of the anterior and posterior synovial
membrane. A variable amount of fluid can be appre-
dated on US as an anechoie area located inside the
bursa. US detection of intra-articular fluid appears
to be more accurate and reproducible than clinieal
detection (HAUZEUR et al. 1999). Care must be taken
when evaluating the presence of a joint effusion by
USo Small effusions can be detected on suprapatellar
longitudinal sonograms only at dynamie examination
obtained during isometrie contraction of the quadri-
ceps. This can be explained by the cranial displace-
ment of the fluid induced by displacement of Hoffa's
fat pad against the femoral condyles and by tighten-
ing of the posterior fasda. Additionally, when fluid
is present in very small amounts, it can be imaged
only on transverse sonograms in the lateral or medial
pouch of the suprapatellar recess. It must be stressed,
however, that a modest amount of intra-artieular fluid
is normal in asymptomatie subjects and that corre-
lation with clinieal data and with the contralateral
knee is mandatory. When analysis of synovial fluid is
essential, for example to rule out an infection or gout,
US-guided aspiration can be easily performed (WANG
SC et al. 1999). Utilisation ofUS for this purpose leads
to careful selection of the site of puncture, particu-
larly in small effusions, and confirms the intra-artic-
ular positioning of the needle tip. Usually US-guided
arthrocentesis is less painful than blind puncture.
 68
The suprapatellar fat pad is located cranial to the
patellar upper pole, posterior to the quadriceps tendon
and anterior to the suprapatellar synovial recess. On
longitudinal sonograms it appears as a triangular
hyperechoic structure. When examining the post-trau-
matic knee, the suprapatellar fat pad must not be con-
fused with free intrasynovial fat floating on the effusion,
as can be seen in lipohaemarthrosis (BIANCHI et al.
1995b). The two conditions can be differentiated using
dynamic USo In lipohaemarthrosis, manual squeezing
of the joint induces mixing of the different components
of the intra-articular effusion and disappearance of the
fat-fluid level. The prefemoral fat appears as a hyper-
echoic fat structure located between the anterior cortex
of the femur and the posterior aspect of the suprapa-
tellar recess. The projection of the hyperechoic femoral
fat inside the suprapatellar bursa is a frequent normal
finding and should not be misdiagnosed as hypertro-
phied synovium or a lipoma arborescens. US findings,
together with knowledge of the clinical data, can help
in differentiating these conditions. The synovium pres-
ents a more hypoechoic appearance and is entirely con-
tained inside the synovial space. A greater size and
a frond-like appearance are typical of lipoma arbore-
scens (LEARCH and BRAATON 2000).
The femoral trochlea and the overlying cartilage
are easily assessed on transverse images obtained with
the knee completely flexed. The cartilage appears as
a hypo- or anechoic band overlying a regular hypere-
choic line, which corresponds to the subchondral bone
(GRASSI et al. 1999). In normal conditions the carti-
lage is thicker at its central portion, presents regular
borders and has a homogeneous echogenicity. Small
hyperechoic cartilage dots are related to chondrocal-
cinosis (COARI et al. 1995). The sulcus angle and the
trochlear depth calculated with US correlate weIl with
CT measurements (MARTINO et al.1998).
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
S. Bianchi et al.
5.3.2
Patellar Region
In the patellar region, the patella, the medial and lat-
eral patellar retinacula and the parapatellar synovi-
al recess can be imaged (Fig. 5.3). These structures
are best examined on transverse images. The dorsal
surface of the patella appears as a regular hyper-
echoic line corresponding to the bone surface. In
asymptomatic older subjects, small irregularities
of the proximal pole of the patella are commonly
found and are related to calcific enthesopathy of the
quadriceps tendon. Superficial to the patella, a thin
hyperechoic band can be identified, corresponding
to the most anterior fibres of the rectus femoris
tendon; after overlying the sesamoid, these fibres
gradually blend into the patellar tendon. In bipartite
patella, an interruption of the anterior cortex, typi-
cally located at the supero-external quadrant, is evi-
dent. Bipartite patella can be differentiated from a
fracture because it is painless upon pressure with
the US probe. Patellar position can be evaluated
by US in both children and adults. In children an
advantage of US is that the interpretation is inde-
pendent of the degree of patellar ossification (Joz-
WIAK and PIETRZAK 1998). From both the internal
and the external side of the patella, the patellar reti-
nacula insert into the femur and act as stabilising
structures. US reveals the retinacula as bilaminar
structures with discrete superficial and deep layers
(STAROK et al. 1997). The medial facet of the patella
can be incompletely imaged by US if the patella
is tilted and displaced internally by the operator's
hand. Unfortunately the lateral facet, wh ich is com-
monly affected by osteoarthritis and chondromala-
cia, cannot be imaged at USo
Ultrasound 69

5.3.3 can demonstrate a small amount of fluid inside the

Infra patellar Region
The infrapatellar region includes the prepatellar and
pretibial synovial bursae, the patellar tendon and
Hoffa's fat pad (Fig.5.4). The prepatellar bursa is
located in the subcutaneous tissues overlying the
lower pole of the patella and the cranial third of the
patellar tendon. In normal circumstances, because of
its thin wall and virtual absence of internal fluid, it
cannot be demonstrated by USo Any internal fluid
must be regarded as a sign of prepatellar bursitis. The
pretibial bursa is a small synovial structure located
between the distal portion of the patellar tendon and
the anterior aspect of the tibial epiphysis. In a sm all
percentage of normal subjects, high-resolution US
bursa. A diagnosis of bursitis can be made in cases
in which a larger effusion is demonstrated or when
pain is elicited by applying pressure with the probe
over the inflamed bursa. Longitudinal and transverse
sonograms must be obtained of each bursa to evalu-
ate the size and internal structure.
The patellar tendon originates from the lower pole
of the patella and inserts into the anterior tibial
tuberosity. The tendon has a flattened appearance,
regular borders and an internal fibrillar structure.
A slightly increased thickness of the distal third is
normal and must not be interpreted as a localised
tendinopathy. On transverse sonograms the charac-
teristic image of two cords separated by a low signal
bridge is observed after post-surgical harvesting of

Fig. 5.4a-f. Anterior aspect.

Infrapatellar region. Patellar
tendon. a Probe positioning
for longitudinal examination
of proximal tendon. b Cor-
responding sonogram. P,
a Patella; HFP, Hoffa's fat pad;
arrowheads, proximal patellar
tendon. cProbe positioning
for longitudinal examination
of distal tendon. d Cor-
responding sonogram. T,
Tibia; TT, tibial tuberosity;
arrowheads, proximal patellar
tendon; arrow, normal wid-
ening of distal tendon. e
Probe positioning for trans-
verse examination. f Cor-
responding sonogram. HFP,
Hoffa's fat pad; arrowheads,
proximal patellar tendon
 70
the mid third of the tendon for anterior cruciate lig-
ament reconstruction. This aspect must not be con-
fused with a longitudinal tendon split (ADRIANI et
al. 1995). Internal hypoechoic areas can be found
in asymptomatic elite athletes (COOK et al. 1998).
This has implications for clinicians managing ath-
letes with anterior knee pain. Calcifications and ossi-
fication of the distal portion of the tendon can be
found in asymptomatic subjects and are usuallyrelat-
ed to sequelae of Osgood-Schlatter disease. Deep to
the patellar tendon, between it and the anterior por-
tion of the condyles, the intracapsular Hoffa's pad
appears as an adipose structure containing internal
fibrous septations.
5.4
Medial Aspect
The medial aspect of the joint is examined by asking
the patient to externally rotate the leg. The medial
collateralligament, internal femorotibial joint space,
medial meniscus and pes anserinus tendons are eval-
uated (Fig. 5.5). The medial collateralligament com-
prises two portions: the superficial portion, which
originates from the superior aspect of the medial
condyle and inserts on the internal tibial metaphy-
sis, and the deep component, which links the internal
meniscus to the femur (meniscofemoral ligament)
and to the tibia (meniscotibialligament). On US, both
superficial and deep ligament components appear
as regular, hyperechoic laminae separated by a
hypoechoic area related to fat and loose connective
a _ _ _ _
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
S. Bianchi et al.
tissue. A synovial bursa can be located between the
two components but cannot be demonstrated by US
in the normal state. Collateralligament injuries can
be diagnosed with US (LEE et al. 1996; MATHIEU et
al.1997). Dynamic US images obtained during valgus
stress improve the assessment of ligament integrity
(DE FLAVIIS et al. 1988). Images are obtained both at
rest and during stress manoeuvres and documented.
Then the distance between the tibia and femur can be
easily measured and compared with the contralateral
knee. Although the possibility of evaluating menis-
cal tears with US has been reported (GERNGROSS and
SOHN 1992), there is now general agreement that the
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
associated with meniscal tears. When meniscal ossicle,
which represents a vestigial structure, is suspected on
standard radiographs, US can confirm the diagnosis.
The ossicle appears as a regular hyperechoic structure
embedded within the posterior horn of the medial
meniscus (MARTINOLI et al. 2000a). The pes anseri-
nus tendons (sartorius, gracilis and semitendinosus
tendons) insert in the anteromedial portion of the
tibial metaphysis. The different tendons can be distin-
guished by US only at a higher level. As the tendons
approach the distal insertion inferiorly, they blend
together and cannot be differentiated. Different syno-
vial bursae located among the tendons and between
them and the tibia lower the friction among these
structures. In normal conditions these bursae cannot
be demonstrated by US, whereas they are readily evi-
dent when distended by an effusion in bursitis.
b
Ultrasound
5.5
Lateral Aspect
The lateral aspect of the joint is examined by asking
the patient to internally rotate the leg. The lateral col-
lateral ligament, external femorotibial joint space, lat-
eral meniscus and distal aspect of the iliotibial band
are evaluated (Fig. 5.6). The lateral collateralligament
appears as a cord-like hyperechoic structure which is
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
limitations as US of the medial meniscus. Examin-
ing the knee in forceful flexion makes meniscal cysts
more apparent. The distal tract of the iliotibial band
is found at the anterior aspect of the lateral face. It
appears as a thin fibrillar structure that inserts in
a tibial tubercle located at the anterolateral aspect
of the tibial epiphysis (BONALDI et al. 1998). The
more distal portion of the band normally widens
just before its insertion into the tibia. This normal
appearance must not be misdiagnosed as localised
tendinopathy. The standard US examination must
also include transverse sonograms of the proximal
tibiofemoral joint and of the antero-external com-
partment of the leg since this is the most frequent
location of intramuscular ganglia. These ganglia orig-
inate from the proximal tibiofemoral joint, extend
inside the anterior tibialis or peroneus muscle and
can cause compression of the peroneal nerve (BIAN-
CHI et al. 1995a).
Fig. 5.6a, b. Lateral aspect. Iliotibial band. a Probe positioning for longitudinal examination. b Corresponding sonogram.
T, tibia; arrowheads, iliotibial band
71
5.6
Posterior Aspect
The the patient is then asked to lie prone with the
knee extended, to allow examination of the posteri-
or region. The US examination is performed in the
transverse and sagittal planes, starting with evalua-
tion of the intern al region, followed by assessment of
the middle and lateral regions.
5.6.1
Internal Region
The structures that can be assessed by US are the
proximal portions of the tendons of pes anserinus,
the semimembranosus tendon and the gastrocne-
mius-semimembranosus synovial bursa (Fig. 5.7).
The sartorius is mainly composed of muscle fibres
and is located internally. The gracilis tendon is locat-
ed posterior to the sartorius and, as its name implies,
it is the thinnest of the internal tendons. The cranial
portion of the semitendinosus tendon is located pos-
terior to the semimembranosus muscle, then, more
caudally, it lies behind the semimembranosus tendon.
The large semimembranosus tendon lies in a more
lateral position and inserts by its direct tendon on
the posteromedial aspect of the tibial epiphysis. The
indirect tendon cannot be detected at USo Between
the tendon of the semimembranosus and the medial
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
 72 s. Bianchi et al.

Fig. 5.7a, b. Posterior

aspect. Medial region.
a Probe positioning for
transverse examination.
b Corresponding sono-
gram. MC, Medial
condyle; asterisks, carti-
lage of medial condyle;
SM, semimembranosus
tendon; MHG, medial
head of gastrocnemius

a ....._ __
b

in adults they are eonneeted with it by a short pedi- 5.6.2

de. The bursa has two eomponents, the deep and the
superficial. The smaHer deep eomponent is loeated
between the medial head and the posterior aspeet of
the knee and eommunicates with the joint through a
defeet in the eapsule. The larger eomponent is loeat-
ed in the subeutaneous soft tissues, superficial to
the medial head. The loeation of the bursa must be
known by sonographers sinee it is the most eommon
loeation of Baker's eysts. Flexion of the knee opens
the eommunication with the joint spaee and ean lead
to a better distension and assessment of the eyst ped-
ide.
The eartilage of the posterior medial eondyle is
weH evaluated by USo Unfortunately, the weight-bear-
ing area loeated at the junetion of the middle and the
posterior third of the eondyle eannot be assessed.
Middle Region
In the eentral third the medial head of the gastroc-
nemius, the popliteal artery and vein, the posterior
tibial nerve and the intercondylar notch are identi-
fied (Fig. 5.8). The medial head of the gastrocnemius
has a triangular shape when examined on transverse
sonograms. Its tendon is located at the medial aspect
of the musde and can be demonstrated as a eomma-
shaped hypereehoic structure. The popliteal artery is
located lateraHy to the medial head. The vein is locat-
ed posterior and lateral to the artery and the nerve
posterior and lateral to the vein. So, on transverse
sonograms, the popliteal artery, vein and nerve are
loeated in a line which runs from anterior to poste-
rior and from medial to lateral. The ealibre, thick-

a L - - - -.......- -

Fig. 5.8a, b. Posterior aspect. Middle region.

a Probe positioning for transverse exam-
ination. b Corresponding sonogram. MC,
Medial condyle; LC, lateral condyle; A, popli-
teal artery; V, popliteal vein; arrow, posterior
tibial nerve; MHG, medial head of gastroc-
nemius; LHG, lateral head of gastrocnemius b
Ultrasound 73

ness and appearance of the wall, as weH as puls at ions

of the popliteal artery, can be well assessed with USo
Since the patient is examined prone, the vein can be
collapsed and difficult to evaluate by USo In this case
a small elevation of the foot, obtained by flexing the
knee, fills the vein and leads to its detection. The col-
lapse of the vein obtained by local compression with
the probe excludes vein thrombosis. Both artery and
vein, however, are beUer evaluated by US colour Dop-
pler sonograms. At the cranial edge of the popliteal
space the sciatic nerve splits into the posterior tibial
nerve and the common peroneal nerve. US can assess
a
the typical fascicular echotexture of peripheral nerves
due to multiple hypoechoic parallel but discontin-
uous linear areas separated by hyperechoic bands
(MARTINOLI et al. 2000b). The US appearance cor-
relates weH with the internal structure composed of
neural fascicles (hypoechoic areas) separated by the
epineurium (hyperechoic bands) (SILVESTRI et al.
1995). The posterior tibial nerve is larger than the
common peroneal nerve and can be followed from its
origin to the ankle region.
The cruciate ligaments are difficult to evaluate by
USo The posterior cruciate ligament can be demon-
strated as a hypoechoic band on longitudinal sono-
grams in normal subjects and confirmed in amputat-
ed knee specimens (WANG TG et al. 1999). The distal b

portion of the ligament and its insertion into the tibia

are more superficial and are more easily assessed
than the proximal deep portion. Due to its internal
location and oblique orientation, the anterior cru-
ciate ligament is barely visible. Indirect US evalu-
ation of the integrity of cruciate ligaments can be
achieved by measuring the tibial subluxation during
stress manoeuvres (GEBHARD et al. 1999). US detec-
tion of a localised haematoma at the posterocranial
portion of the anterior cruciate ligament correlates
well with acute te ars (PTASZNIK et al. 1995).
5.6.3
Lateral Region
The lateral region is examined for the evaluation of
the biceps and gastrocnemius muscles and tendons,
common peroneal nerve and lateral condylar carti-
lage (Fig. 5.9). The biceps muscle is composed of a
long head which originates from the ischial tuberos-
ity, and a short head which arises from the femoral
shaft. The two heads join at the superior edge of
the popliteal space to form a strong tendon which
inserts into the fibular head. The biceps muscle and
tendon are weH evaluated by longitudinal and trans-
Fig. 5.9a, b. Posterior aspect. Lateral region. a Probe position-
ing for transverse examination. b Corresponding sonogram.
LC, Lateral condyle; asterisks, cartilage of lateral condyle; LHG,
lateral head of gastrocnemius; BT, biceps tendon; arrow, pero-
neal nerve
verse sonograms. Cranial images must include evalu-
ation of the myotendinous junction of the long head
since this is a common site of sport-related tears.
Transverse sonograms show both the lateral collat-
eralligament and the biceps tendon and allow their
differentiation. The common peroneal nerve arises
from the sciatic nerve and reaches the posteromedi-
al border of the biceps muscle. The nerve is smaller
than the posterior tibial nerve and presents the typ-
ical fascicular pattern. At the level of the fibular
head the nerve is located in the subcutaneous tissues
between the skin and the bone cortex (MARTINOLI
et al. 2000). This area must be accurately examined
since this is the typicallocation at which the nerve
can be injured by local trauma. The bifurcation of
the common peroneal nerve into the superficial and
deep peroneal nerves can also be imaged by USo In a
deeper location, the lateral head of the gastrocnemius
muscle, the posterior portion of the lateral condyle
74
and the hyaline eartilage ean be assessed. The lateral
head of the gastroenemius muscle is smaller than the
medial head. Its tendon may eontain the fabella, a ses-
amoid bone that appears as a eurvilinear hypereeho-
ie strueture showing adefinite posterior shadowing.
Care must be taken not to eonfuse the fabella with an
intra-artieular loose body. The intratendinous loea-
tion and the fixed position during dynamie examina-
tion exclude an intra-articular fragment. Differentia-
tion from a eapsular ealcifieation is best aehieved by
lateral radiographs.
References
Adriani E, Mariani PP, Maresca G et al (1995) Healing of the
patellar tendon after harvesting of its mid-third for ante-
rior cruciate ligament reconstruction and evolution of the
unclosed donor site defect. Knee Surg Sports Traumatol
Arthrose 3:138-143
Bianchi S, Martinoli C (1999) Detection of loose bodies in
joints. Radiol Clin North Am 37:679-690
Bianchi S, Zwass A, Abdelwahab IF et al (1994) Diagnosis of
tears of the quadrieeps tendon of the knee: value of sonog-
raphy.AJR 162:1137-1140
Bianchi S, Zwass A, Abdelwahab IF et al (1995a) Sonographie
evaluation of intramuscular ganglia. Clin Radiol
50:235-236
Bianchi S, Zwass A, Abdelwahab IF et al (1995b) Sonographie
evaluation of lipohemarthrosis: clinieal and in vitro study.
J illtrasound Med 14:279-282
Bonaldi VM, Chhem RK, Drolet R et al (1998) Iliotibial band
frietion syndrome: sonographie findings. J illtrasound Med
17:257-260
Coari G, lagnocco A, Zoppini A (1995) Chondrocalcinosis:
sonographie study of the knee. Clin RheumatolI4:511-514
Cook JL, Khan KM, Harcourt PR et al (1998) Patellar tendon
ultrasonography in asymptomatie active athletes reveals
hypoechoie regions: a study of 320 tendons. Vietorian
Institute of Sport Tendon Study Group. Clin J Sport Med
8:73-77
De Flaviis L, Nessi R, Leonardi M et al (1988) Dynamie ultra-
sonography of capsulo-ligamentous knee joint traumas. J
Clin Ultrasound 16:487-492
Fornage BD (1987) The hypoechoie normal tendon. A pitfall. J
illtrasound Med 6:19-22
Gebhard F, Authenrieth M, Strecker W et al (1999) illtrasound
evaluation of gravity induced anterior drawer following
anterior cruciate ligament lesion. Knee Surg Sports Trau-
matol Arthrose 7:166-172
Gerngross H, Sohn C (1992) illtrasound scanning for the diag-
nosis of meniscal lesions of the knee joint. Arthroscopy
8:105-110
S. Bianchi et al.
Grassi W, Lamanna G, Farina A et al (1999) Sonographie imag-
ing of normal and osteoarthritie cartilage. Semin Arthritis
Rheum 28:398-340
Grobbelaar N, Bouffard JA (2000) Sonography of the knee, a
pietorial review. Semin illtrasound CT MR 21:231-274
Hauzeur JP, Mathy L, De Maertelaer V (1999) Comparison
between clinieal evaluation and ultrasonography in detect-
ing hydrarthrosis of the knee. J Rheumato126:2681-2683
Jozwiak M, Pietrzak S (1998) Evaluation of patella position
based on radiologie and ultrasonographie examination:
comparison of the diagnostie value J Pediatr Orthop
18:679-682
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
Malghem J, Vande Berg B, Noel H et al (1992) Benign osteo-
chondromas and exostotie chondrosarcomas: evaluation
of cartilage cap thiekness by ultrasound. Skeletal Radiol
21:33-37
Martino F, De Serio A, Macarini L et al (1998) illtrasonogra-
phy versus computed tomography in evaluation of the fem-
oral trochlear groove morphology: a pilot study on healthy,
young volunteers. Eur Radio18:244-247
Martinoli C, Bianchi S, Derchi LE (1999) illtrasound of tendon
and nerves. Radiol Clin North Am 37:691-711
Martinoli C, Bianchi S, Spadola L et al (2000a) Multimodality
imaging assessment of meniscal ossicle. Skeletal Radiol
29:481-484
Martinoli C, Bianchi S, Gandolfo N et al (2000b) US of nerve
entrapments in osteofibrous tunnels of the upper and lower
limbs. Radiographies 20 [Spec]:SI99-S217
Mathieu P, Wybier M, Busson J et al (1997) The medial col-
lateral ligament of the knee. Ann Radio140: 176-181
Peetrons P,Allaer D, Jeanmart L (1990) Cysts of the semilunar
cartilages of the knee: a new approach by ultrasound imag-
ing. A study of six cases and review of the literature. J Ultra-
sound Med 9:333-337
Ptasznik R, Feller J, Bartlett J et al (1995) The value of sonog-
raphy in the diagnosis of traumatic rupture of the anterior
cruciate ligament of the knee. AJR 164: 1461-1463
Rutten MJ, Collins JM, van Kampen A et al (1998) Meniscal
cysts: detection with high-resolution sonography. AJR
171 :491-496
Silvestri E, Martinoli C, Derchi LE et al (1995) Echotexture
of peripheral nerves: correlation between US and histo-
logie findings and criteria to differentiate tendons. Radiol-
ogy 197:291-296
Starok M, Lenchik L, Trudell D et al (1997) Normal patellar
retinaculum: MR and sonographie imaging with cadaverie
correlation.AJR 168:1493-1499
Wang SC, Chhem RK, Cardinal E et al (1999) Joint sonography.
Radiol Clin North Am 37:653-668
Wang TG, Wang CL, Hsu TC et al (1999) Sonographie evalua-
tion of the posterior cruciate ligament in amputated speci-
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

Fig.6.1. a Calculation of varus angulation on an AP

radiograph. Lines are drawn along the mid shaft of the
femur and tibia, and the angle of intersection is calcu-
lated. b An AP radiograph demonstrating varus angu-
c lation in a l-year-old. c An AP radiograph demonstrat-
ing valgus angulation in a 5-year-old

6.2.2 in rickets), severe obesity or ligamentallaxity. Anteri-

Genu Valgum (Knock Knees)
Genu valgum that persists or worsens during the
child's development, particularly beyond the age of 12
years, is considered abnormal. Clinically the degree
of valgus is assessed with both patellae facing for-
ward and touching; the distance between the medial
malleoli is then measured for assessment. A distance
of less than 10 cm requires no radiological evalua-
tion, but a distance of more than 10 cm that persists
over 3 years should be investigated for an underly-
ing abnormality (SHARRAD 1976). Valgus deformity
may be due to a developmental or neuromuscular
abnormality such as cerebral palsy or a bone dyspla-
sia; alternatively it may be the result of a tibial growth
plate injury, reduced bone mineral density (such as
or-posterior radiographs show abduction of the tibia
with evidence of remodelling with bowing and corti-
cal sclerosis along the medial tibial border. There is
underdevelopment of the lateral femoral condyle and
in severe ca ses, widening of the medial aspect of the
distal femoral growth plate (MACEwAN and DUNBAR
1958).
6.2.3
Genu Varum (Bow Legged)
The commonest cause for genu varum is tibia vara
(Blount's disease) (BLOUNT 1937; LAVILLE et al. 1999),
which is divided into two forms, the infantile and the
late onset (adolescent).
Congenital and Developmental Abnormalities of the Knee 79

The infantile form typically presents between 2

and 3 years of age. It is usually bilateral but may be
asymmetrical and is more common in overweight
children (DIETZ et al. 1982).1t is thought to be due to
abnormal pressure loading against the tibial physis
resulting in abnormal endochondral ossification and
disorganised physeal zones (CARTER et al. 1988).
The adolescent form affects children between 6
and 15 years of age, is less common than the infan-
tile form and is usually unilateral. It is believed to
be due to segmental arrest of medial physeal junc-
tion (LODER et al. 1991) and has been seen following
osteomyelitis and trauma; however, a few cases with-
out any underlying factor have been reported (BATH-
FIELD and BEIGHTON 1978). This idiopathic form is
commoner in obese males (HENDERSON et al. 1990).
In the early stages, distinguishing abnormal tibia
vara in a young infant from normal physiological
bowing can be problematic. Use of the metaphyseal-
diaphyseal angle may help predict the development
of infantile Blount's disease (LEVINE and DRENNAN
1982). The metaphyseal angle is determined from a
line through the proximal tibia metaphysis and one
perpendicular to the long axis of the tibia. An angle Fig.6.2. Calculation of the metaphyseal angle on an AP radio-
graph of the knee. A line is drawn through the proximal tibial
of greater than 11 0 is associated with Blount's disease
metaphysis and another is drawn perpendicular to the long
while one less than 11 0 suggests physiological bowing axis of the tibia. The angle of intersection is the metaphyseal
that will spontaneously correct (Fig. 6.2). The tibial- angle. Less than 11 % is normal. In this patient with Blount's
femoral angle, determined from a line along the long disease the metaphyse al angle demonstrated is 23%
axis of these bones, is not significantly different in
patients with Blount's disease and those who have
physiological bowing.
Radiographs in Blount's disease show the tibia is
in varus due to angulation within the metaphysis.
There is thickening of the medial cortices of the femur
and tibia with formation of a prominent beak on the
medial tibial metaphysis. This beak may hypertrophy,
causing irregular ossification leading to fragmenta-
tion and depression of the beak and tibial metaphy-
sis. The tibia can sublux laterally, causing the medial
tibial plateau to lie in the femoral intercondylar notch.
As the weight-bearing pressure continues, the medial
tibial physis becomes narrowed and premature fusion
occurs with compensatory widening of the lateral
physis of the tibia and femur (CURRANIO and KIRKS
1977). Both magnetic resonance (MR) imaging and
arthrography have demonstrated enlargement of the
epiphyseal cartilage and medial meniscus (DALINKA
et al. 1974) (Fig. 6.3). The medial meniscus enlarges
in both height and width and becomes hypermobile.
Premature degeneration occurs within the meniscus.
Fig. 6.3. Coronal three-dimensional fast spoiled gradient-
MR imaging can demonstrate the integrity of the recalled echo sequence (3D-SPGR) with fat suppression, MR
physis, the irregular ossification and premature phy- image of the knee in Blount's disease. Poorly developed medial
seal closure (Ducou LE POINTE et al. 1995). Docu- epiphysis and overgrowth of the cartilage
80
mentation of unsuspected physeal bars can be helpful
for surgical planning, and assessment of the physeal
growth plate with 3D reformatting further aids surgi-
cal intervention (CRAIG et al. 1999).
The differential diagnosis for genu varum includes
unresolved and severe physiological bowing, rickets,
osteogenesis imperfecta and other congenital bone
dysplasias. In the majority of bone dysplasias, the
genu valgum is the result of abnormal ossification
and bone remodelling.
6.3
Hyperextension Deformities of the Knee
6.3.1
Normal Development
The angle of extension of the knee joint is measured
from the intersection of lines drawn along the shaft
of the femur and the tibia. Hyperextension of the
knee joint of up to 20° is a normal variant at birth.
This hyperextension is completely stable, there is no
limitation of flexion and the integrity of the joint
and surrounding structures is maintained. A wide
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).
6.3.2
Genu Recurvatum
Genu recurvatum describes extreme hyperextension
with some restriction of flexion. This condition is
more often unilateral. Radiographically there is close
contact of the articular surfaces of the tibia and
femur, which differentiates it from congenital sublux-
ation. The femoral condyles are enlarged posteriorly,
which causes the reduced degree of flexion. There is
an association with fibrous contracture of the quad-
riceps muscle, particularly the rectus femoris, and
shortening and tightening of the posterior cruciate
ligament (AUSTWICK and DANDY 1983).
6.3.3
Congenital Subluxation/Dislocation
Congenital subluxation and dislocation are rare
deformities, occurring 40-80 times less frequently
K. Johnson and A. M. Davies
than congenital dislocation of the hip (BENSAHEL
et al. 1989; NIEBAUER and KING 1960). The deformi-
ty commonly involves both knees and usually there
are other associated congenital abnormalities. These
abnormalities include hip dislocation, arthrogrypo-
sis multiplex, elbow dislocation and club-foot defor-
mity (CURTIS and FISHER 1969; BELL et al. 1987).
An isolated form of anterior tibial displacement with-
out hyperextension has been described (CURTIS and
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).
Anterior radiographs mayaiso demonstrate rotatory
and lateral movement. Serial radiographs will show
a delay in the ossification of both the femoral and
the tibial epiphysis (CARLSON and O'CONNOR 1976).
Other radiological findings include patella baja, patel-
la alta, genu valgum and hypoplasia of the lateral
femoral condyle (BENSAHEL et al. 1989).
Arthrography and MR imaging demonstrate flat-
tening of the inferior portion of the femoral con-
dyles. Fibrous contractures of the quadriceps muscles
have been demonstrated and the lack of a suprapatel-
lar pouch is believed to be a bad prognostic indica-
tor.
6.4
Congenital Ankylosis of the Knee
Congenital ankylosis of the knee is a rare condition
(RYAN et al. 1978) and may cause difficulty in labour.
It is usually encountered at birth due to difficulty in
labour. The joint is usually fixed in some degree of
flexion and radiographs demonstrate fusion of the
lower femora with the upper tibia. There is severe
atrophy of the extensor muscles.
6.5
TibiallFibular Bowing and
Pseudo-arthrosis
Bowing of the tibia and fibula may occur together or
in isolation. Such bowing may be unilateral or bilater-
al, and symmetrical or asymmetrical. The condition
can occur in the absence of any other abnormality
or be associated with an underlying condition (BOYD
1982). Isolated tibial bowing is usually associated
with intrauterine position (CAFFEY 1947). Associat-
Congenital and Developmental Abnormalities of the Knee 81

Congenital pseudo-arthrosis may be present at

birth but the majority of cases develop as a result of
severe anterior bowing and fracturing through the
tibia or fibula. Fractures leading to infantile pseudo-
arthrosis occur most commonly in the first year of
life (RATHGEB et al. 1974) but have been described as
late as 12 years of age. Congenital pseudo-arthrosis
usually occurs in the middle to lower third tibia
and is associated with cupping at the proximal bone
ends and pointing at the distal ends. Sclerosis, cystic
change within the medulla and narrowing of the
bone at the maximum point of angulation predis-
pose to pseudo-articulation (Fig.6.5). There is a
very strong association with neurofibromatosis, but
pseudo-arthrosis is also seen in fracture non-union,
osteogenesis imperfecta and fibrous dysplasia (BOYD
and SAGE 1958).

Fig.6.4. Lateral radiograph of the knee demonstrating con-

genital dislocation of the knee. There is no contact between
the distal femoral epiphysis and the proximal tibia. There is a
bony spur arising from the distal femur

ed conditions include neurofibromatosis, hypophos-

phataemic rickets, osteogenesis imperfecta and other
skeletal dysplasias. The bowing may occur at the site
of local bone pathology such as fibrous dysplasia
(RESNICK et al. 1990).
The curvature of the bowed tibia may be in either
an anterior or a posterior direction, and there may
be either an anterior or a lateral component. Poste-
rior bowing causes shortening and thickening of the
cortex on the inner aspect of the curve and usually
there is no dysplastic change within the medullary
bone. Posterior bowing does not progress to pseudo-
articulation (HEYMAN and HERNDoN 1949; PAPPAS
1984). There is an association with club-foot defor-
mity (HEYMAN et al. 1959).
Anterior bowing is a more progressive condition,
particularly if there is narrowing and dysplasia of the
medullary canal (NEWELL and DURBIN 1976). Med-
ullary narrowing, cortical sclerosis and reduced tibial
diameter can progress and lead to fracturing of the
Fig. 6.5. AP and lateral radiographs showing congenital antero-
tibia. The fracture may not unite completely, leading
medial bowing of the tibia. At the maximum point of angula-
to pseudo-arthrosis. Tibial bowing associated with tion of the tibia there is increased sclerosis and narrowing
neurofibromatosis is more likely to progress and lead of the medullary canal. There is a pseudo-articulation of the
to pseudo-articulation. distal fibula
82 K. Johnson and A. M. Davies

6.6
Limb Deficiencies

Congenitallimb deformities may be classified as lon-

gitudinal abnormalities, which include the fibular and
tibial hemimelias, and transverse anomalies, such as
in the constriction band syndrome (LAOR et al. 1996).
The majority of this classification is based on con-
ventional radiographs (FRANTz and O'RAHILLY 1961)
but the use of MR imaging has added significant
information about the underlying cartilaginous and
vascular abnormalities (LAOR et al.1996).
There is a very wide spectrum of limb deficiencies
that range from complete absence of the whole limb
to partial development of a single bone. MR imaging
has been able to evaluate the nature and structure of
joints around the abnormal skeleton.

6.6.1
Fibular Hemimelia

Fibular hemimelia is the commonest long-bone

deficiency and has a spectrum of abnormalities rang-
ing from minimal shortening (type I) to complete
absence (type III) (ACHTERMAN and KALAMCKE 1979)
(Fig. 6.6). Around the knee joint there is an association
with a small, high-riding patella with orwithout lateral
subluxation, and absence of the patellar tendon (LAOR
et al. 1996). The femoral and tibial epiphysis may be
hypoplastic. Within the joint there may be absence of
the menisci and cruciate and collateralligaments with
associated anteroposterior laxity. Other associations
include tibial shortening and bowing that may cause
leg length discrepancy, fusion or absence of the later-
al rays of the foot and deformities of the tarsal bones
(CANALE and BEATY 1991).
6.6.2
Tibial Hemimelia
Fig.6.6. AP and lateral radiographs demonstrating fibular
aplasia
6.6.3
Femoral Absence
Congenital absence of the femur is commonly a prox-
imal deficiency affecting the development of the ace-
tabulum and pelvis. In complete absence of the femur,
the tibia may articulate with the pelvis and the fibula
may be absent (PAPPAS 1983).Absent patella and knee
joint instability have been described in a number of
cases of femoral deficiency.

Tibial hemimelia is less common than fibular hemi-

melia. It also has a wide variation in severity, from
mild shortening to complete absence. There is an
association with abnormalities around the knee joint,
including epiphyseal irregularities and absent menis-
ci and cruciate ligaments (LAOR et al. 1996).
6.7
Femoral Bifurcation/Duplication
Congenital femoral duplication and femoral bifur-
cation are two distinct conditions. Femoral bifur-
cation of the distal portion of the femur produces a
Y-shaped defect; the proximal femur is normal with a
well-developedacetabulum (WOLFGANG 1984).Dupli-
cation is a rare but distinct abnormality (Fig. 6.7).
Congenital and Developmental Abnormalities of the Knee
6.8
Discoid Meniscus
In cases of discoid meniscus there is loss of the normal
triangular appearance of the meniscus with thickening
and widening in one or all dimensions (see Chap.9).
The commonest type has a "slab" appearance, but
biconcave and wedge shapes occur. The lateral menis-
cus is the most commonly affected site (SILVERMAN
et al. 1989). On MR imaging with sagittal 3- or 4-mm
slices there is loss of the normal "bow-tie" appearance,
and on coronal images the meniscus extends into the
joint space. The discoid meniscus may cause knee pain
and a snapping sensation on movement and is more
prone to tear and degeneration (STARK et al.1995).
6.9
Congenital Absence of the Anterior Cruci-
ate Ligament
Congenital absence of the anterior cruciate ligament
(ACL) is a rare condition and commonly occurs with
other limb anomalies (THOMAS et al. 1985; BARNES
et al. 1988). Up to half the reported patients with this
condition are asymptomatic, while others complain
of the knee giving way and of intermittent pain.
Associated anomalies include short femur, absence
of the menisci, tibial/fibular dysplasia and patellar
hypoplasia which may be dislocated or subluxed
(JOHANSSON and ASARISI 1983; MALUMED et al.1999)
83
Fig.6.7. AP radiograph of the lower
body in a neonate where there is dupli-
cation of the acetabulum and lower
limbs. In the distal duplicated limbs
there is tibial hypoplasia and fibular
aplasia. There is abnormal development
of the pelvis
ANDERSSON and ELLITSGAARD (1992) have report-
ed a patient in whom congenital absence of the ACL
occurred as an isolated finding. This patient did not
have a positive draw test because of compensating
posterior cruciate hypertrophy.
Radiological signs suggesting ACL deficiency
include hypoplasia of one or both tibial spines, hypo-
plasia of the lateral femoral condyle and a 'V' shaped
or narrowed intercondylar notch.
6.10
Arthrogryposis Multiplex
Arthrogryposis multiplex congenita (AMC) is a rela-
tively common congenital disorder characterised by
multiple joint contractures and fibrosis of skeletal
muscle (SÖDERGARD and Ryoppy 1990). Knee involve-
ment is relatively common, the deformity being one
of either extension or flexion. AMC is only a relatively
descriptive diagnosis and the aetiology is not uniform,
encompassing causes such as neuronal migrational
disorder in utero and spinal cord atrophy.
Radiologically there is a wide spectrum of abnor-
mality. Depending on the severity of the disorder,
bone alignment around joints may be normal or there
may be mild or severe incongruence. Serial exami-
nations of patients with AMC show early osteoarthri-
tis particularly of the patellofemoral or the medial
and lateral tibial femoral joints. The development of
osteoarthritis appears to be related to the degree
84 K. Johnson and A. M. Davies

of joint surface incongruity and is more common 6.11.4

in the extensor group (SÖDERGARD and Ryoppy
1990). Overall, the prognosis for extensor contrac-
tures seems to be slightly better than that of flexion
contractures. Some patients with AMC mayaiso have
agenesis of the anterior cruciate ligament.
6.11
Transplacental Acquired Infections
Transplacentally acquired viral infections include the
TORCH complex [toxoplasmosis, rubella, cytomega-
lovirus and herpes simplex], syphilis and the human
immunodeficiency virus (HIV).
6.11.1
Rubella
Congenital Syphilis
The manifestations of congenitally acquired syphilis
are due to maternal infection in the second or third
trimester. Skeletal involvement occurs in up to 95%
of cases of overt disease, but many infections are
asymptomatic. Other manifestations of infection are
rash, anaemia, rhinorrhoea and nephritic syndrome.
Within the long bones there are lucent metaphyseal
bands with fragmentation and destruction of the
distal ends. The characteristic Wimberger sign refers
to metaphyseal destruction in the upper medial
aspect with relative sparing of the recently formed
metaphyseal bone, the so-called Laval-Jeantet collar
(Fig.6.8). Due to the reduction in ossified callus
there may be pathological fractures of the long bones
(SACHDEV et al. 1982).

Congenitally acquired rubella is the commonest

agent and causes a syndrome that includes intra-uter-
ine growth retardation, thrombocytopenia, anaemia,
hepatosplenomegaly, cataracts and other eye abnor-
malities, patent ductus arteriosus and aortic stenosis.
Skeletal manifestations occur in 23%-50% of cases,
and the bone changes are thought to be related to the
virulence of the virus. The distal femur is the most
commonly involved site, with the classical feature of
dense longitudinal metaphyseal bands (celery stalk
appearance). Epiphyseal maturation is delayed and
there may be metaphyseal irregularity (KUHNS et al.
1977). The bones may become osteopenic and devel-
op metaphyseal cupping with irregular ossification
centres.

6.11.2
Cytomegalovirus

The bone changes are similar to those of rubella but

the delay in epiphyseal maturation is not a feature
(KUHNS et al. 1977).

6.11.3
HIV Infection

Infection with the HIV is rare and there are no spe-

cific skeletal manifestations. The child may develop
superadded infection such as osteomyelitis, septic
arthritis or lymphoma (HARTY et al. 1994).
Fig.6.8. AP radiograph of the tibia in congenital syphilis.
There is a symmetrical periosteal re action along the tibial
diaphyses. A focal bony defect is present on the medial aspect
of the proximal tibial metaphysis (Wimberger sign)
Congenital and Developmental Abnormalities of the Knee 85

6.12
Neurofibromatosis

There are two forms of neurofibromatosis, both of

which are autosomal dominant disorders. Type I is
the peripheral form and is due to a gene defect on
chromosome 17, while type 2 is the rarer central form
due to a gene defect on chromosome 22.
The clinical features of type 1 include cafe-au-lait
spots, palpable neurofibromas of peripheral nerves,
neurofibromas, optic gliomas and Lisch spots on
the iris. A wide spectrum of skeletal manifestations
occur, such as overgrowth or undergrowth and over-
or undertubulation due to cortical thickening of the
long bones. There may be subperiosteal resorption
due to the pressure effect of soft tissue neurofibroma.
In the tibia there will be anterior and lateral bowing,
leading to irregular periosteal thickening and pseu-
do-arthrosis. These tibial changes usually appear by
the first year oflife (KLATTE et al. 1976).

6.13
Haematological Disorders

There are numerous clotting or platelet disorders

Fig. 6.9. AP radiograph of the knee in haemophilia. There is
that can cause an arthropathy due to repeated spon- generalised loss of joint space, marginal erosions and widen-
taneous haemarthroses, typically affecting the knee. ing of the intercondylar noteh. Both the femoral and the tibial
metaphysis are widened

6.13.1
Haemophilia
The most common disorder is haemophilia, which is
X-linked and occurs in 1 in 10,000 male births. Hae-
mophilia is due to a deficiency of factor 8, a constituent
of the clotting cascade; there is also factor 9 deficiency
(Christmas disease) and factor 11 deficiency.
Soft tissue haemorrhage around the knee joint
is relatively common and will cause distortion and
obliteration of fat and fascial planes on radiographs.
MR imaging is able to accurately localise the site
and soft tissue compartment of this haemorrhage.
Repeated soft tissue bleeding can lead to myositis
ossificans.
Later radiographic signs are of peri-articular osteo-
porosis and accelerated maturation, with epiphyse al
overgrowth and widening of the intercondylar notch
(Fig. 6.9). These signs reflect the localised hyperaemia
and synovial proliferation caused by repeated haem-
arthroses. Radiographs are not able to differentiate
between fresh haemorrhage and increased synovium
in a swollen joint. In the advanced stages there is carti-
lage loss and articular destruction which will be seen
as loss of joint space and peri-articular erosions.
MR imaging is able to delineate hypertrophic pannus
from fresh haemorrhage and is more sensitive in
detecting early subcortical cysts and cartilage destruc-
tion (HERMANN et al.1992). Gadolinium enhancement
of the synovium is less intense than with juvenile idio-
pathic arthritis, probably due to the hypovascular con-
nective tissue and haematological degradation prod-
ucts in the synovium (NAGELE et al.1995).
Another feature of the disease is haemophilic
pseudotumours, which are lytic expansile lesions that
occur in the ilium, femur and tibia. They may be
intra-osseous, where they are seen as a well-defined
medullary lucency with a thin sclerotic rim, or sub-
periosteal, where there is extensive subperiosteal
re action and resorption of cortex with a soft tissue
component. Discrimination of pseudotumours may
be improved by MR imaging as it will detect any
blood degradation products (GAARY et al. 1996).
86 K. Johnson and A. M. Davies

6.13.2
Thalassaemia

Thalassaemia is due to a dis order of the alpha chains

that form the haemoglobin in red blood cells. The
heterozygous form is thalassaemia minor, a relatively
mild condition. The homozygous form is thalassae-
mia major, which is a severe disabling disorder.
Radiological abnormalities result from marrow
hyperplasia and cause bone expansion, commonly
around the distal femur and proximal tibia. There is
an increased cortical trabecular pattern within the
bone and widened medullary cavities. In the long
bones this may lead to the Erlenmeyer flask deformi-
ty (MOSELEY 1974).

6.14
Diaphyseal Aclasis
Diaphyseal aclasis is one of the multiple pedunculat-
ed bony exostoses that commonly occur around the
faster growing joints. There is a familial preponder-
ance, with some cases being autosomal dominantly
inherited with variable penetrance. The exostoses
appear within the first 10 years of life, are often bilat-
eral, can be symmetrical and are assodated with
short stature (Fig. 6.10). The bony prominence points
away from the nearest joint and is covered with a car-
tilage cap to produce a lobulated outline on radio-
graphs (SHOGRY and ARMSTRONG 1990).
Around the knee joint the exostosis may causes
localised pain and swelling due to the development
of a tendonitis or bursitis. Nerve compression effects
may occur, as can pseudo-aneurysm formation and
venous compression within the popliteal fossa. Malig-
nant change can occur; the rate has been reported to
be between 1% and 25% (WILLMS et al. 1997; WUIS-
MAN et al. 1997; VOUTSINAS and WYNNE-DAVIES
1983).
Fig.6.10. A 14-year-old boywith diaphyseal aclasis who clini-
cally has multiple subcutaneous bony protuberances. An AP
radiograph of the knee shows multiple bony exostoses which
are directed away from the joint
LENCE et al. 1979; SPRANGER et al. 1982). All the
types have reduced bone mineral density and are
prone to repeated fracturing which causes metaph-
yseal irregularity, epiphyseal widening, widespread
bone remodelling and deformity and excessive callus
formation (Fig. 6.11). Traumatic fragmentation of the
physes may lead to "popcorn" calcification of the
epiphysis.
Type II is invariably lethal. Distinguishing features
among the other types are the age at presentation, the
propensity to fracture, the state of dentition and the
colour of the sclera.
6.16
Modelling Deformities

6.15 Modelling deformities that may affect the distal

Osteogenesis Imperfecta
Osteogenesis imperfecta (OI) comprises a wide spec-
trum of skeletal dis orders that arise from abnormal
type 1 collagen formation. Multiple genetic Iod have
been isolated to account for the various clinical mani-
festations of the disease. OI has been classified into
four types, each with different subgroupings (SIL-
femur and proximal tibia may be an isolated finding
or be related to an underlying bone dysplasia or met-
abolic or endocrine disorder. Many of the dis orders
are related to abnormal ossification or bone mineral-
isation. The knee joint is relatively commonly affect-
ed as it is a site of rapid growth and bone turnover.
The modelling deformities may be characteristic for
the underlying disorder but are not diagnostic. Not
Congenital and Developmental Abnormalities of the Knee 87

spread metaphyse al deformity with extended bone

ends, most marked in the distal femur and the proxi-
mal tibia and fibula (GIRDWOOD et al. 1969). Metaph-
yseal irregularities also occur in the proximal humer-
us, distal radius and ulna. The expanded bone is rel-
atively lucent and there is sclerosis around the mid
diaphysis. Bone sclerosis occurs around the skull base
and may cause cranial nerve impingement (HESEL-
SON et al.1979).

6.16.1.2
Craniometaphyseal Dysplasia

The appearances are similar to Pyle's disease, but the

distal femur becomes more club shaped in adulthood.
There are no cranial nerve abnormalities (MACPHER-
SON 1974).

6.16.1.3
Otopalatal Digital Syndrome

The bony findings are very similar to Pyle's disease.

There is widening of the metaphyses and eventually
there may be curvature of the long bon es. The hands
Fig. 6.11. An AP radiograph of the legs in osteogenesis imper- are affected, with shortening of the first digit and
fecta (type 1). There is generalised loss of bone density, with
widespread modelling deformities. There is evidence of mul-
tiple previous fractures which show different stages of healing

all the causes for each deformity are congenital: some

are acquired and some may be dietary.

6.16.1
Erlenmeyer Flask Deformity

The Erlenmeyer flask deformity describes the distal

expansion of long bones, most commonly the femur,
and is seen in a variety of disorders, including scle-
rosing skeletal dysplasias (Fig. 6.12), metabolie disor-
ders, Gaucher's disease and some neurological con-
ditions. The deformity is named after the Erlenmey-
er flask, which is a wide-neck glass container used
in chemicallaboratories at the end of the nineteenth
and in the early twentieth century. The flask itself
Fig. 6.12. There is
was named after Richard August Karl Erlenmeyer
remodelling and wid-
(1825-1907), an eminent German chemist. ening of the distal femo-
ral metaphysis indicative
6.16.1.1 of the Erlenmeyer flask
Pyle's Disease (Metaphyseal Dysplasia) deformity. The increased
bone density is part of
the spectrum of abnor-
Pyle's disease is an autosomal recessive condition in mality seen in osteope-
which there is normallife expectancy. There is wide- trosis
88 K. Johnson and A. M. Davies

metaphyseal widening of the small bones of the
hand. The condition is associated with frontal and
facial hypoplasia, cleft palate and conductive deaf-
ness (GALL et al. 1972).
6.16.1.4
Osteodysplasty (Melnick-Needles Syndrome)
These patients do not usually present until early
infancy and nearly all reported cases have been
female (KAUFMAN 1973). As the child grows there is
involvement of alliong bones, which develop ribbon
waviness, cortical irregularity and S-shaped bowing
(MELNICK and NEEDLES 1966).
6.16.1.5
Gaucher's Disease
intercondylar noteh. On a lateral radiograph of the
knee the irregularity is usually seen posterior to the
intercondylar eminen ce.
6.16.2.2
Chondrodysplasia Punctata
Autosomal recessive and dominant forms of chon-
drodysplasia punctata are recognised, and the disor-
der occurs in 1 in 110,000 births. There is small punc-
tate calcification of varying size in the epiphysis that
occurs before the ossification centres. The dominant
form (Conradi-Hunerman) involves the long bones
and the spine. It is non-Iethal and there is asymmetri-
cal shortening of the limbs (Fig. 6.13).

Gaucher's disease is an autosomal recessive condi-

tion with a maximum prevalence in Ashkenazi Jews,
among whom its incidence is as high as 1 in 2,500.
The dis order is due to adefeet in the activity of the
enzyme ß-glucosidase and three forms are recog-
nised. The infantile and juvenile forms are lethai,
while the adult type is compatible with a normallife
span. Presentation is usually after birth with spleno-
megaly and neuropathie problems. Skeletal changes
do not occur until later in childhood, and involve
metaphyseal widening, necrosis of femoral heads
and infarction of the long bones (HODSON et al.
1976).

6.16.1.6
Niemann-Pick Disease
Niemann-Pick disease occurs due to abnormal accu-
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).
6.16.2
Epiphyseallrregularity
6.16.2.1
Normal Variant
Irregular ossification of the epiphyses around the
knee joint is common, occurring in approximately
80% of children of about 4 years of age. The changes
may be unilateral or bilateral and are more frequent
in girls. The irregularity is most often seen on the
lateral aspect of the knee but very rarely involves the
Fig.6.13. AP radiograph of the knee in chondrodysplasia
punctata. There are stippled epiphyses, with multiple calcified
punctata of varying size
The recessive form shows symmetrical rhizomelic
shortening of the limbs with mild spinal involve-
ment. There are facial and ocular abnormalities and
death usually occurs within 2 years. There is diaph-
yseal thickening and metaphyseal splaying, humeral
involvement being more severe than in the femur
(SHEFFIELD et al. 1976).
6.16.2.3
Hypothyroidism
Congenital hypothyroidism is usually due to absence
of functioning thyroid tissue. There is delayed skel-
etal maturation in both the appearance and the ossi-
fication of the epiphysis. The epiphyses are fragment-
Congenital and Developmental Abnormalities of the Knee
ed and stippled. The stippling is coarser with larger
fragments than in chondrodysplasia punctata (CHEW
1991).
6.16.2.4
Multiple Epiphyseal Dysplasia
Multiple epiphyseal dysplasia is the commonest con-
genital skeletal dysplasia, and invariably involves the
proximal femoral epiphysis if epiphyses around the knee
are affected. There is shortening of the limbs and irregu-
lar mottled calcification of the epiphyses. The epiphyseal
irregularity may cause premature degenerative disease
(SPRANGER 1976; VAN MOURIK et al. 2001).
6.16.2.5
Spondylo-epiphyseal Dysplasia
There are five subcategories of spondylo-epiphyseal
dysplasia (SED): SED congenita (mild coxa vara), SED
congenita (severe coxa vara), SED tarda (autosomal
recessive and dominant), SED tarda (x-linked) and SED
tarda (with progressive arthropathy) (WYNNE-DAVIES
et al. 1985). There is wide variation in the presentation
of the disorder, with varying severity. The epiphyseal
appearance can varyfrom normal to severe fragmenta-
tion and there also may be some metaphyseal irregular-
ity (SPRANGER and LANGER 1970).
6.16.2.6
Fetal Warfarin Syndrome
Fetal warfarin syndrome results from maternal inges-
tion of warfarin, whieh causes bleeding in the primor-
dial cartilage in the fetus. At birth there is stippling
of the unossified epiphysis, most commonly seen in
the proximal femur and calcaneum. The appearances
disappear by 1 year of age (JOHNSON 1979).
6.16.2.7
Trisomy18
Trisomy 18 is more commonly seen in females. The
child is usually hypotonie with abnormal facies. The
child does not usually survive beyond 6 months. There
are associated cardiac and renal anomalies. The epiph-
yses are stippled (FRANCESHINI et al. 1974).
6.16.2.8
Trisomy21
Trisomy 21 is the commonest chromosomal defect,
classieally presenting with mental retardation, abnor-
89
mal facies and hypotonia. There are widespread skull,
ehest, gastrointestinal and cardiac anomalies. There
is metaphyseal splaying of the long bones and irregu-
lar stippling of the epiphyses.
6.16.2.9
Zellweger (Cerebrohepatorenal) Syndrome
Zellweger syndrome is an autosomal recessive con-
dition with muscle hypotonia, hepatomegaly, renal
cysts and facial and cerebral dysgenesis. There is stip-
pling of the epiphysis. Death occurs in early infancy
(BAROLETTI et al.1978).
6.16.2.10
Transplacental Acquired Infections
See Sect. 6.11.
6.16.3
Metaphyseal Widening and Splaying
6.16.3.1
Rickets
Deficiency of vitamin D or its metabolie derivatives
causes failure in bone mineralisation and in cartilage,
causing riekets (RUSSELL and HILL 1974). Such defi-
ciency may be dietary or be due to familial vitamin
D resistance (x-linked hypophosphataemia). The fea-
tures manifest first in the rapidly growing bone ends,
particularly around the wrist, knee and proximal
humerus. There is widening of the growth plate, cup-
ping, fraying and irregularity of the distal metaphyses
and generalised loss ofbone density (Fig. 6.14). Over-
all skeletal maturation may be delayed (SWISCHUK
and HAYDEN 1979).
6.16.3.2
Trisomy21
See Sect. 6.16.2.8.
6.16.3.3
Hypophosphatasia
Hypophosphatasia is an autosomal recessive condi-
tion due to a deficiency of serum and alkaline phos-
phatase, with increased urinary phospho-ethanol-
amine. There are four forms that decrease in the
severity of symptoms with increasing age. In the neo-
natal form, death usually occurs within 6 months of
birth. There is severe hypotonia, and radiologieally
90
Fig. 6.14. AP radiograph of the knees in rickets. There is genu
varum. The metaphyses are splayed and irregular, with widen-
ing of the growth plates
bone mineral density is reduced with increased lia-
bility to fracture. With the infantile form symptoms
occur within 6 months; there is cupping and fraying
of the metaphyses and widening of the growth plate.
Bone density is reduced, but there may be premature
sutural fusion in the skull. The childhood form pres-
ents between 6 months and 2 years with bowed legs,
ge nu valgum and dental problems. Radiographically
there is widening of the growth plate and metaphy-
seal splaying (similar to rickets), but no craniosteno-
sis. Osteomalacia is the only sign of the adult form
(JAMES and MOULE 1966).
6.16.3.4
Copper Deficiency
Risk factors for copper deficiency are prematurity,
total parental nutrition, malabsorption and low
copper diet. Normal copper and caeruloplasmin levels
do not exclude the diagnosis. Copper deficiency is
associated with shaft fractures, abnormal splayed
metaphyses, periosteal re action along the shaft of the
long bones and osteopenia (SHAW 1988).
6.17
Patella
6.17.1
Patellar Abnormalities
Abnormalities of the patella range from complete
absence to variations in ossification. Ossification of
the patella begins at 2-3 years of age in girls and 4
K. Johnson and A. M. Davies
years in boys. Once ossified, the bony patella has two
facets that articulate with the femoral condyles. The
lateral facet is the larger but is less steeply angled.
Imaging prior to ossification needs to be with
either ultrasound or MR imaging (WALKER et al.
1991; BAR-ON et al. 1995). Ultrasound has the advan-
tage that sedation is not required and the examina-
tion can involve kinematic tracking of the unossified
patella (MILLER et al. 1998).
The commonest ossification variant is the bipar-
tite patella, which occurs in between 1% and 6% of
the population. It is slightly more often unilateral
than bilateral and can be asymmetrical. The extra
ossification centre is usually superolateral but it can
occur laterally, inferiorly or medially (OGDEN et al.
1982). The accessory ossification centres are attached
to the main body of the patella by fibrocartilaginous
tissue. Radiographically, differentiation from a frac-
ture can be difficult but is aided by the smooth
semilunar interface between the ossification frag-
ments and by the characteristic superolateral posi-
tion. These accessory ossification centres may be
asymptomatic or be associated with pain and dis-
comfort or may represent stress fracture (LAWSON
1985; GREEN 1975).
6.17.2
Absent Patella
Complete absence of the patella rarely occurs in iso-
lation (BERNHANG and LEVINE 1973); rather,it usual-
ly forms apart of the nail-patella syndrome (onycho-
osteodysplasia). This syndrome comprises a combi-
nation of nail abnormalities, radial head dislocation,
capitellar deformity, iliac horns and renal disease.
The patella is absent in 14% of cases and dysplastic in
up to 87% (FAURE and PETREL 1968). Absence of the
patella can be associated with wasting of the quad-
riceps muscles and hypoplasia of the femoral con-
dyles.
6.18
Congenital Dislocation of the Patella
Prior to determining whether the patella is dislocat-
ed, an assessment of the normal position needs to be
made. The position of the patella prior to the age of 3
years can only be done with either ultrasound or MR
imaging, and as yet there are no published data on
normallandmarks. Ultrasound can show complete
Congenital and Developmental Abnormalities of the Knee
dislocation and loss of congruity of the articulation
with the femur, but no assessment of the ability of
ultrasound to detect mild patellar subluxation has
been published (MILLER et al.1998).
In the skeletally mature knee, the lateral radio-
graph can be used to assess the normal position of
the patella. The position is obtained from the ratio
of the length of the patellar tendon to the diagonal
length of the patella (INsALL and SALVATI 1971), the
normal ratio being approximately 1, i.e. the diagonal
length of the patella should equal the length of the
patellar tendon.
In the skeletally immature knee this ratio is inac-
curate owing to the underestimation of the size of the
unossified patella. A separate ratio has been derived
for use in the immature skeleton (KOSHINO and SUGI-
MOTO 1989) using the midpoint of the epiphyseallines.
The length between the midpoints of the distal femoral
and proximal tibial epiphyseal growth plates is calcu-
lated; this is the femoral-tibial distance (FT). The patel-
lar axis (PT) is measured from the tibial midpoint
to the centre of the patella. A PT/FT ratio of 0.9-1.3
is considered normal. A deviation of more than 20%
from normal is considered significant (Fig. 6.15).
Fig.6.15. Demonstration of the size of the patella in ossified
skeleton from a lateral radiograph of the knee. The midpoint
of the growth plate of both the tibia and the femur is calcu-
lated. A connecting line is drawn between these two points
(this is the femoral-tibialline, PD. A line from the midpoint
of the patella to the midpoint of the tibia is drawn (this is the
patella-tibialline, PD. The PT1FT ratio is calculated; a ratio of
0.9-1.3 is considered normal
91
A high-Iying patella (patella alta) is associated
with contractures of the quadriceps muscles, which
can be seen as abnormal signal on MR imaging, patel-
lar subluxation and Sinding-Larsen-Johansson dis-
ease.An abnormally low-Iying patella (patella baja) is
associated with achrondroplasia and poliomyelitis.
On the axial projection there is some dispute over
the variation in angulation that can be considered
to be within normal limits. This is particularly true
if radiographs and axial CT measurements are com-
pared (MARTINEZ et al. 1983). On the axial projec-
tion, lines are drawn along the patellofemoral sulcus.
The angle of intersection is known as the sulcus
angle. The average value is 1420 on plain radiographs
(BRATTSTRÖM 1964) and 121 0 on CT (STANDFORD et
al. 1988). An abnormally shallow sulcus is associated
with an increased incidence of patellar subluxation
(BRATTSTRÖM 1964).A congruence angle is obtained
from two lines: one bisecting the sulcus angle and
one from the sulcus angle to the apex of the patella.
This congruence angle should be less than 160 on
plain radiographs (MERCHANT et al. 1974) and no
more than 100 on CT (STANDFORD et al. 1988)
(Fig.6.16).
Fig.6.16. Skyline views of the patella. The left knee dem-
onstrates the sulcus angle. This is the angle of intersection
of lines drawn along the femoral condyles. The right knee
demonstrates the congruence angle. This angle is formed
from a line bisecting the sulcus angle and another drawn
from the apex of the sulcus angle through the inferior point
of the patella. The congruence angle should be less than
16°
 92
a tibia will abduct and flex and an abnormal patello-
b ~------=
Fig. 6.17. Bilateral lateral subluxation of the patella with AP
(a) and skyline (b) views of the knee
Dislocation of the patella may be congenital, recur-
rent or habitual. Congenital dislocation occurs at
birth but may not be detected until infancy. Recur-
rent dislocation occurs more commonly in adoles-
cence and habitual dislocation is associated with con-
tractures of the quadriceps muscles with the patella
dislocating on knee flexion but returning to anormal
position on full extension (Fig. 6.17).
Congenital dislocation is rare. It usually occurs
in a lateral direction, and can be unilateral or bilat-
eral. There is a familial tendency (GREEN and WAUGH
1968; MACNAB 1952).It is associated with othercondi-
tions such as Down's syndrome, arthrogryposis mul-
tiplex and spinal abnormalities (LALAIN et al. 1987;
DUGDALE and RENSHAW 1986; ANDERSEN 1958). The
dislocation may reflect the primary congenital abnor-
mality but in some cases fibrous bands inserting into
the lateral aspect of the patella and a tight lateral cap-
sule along with abnormalities of the quadriceps have
been described (STANISAVJEVIC et al.1976). It should
be considered when there is a fixed flexion deformity
at birth, particularly in the presence of lateral tibial
rotation (GREEN and WAUGH 1968). Ultrasound can
detect the dislocation at an early age (WALKER et al.
1991). Sonographically there is deficiency of the fem-
oral condyles and a shallow femoral groove, with the
unossified patella in a lateral position. Uhrasound
may be used to dynamically track the position of the
K. Johnson and A. M. Davies
patella in suspected cases and in those with other
congenital abnormalities (MILLER et al. 1998). Radio-
graphically the patella is smaller than expected, with
no patella ridge and a less obvious quadriceps soft
tissue shadow. The patella may be normally sited on
extension views and only dislocate on flexion. When
dislocated, the patella may rotate up to 90°. There is
lateral rotation of the tibia so that a lateral view of the
femur will give an anteroposterior projection of the
tibia. The lateral tibial rotation has been confirmed
with CT (LALAIN et al.1987). An axial projection will
show a shallow sulcus and possible hypoplasia of the
femoral condyles. If the condition is undiagnosed, the
femoral joint may develop on the lateral aspect of the
femur (STERN 1964).
The first episode of patellar subluxation or dislo-
cation in adolescence is nearly always laterally. It is
often traumatic with disruption of the medial reti-
naculum but can occur spontaneously due to a con-
genital, developmental or acquired abnormality of
the knee or quadriceps. The unpredictable sponta-
neous dislocation occurs around the age of 15 years
and is more common in girls. Predisposing factors
are increased ligamentallaxity in particular associ-
ated with Down's and Marfan's syndromes, increased
genu valgum, an abnormally located lateral tibial
tuberosity, a shallow patella sulcus or an abnormally
shallow sloping lateral patella. If the lateral patellar
retinaculum is too tight or the patella tendon too
long, there will be abnormal pull of the quadriceps
muscles, tending to cause the patella to dislocate.
There is sometimes a family history of instability.
Lateral radiographs may show patella aha, while the
anterior-posterior radiographs may show the patella
to be normally or laterally sited. Axial views demon-
strate flattening of the lateral patella edge and a shal-
low patellofemoral sulcus (KUJALA et al. 1989; Ko et
al. 1999; INsALL et al. 1972). There is a higher inci-
dence of osteochondral fractures in these children.
Habitual dislocation usually presents by 10 years
of age, the initial history being of the knee giving way
or of limited knee flexion. The dislocation is due to
disordered quadriceps contraction and there may be
a his tory of intramuscular injections.
6.19
Fabella
The fabella is a sesamoid bone in the tendon of the
lateral gastrocnemius muscle. It may be oval or mul-
Congenital and Developmental Abnormalities of the Knee
tipartite in appearance and is estimated to occur in
approximately 12% of the population, being either
unilateral or bilateral. It is most commonly an inci-
dental finding but has been known to cause pain and
tenderness in some teenagers (WEINER et al. 1977).
6.20
Popliteal Pterygium Syndrome
Popliteal pterygium syndrome is a disease that only
affects the knees; it is usually bilateral but can be
unilateral. There are flexion contractures of the knee
associated with webbing. There is an association with
cleft lip and palate, genital abnormalities and syndac-
tyly of the toes. The popliteal pterygium syndrome
has an autosomal dominant inheritance, as compared
with autosomal recessive for the multiple pterygium
syndrome.
6.21
Congenital Bone Dysplasias
The classification of skeletal dysplasias lists over 200
different syndromes and conditions (WYNNE-DAVIS et
al.1985; SPRANGER et al.1974). The dysplasias include
disorders of bone growth, maturation and structure.
If the spine or limb length is below the third cen-
tile, it is classified as a dwarfism. Dwarfism is sub-
divided into rhizomelic (proximallimb shortening),
mesomelic (middle segment shortening) and acro-
melic (distallimb shortening).
As well as by limb length, the dysplasia may be
eategorised as lethai or non-Iethal and by the site of
abnormality, namely epiphyseal, metaphyseal, spon-
dylo-epiphyseal and spondylometaphyseal. The clas-
sifieation also includes the numerous storage disor-
ders. Other non-skeletal manifestations should also
be included in the classifieation process, such as
maturity, abnormal fades and metabolie investiga-
tions.
For radiological diagnostic purposes alimited skel-
etal survey is needed that includes the lateral skulI,
anterior-posterior and lateral spine, ehest (including
shoulders), pelvis/hips, one forearm, posterior-ante-
rior hand/wrist on single film (WYNNE-DAVIS et al.
1985; SPRANGER et al. 1974).
93
References
Achterman C, Kalamche A (1979) Congenital defidency of the
fibula. J Bone Joint Surg 61B:133-137
Andersen PT (1958) Congenital deformities of the knee joint in
dislocation of the patella and achondroplasia. Acta Orthop
Scand 28:27-50
Andersson AP, Ellitsgaard N (1992) Aplasia of the anterior cru-
date ligament with a compensating posterior crudate liga-
ment. Acta Orthop Belg 2:240-242
Austwiek DH, Dandy DJ (1983) Early operation for congenital
subluxation of the knee. J Paediatr RadioI3:85-87
Barnes CL, McCarthy RE, VanderSchildren JL et al (1988) Dis-
coid lateral meniscus in a young child: case report and
review of the literature. J Pediatr Orthop 3:85-87
Bar-On E, Howard CB, Porat S (1995) The use of ultrasound in
the diagnosis of atypieal pathology in the unossified skel-
eton. J Pediatr Orthop 15:817-820
Bartoletti S, Armfield SC, Lesema-Medina J (1978) The cere-
brohepatorenal (Zellweger) syndrome: report of four cases.
Radiology 127:741-745
Bathfield CA, Beighton PH (1978) Blount disease: a review
of the etiologieal factors in 11 0 patients. Clin Orthop
135:29-33
Bell MJ, Atkins RM, Sharrard WJW (1987) Irredudble congeni-
tal dislocation of the knee. Aetiology and management. J
Bone Joint Surg 69B:403-406
Bensahel H, Hjelmstedt A, Bjerkreim I et al (1989) Congenital
dislocation of the knee. J Pediatr Orthop 9: 174-177
Bernhang AM, Levine SA (1973) Familial absence of the patel-
la. J Bone Joint Surg 55A:I088-1090
Blount WP (1937) Tibia vara: osteochondrosis deformans
tibiae. J Bone Joint Surg 19 A: 1-28
Boyd HB (1982) Pathology and natural history of congenital
pseudarthrosis of the tibia. Clin Orthop 166:5-13
Boyd HB, Sage FP (1958) Congenital pseudoarthrosis of the
tibia. J Bone Joint Surg 40A:1245-1270
Brattstöm H (1964) Shape of the intercondylar groove normal-
ly and in recurrent dislocation of the patella. A clinieal and
x-ray anatomieal investigation. Acta Orthop Scand Suppl
68:1-138
Caffey J (1947) Prenatal bowing and thiekening of tubular
bones, with multiple cutaneous dimples in arms and legs. A
congenital syndrome of mechanical origin. Am J Dis Child
74:543-562
Canale ST, Beaty JH (eds) (1991) Pediatrie operative orthopae-
die. Mosby Year Book, St Louis, pp 188-212
Carlson DH, O'Connor J (1976) Congenital dislocation of the
knee. AJR 127:465-468
Carter JR, Leeson MC, Thompson GH et al (1988) Late onset
tibia vara: a histopathologieal analysis: a comparative eval-
uation with infantile tibia vara and slipped capital femoral
epiphysis. J Pediatr Orthop 8:187-195
Chew FS (1991) Radiologieal manifestation in the musculosk-
eletal system of miscellaneous endrocrine disorders. Radiol
Clin North Am 29:135-147
Craig JG, Cramer KF, Cody DD et al (1999) Premature
partial closure and other deformities of the growth plate:
MR imaging and three-dimensional modelling. Radiology
210:835-843
Curranio G, Kirks DR (1977) Lateral widening of epiphyseal
plates in children with bowed legs. AJR 129:309--312
Curtis BH, Fisher RL (1969) Congenital hyperextension with
94
anterior subluxation of the knee. J Bone Joint Surg
51A:255-269
Curtis BH, Fisher RL (1970) Heritable congenital tibiofemoral
subluxation. Clinieal feature as surgieal treatment. J Bone
Joint Surg 52A:1104
Dalinka MK, Coren G, Hensinger R et al (1974) Arthrography
in Blount's disease. Radiology 113:161-164
Dietz WH, Gross WL, Kirkpatriek JA (1982) Blount disease
(tibia vara): Another skeletal disorder associated with child-
hood obesity. J Pediatr 101:735-737
Ducou Le Point H, Mousselard H, Rudelli A et al (1995)
Blount's disease: Magnetie resonance imaging. Pediatr
Radiol25: 12-14
Dugdale TW, Renshaw TS (1986) Instability of the patel-
Iofemoral joint in Down's syndrome. J Bone Joint Surg
68A:405-413
Faure C, Petrel P (1968) Hereditary onycho-osteodysplasia.
Ann Radiol11:376-388
Foreman KA, Robertson WW (1985) Radiographie measure-
ment of infantile tibia vara. J Pediatrie Orthop 5:452-455
Frantz CH, O'Rahilly R (1961) Congenital skeletallimb defi-
ciencies. J Bone Joint Surg 43A:1202-1224
Franceshini P, Fabris C, Ponzone A et al (1974) Skeletal alter-
ations in Edwards disease (trisomy 18 syndrome). Ann
RadioI17:361-367
Gaary E,Grolin JB,Jaramillo D (1996) Pseudotumour and arthrop-
athy in the knees of a haemophiliac. Skeletal RadioI25:85-87
Gall JC, Stern AM, Poznanski AK et al (1972) Oto-palatal-
digital syndrome. Comparison of c1inieal and radiologieal
manifestations in males and females. Am J Hum Genet
24:24-36
Girdwood TG, Gibson WJA, MacIntosh TF (1969) Craniome-
taphyseal dysplasia congenita - Pyle's disease in a young
child. Br J RadioI42:299-303
Green JP, Waugh W (1968) Congenitallateral dislocation of the
patella. J Bone Joint Surg 50B:285-289
Green WT (1975) Painful bipartite patella: areport of three
cases. Clin Orthop 110: 197-200
Harty MP, Markowitz RZ, Rustien RM et al (1994) Imaging
features of human immunodeficiency virus infection in
infants and children. Semin RoentgenoI29:303-314
Henderson Re, Lechner CT, DeMasai RA et al (1990) Variability
in radiographie measurement of bowlegged deformity in
children. J Pediatr Orthop 10:491-494
Hermann G, Gilbert MS, Abdelwahab IF (1992) Hemophilia:
evaluation of musculoskeletal involvement with CT, Sonog-
raphy and MR imaging.AJR 158:119-123
Heyman SH, Herndon CH (1949) Congenital posterior angula-
tion of the tibia. J Bone Joint Surg 41A:476-488
Heselson NG, Raad MS, Hamersma H et al (1979) The radio-
logieal manifestation of the metaphyseal dysplasia (Pyle's
disease). Br J RadioI52:431-440
Heyman CH, Herndon CH, Heiple KG (1959) Congenital pos-
terior angulation of the tibia with talipes calcaneus. A
long-term report of eleven patients. J Bone Joint Surg
41A:476-88
Hodson P, Goldblatt J, Beiehton P (1976) Non-neuropathic
Gaucher disease presenting in infancy. Br J Radiol
49:1018-1022
Insall J, Salvati E (1971) Patella position in the normal knee
joint. Radiology 101:101-112
Insall J, Goldberg V, Salvati E (1972) Recurrent dislocations and
the high riding patella. C!in Orthop 88:67-69
K. Johnson and A. M. Davies
James W, Moule B (1966) Hypophosphatasia. C!in Radiol
17:368-376
Johansson E, Aparisi T (1983) Missing cruciate ligament in
congenital short femur. J Bone Joint Surg 65A:ll09-1115
Johnson JF (1979) Case report. Coumadin embryonopathy.
Skeletal RadioI3:244-246
Kaufmann KJ (1973) Osteodysplasia (Melniek-Needles). Semin
RoentgenoI8:182-186
Klatte EC, Franken EA, Smith JA (1976) The radiographie spec-
trum in neurofibromatosis. Semin Roentgenol11:17-33
Ko JY, Shih CH, Wenger DR (1999) Congenital dislocation of
the knee. J Pediatr Orthop 19(2):252-259
Koshino T, Sugimoto K (1989) New measurement of patella
height in the knees of children using the epiphyseal line
midpoint. J Pediatr Orthop 9:216-219
Kujala UM, Osterman K, Kurmanu M (1989) Patellofemoral
relationship in recurrent patellar dislocation. J Bone Joint
Surg 71A:788-792
Kuhns LR, Slovis T, Hernandez R et al (1977) Knee matura-
tion as a differentiating sign between congenital rubella
and cytomegalovius infection. Pediatr RadioI6:36-38
Lachman R, Crocker A, Schulman J et al (1973) Radiologieal
findings in Niemann-Piek disease. Radiology 108:659-664
Lalain JJ, Lerat JL, Moyen B et al (1987) Les luxations congeni-
tales de rotule. Apport du scanner-indications chirurieales.A
propos de 22 cases Rev Chir Orthop 73 [SuppI2]:152-156
Laor T, Jaramillo D, Hoffer FA et al (1996) MR imaging in con-
genitallower limb deformities. Pediatr RadioI26:381-387
Laurence M (1967) Genu recurvatum congenitum. J Bone Joint
Surg49B:121-134
Laville JM, Chau E, Willemen L, Kohler R, Garin C (1999)
Blount's disease: c1assification and treatment. J Pediatr
Orthop 8B:19-25
Lawson JP (1985) Symptomatie radiographie variants in
extremities. Radiology 157:625-631
Levine AM, Drennan JC (1982) Physiologieal bowing and tibia
vara: the metaphyseal-diaphyseal angle in the measurement
of bowleg deformities. J Bone Joint Surg 64A: 1158-1163
Loder RT, Schaffer JJ, Bardenstein MB (1991) Late-onset tibia
vara. J Paediatr Orthop 11: 162-167
MacEwan DW, Dunbar JS (1958) Radiology study of physiolog-
ie knock knees in childhood. J Can Assoc RadiollO:59-63
Macnab I (1952) Recurrent dislocation of the patella. J Bone
Joint Surg 34A:957-967
MacPherson RI (1974) Craniodiaphyseal dysplasia, a disease
or group of diseases? J Can Assoc RadioI25:22-33
Martinez S, Korokobin M, Fondren FB, Hedlund LW, Goldner JL
(1983) Diagnosis of patellofemoral malalignment by com-
puted tomography. J Comput Assist Tomogr 7:1050-1053
Malumed J, Hudanieh R, Collins M (1999) Congenital absence
of the anterior and posterior cruciate ligaments in the
presence of bilateral absent patellae. Am J Knee Surg
12:241-243
Melniek JC, Needles CF (1966) An undiagnosed bone dyspla-
sia; a 2 family study of 4 generations and 3 generations. AJR
97:39-48
Merchant AC, Mercer RL, Jacobsen RH, Cool CR (1974) Roent-
genographie analysis of patellofemoral congruence. J Bone
Joint Surg 56A:1391-1396
Miller TT, Shapiro MA, Schultz E et al (1998) Sonography of
patella abnorma!ities in children.AJR 171:739-742
Moseley JE (1974) Skeletal change in the anaemias. Semin
RoentgenoI9:169-184
Congenital and Developmental Abnormalities of the Knee
Nagele M, Bruning R, Kunze V et al (1995) Hemophilic arthrop-
athy of the knee joint. Static and dynamic Gad-DPTA
enhanced MRI. Eur RadioI3:547-552
Newell RLM, Durbin FC (1976) The aetiology of congenital
angulation of tubular bones with constriction of the med-
ullary canal and its relationship to congenital pseudoar-
throsis. J Bone Joint Surg 58B:444-447
Niebauer JJ, King DE (1960) Congenital dislocation of the knee.
J Bone Joint Surg 42A:207-225
Ogden JA, McCarthy SM, Jokl P (1982) The painful bipartite
patella. J Pediatr Othop 2:263-269
Pappas AM (1983) Congenital abnormalities of the femur and
related lower extremity malformations: classification and
treatment. J Pediatr Orthop 3:45-60
Pappas AM (1984) Congenital posteromedial bowing of the
tibia and fibula. J Pediatr Orthop 4:525-531
Rathgeb JM, Ramsey PL, Cowell HR (1974) Congenital kypho-
scoliosis of the tibia. Clin Orthop 103:178-190
Resnick CS, Young JWR, Levine AM et al (1990) Osteofibrous dys-
plasia (ossifying fibroma) of the tibia. Skeletal Radio119:217
Russell JGB, Hill LF (1974) True fetal rickets. Br J Radiol
47:732-734
Ryan JR, Perrin JC, Morewa LG (1978) Congenital synostosis
of the knee. Clin Orthop 135:34-35
Sachdev M, Berry K, Chawla S (1982) Osseous manifestations
in congenital syphilis. A study of 55 cases. Clin Radiol
33:319-323
Salenius P,Vankka E (1975) The development of the tibiofemo-
ral angle in children. J Bone Joint Surg 57 A:259-269
Sharrad WJW (1976) Problems of childhood. Knock knees and
bow legs. Br Med J i:826-827
Shaw JCC (1988) Copper deficiency and non-accidental injury.
Arch Dis Child 63:448-455
Sheffield LJ, Danks DM, Mayne V et al (1976) Chrondysplasia
punctata - 23 cases of a mild and relatively common vari-
ant. J Paediatr 89:916-923
Shogry MEC,Armstrong P (1990) Reactive bursa formation sur-
rounding an osteochondroma. Skeletal RadioI19:465-467
Shopfner CF, Coin CG (1969) Genu varus and valgus in chil-
dren. Radiology 92:723-732
Sillence DO, Senna, Danks M (1979) Genetic heterogeneity in
osteogenesis imperfecta. J Med Genet 16: 101-116
Silverman JM, Mink JH, Deutsch AL (1989) Discoid menisci of
the knee: MR imaging appearance. Radiology 173:351-354
Södergard J, Ryoppy S (1990) The knee in arthrogryposis mul-
tiplex congenital. J Pediatr Orthop 10:177-182
Spranger J (1976) The epiphyse al dysplasias. Clin Orthop Relat
Res 114:46-60
95
Spranger J, Langer LO, Wiedman HR (1974) Bone dysplasias.
An atlas of constitutional disorders of skeletal develop-
ment. Saunders, Philadelphia
Spranger J, Cremin B, Beighton P (1982) Osteogenesis imper-
fecta congenita: features and prognosis of a heterogeneous
condition. Pediatr RadioI12:21-27
Spranger JW, Langer CO (1970) Spondylepiphyseal dysplasia
congenital. Radiology 94:313-322
Stanford W, Phelan J, Kathol MH et al (1988) Patellofemoral
joint motion: evaluation by ultrafast computed tomogra-
phy. Skeletal RadioI17:487-492
Stanisavljevic S, Zemenick G, Miller D (1976) Congenital irre-
ducible, permanent lateral dislocation of the patella. Clin
Orthop 116:190-199
Stark JE, Siegel MJ, Weinberger E, Shaw DW (1995) Discoid
menisci in children. MR feature. J Comput Assist Tomogr
19:608-611
Stern M (1964) Persistent congenital dislocation of the patella.
Report of a case. J Int Coll Surg 41:654-656
Swischuk LF, Hayden CK (1979) Rickets. A roentgenographic
scheme for diagnosis. Pediatr RadioI8:203-208
Thomas NP, Jackson AM, Aichroth PM (1985) Congenital
absence of the anterior cruciate ligament. J Bone Joint Surg
67B:572-575
van Mourik JBA, Weerdenburg JP, Verhar JAN (2001) Magnetic
resonance imaging of multiple epiphyseal dysplasia (type
2) of the knee. Radiography 7:61-64
Voutsinas S, Wynne-Davies R (1983) The frequency of malig-
nant disease in diaphyseal aclasis and neurofibromatosis. J
Med Genet 20:345-349
Walker J, Rang M, Daneman A (1991) Ultrasonography of
the unossified patella in young children. J Pediatr Orthop
11:100-102
Weiner D, MacNab I, Turner M (1977) The fabella syndrome.
Clin Orthop 126:213-215
Willms R, Hartwig CH, Bohm P, Sell S (1997) Malignant trans-
formation of a multiple cartilaginous exostosis - a case
report. Int Orthop 21:133-136
Wolfgang GL (1984) Complex congenital abnormality of the
bone extremities: femoral bifurcation, tibial hemimelia, and
diastasis of the ankle. Case report and review of the litera-
ture. J Bone Joint Surg 66A:453-458
Wuisman PIJM, Jutte PC, Ozaki T (1997) Secondary chondro-
sarcoma in osteochondromas: medullary extension in 15 of
45 cases. Acta Orthop Scand 68:396-400
Wynne-Davies R, Hall CM, Apley AG (1985) Atlas of skeletal
dysplasias. Churchill Livingstone, Edinburgh
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

A.M. Davies et al. (eds.), Imaging of the Knee

© Springer-Verlag Berlin Heidelberg 2003
98
Table 7.1. Roles of the radiologist
To identify the radiologieal findings
To differentiate abnormalities from normal variation in the
growing skeleton
To differentiate traumatie from non-traumatie abnormalities
To suggest a mechanism of causation for traumatie lesions
To date the injury
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
degree of force is less than would have been expected to
cause the injury
cortices, mature bone becomes dominant. This early
bone is very vascular and more porous than the
maturing bone of older children and adults and
better able to tolerate deformation. Pores retard pro-
gression of a fracture line but increase the likeli-
hood of compression, and it is this that makes green-
stick and torus, or buckle, fractures more common in
childhood (OGDEN 1990). Immature periosteum also
differs from that in the adult, being thieker and more
loosely attached to the underlying cortex. It minimis-
es displacement of diaphyseal fractures and is capa-
ble of faster callus and membranous bone formation
(CRILLY 1972). Towards the end of the bone, adjacent
to the growth plate, is the metaphysis, a site charac-
terised by decreased cortical bone and increased tra-
becular bone. The trabecular bone undergoes exten-
sive remodelling as the primary spongiosa is trans-
formed into more mature secondary spongiosa. This
area of transformation is structurally weak during
the period of rapid growth during infancy (ALEXAN-
DER 1976).
Due to the child's more flexible skeleton and the
increased amount of stretch allowed by the soft tis-
sues, the majority of childhood fractures are undis-
placed. Greater strength of the joint capsule and lig-
aments compared with the metaphyseal bone also
me ans that fractures occur much more readily than
dislocations.
7.3
Patterns of Skeletallnjury
With broadening of the definition of child abuse, the
percentage of children with fractures has decreased;
reported frequencies vary from 11 % to 55%, with
age having a significant influence (EBBIN et al. 1969;
GALLENO and ÜPPENHEIM 1982; HERNDON 1983;
S. Chapman
KOGUTT et al. 1974; LAUER et al. 1974; MERTEN et al.
1983; RYAN et al. 1977). EBBIN et al. (1969) did not
find any fractures in children over the age of 5 years.
AKBARNIA et al. (1974), in aseries of 217 abused
infants, found that 78% of fractures occurred below
the age of 3 years and 50% occurred below 1 year of
age.
7.3.1
Subperiosteal New Bone Formation
In the infant, as opposed to the older child, the peri-
osteum is loosely adherent to the shaft of a long bone,
but is strongly attached to the physeal cartilage. Sub-
periosteal haemorrhage results in separation of the
osteogenie periosteum from the bony cortex, but the
perichondral attachment is usually preserved, result-
ing in maximum thiekness of haemorrhage along the
shaft and tapering towards the epiphysis. Subperi-
osteal new bone formation (SNBF) is non-specific,
being seen in infectious, metabolie and neoplastic
disease processes; the importance and frequency of
physiologieal SNBF (see below) should also be recog-
nised. SNBF may be the clue to a subtle underlying
fracture and it should prompt the request for coned
radiographs, other views or follow-up radiographs.
SNBF without a fracture is probably the result of an
applied tractional or torsional force, such as when a
child is grabbed by the leg, which displaces the loose-
ly adherent periosteum.
7.3.2
Metaphyseal Fractures
For many years the mechanism of injury at this site
was believed to be that proposed by CAFFEY (1957).
He stated that indirect forces transmitted through
the periosteum produced avulsions of the metaphy-
sis where it was tightly adherent. When a small frag-
ment of bone was avulsed the radiographie finding
of the "corner" fracture resulted (Fig. 7.1). When a
larger crescentic fragment of bone was avulsed and
displaced or viewed obliquely the appearance was
that of the "bucket-handle" fracture (Figs. 7.1, 7.2).
KLEINMAN and co-workers have thrown new light
on this classie lesion associated with abuse (KLEIN-
MAN et al. 1986; KLEINMAN and MARKS 1995,1996,
1998). They have shown that the fundamental abnor-
mality is a complete or incomplete transmetaphy-
seal fracture through the most immature metaphy-
seal primary spongiosa. Furthermore, as the fracture
Non-accidental Injury
line passes peripherally towards the cortex it deviates
away from the physis to undercut a thieker peripher-
al segment that encompasses the subperiosteal bone
collar (Fig. 7.2). Thus, there is separation of a disk of
bone, whieh is thin centrally and thicker peripher-
ally (Fig. 7.3). The radiographie appearance is typi-
cal, but radiographically variable. When the separat-
ed ossified disk is very thin, the radiograph is normal,
although histologieal examination of the resected
bone may demonstrate the lesion. When the fracture
Fig.7.1. Radiograph of the knee showing corner fractures of
the distal femoral metaphysis and a bucket-handle fracture of
the proximal tibia
Fig. 7.2. A large bucket-handle fracture of the proximal tibia with
a markedly thickened peripheral segment on the medial side
99
line is wider there will be alucent line extending
across the metaphysis (Fig. 7.4); the fracture line may
be visible in only one plane. If the radiographie pro-
jection is such that only the thicker peripheral seg-
ment is demonstrable, then the resulting radiograph-
ie finding is the corner fracture (Figs. 7.1, 7.5a). When
the entire fracture is visible and viewed obliquely, the
disk appears to be further separated from the adja-
cent metaphysis and the bucket-handle appearance
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. 7.3. Radiograph of the distal femur showing a radiograph-
ically incomplete thin disk of bone centrally with thicker seg-
ments peripherally
Fig.7.4. A metaphyseal fracture of the proximal tibia, which is
visible only as a thin radiolucent band
 100
a
Fig. 7.5a, b. A metaphyseal fracture of the distal femur. a In the
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
evidence of bony healing. b A follow-up film 8 weeks later is
entirely normal with no evidence of the previous injury
Fig.7.6. Well-advanced healing of a metaphyseal fracture.
There is undulation of the metaphyseal margin due to
ingrowth of cartilage
S.Chapman
As CAFFEY originally pointed out, metaphyseal
fractures are the result of tractional and torsional
forces applied to a limb. This can occur when a limb
is used as a handle or when centrifugal forces act
during violent shaking. Rarely, an isolated metaph-
yseal fracture may be found and correlated with a
plausible accidental event. However, in all other cir-
cumstances this injury in an otherwise normal child
is diagnostic of non-accidental injury.
7.3.3
Epiphyseal Separations
Fractures which involve the growth plate, with or
without involvement of the adjacent epiphysis and/or
metaphysis, as classified by Salter and Harris, occur
most commonly as accidents. The forces which pro-
duce these physeal separations are similar to those
which produce the classic metaphyseal fracture but
their magnitude is greater (KLEINMAN 1998).
7.3.4
Shaft Fractures
In general, shaft fractures are not specific for abuse.
The most common sites are the femur and humerus
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
Children who present after accidental trauma virtu-
ally always have a clear history of the events associ-
ated with the injuries. When there is no history (from
an infant or unhelpful parents) or when there are
suspicious circumstances, one of the many pieces of
the jigsaw that leads to, or excludes, a diagnosis
of non-accidental injury is the dating of those inju-
ries (CHAPMAN 1992). The most comprehensive dis-
cussion of the dating of fractures in non-accidental
injury is by O'CONNOR and COHEN (1998) and the
changes are summarised in Table 7.2.
Non-accidental Injury 101

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

Radiological change Earliest Peak Latest

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

7.4.1 fractures in children are not associated with clinical

Radiological Changes in the Soft Tissues
Haemorrhage and inflammatory exudate begin
immediately after injury. They obliterate the normal
lucent fat planes and well-defined muscle boundar-
ies but for bones surrounded by a large muscle bulk
there may be a delay of many hours before these
changes are apparent radiologically (Fig. 7.7a). In the
absence of a major underlying fracture, these chang-
es pers ist for a few days only, but more severe inju-
ries may be associated with soft tissue changes which
persist for longer. Interestingly, the majority of limb
bruising, either at diagnosis or in the first week after
injury, and the absence of this sign should not taken
as evidence that a minimal force produced a fracture
in a "brittle"bone (MATTHEW et al. 1998).
7.4.2
Response of the Periosteum to Injury
Elevated periosteum is not visible radiographically
until calcium accumulates in the subperiosteal hae-
matoma and the new bone. The first appearance of

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
periosteal new bone is dependent on the age of the
child: the younger the child, the earlier it appears. In
the infant it can be seen radiographically as early as 4
days after injury, but an interval of7-14 days is more
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-
sibility of abuse. Beyond the neonatal period, perios-
teal reaction delayed longer than 21 days may be a
consequence of poor nutrition or vitamin D deficien-
cy. The amount of subperiosteal new bone increases
with repeated trauma and continuing subperiosteal
haemorrhage (leading to exuberant callus forma-
tion), but chronic repetitive trauma may be of such
severity as to destroy periosteal new bone as it devel-
ops, with resultant failure of the usual sequence of
changes with healing.
Ultrasonography has been used to demonstrate
subperiosteal haemorrhages and occult long bone
fractures (GRAIF et al. 1988) before radiographs
confirm the presence of bony injury and may be a
useful supplementary investigation when attempt-
ing to date an acute injury. Radionuclide bone scan-
ning is popular in some centres but its role is
controversial. False-negative rates range from 0.8%
(STY and STARSHAK 1983) to 27% (MERTEN et al.
1983). From the data of these and others (HAASE
et al. 1980; ]AUDES 1984; SMITH et al. 1980) we can
conclude that scintigraphy demonstrates superior
sensitivity for rib fractures, especially at the cos-
tovertebral junction, undisplaced shaft fractures
and subperiosteal haemorrhage. ROSENTHAL et al.
(1976) reported that the earliest fracture demon-
strated with radionuclides was 7 h after the injury
and that if a bone scan obtained 3 or more days
after injury fails to reveal a focallesion, then a frac-
tu re has been excluded. A positive scan steadily
becomes less so as the age of a fracture increases,
but there is no definite time course. The pitfalls of
imaging with a bone-seeking isotope are that sym-
metrical metaphyseal injuries adjacent to normal-
ly "hot" epiphyseal growth plates, some vertebral
body fractures, and many skull fractures may be
undetectable.
7.4.3
Fracture Une Definition
Fresh fractures, including metaphyse al fractures, have
sharply defined margins. With the development of
an osteoclastic response to necrotic bone, the frac-
ture ends become less well defined and the fracture
S.Chapman
widens (Fig. 7.7b). There are no objective criteria to
estimate fracture healing using this sign but it is not
apparent radiologically before 10 days and reaches a
peak at 2-3 weeks.
7.4.4
Callus Formation
With the production of osteoid and its subsequent
calcification and ultimate transformation into bone
there is an increase in density along the fracture line.
With impacted fractures and other fractures where
the periosteum remains intact or little disturbed, this
may be the only evidence of injury and it is, there-
fore, a less valuable sign than subperiosteal new bone
and fracture line clarity. Again, movement may inhib-
it or destroy endosteal callus and a fracture line may
remain clearly visible after a bone has united by peri-
osteal new bone. Endosteal soft callus is first visu-
alised as an ill-defined increase in density and begins
to form soon after the first appearance of periosteal
new bone, i.e. at about 10-14 days. When lamellar
bone bridges the fracture, about 1 week later, the
stage of hard callus has begun. The chronology of
callus formation and fracture consolidation is depen-
dent on age.A birth-related fracture may be united at
1 month, but a similar fracture in an 8-year-old will
be united at 8 weeks.
7.4.5
Remodelling
Estimation of the age of a fracture by this criterion
is very difficult because initial deformity, the volume
of callus produced and the age of the child are major
variables. In the young infant with a stable undis-
placed fracture, the remodelling process may be com-
plete at 3 months, while in the older child with an
angular deformity or a markedly displaced fracture,
remodelling may continue for up to 2 years.
7.5
The Distal Femur
The overall prevalence of femoral fractures in abused
children with fractures is 20% (AKBARNIA et al. 1974;
GALLENO and OPPENHEIM 1982; HERNDON 1983;
O'NEILL et al. 1973; ROSENBERG and BOTTENFIELD
1982).
Non-accidental Injury
7.5.1
Shaft Fractures
Femoral shaft fractures have been estimated to
account for 1.6% of all fractures in children (LANDIN
1983) (Fig. 7.7). There is no association between the
morphological features of the fracture and the likeli-
hood of abuse (REx and KAY 2000). Although there is
some disagreement in the published literature, trans-
verse and oblique fractures seem to be the most
common type, both in abuse and in accidental injury;
spiral fracture is not suggestive of abuse (KING et
al. 1988; LODER and BOOKOUT 1991; SCHERL et al.
2000).
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-
dren under 2 years of age, 19 (79%) had been abused.
BEALS and TuFTS (1983) suggested that as many as
30% of femoral fractures in children less than 4 years
of age were the result of a non-accidental injury.
THOMAS et al. (1991) found that 60% of femoral frac-
tures in children less than 1 year of age were due to
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
of WELLINGTON and BENNET (1987) in which there
were only 14 fractures (36%) due to abuse in 39 chil-
dren under 1 year of age; furthermore, only two were
considered definite abuse and 12 were suspected.
7.5.2
Metaphyseal Fractures
The distal femur is a common site for metaphyseal
fractures and approximately 25% of these fractures
are to be found at this site (KLEINMAN and MARKS
1998; LODER and BOOKOUT 1991; WORLOCK et al.
1986). They can be bilateral or unilateral and are
commonly associated with fractures of the ipsilateral
proximal tibial metaphysis.
103
7.6
The Proximal Tibia
The tibia is a common site of fracture in non-acci-
dental injury, accounting for 7%-18% of the total
(AKBARNIA et al. 1974; GALLENO and ÜPPENHEIM
1982; HERNDON 1983; KING et al. 1988; LEvENTHAL
et al. 1993; LODER and BOOKOUT 1991; WORLOCK
et al. 1986). Metaphyseal fractures considerably out-
number shaft fractures; indeed, the proximal tibial
metaphysis is the most common site for this partic-
ular fracture. There are no specific points relating
to metaphyseal fractures at this site (KLEINMAN and
MARKS 1996).
7.7
The Proximal Fibula
Metaphyseal fractures are less common in the proxi-
mal fibula than in the proximal tibia. When present,
there is usually an associated metaphyseal fracture of
the proximal tibia.
7.8
Differential Diagnosis
7.8.1
Normal Variants
7.8.1.1
Physiological Subperiosteal New Bone Formation
Radiographically, physiological subperiosteal new
bone formation presents as a thin well-defined densi-
ty,2 mm or less in thickness, separate from the cortex
of the bone and restricted to the diaphysis (Fig.7.8)
(KLEINMAN and KWON 1998). The tibia is more com-
monly affected than the femur. The changes are most
often bilateral and symmetrical and are unusual
above the age of 6 months (SHOPFNER 1966).
104
Fig. 7.8. Physiological subperiosteal
new bone formation
7.8.1.2
Metaphyseal Variants
There are many variations which affect the growing
metaphysis and which cause confusion with the inter-
pretation of the c1assic metaphyseal fracture. The
most important are the "step-off" configuration, the
metaphyseal spur and the metaphyseal beak. Four
percent of infants may demonstrate focal cortical
irregularity along the medial aspect of the proximal
tibial diametaphysis. It is bilateral in 25% of cases
(KLEINMAN et al. 1991).
S.Chapman
7.8.2
Birth Trauma
The femur is the third most common bone, after
the c1avic1e and humerus, to suffer a fracture at
birth. Shaft fractures are usually oblique and are
caused by the torsional force applied during diffi-
cult vaginal breech delivery or caesarean section.
The incidence of fractures is higher in infants with
intrinsically osteopenic bones or neuromuscular
disease.
7.8.3
Osteogenesis Imperfecta
Osteogenesis imperfecta (01) is an inherited dis order
of connective tissue resulting from abnormal quanti-
ty and/or quality of type I collagen, the major protein
of bone. The phenotypic presentation is enormously
varied, ranging from perinatal death to normallifes-
pan complicated by only a few fractures. Because
fractures are a feature of the condition, it must be
given serious consideration in any child with unex-
plained fractures. There are four major types in the
Sillence c1assification (SILLENCE et al. 1979), but it
should be noted that there are no strict boundaries
(see Chap.6). Type IV is an uncommon type of
01 (only 5% of the patients in the Sillence study)
with mild to severe bone disease. Patients with this
type neady always have a positive family history,
but because of white sc1erae it may be considered as
an explanation for the child with unexplained frac-
tures.
Although many regard 01 IV as rare, one group
of authors in particular has suggested a higher
prevalence (PATERSON et al. 1983), including those
with normal teeth, type IVA (PATERSON et al. 1987).
The latter study of 78 patients with 01 subtype IVA
has, however, been criticised by others (ABLIN et al.
1990; CARTY and SHAW 1988) who have raised the
possibility that the population inc1udes individuals
with non-accidental injury mistakenly diagnosed
as 01.
New genetic mutations for 01 have been docu-
mented, so it is possible for a child to have a new
mutation for type IV 01, i.e. to have no relevant
family history and also not to have any of the other
signs of 01, such as osteoporosis and multiple wor-
mian bones in the skulI. TAITZ (1987) showed
that the incidence of type IV 01 with fractures
under 1 year of age, no family his tory, otherwise
normal radiology (including no wormian bones)
Non-accidental Injury
and normal teeth is between 1 in 1 million and 1
in 3 million. Thus in a city of 500,000 people with
6,000 births per annum, the incidence would be
one case every 100-300 years. The number of cases
of non-accidental injury with fractures would be
expected to be 15 cases per year. Metaphyseal frac-
tures do occur in 01, but only in the presence of
obvious bone disease with radiologically abnormal
bones (ASTLEY 1979).
Collagen abnormalities can be found after skin
fibroblast culture in approximately 85% of patients
with 01 (BYERS 1993). The true prevalence of bio-
chemical collagen abnormalities in the paediatric
age group is not known and until such data are
available, the results of skin fibroblast culture can
only provide one piece of the jigsaw. Furthermore,
the abnormalities detected in skin fibroblasts may
have nothing to do with alterations in bone and
increased bone fragility. Recent court judgements
in the United Kingdom have upheld the diagnosis
of non-accidental injury even in the presence of
abnormal collagen, when the radiological evidence
has been strongly in favour of abuse. So which chil-
dren should undergo skin biopsy and fibroblast
culture? Three difficult situations in which it may
be appropriate are: (a) when the fracture site is
consistent with the history but the mode of injury
seems too minor to have caused the injury, (b)
when fractures recur in a protected environment,
and (c) when there are no extern al signs of abuse
and the diagnosis seems unlikely. Fibroblast culture
takes at least 3 months and it is important that the
child is protected during this period. This test is
currently not provided as a routine service in the
United Kingdom.
There are case reports of 01 that has initially been
misdiagnosed as non-accidental injury (AUGARTEN
et al. 1993; GAHAGAN and RIMSZA 1991; OJIMA et
al. 1994; PATERSON and McALLION 1987; WARDIN-
SKY 1995; WARDINSKY et al. 1995). However, these
cases have exhibited clinical and/or radiological
abnormalities suggestive of 01 or a missed posi-
tive family history, wh ich should have enabled the
correct diagnosis to be made. These sporadic case
reports do not provide significant evidence for the
view that there is a large group of infants with clini-
cally and radiologically occult bone fragility. The
differentiation of child abuse from 01 has been the
subject of recent reviews (ABLIN et al. 1990; ABLIN
and SANE 1997; CARTY 1998; CHAPMAN and HALL
1997; SMITH 1995).
105
7.8.4
Rarer Bone Dysplasias with Bone Fragility
There are other rare causes of bone fragility in
infancy. They all have recognisable clinical, genetic
and/or radiological manifestations which facilitate
their diagnosis and they should not be confused
with the "normal" infant with a fracture. The list
includes achondrogenesis, osteopetrosis, pycnodys-
ostosis, mucolipidosis II (I-cell disease), metaphyse-
al dysplasia (Jansen type), hypophosphatasia, homo-
cystinuria, Menkes' syndrome (kinky hair disease)
and glycogen storage disease. For a fuller discussion
of these conditions, see Chap. 6.
7.8.5
Temporary Brittle Bone Disease
PATERSON et al. (1993) described 39 children, collect-
ed over a lO-year period, with fractures only in the
first year oflife. The authors speculated that the infants
suffered a temporary collagen defect due to transient
copper deficiency or another metalloenzyme deficien-
cy, although this remains unproved. This hypothetical
condition, as described, bears a striking similarity
to many cases of non-accidental injury (CHAPMAN
and HALL 1997).1t could be argued that if the carers
offer an inappropriate explanation for the injuries or,
indeed, if there is no explanation, then currently held
opinion would be that the child should be considered
as suffering from abuse. If, on the other hand, the
carers are believed, then there must be a medical diag-
nosis and "temporary brittle bone disease" becomes
an acceptable label, even though there is no proof
as to its existence. It is suggested that non-accidental
injury and temporary brittle bone disease are indeed
the same condition but with different labels which are
dependent on the credibility of the carer's explanation.
If there are further attempts to define this condition
it is important that solid medical evidence is used. It
is unfortunate that cases which have been debated in
the courts by medical experts of differing opinions and
which the courts have found to be proven non-acci-
dental injury may be included as further cases of tem-
porary brittle bone disease (WALL 1995).
7.8.6
Other Differential Diagnoses
Other differential diagnoses are summarised in
Table 7.3 and discussed by Brill et al. (1998)
106 S. Chapman

Table 7.3. Differential diagnoses of the bony lesions in non-accidental injury

Condition Shaft SNBF Metaphyseal Osteopenia Comments

fractures abnormalities

Riekets + + + + Earliest signs are widening of growth plates

and cupping and fraying of the metaphyses.
Elevated alkaline phosphatase
Bone disease + + + + Most common in very low birth weight infants.
of prematurity More likely with a his tory of frusemide
administration, prolonged ventilation and/or
prolonged jaundiee
Myelomeningocoele + + + +1-
Congenital insensi- + + +
tivity to pain
Leukaemia + + +
Osteomyelitis + + Localised May be multifocal
Prostagiandin EI treatment +
Congenital syphilis + +
Scurvy + + ++ Not in first 6 months of life
Vitamin A intoxication +
Menkes' syndrome + + +1- Males with wormian bones, low serum copper
(kinky hair disease) and tortuous intracranial vessels
Copper deficiency + + + + Unlikely in the absence of low birth weight,
total parenteral nutrition, low-copper diet or
in a term infant in the first 6 months of life

References Caffey J (1957) Some traumatie lesions in growing bones other

Ablin DS, Sane SM (1997) Non-accidental injury: confusion
with temporary brittle bone disease and mild osteogenesis
imperfecta. Pediatr RadioI27:111-113
Ablin DS, Greenspan A, Reinhart M, Grix A (1990) Differen-
tiation of child abuse from osteogenesis imperfecta. AJR
154:1035-1046
Akbarnia B, Torg JS, Kirkpatriek Jet al (1974) Manifestations
of the battered-child syndrome. J Bone Joint Surg [Am]
56:1159-1166
Alexander CI (1976) Effect of growth rate on the strength of
the growth plate shaft junction. Skeletal Radioll:67-76
Anderson WA (1982) The significance of femoral fractures in
children. Ann Emerg Med 11: 174-177
Astley R (1979) Metaphyseal fractures in osteogenesis imper-
fecta. Br J Radio152:441-443
Augarten A, Laufer J, Szeinberg A et al (1993) Child abuse,
osteogenesis imperfecta and the grey zone between them.
J Med 24:171-175
Beals RK, Tufts E (1983) Fractured femur in infancy: the role
of child abuse. J Pediatr Orthop 3:583-586
Brill PW, Winchester P, Kleinman PK (1998) Differential diag-
nosis I. Diseases simulating abuse. In: Kleinman PK (ed)
Diagnostie imaging of child abuse, 2nd edn. Mosby, St
Louis, pp 178-196
Byers PH (1993) Osteogenesis imperfecta. In: Royce PM, Stein-
mann B (eds) Connective tissue and its heritable disorders.
Wiley/Liss,NewYork,pp 317-350
Caffey J (1946) Multiple fractures in the long bones of
infants suffering from chronic sub dural haematoma. AJR
56:163-173
than fractures and dislocations: clinieal and radiologieal
features. Br J RadioI30:225-238
Carty H (1988) Brittle or battered. Arch Dis Child 63:350-352
Carty H, Shaw D (1988) Osteogenesis imperfecta type IV A
(letter). Acta Paediatr Scand 77:752-754
Chapman S (1992) The radiologieal dating of injuries. Arch
Dis Child 67:1063-1065
Chapman S, Hall CM (1997) Non-accidental injury or brittle
bones. Pediatr RadioI27:106-110
Crilly R (1972) Longitudinal overgrowth of the chieken radius.
J Anat 112:11-18
Cumming WA (1979) Neonatal skeletal fractures. Birth trauma
or child abuse? J Can Assoc Radio130:30-33
Ebbin AJ, Gollub MH, Stein AM et al (1969) Battered child syn-
drome at the Los Angeles County General Hospital. Am J
Dis Child 118:660-667
Gahagan S, Rimsza ME (1991) Child abuse or osteogenesis
imperfecta: how can we tell? Pediatries 88:987-992
Galleno H, Oppenheim WL (1982) The battered child syn-
drome revisited. Clin Orthop 162:11-19
GraifM,Stahl-KentV,Ben-Ami T et al (1988) Sonographie detec-
tion of occult bone fractures. Pediatr RadioI18:383-385
Haase GM, Ortz VN, Stakianakis GN et al (1980) The value
of radionuclide bone scanning in the early recognition of
deliberate child abuse. J Trauma 20:873-875
Herndon WA (1983) Child abuse in a military population. J
Pediatr Orthop 3:73-76
Jaudes PK (1984) Comparison of radiography and radionu-
clide bone scanning in the detection of child abuse. Pediat-
ries 73:166-168
Kempe CH, Silverman FN, Steele BF et al (1962) The battered
child syndrome. JAMA 181:17-24
Non-accidental Injury
King J, DiefendorfD,Apthorp J et al (1988) Analysis of 429 frac-
tures in 189 battered children. J Pediatr Orthop 8:585-589
Kleinman PK (1998) Diagnostie imaging of child abuse, 2nd
edn. Mosby, St Louis, p 22
Kleinman PK, Kwon DS (1998) Differential diagnosis IV:
normal variants. In: Kleinman PK (ed) Diagnostie imaging
of child abuse, 2nd edn. Mosby, St Louis, pp 225-228
Kleinman PK, Marks SC Jr (1995) Relationship of the sub-
periosteal bone collar to the metaphyseallesions in abused
infants. J Bone Joint Surg [Am] 77:1471-1476
Kleinman PK, Marks SC Jr (1996) A regional approach to the
classie metaphyseallesion in abused infants: the proximal
tibia. AJR 166:421-426
Kleinman PK, Marks SC Jr (1998) A regional approach to
the classie metaphyseallesion in abused infants: the distal
femur. AJR 170:43-47
Kleinman PK, Marks SC, Blackbourne B (1986) The metaphy-
seallesion in abused infants: a radiologie-histopathologie
study. AJR 146:895-905
Kleinman PK, Belanger PL, Karellas A et al (1991) Normal
metaphyseal radiologie variants not to be confused with
findings of infant abuse. AJR 156:781-783
Kogutt MS, Swischuk LE, Fagan CJ (1974) Patterns of injury
and significance of uncommon fractures in the battered
childsyndrome.AJR 121:143-149
Kowal-Vern A, Paxton TP, Ros SP et al (1992) Fractures
in the under-3-year-old age cohort. Clin Pediatr (Phila)
31:653-659
Krugman RD (1997) Child protection poliey. In: Helfer ME,
Kempe RS, Krugman RD (eds) The battered child. Univer-
sity of Chieago Press, Chieago
Landin LA (1983) Fracture patterns in children. Analysis of
8,682 fractures with special reference to incidence, etiol-
ogy, and secular changes in a Swedish urban population
1950-1979. Acta Orthop Scand SuppI202:1-109
Lauer B, Broeck ET, Grossman M (1974) Battered child syn-
drome: review of l30 patients with controls. Pediatries
54:67-70
Leventhal JM, Thomas SA, Rosenfield NS, et al. (1993) Frac-
tures in young children. Distinguishing child abuse from
unintentional injuries. Am J Dis Child 147:87-92
Loder RT, Bookout C (1991) Fracture patterns in battered chil-
dren. J Orthop Trauma 5:428-433
Matthew MO, Ramamohan N, Bennet C (1998) Importance of
bruising associated with paediatrie fractures: prospective
observational study. Br Med J 317:1117-1118
Merten DF, Radkowski MA, Leonidas JC (1983) The abused
child. A radiologie al reappraisal. Radiology 146:377-381
O'Connor JF, Cohen J (1998) Dating fractures. In: Kleinman PK
(ed) Diagnostie imaging of child abuse, 2nd edn. Mosby, St
Louis, pp 168-177
Ogden JA (1990) Injury to the immature skeleton. In: Toulouki-
an RJ (ed) Pediatrie trauma. Mosby Year Book, St Louis,
pp 399-439
Ojima K, Matsumoto H, Hayase T et al (1994) An autopsy
case of osteogenesis imperfecta initially suspected as child
abuse. Forensie Sci Int 65:97-104
O'Neill JA Jr, Meacham WF, Griffin PP et al (1973) Patterns
of injury in the battered child syndrome. J Trauma
l3:332-339
Paters on CR, McAilion SJ (1987) Child abuse and osteogenesis
imperfecta. Br Med J 295:1561
107
Paterson CR, McAilion S, Miller R (1983) Osteogenesis imper-
fecta with dominant inheritance and normal sclerae. J Bone
Joint Surg [Br] 65:35-39
Paterson CR, McAllion S, Shaw J (1987) Clinieal and radiologi-
cal features of osteogenesis imperfecta type IV A. Acta Pae-
diatr Scand 76:548-552
Paterson CR, Burns J, McAllion SJ (1993) Osteogenesis imper-
fecta: the distinction from child abuse and the recognition
of a variant form. Am J Med Genet 45: 187 -192
Rex C, Kay PR (2000) Features of femoral fractures in nonac-
cidental injury. J Pediatr Orthop 20:411-4l3
Riehman HA (2000) From a radiologist's judgement to public
policy on child abuse and neglect: what have we wrought?
Pediatr RadioI30:219-228
Rosenberg N, Bottenfield G (1982) Fracture in infants: a sign
of child abuse.Ann Emerg Med 11:178-180
Rosenthal L, Hili RO, Chuang S (1976) Observations of the
use of 99ffiTc-phosphate imaging in peripheral bone trauma.
Radiology 119:637-641
Ryan MG, Davis AA, Oates RK (1977) One hundred and
eighty-seven cases of child abuse and neglect. Med JAust
2:623-628
Scher! SA, Miller L, Lively N et al (2000) Accidental and
nonaccidental femur fractures in children. Clin Orthop
376:96-105
Shopfner CE (1966) Periosteal bone growth in normal infants.
A preliminary report. AJR 97:154-163
Sillence DO, Senn A, Danks DM (1979) Genetic heterogeneity
in osteogenesis imperfecta. J Med Genet 16: 101-116
Smith FW, Gilday DL, Ash JM et al (1980) Unsuspected costo-
vertebral fractures demonstrated by bone scanning in the
child abuse syndrome. Pediatr Radioll 0: 103-1 06
Smith R (1995) Osteogenesis imperfecta, non-accidental injury
and temporary brittle bone disease. Arch Dis Child
72:169-176
Sty JR, Starshak RJ (1983) The role of bone scintigraphy in
the evaluation of the suspected abused child. Radiology
146:369-375
Taitz LS (1987) Child abuse and osteogenesis imperfecta. Br
Med J 295:1082-1083
Tardieu A (1860) Etude medieo-legale sur les seviee et mauvais
traitments exerces sur des enfants. Ann Hyg Publ Med Leg
l3:361-398
Thomas SA, Rosenfield NS, Leventhal JM et al (1991) Long-
bone fractures in young children: distinguishing accidental
injuries from child abuse. Pediatries 88:47l-476
Wall (1995) Re AB (Child abuse: expert witnesses). FLR
1:181-200
Wang C-T (1997) Current trends in child abuse reporting and
fatalities: the results of the Annual Fifty State Survey. Center
for Child Abuse Prevention Research, Prevent Child Abuse
Ameriea, Chieago
Wardinsky TD (1995) Genetic and congenital defect condi-
tions that mimie child abuse. J Farn Pract 41:377-383
Wardinsky TD, Vizcarrondo FE, Cruz BK (1995) The mistaken
diagnosis of child abuse: a three-year USAF medieal centre
analysis and literature review. Milit Med 160:15-20
Wellington P, Bennet GC (1987) Fractures of the femur in
childhood. Injury 18:103-104
Wor!ock P, Stower M, Barbor P (1986) Patterns of fractures
in accidental and non-accidental injury in children: a com-
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
8.2
Indications for Radiography in
Suspected Knee Injury
Acute knee trauma is a common injury. Determining
the need for radiographie examination of the knee fol-
lowing injury is largely directed by the clinical histo-
ry and findings on examination. The forces involved,
whether direct, due to impaction, or indirect, due to
excessive movements in flexion, extension, varus or
valgus, together with any rotational or twisting com-
ponent, need to be considered. On examination, the
presence of an effusion, point tenderness, limitation
or excess movement is important.
A number of studies have been carried out to try
and assess the relative importance of certain features
in predieting the presence of positive findings on the
conventional radiograph. The Ottawa Knee Rule was
devised in 1995 (STIELL et al. 1995). This was based
on the assessment of 1,047 adult knee injury patients
assessed for 23 standardised clinieal findings. It was
shown that a knee radiographie series was only
required in knee injury patients with at least one of
the following findings: (a) age 55 years or older, (b)
isolated tenderness of the patella, (c) tenderness at
the head of the fibula, (d) in ability to flex the knee to
90°, (e) inabilityto bear weight both immediately and
in the emergency department (four steps; unable to
transfer weight twiee onto each lower limb regardless
of limping). Subsequent studies later confirmed the
validity of this rule (STIELL et al. 1996, 1997; TIGGES
et al. 1999). A later study highlighted the significance
of the presence of an effusion or haemarthrosis in
predieting significant bone injury (FAGAN and DAVIS
2000). In this study, these features and bony tender-
ness were considered to be the most significant char-
acteristies and truly independent variables associat-
ed with fracture. An effusion in the knee joint is diag-
nosed when a soft tissue shadow, usually ovoid in
outline, is visible in the suprapatellar region (owing
to fluid within the suprapatellar pouch) as viewed on
a lateral projection. This is associated with oblitera-
tion of the suprapatellar fat line (BUTT et al. 1983). If a
lipohaemarthrosis (fat and blood within the joint) is
suspected, a horizontal beam lateral projection should
be obtained. A line separating the two densities of
blood on the bottom and the lower density fat on the
top of the "effusion" confirms this diagnosis and is
highly suspicious of an intra-artieular fracture.
When radiographs are completely normal in this
scenario, and additional projections, including inter-
nal and external oblique and stress views, have also
yielded normal findings, then soft tissue injury needs
P. N. M. Tyrell and V. N. Cassar-Pullicino
to be considered, possibly in association with micro-
trabecular or chondral injury not visible on the plain
radiographs. In this situation, MR imaging should be
undertaken.
8.3
Extra-articular Fractures
8.3.1
Supracondylar Fractures
Supracondylar fractures occur just above the level
of the condyles, are frequently slightly transverse
oblique, may be comminuted and not infrequently
extend into the joint. The distal fragment is frequent-
ly angulated posteriorly owing to the pull of the gas-
trocnemius muscle attached posteriorly just above
the femoral condyles. There is also a risk of associ-
ated vascular injury to the popliteal vessels behind
the knee joint. Fractures in this location can be asso-
ciated with fractures and fracture-dislocation of the
hip. There is also an association with fracture of the
tibial shaft when there is a high-impact injury, result-
ing in a floating knee.
Various classification systems have been used for
fractures ofthe distal femur (NEER et al. 1967; STEW-
ART et al. 1966; SCHATZKER and TILE 1987; SEIN-
SHEIMER 1980; MÜLLER et al. 1979). That of MÜLLER
et al. indieates three general patterns (supracondylar
without joint involvement, a two-part fracture with
intra -artieular extension, and an inverted Y type frac-
ture with intra-artieular extension) with three sub-
groups, based on radiographie appearance. There are
a very large number of fracture patterns, many of
whieh will not fit neatly into a classification scheme.
Hence, although classification systems are useful
when comparing and analysing different types of
treatment, in practical terms each case must be indi-
vidually assessed.
As regards management of supracondylar frac-
tures, in general, closed reduction can be employed
for fractures without joint involvement. Open reduc-
tion and internal fixation will be required for those
fractures with intra-artieular extension.
8.3.2
Distal Femoral Physeal Fractures
Distal femoral physeal (and proximal tibial physeal)
injuries are usually due to a valgus or hyperextension
Bone Trauma III

force at the knee. Since physeal injuries occur prior to

skeletal fusion, such injuries can be associated with
a high incidence of growth disturbance due to post-
traumatic tethering across the physis. Distal femoral
physeal fractures are much rarer than those occur-
ring at the ankle or affecting the upper limb. They
are associated with ligamentous injuries of the knee
joint. They are classified according to the Salter
Harris classification of physeal injuries (SALTER and
HARRIS 1963) (Fig. 8.1).
- Salter Harris (SH) type I is aseparation through
the distal femoral physis, without fracture of the
adjacent bone.
- SH type II is a widening of the growth plate with
oblique extension of the fracture line through
one corner of the metaphysis. This is the most
common type of fracture at this location. The frac-
ture is usually displaced laterally as the result of a
valgus force, or anteriorly as the result of a hyper-
extension force.
a
- SH type III is a widening of the physis together
with a vertical epiphyseal fracture line extending
from the physis down to the articular surface.
Although the fracture line can occur anywhere,
it is usually through the intercondylar region
(Fig.8.2a). The widening of the physis usually

11 111

IV v

Fig. 8.1. Line diagram of distal femoral physeal injuries (types

1- V). See text for details

Fig.8.2. a AP radiograph demonstrating an SH type III frac-

ture of the distal femur. b Coronal STIR image demonstrates
a large effusion and highlights the degree of displacement of
the medial fragment. c Axial T 1-weighted image highlights the c
displaced fragments and reveals the posterior extension into
the intercondylar notch
112
affects the medial condyle. This is the most
common undisplaced spontaneously reduced
epiphyse al injury at the knee. It may be difficult
to appreciate on initial examination. A tear of the
anterior cruciate ligament (ACL) can be associated
with this type of fracture.
- SH type IV consists of a vertical fracture line
extending down through the metaphysis, across
the physis and through the epiphysis to the articu-
lar surface. The fracture usually involves the lateral
metaphysis extending into the intercondylar notch
with separation of the lateral femoral condyle.
- SH type V is physeal impaction and has a high
association with consequent growth retardation.
These are often found in association with frac-
tures of the proximal tibia.
Avulsion injuries from the edge of the physis
including a portion of the perichondrium and
attached bone may occur when there is avulsion of a
collateralligament.
Separate from the Salter Harris classification of
physeal injury, injuries can be described relative to
the type of displacement, and this in turn is asso-
ciated with the mechanism of injury. Anterior dis-
placement of the epiphysis is due to a hyperextension
injury to the knee. This is associated with a risk of
neurovascular damage posteriorly. Medial or lateral
displacement also can occur in association with SH
type II injury.
In children, because the physis is radiolucent,
appreciation of injury is determined by observation
of widening, displacement or adjacent bony disrup-
tion (OZONOFF 1979). Standard anteroposterior (AP)
and lateral radiographs may overlook injury and
oblique views may be required. In an infant, only the
centre of the distal femoral epiphysis is ossified. This
should be in line with the long axis of the femur on
both AP and lateral projections. Comparative views
with the opposite knee in cases of doubt may be help-
ful. The radiolucent line representing the physis on
the AP projection usually measures 3-5 mm until
adolescence. Stress views mayaiso be necessary, and
can help differentiate between ligamentous and phy-
seal injury (ROGERS 1992).
In general terms, SH I and II injuries can be
reduced by closed me ans. This can fail, however, often
owing to a sleeve of periosteum interposed within the
physis, in which case open reduction becomes nec-
essary (SPONSELLER and BEATY 1996). SH III and
IV almost always require open reduction and inter-
nal fixation to minimise disruption of the articular
surface and to decrease the likelihood of premature
P. N. M. Tyrell and V. N. Cassar-Pullicino
growth arrest. SH II injuries have a high association
with subsequent growth disturbance. This retarda-
tion of growth usually occurs on the side opposite
to the metaphyseal fracture since this is the point
of maximal injury to and separation of the physis.
The radiographic signs of premature closure usually
become evident within 6 months after injury.
8.3.3
Fractures of the Proximal Fibula
Fractures of the head and neck of the fibula rarely
occur in isolation and are usually associated with
fracture of the lateral tibial condyle, ligamentous
injuries of the knee or fracture of the ankle.
Fractures of the head of the fibula are of three
types:
1. Direct impaction
2. Valgus injuries, usually associated with fracture of
the lateral tibial condyle and medial collateralliga-
ment injury
3. Varus injuries in which there is avulsion of the sty-
loid process of the fibula at the site of insertion of
the lateral collateralligament and biceps femoris
tendon. These can be associated with injury to the
common peroneal nerve. There mayaiso be signif-
icant damage to the lateral capsule and ligamen-
tous structures (Fig. 8.3).
When injuries of the head and neck of the fibula
occur a little more distally, they are often associated
with external rotational injury at the ankle.
8.3.4
Proximal Tibial Physeal Fractures
The Salter Harris classification applies to this area
(SALTER and HARRIS 1963). In the radiographic eval-
uation of this type of fracture it is paramount to
remember that even minimally displaced fractures
may have been severely displaced before recoiling to
a nearly reduced location as seen on the radiographs.
Most physeal injuries here are types I and II. Sixty-
six percent of SH type II injuries are displaced. The
metaphyseal fracture is usually medial and the sepa-
ration lateral, resulting in a valgus deformity. In SH
type III injury, the fracture usually involves the later-
al epiphysis and the medial collateralligament (MeL)
is torn. In SH type IV injuries, the medial or lateral
tibial condyle can be affected. This type of injury
is particularly associated with avulsion of the ACL
Bone Trauma 113

attachment. SH type V injuries occur particularly in

association with fractures of the tibial shaft. This
type of injury can also be associated with genu recur-
vatum as a result of a dashboard injury, with involve-
ment of the antero-inferior part of the epiphysis and
the growth plate.
üf all the physeal fractures in the long bones of
children, proximal tibial physeal fractures have the
greatest potential for disastrous neurovascular com-
plications. The proximal tibial fractures can also be
classified according to the direction of displacement.
A hyperextension injury leads to posterior displace-
ment and angulation of the metaphysis (Fig.8.4).
A hyperflexion injury leads to anterior angulation.
Varus and valgus types result from abduction or
adduction forces.
SH type land II injuries can generally be reduced
by closed means. SH type III and IV usually require
open reduction and internal fixation (SPONSELLER
and BEATY 1996). When closed reduction is under-
taken, appropriate care needs to be employed to avoid
a
inadvertent injury to the popliteal vessels posteriorly
and the peroneal nerve laterally.

Fig. 8.4. Line diagram of proximal tibial physeal fracture due

to hyperextension. There is risk of vascular damage

..
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
8.3.5
Fractures of the Patella
Patellar fractures are less common in children than
in adults. This may be related to the cushioning effect
of the surrounding cartilage, protecting it from direct
injury, and also due to the smaller musde mass acting
on the extensor mechanism, which has the conse-
quence that less force is applied in the severe contrac-
tion associated with indirect injury.
Patellar fractures occur from direct or indirect
forces. Direct force is the most common and can
occur in all age groups. Fractures of the patella can
be dassified into:
- Marginal (vertical)
- Stellate
- Transverse
- Avulsion
- Comminuted
Vertical fractures in adults are usually the result
of direct injury and involve the lateral facet. They
indude the entire thickness of the patella. A medial
marginal fracture in children may traverse the entire
thickness of bone or there may be a medial tangen-
tial osteochondral fracture. These longitudinally ori-
ented fractures are usually undisplaced and are best
seen on axial or oblique views.
Patellar fractures occurring due to indirect forces
result when the intrinsic strength of the patella
Fig.8.5. Lateral radiograph demonstrating a fracture of the
patella
P. N. M. Tyrell and V. N. Cassar-Pullicino
is exceeded by the pull of the musculotendinous
expansion attaching to it. This can occur with falls,
with the knee in flexion and severe contraction
of the quadriceps tendon. After the patella frac-
tures, continuing quadriceps musde action results
in tearing of the medial and lateral quadriceps
expansions. The typical fracture in this situation is
a transverse fracture. Transverse fractures are best
seen in a lateral projection (Fig. 8.5). The bone only
may be fractured, the articular cartilage remaining
intact with the gap maximal anteriorly and least
posteriorly. Even if a fracture is comminuted, the
fragments will be dose together if the surrounding
soft tissues of the quadriceps expansion are intact
(Figs. 8.6, 8.7).
Avulsion injury of the medial margin of the patel-
la is more common in children than in adults and
occurs in association with lateral dislocation. A sleeve
fracture of the patella occurs in young children.
This is an avulsion of a small bony fragment from
the distal pole of the patella together with a sleeve
of articular cartilage, periosteum and retinaculum
pulled off the remaining main body of the patella.
Conventional radiography can underestimate this
injury since if a bone fragment is visible, the carti-
lage component greatly exceeds its size. Ultrasound
has been used to confirm this injury, and also allows
assessment of fracture separation and displacement,
which is helpful in subsequent management (DITCH-
FIELD et al. 2000).
Fig.8.6. Lateral radiograph of the knee demonstrating a com-
minuted fracture of the patella. There is depression of the
articular surface but otherwise the fracture is non-displaced.
Note the effusion
 Bone Trauma 115

Fig.8.7. a Comminuted fracture of the patella b

in the AP and b lateral projections

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

Femoral condylar fractures may involve one or

both condyles. Isolated femoral condyle fractures are
uncommon. The fractures may be T- or Y-shaped, Fig.8.8. AP radiograph ofknee demonstrating abipartite patella
116
These fractures can often be visualised on AP
and lateral projections but oblique views (in varying
degrees of internal and external rotation) are some-
times required. These can demonstrate the degree of
separation of fracture fragments, which is important
in the planning of treatment. CT is also particularly
good in this area, since the relationship of the frac-
ture line to the intercondylar notch can be identified
and related to the cruciate ligament attachment sites.
In this type of fracture, because of anticipated exten-
sive soft tissue injury, MR imaging is often valuable
to confirm the ligaments involved and to facilitate the
management of the injury. Non-operative treatment
is confined to patients with incomplete or non-dis-
placed fractures. When the fracture is displaced, then
open reduction and internal fixation to achieve joint
congruity and stabilisation are indicated.
8.4.2
Fractures of the Tibial Plateau
Fractures of the tibial plateau occur due to a com-
bination of valgus and compression forces generated
by impaction of the femoral condyles against the
tibial plateau. The lateral tibial condyle and plateau
are weaker than the medial plateau, and fractures are
thus more likely to occur here. Also the femoral con-
dyles are stronger than the tibial plateau, and thus
with impaction injury the plateau is more vulnerable.
Varus stresses and adduction of the tibia occur much
less frequently than valgus stress; hence the greater
propensity for injury to involve the lateral condyle.
A number of classifications of fractures of the
tibial plateau have been used. That given here is the
one proposed by SCHATZKER et al. (1979) (Fig. 8.9).
11
IV V VI
P. N. M. Tyrell and V. N. Cassar-Pullicino
- Type I is a split fracture of the lateral tibial condyle.
This is usually a vertical fracture of the condyle
with no depression and tends to occur in young
adults.
- Type 11 is a split fracture of the lateral tibial con-
dyle with depression. It occurs as a result of a lat-
eral bending force with axialloading and there is
usuallyan associated tear of the MCL.
- Type III is a depressed fracture of the lateral tibial
plateau.
- Type IV is a fracture of the medial tibial plateau.
There is a high association with soft tissue injury
as a result of high impact, related to varus and
axialloading.
- Type V is a bicondylar fracture with varying
degrees of articular depression and displacement
(these fractures may be T- or Y-shaped).
- Type VI is a bicondylar fracture with metaphyseal!
diaphyseal dissociation, and usually results from a
high impact injury.
A vertical fracture through the posterior margin
of the condylar surface mayaiso occur.
In the conventional radiographic evaluation of the
normal tibia, the condylar region slopes downwards
by approximately 15° posteriorly, and therefore in
the standard AP projection the tibial plateau is not
seen tangentially. Condylar/plateau fractures, par-
ticularly depressed fractures involving the anterior
plateau margin, can be overlooked on account of
this. Fractures of the lateral tibial plateau are most
common in the elderly. They may occur from
pedestrian/automobile accidents, but are more usu-
ally due to twisting falls. In undisplaced fractures,
the fracture line may be obliquely oriented and not
well seen on standard AP and lateral projections. An
111
Fig.8.9. Line diagram of tibial plateau fractures
(types I-VI). See text for details
Bone Trauma
AP projection with 15 0 caudad angulation produces
a tangential view of the tibial plateau and facilitates
observation of depressed fractures. The depression
may be gentle or marked, with bone fragments being
driven deep into the condyle (often associated with a
vertical split fracture) (Fig. 8.lOa, b). Minimally dis-
placed fractures of the lateral tibial condyle are often
seen only in the oblique view.
Conventional tomography was previously used in
the assessment of tibial condylar and plateau fractures,
but has been superseded by CT. CT can be carried out
through pIaster casts, sagittal and coronal (Fig.8.lDc)
reconstructions can be obtained, and accurate mea-
surements of the degree of depression of the plateau
and the degree of separation of fragments can be
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-
tion, in improving the accuracy of the classification
of the fracture (WICKY et al. 2000). This assists the
surgeon and results in improved management of the
injury. The management of tibial plateau fractures is
controversial and beyond the scope of this review. Ulti-
mately, the aim of treatment, as with fractures on the
femoral side of the joint, is to achieve, as far as is pos-
sible, a stable knee which is mobile, painless and satis-
factorily aligned, with a minimal risk of post-traumat-
ic osteoarthritis (SPONSELLER and BEATY 1996).
When there has been marked soft tissue injury to
the knee, such as with an ACL tear, there is often asso-
ciated bone bruising/microtrabecular injury which is
only detectable on the subsequent MR imaging study
(Fig. 8.11). Such injury not infrequently occurs in the
117
posterolateral aspect of the tibial condylar region.
Some cruciate ligament injuries can be associated with
pure hyperextension and this mechanism of injury is
associated with bruising ofboth the medial and the lat-
eral condyle. Ligamentous injury occurs in 10%-12%
of plateau fractures and is more likely due to the valgus
mechanism of injury rather than axial loading. The
ACL and MCL are most likely to be injured. The MCL
injury most often occurs in association with a split or
split-depression type of fracture of the lateral tibial
plateau. Where an injury of the MCL occurs in associa-
tion with valgus stress, there may be an avulsion of the
MCL as opposed to a tear. This avulsion can be associ-
ated with a flake fracture from the medial aspect of the
medial femoral condyle.
8.4.3
Fractures of the Tibial Spine and
Intercondylar Eminence
Fractures of the tibial spine and intercondylar emi-
nence are more common in children than in adults.
The injury is most commonly seen in conjunction
with high-energy plateau fractures and is frequently
avulsion in nature, typically occurring at the site
of origin of the ACL. In older children and adults,
the ACL is usually tom within its substance without
an associated avulsion fracture. The mechanism of
injury is thought to be hyperextension combined
with strong rotational forces. There is an incidence of
associated ligamentous injuries and this is greater in
adults than children.
Fig. 8.10. a Depressed fracture of the lateral tibial
plateau with inverted extension distally. Also note
the fracture of the head and neck of the fibula.
Note that the bones are osteoporotic. b Horizon-
tal beam lateral projection of the same patient as
in a. Note lipohaemarthrosis. c Coronal CT recon-
struction of a fracture through the lateral tibial
plateau (different patient to a and b)
 118
a b
The classification of these fractures as proposed
by MEYERS and McKEEVER (1959) is related to the
degree of displacement of the eminence (Fig. 8.12).
- Type I: A horizontal fracture line at the base of the
anterior portion of the tibial spine. Elevation of
only the anterior edge of the fracture.
- Type II: The eminence is elevated, angulated and
hinged posteriorly.
- Type IIIA: The entire eminen ce has been avulsed
from the parent bone.
- Type IIIB: The entire eminence has been avulsed,
and is elevated and rotated.
Fractures may be viewed on standard AP or lateral
views (Figs.8.13, 8.14), but a notchltunnel projec-
tion gives a better view of the intercondylar region
and oblique views may be helpful. In children, the
avulsed fragment may be mainly unossified cartilage
with only a very small thin ossified fragment. The
margins of the bones should be scrutinised for fine
avulsion fractures and oblique views may be helpful
in addition. MR imaging will allow evaluation of
associated soft tissue injuries.
11 lilA
P. N. M. Tyrell and V. N. Cassar-Pullicino
Fig.8.11. a Sagittal STIR image reveals
extensive soft tissue trauma, with rup-
ture of the ACL and posterior capsule. b
Sagittal STIR image shows bone bruis-
ing of the lateral femoral condyle, typi-
cally associated with ACL rupture
Avulsion of the posterior tibial eminen ce can
occur, at the tibial insertion site of the posterior cru-
ciate ligament (PCL) (TORISU 1977; Ross and CHES-
TERMAN 1986). This is, however, rare.
For non-or minimally displaced fractures (types I
and II), management is usually non-operative. Open
reduction and internal fixation is required for type
III displaced fractures.
8.4.4
Avulsion of the Tibial Tuberosity
Developmentally the tibial tubercle passes through
four stages (SPONSELLER and BEATY 1996):
1. The cartilaginous stage, in which no ossification
centre in the cartilaginous anlage of the tibial
tubercle is present.
2. The apophyseal stage, in which the ossification
centre for the tubercle appears (occurs at 8-12
years in females and 9-14 years in males).
3. The epiphyseal stage, in which the secondary ossi-
fication centres of the tubercle and the proximal
tibial epiphysis coalesce to form a tongue of bone
continuous between the tubercle and proximal
tibial epiphysis.
4. The final bony stage, in which the epiphysealline
is closed between the fully ossified tuberosity and
the tibial metaphysis.
The ligamentum patellae (patellar tendon) is
attached to the tubercle. Avulsion of the tubercle
can occur when the patellar tendon traction exceeds
the combined strength of the physis underlying
Bone Trauma
the tubercle, the surrounding perichondrium and
the adjacent periosteum. Avulsion can be associated
with sudden acceleration or deceleration of the knee
extensor mechanism.
WATSON-JONES (l955) described three types of
avulsion fracture of the tibial tuberosity. OGDEN and
co-workers (l980) later refined this, describing three
types depending on the distance of the fracture from
the distal tip of the tubercle (Fig. 8.15).
- Type I: The separation through the distal portion
of the physis under the tubercle breaks up proxi-
maHy through the secondary ossification centre of
the tubercle.
- Type II: The separation extends anteriorly through
the area bridging the ossification centres of the
tibial tubercle and the proximal tibial epiphysis.
- Type III: A fracture/separation of the entire tuber-
119
Fig.8.13. a Lateral radiograph demon-
strating an avulsion fracture of the tibial
intercondylar eminence (arrow). Note
the effusion. b AP radiograph showing
the avulsion fracture. cSagittal MR T2-
weighted (gradient echo) image dem-
onstrating the fracture Hne through the
tibial intercondylar eminence (anterior
to the ACL attachment site) and extend-
ing posteriorly through the epiphysis
into the physis. d Sagittal MR Tl-
weighted image through the same level
as in c. e Coronal Tl-weighted image
(same patient). Note the fracture Hne
ele extends into the proximal tibial epiphysis and
the knee joint.
These fractures can be further classified according
to the presence or absence of displacement.
On a standard AP radiograph, the tubercle lies just
lateral to the midline of the tibia and therefore the
best profile is obtained with the tibia slightly inter-
nally rotated (Fig. 8.16).
- Type I fractures are usually treated with closed
reduction. Open reduction and internal fixation is
required for type II and III injuries.
Acute avulsion of the tibial tubercle differs from the
Osgood-Schlatter lesion (OSGOOD 1903; SCHLAT-
TER 1903), which is a more chronic lesion but is feit
to represent one end of the spectrum of avulsive
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

Fig.8.16. a Lateral radiograph of the

knee in an adolescent demonstrating
a fracture of the tibial apophysis with
extension into the epiphysis. b Lateral
internal oblique projection better visu-
alises the fractured apophysis extend-
ing into the epiphysis, with elevation
of the apophysis, posterior hinging and
depression of the epiphysis
Bone Trauma
8.4.5
Osgood-Schlatter Lesion
The Osgood-Schlatter lesion occurs when the tibial
tubercle is in the apophyseal stage and when the sec-
ondary ossification centre has appeared. The carti-
lage overlying the ossification centre anteriorly and
posteriorly can resist tension forces better than the
bone of the ossification centre. Usually no single
acute traumatic event is responsible for the devel-
opment of the Osgood-Schlatter lesion; rather repeat-
ed normal stresses or overuse (chronic tug) produce
minor disruption of the bony apophysis (OGDEN and
SOUTHWICK 1976). On radiographs, best taken with
a soft tissue technique, the edges of the distal patel-
lar ligament are blurred and there may be tiny ßake
ossific fragments either anterior or superior to the
main ossification centre; these fragments, however,
can be difficult to differentiate from normal mul-
ticentric ossification of the tubercle. Later in the
course, the tiny displaced fragments may enlarge and
may fail to unite and remain as separate ossicles
detached from the mature tubercle. The smooth scle-
rotic margin of these ossicles indicates that they
are long-standing in nature, and not acute. On CT,
increased width and decreased attenuation of the
tendon is observed, and on MR imaging there is
enlargement and increased signal intensity of the
tendon. Deep and superficial infrapatellar bursae are
often noted. Bone abnormalities are seen less fre-
quently. On MR imaging an ossicle may be seen, and
on occasion a matching defect can be seen in the
tibial tubercle, suggesting that a bone fragment has
detached from the tuberosity. Sometimes, this ossicle
may fuse. On MR imaging, decreased signal intensity
has been reported in the marrow of the tibial tuberos-
ity and tibial epiphysis on Tl-weighted images, with
increased signal intensity on T2-weighted images,
implying the presence of marrow oedema (ROSEN-
BERG et al. 1992). These appearances contrast with
those of acute fracture-separation of the tubercle
physis, caused by violent contraction of the quadri-
ceps mechanism, which tends to occur during the
epiphyse al stage of development of the tubercle, when
the tubercle ossification centre has coalesced with
that of the proximal tibial epiphysis .. The manage-
ment of Osgood-Schlatter lesion is symptomatic and
supportive.
121
8.4.6
Sinding-Larsen-Johansson Disease
SINDING-LARSEN (1921) and ]OHANSSON (1922)
independently described a condition seen most
commonly in adolescents that consists in pain, ten-
derness and soft tissue swelling over the lower pole
of the patella associated with bony fragmentation.
The lesion is likely due to a traction phenomenon
in which repeated minor trauma at the proximal
attachment of the patellar tendon, initially com-
mencing as inßammation, is followed by calcifica-
tion or ossification (MEDLAR and LYME 1978), or
in wh ich patellar fracture or avulsion produces one
or more distinct ossification sites. Similar findings
are seen in athletes with «jumper's knee" (HECKMAN
and ALKIRE 1984).
The radiograph demonstrates small bony frag-
ments adjacent to the distal surface of the patella with
overlying soft tissue swelling. The bone fragments
may eventually coalesce and become incorporated
into the patella. Ultrasound (DE FLAVIIS et al. 1989)
and MR imaging may help to confirm the diagnosis.
The pathogenesis of this dis order is similar to that
of Osgood-Schlatter disease, and the two conditions
have been described in association (TRAVERSO et al.
1990).
8.4.7
Marginal Avulsion Fractures
Avulsion fractures occur at characteristic sites relat-
ed to ligamentous attachments and, because of these
attachments, they usually do not migrate. There is
usually an associated cortical irregularity of the
parent bone although this may be very subtle, and not
always detectable on plain radiographs.
The most common avulsion is from the anterior
tibial eminence, at the site of attachment of the ACL
(LILEY and BAxTER 1990) (Fig.8.17). Very rarely,
there may be an avulsion from the posterior aspect
of the intercondylar eminen ce at the site of attach-
ment of the PCL (TORISU 1977; Ross and CHESTER-
MAN 1986). A cortical fragment may be avulsed from
the femoral attachment site of the MCL. If there is
injury to the MCL at this site, without cortical avul-
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
a b
Fig.8.18. AP radiograph demonstrating a Pellegrini-Stieda
lesion (arrow) (calcification at the femoral attachment site of
the medial collateralligament)
The Segond fracture is an avulsion fracture of the
lateral tibia, posterior and slightly proximal to the
insertion of the iliotibial tract on Gerdy's tubercle
located on the anterolateral aspect of the tibial con-
dyle (ROGERS 1992) (Fig. 8.19). It is due to tension on
the lateral capsular ligament, which can occur as a
result of internal rotation of the tibia with the knee in
flexion (DIETZ et al. 1986; IRVINE et al. 1987). On an
AP radio graph, the bone fragment is visible adjacent
to adefeet on the lateral tibial condyle. It can be dif-
P. N. M. Tyrell and V. N. Cassar-Pullicino
Fig. 8.17. a AP radiograph demonstrat-
ing Hake fragment just above tibial emi-
nence (arrow). b Lateral radiograph
again demonstrating Hake fragment
anterior to the femoral intercondylar
region (white arrow) -of unclear origin.
Note deep depression (impaction
injury) on the anterior aspect of lateral
femoral condyle (black arrow), a feature
recognised to be associated with ACL
tear
ferentiated from an avulsion injury at the Gerdy's
tubercle since in the latter case there is no osseous
defect on the adjacent tibia. The importance of the
Segond fracture is that 75%-100% of such fractures
are associated with rupture of the ACL. Medial liga-
mentous damage mayaiso be present.
8.4.8
Osteochondral Fractures
Osteochondral fractures involve the medial or lat-
eral femoral condyle or the patella. These fractures
are usually identified on the basis of a small bone
fragment located intra-artieularly. They can, howev-
er, sometimes be very difficult to appreciate since
even a large fragment may eontain only a very small
ossified eomponent that is difficult to see on plain
radiographs. The fragment itself is more likely to be
identified than its actual point of origin.
Two mechanisms of injury leading to this type of
fracture were described by KENNEDY et al. (1979): (a)
exogenous, due to a direct blow with a shearing force
to either the medial or the lateral condyle, and (b)
endogenous, a flexion rotation injury in which contact
between the tibia and femoral condyle causes the frac-
ture. The endogenous meehanism also accounts for
the osteochondral injury which results from patellar
dislocation when the patella contacts with the lateral
femoral condyle (RORABECK and BOBECHKO 1976).
Because detection of these fragments may be dif-
ficult, the AP and lateral radiographs may need to
be supplemented with a notch view. The radiographs
Bone Trauma 123

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
should be scrutinised; the fragment may be located
anywhere, including high in the suprapatellar pouch.
Arthrography with or without CT, CT alone or MR
imaging may be required in further evaluation. See
Chaps. 2 and 3 for more details.
8.5
Dislocation
8.5.1
Knee
Dislocation of the knee at the femorotibial joint is a
rare injury. It is associated with a considerable force,
such as may occur with a severe automobile injury.
There is a high association with injury to the popli-
teal artery (CONWELL and ALDREDGE 1937; MONT-
GOMERY 1987).
The dislocation is described according to the direc-
tion of displacement of the tibia relative to the femur.
Anterior dislocation is the most common type and
occurs as a result of acute hyperextension. It is asso-
ciated with disruption of the posterior capsule and
PCL. It also has the highest association with injury
to the popliteal artery. Posterior dislocation is less
common but can also be associated with arterial
injury. It usually occurs due to a crushing blow to the
leg. Rotatory or posterolateral dislocation is associ-
ated with forced abduction and internal rotation. It
is associated with disruption of the medial capsular
mechanism. The medial femoral condyle is displaced
through a rent in the medial capsule and because of
this, the knee cannot be reduced by closed means. On
the lateral radiograph, while the femoral condyles are
in profile, the tibia is rotated posterolaterally and the
proximal tibiofibular joint is seen in its entirety. On
the AP radio graph, the lateral tibial condyle lies later-
al to the lateral femoral condyle, but never more than
one-quarter the width of the condyle. The medial
joint space is also widened. The incidence of popli-
teal artery injury in this type of injury is usually low
relative to anterior and posterior dislocations. Frac-
tures of the knee in association with dislocation are
uncommon. Occasionally there may be an avulsion of
the anterior tibial eminence related to an ACL avul-
sion.
P. N. M. Tyrell and V. N. Cassar-Pullicino
8.5.2
Patellar Dislocation
It is rare for patients to present with acute dislocation
of the patella. Rather, most give a history suggestive
of dislocation of the patella that has either relocated
spontaneously or has been reduced by the patient. On
occasion there may be radiographic evidence of an
osteochondral fracture (Fig. 8.20). A small flake frag-
ment may be visible, or, on an axial view, an actual
defect in the patella may be visualised. Not infre-
quently, however, the conventional radio graph may
demonstrate an effusion only. MR imaging is perhaps
the best modality for confirming this diagnosis by
revealing a number of features characteristic of the
injury (VIROLAINEN et al. 1993; KIRSCH et al. 1993;
SPRITZER et al.1997) (Fig. 8.2l). These include osteo-
chondral injury to the medial aspect of the patella
and to the lateral femoral condyle. The cartilage of
the lateral femoral condyle may well remain intact
but there is often high signal within the lateral aspect
of the lateral femoral condyle consistent with micro-
trabecular injury or bone bruising. The medial patel-
lar retinaculum is also traumatised, with associated
soft tissue swelling. There may be buckling of the
lateral patellar retinaculum. Arecent study has also
revealed extensive signal change in Hoffa's fat pad
(ApOSTOLAKI et al. 1999).
8.5.3
Proximal Tibiofibular Joint
Dislocation of the proximal tibiofibular joint is usu-
ally classified according to the direction of displace-
ment of the fibular head. Dislocation may occur
anterolaterally, posteromedially or superiorly.
Anterolateral dislocation is the most common and
usually occurs as a result of a twisting fall, the foot
being fixed in inversion while the body rotates exter-
nally. On an AP radiograph, the head of the fibula is
seen almost in its entirety while on the lateral view
it is completely overlaid by the tibial condyle. In pos-
teromedial dislocation, on an AP view, the head is
completely overlaid by the lateral tibial condyle, but
in the lateral view the fibular head is displaced poste-
riorly and visualised almost in its entirety. In superior
dislocation, which occurs when the tibia is foreshort-
 Bone Trauma 125

Fig.8.20. a AP radiograph demonstrating a bone fragment in the intercondylar

region (arrow). The source of the fragment is not visible. b Lateral projection again
demonstrates a bone fragment (arrow), but the origin is still unclear. c Intercondylar
notch view better demonstrates the free bone fragment. d Axial MR STIR image
demonstrating typical features of patellar dislocation (see text) and probable osteo-
chondral injury, accounting for the bone fragment seen on plain radiographs

Fig. 8.21. a Axial MR STIR image demon-

strates extensive bone bruising/marrow
oedema along the lateral aspect of the
lateral femoral condyle, and oedema of
the medial patellar retinacular tissues.
Appearances consistent with previous
patellar dislocation. b Axial MR STIR
image (same case as in a) demonstrat-
ing irregularity of the patellar articular
cartilage, a large effusion and oedema of
the medial patellar retinaculum
126
ened as a result of a shaft fracture with overlap, the
fibular head lies at the level of the knee joint. Supe-
rior dislocations can usually be readily appreciated
on plain radiographs.
Anterolateral dislocations can be managed with
elose reduction. The posteromedial variant usually
requires surgery since it is usually associated with
very significant injury, ineluding disruption of the lat-
eral joint capsule and the lateral collateralligament.
8.6
Fatigue Fractures
8.6.1
Stress Fractures
Stress fractures most commonly occur in young fit
adolescents or young adults as a result of repetitive
stress to metabolically normal bone (DAFFNER and
PAVLOV 1992). The most common site about the knee
is the proximal shaft of the tibia. Other sites may
involve the neck and proximal shaft of the fibula.
Stress fractures mayaiso involve the physis of either
the distal femur or the proximal tibia. These may be
manifest by widening of a portion of the growth plate
with selerosis at both epiphyseal and metaphyseal
margins.
8.6.2
Insufficiency Fractures
Insufficiency fractures can occur as a result of normal
stresses in metabolically abnormal bone such as
osteoporosis. They are therefore most common in
elderly females. The tibial plateau, most commonly
the medial plateau, is usually affected, although frac-
tures may occur a little more distally and involve
the medial cortex (MANCO et al. 1983). This type
of fracture should be strongly suspected in an elder-
ly female with osteoporosis who presents with knee
pain. Radiographs may be normal initially, but radi-
ography performed 3-4 weeks later may demonstrate
a characteristic linear density in the tibial plateau. If
bone scintigraphy is performed early on, when radio-
graphs are negative, intense uptake at the site of frac-
ture may be demonstrated.
A similar type of insufficiency fracture mayaiso
occur in the distal femoral metaphysis.
P. N. M. Tyrell and V. N. Cassar-Pullicino
8.7
Conclusion
Bone injury to the knee can often be diagnosed
purelyon the basis of plain radiographs, sometimes
employing additional views to those which are rou-
tinely taken. Classification of this type of injury is
important since it can direct management of the
injury. Injury to the fused and unfused skeleton dif-
fers since in the unfused skeleton, consideration will
elearly need to be given to potential disturbance of
the growth plate and to subsequent management of
any resultant deformity. CT and MRI both have a
role in the complete evaluation of complex injuries
involving significant soft tissue structures and also in
the evaluation of suspected occult injury.
References
Apostolaki E, Cassar-Pullicino VN, Tyrrell PNM et al (1999)
MR imaging in the diagnosis and management of occult
traumatic patellar dislocation. Clin Radio154:743-747
Butt WP, Lederman H, Chuang S (1983) Radiology of the supra-
patellar region. Clin Radio134:511-522
Conwell HE, Alldredge RH (1937) Complete dislocations of the
knee joint. Areport of 7 cases with end results. Surg Gyne-
col Obstet 64:94
Daffner R, Pavlov H (1992) Stress fractures: current concepts.
AJR 159:245-252
De Falviis L, Nessi R, Scaglione P et al (1989) Ultrasonic diag-
nosis of Osgood-Schlatter and Sinding-Larsen-Johansson
diseases ofthe knee. Skel Radio118:193-197
Dietz GW, Wilcox DM, Montgomery JB (1986) Segond tibial
condyle fracture: lateral capsular ligament avulsion. Radi-
ology 159:467-469
Ditchfield A, Sampson MA, Taylor GR (2000) Ultrasound
diagnosis of sleeve fracture of the patella. Clin Radiol
55:721-722
Fagan DJT, Davis S (2000) The clinical indications for plain
radiography in acute knee trauma injury. Int J Care Injur
31:723-727
Heckman JD, Alkire CC (1984) Distal patellar pole fractures. A
proposed common mechanism of injury. Am J Sports Med
12:424-428
Irvine GB, Dias JJ, Finlay DBL (1987) Segond fractures of
the lateral tibial condyle: brief report. J Bone Joint Surg
69B:613-614
Johansson S (1922) En forut icke beskriven sjukdom i patella.
Hygiea 84:161
Kennedy JC (1979) The injured adolescent knee. Williams and
Wilkins, Baltimore
Kirsch MD, Fitzgerald SW, Friedman H et al (1993) Transient
lateral patellar dislocation: diagnosis with MR imaging.
AJR 161:109-1l3
Liley JJ, Baxter MP (1990) Tibial spine fractures in children.
Clin Orthop 225:54-60
Bone Trauma
Manco LG, Schneider R, Pavloc H (1983) Insufficiency frac-
tures ofthe tibial plateau.AJR 140:1211-1215
Medlar RC, Lyme ED (1978) Sinding-Larsen-Johansson dis-
ease. Its etiology and natural history. JBJS 60A:I113-1116
Meyers MH, McKeever FM (1959) Fracture of the intercondylar
eminence of the tibia. J Bone Joint Surg 41A:209-222
Montgomery JB (1987) Dislocation of the knee. Orthop Clin
North Am 18:149-156
Müller ME, Allgower M, Schneider R, Willenegger H (1979)
Manual of internal fixation. Springer, Berlin Heidelberg
NewYork
Neer CS, Grantham SA, Shelton ML (1967) Supracondylar frac-
ture ofthe adult femur. J Bone Joint Surg 49A:591-613
Ogden JA,Southwick WO (1976) Osgood-Schlatter's disease and
tibial tuberosity development. Clin Orthop 116:180-189
Ogden JA, Tross RB, Murphy MJ (1980) Fractures of
the tibial tuberosity in adolescents. J Bone Joint Surg
62A:205-215
Osgood RB (1903) Lesions of the tibial tubercle occurring
during adolescence. Boston Med Surg J 148: 114-117
Ozonoff MB (1979) Pediatric orthopaedic radiology. Saunders,
Philadelphia
Rogers LF (1992) The knee and shafts of the tibia and fibula.
In: The radiology of skeletal trauma. Churchill Livingstone,
NewYork
Rorabeck CH, Bobechko WP (1976) Acute dislocation of the
patella with osteochondral fracture: a review of 18 cases. J
Bone Joint Surg 58B:237-240
Rosenberg ZS, Kauelblum M, Cheung YY et al (1992) Osgood-
Schlatter lesion: fracture or tendinitis? Scintigraphic, CT
and MR imaging features. Radiology 185:853-858
Ross AC, Chesterman PJ (1986) Isolated avulsion of the tibial
attachment of the posterior cruciate ligament in childhood.
J Bone Joint Surg 68B:747
Salter RB, Harris WR (1963) Injuries involving the epiphyseal
plate. J Bone Joint Surg 45A:587-622
Schatzker J, Tile M (1987) The rationale of operative fracture
care. Springer, Berlin Heidelberg New York
Schatzker J, McBroom R, Bruce D (1979) The tibial plateau
fracture: the Toronto experience. 1968-1975. Clin Orthop
138:94-104
127
Schlatter C (1903) Verletzurgen des Schnabelformigen Fortsatz-
es der Oberen Tibiaepiphyse. Beitr Klin Chir 38:874-887
Seinsheimer F III (1980) Fractures of the distal femur. Clin
Orthop 153:169-179
Sinding-Larsen MF (1921) A hitherto unknown affection of
the patella in children. Acta Radioli: 171
Spritzer CE, Courneya DL, Burk DL Jr et al (1997) Medial reti-
nacular complex injury in acute patellar dislocation: MR
findings and surgical implications. AJR 168: 117-122
Sponseller PD, Beaty JH (1996) Fractures and dislocations about
the knee. In: Rockwood CA Jr, Wilkins KE, Beaty JH (eds)
Fractures in children,4th edn. Lippincott-Raven, Philadelphia
Stewart MJ,Sisk TD, Wallace SL Jr (1966) Fractures ofthe distal
third of the femur. J Bone Joint Surg 48A:784-807
Stiell IG, Greenberg GH, Wells GA et al (1995) Derivation of
adecision rule for the use of radiography in acute knee
injuries. Ann Emerg Med 26:405-413
Stiell IG, Greenberg GH, Wells GA et al (1996) Prospective vali-
dation of adecision rule for the use of radiography in acute
knee injuries. JAMA 275:611-615
Stiell IG, Wells GA, Hoag RH et al (1997) Implementation of
the Ottawa knee rule for the use of radiography in acute
knee injuries. JAMA 278:2075-2079
Tigges S,Pitts S,Mukundan S Jr et al (1999) External validation
of the Ottawa knee rules in an urban trauma center in the
United States. AJR 172: 1069-1 071
Torisu T (1977) Isolated avulsion fracture of the tibial attach-
ment of the posterior cruciate ligament. J Bone Joint Surg
59 A:68-72
Traverso A, Baldari A, Catalari F (1990) The coexistence of
Osgood-Schlatter's disease with Sinding-Larsen-Johans-
sons disease. Case report in an adolescent soccer player. J
Sports Med Phys Fitness 30:331-333
Virolainen H, Visuri T, Kuusela T (1993) Acute dislocation of
the patella: MR findings. Radiology 189:243-246
Watson-Jones R (1955) Fractures and joint injuries, 4th edn.
Livingstone, Edinburgh
Wicky S, Blaser PF, Blanc CH et al (2000) Comparison between
standard radiography and spiral CT with 3D reconstruc-
tion in the evaluation, classification and management of
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
fast spin echo techniques (FSE). It is convenient to
obtain a proton density-weighted and T2-weighted
set of images in one plane (e.g., 2,000120-80 TR/TE),
frequently the sagittal. Proton density images are
probably the most sensitive for detection of menis-
cal tears. The axial plane can be imaged using Tl-
weighted pulse sequences (e.g., 500/20 TR/TE) or
alternativelyT2* -weighted gradient echo techniques,
which allow good visualization of the articular car-
tilage of the patellofemoral joint (e.g., 400/9120,TR/
TE/flip angle). The same T2* sequence can be used
in the coronal plane. Coronal proton density fat-sup-
pressed high-resolution images provide high sensi-
tivity for bone marrow lesions in addition to menis-
cal tears.
The value of FSE pulse sequences versus conven-
tional spin echo (SE) techniques has been discussed
extensively in the radiological literature (RUBIN et
al. 1994; ESCOBEDO et al. 1996; WHITE et al. 1996;
CHEUNG et al. 1997; EUSTACE et al. 1999). RUBIN et
al. (1994) evaluated 129 menisci in 66 consecutive
patients with suspected meniscal tears using conven-
tional and FSE sequences. They concluded that FSE
showed only 65% of meniscal tears shown on conven-
tional SE images and therefore did not recommend
FSE for evaluation of meniscal tears. These authors
attributed the blurring artifact of FSE to the low sen-
sitivity of this technique. Blurring artifact in FSE
sequences is related to the use of short effective TE,
and it is related to the attenuation of the T2 decay
that occurs at the edges of the K-space when the later
echo es of an echo train chain are used to collect the
high-spatial-frequency data (MELK I et al. 1991). Tis-
sues with short T2 and pulse sequences using short
effective TE, large ETL and small acquisition matrices
will contribute to the blurring effect of FSE sequenc-
es. Decreased signal intensity of the menisci sec-
ondary to increased magnetization transfer that
occurs with FSE sequences mayaIso contribute to the
decreased accuracy of FSE sequences for the detec-
tion of meniscal tears (LISTERUD et al. 1992). Similar
results were published by WHITE et al. (1996). These
authors compared the accuracy of T2-weighted FSE
and proton density SE imaging in the detection of
meniscal tears in 152 menisci and found FSE to be
slightly less accurate that proton density SE.
On the other hand, ESCOBEDO et al. (1996) evaluat-
ed 80 menisci with conventional SE and FSE sequenc-
es and obtained the same sensitivity and similar
specificity for both sequences for the detection of
meniscal tears. They concluded that proton density
FSE imaging with a short ETL is a reasonable sub-
stitute for conventional SE imaging of the knee.
J. Beltran and S. Shankman
CHEUNG et al. (1997) evaluated double FSE sagittal
images for the detection of meniscal tears in a group
of 293 patients, and obtained a sensitivity and spec-
ificity similar to those of conventional SE. In their
imaging protocol they included TR of 4,000 ms, an
ETL of 8 and an effective TE "as short as possible", in
the range of 18 ms. They also concluded that FSE is
an alternative to conventional SE for knee imaging.
More recently, EUSTACE et al. (1999) compared
Tl-weighted conventional and turbo SE (TSE) or
hybrid rapid relaxation enhancement (RARE) pulse
sequences in patients with acute knee trauma (60
menisci). Their rationale was to take advantage of
the fast imaging time that can be obtained with TSE
(1 min 20 s), as compared with SE (4 min 10 s), in
an acute trauma situation. Their results indicate that
both techniques are similarly accurate.
Three-dimensional Fourier transform (3DFT)
volume acquisition techniques are popular owing to
software improvements and availability of worksta-
tions. These techniques are performed using gradi-
ent echo pulse sequences with short TR/TE (e.g.,
18/9/30, TR/TElflip angle). The advantages of 3DFT
data acquisition include improved signal-to-noise
ratio and reconstruction of true contiguous slices of
about 1 mm in infinite planes of section, with mini-
mal distortion. When used in the axial plane, 3DFT
gradient echo sequences are capable of displaying the
menisci in more than one consecutive image. This
is helpful for certain types of meniscal tears such as
radial tears. These techniques became popular fol-
lowing the introduction of clinical MR applications
for musculoskeletal imaging. HAGGAR et al. (1988)
used 3DFT GRASS (gradient recalled acquisition in
the steady state) sequences for the evaluation of
meniscal tears in a group of 35 patients. They com-
pared the accuracy of this technique with convention-
al SE imaging, using 3-mm partitions, and observed
equal detection of meniscallesions with both tech-
niques.
Different forms of fat suppression are frequently
used for the evaluation of musculoskeletal lesions
(MARTy-DELFAuT and BELTRAN 1999). In the knee,
the main value of fat suppression has been in the
study of bone contusions (LAL et al. 2000) and carti-
lage lesions (TRATTNIG et al. 1998), and for this pur-
pose, fat suppression is often added to the arma-
mentarium of pulse sequences for general evaluation
of knee disorders. For the assessment of meniscal
lesions, fat suppression techniques are mostly used
in combination with direct or indirect MR arthrog-
raphy, to enhance the contrast between the injected
Gd-DTPA and the adjacent pericapsular fat.
The Menisci 131

9.2.2
Artifacts

Artifacts related to imaging techniques can occasion-

ally produce confusion and pitfalls. Two imaging arti-
facts need to be taken into consideration when evalu-
ating meniscal pathology: the truncation artifact and
the magie angle phenomenon (Table 9.1). The trunca-
tion artifact has been weIl described in the literature
(CZERVIONKE et al. 1988; LUFKIN et al. 1986; WOOD
and HENKELMAN 1985) and it is related to the Fou-
rier transform methods of constructing MR images
at boundaries between tissues of high signal intensity
difference, such as the boundaries between articular
cartilage (bright) and meniscus (dark). Under these
circumstances, "overshoots" and "undershoots" occur
in alternate pixels moving away from the boundary
(tuRNER et al.1991). The end result is the presence of Fig.9.1. Truncation artifact. Sagittal proton density (SE
bands or lines of high and low signal intensity adja- 2000/20) image demonstrating two hyperintense lines cross-
cent and parallel to the high-contrast boundary. These ing the posterior horn of the lateral meniscus, simulating
high signal intensity lines, projected over the menis- meniscal tears
cus, may produce an image very similar to a menis-
cal tear (Fig. 9.1). tuRNER et al. (1991) described the
impact of the truncation artifact in a cadaver human the superior-inferior direction. The use of an acquisi-
knee and 83 patients. They found that the artifact tion matrix of 256x256, and orientation of the phase
was more prominent when the acquisition matrix encoding gradient in the anterior-posterior direction,
was 128x256 and the phase-encoded gradient was in reduced or nearly eliminated the artifact.

Table 9.1. Meniscal tears: pitfalls in interpretation

Cause Appearance Location Plane How to avoid it

Truncation artifact Horizontalline Periphery of Sagittal Increase matrix

body
Swap gradients
Magie angle Diffuse Post. horn of Sagittal Increase TE
artifact hyperintensity lateral meniscus
Coronal
Anterior insertion Vertical Ant. horn of Sagittal Knowledge of anatomy
of lateral meniscus striations lateral meniscus
Attachments of Vertiealline Ant. horns of Sagittal Inspect contiguous
transverse ligament both menisci sections
Meniscofemoral Vertiealline Post. horn of Sagittal Inspect contiguous
ligament of Wrisberg lateral meniscus sections
Volume averaging Horizontalline Periphery, body Sagittal Thinner sections
Popliteus tendon Vertiealline Post. lateral Sagittal Knowledge of anatomy
meniscus
Coronal
Vacuum phenomenon Horizontalline Both menisci Sagittal Correlation with
radiographs
Coronal
Meniscal ftounce Waviness Free edge of Sagittal Knowledge of variant
menisci
Previous Irregular, Free edge of Sagittal Proper communieation
meniscectomy small meniscus menisci with surgeon
Coronal

Post., Posterior; ant.) anterior

132 J. Beltran and S. Shankman

The magie angle artifact has also been weIl

described in the literature (ERIKSON et al. 1991, 1993;
RUBENSTEIN et al. 1993). This phenomenon was first
described in tendons (FULLERTON et al. 1985) but
it can be seen in any collagen-containing tissues.
The magic angle phenomenon refers to the increased
signal intensity that occurs when the collagen fibers
of these tissues are oriented at 55° with the statie
magnetie field (B o). At this angle, the dipole-dipole
interactions that contribute to T2 relaxation among
water protons constrained between collagen fibers
are nulled, resulting in an increase in the T2 relax-
ation time of the tissue and thus increased signal
intensity when using TEs of less that 20 ms. PETERFY Fig.9.2. Computer-generated three-dimensional surface ren-
et al. (l994) evaluated this phenomenon in 58 knee dering of the medial meniscus
MR studies and found increased signal intensity of
the medial segment of the posterior horn of the later-
al meniscus in 42 examinations. They concluded that
this was related to the magic angle phenomenon sec- MR arthrography with intra-artieular injection of
ondary to the angular orientation of the meniscus at diluted Gd-DTPA has been used successfully in the
this partieular location and should not be interpreted shoulder for the detection of subtle findings. In the
as a tear or degeneration of the meniscus. knee, its use has been limited to the postoperative
situation (Fig.9.3) (ApPELGATE et al. 1993). Nonen-
hanced MR imaging of the virgin knee has reached
9.2.3 such a level of accuracy that the added patient dis-
Post Processing comfort and potential complications of intra-artie-
ular contrast injection is in general considered not
Post processing of the MR images plays an important worthwhile. However, early experience with intrave-
role in the evaluation of meniscal tears. Two tech- nous injection of Gd-DTPA (indirect MR arthrogra-
niques deserve some attention. The first relates to phy) (VAHLENSIECK et al. 1998) seems to indieate that
the capability of imaging software to render three- it may be a valid substitute for intra-articular injec-
dimensional surface reconstruction of the meniscus tion. Further clinieal testing is necessary to deter-
from 3D or 2D data sets. DISLER et al. (l993) com- mine the future role of this technique.
pared the accuracy of 3D reconstruction of 24 menisci
versus conventional 2D display, and found improved
accuracy for detection of meniscal tears when the 3D
reconstructions were used (Fig.9.2). Although this
post processing technique is available in many state-
of-the-art MR scanners, its use has not become widely
accepted because of the time required to obtain the
reconstructions and the excellent spatial resolution
obtainable with current scanners and surface coils, a
which makes post processing unnecessary.
The second technique refers to the imaging of
the menisci using narrow windows ("meniscal win-
dows") instead of conventional windows. This prac-
tiee is relatively common because it is thought to
improve the identification of meniscal tears (STOLLER
~ ____ ~ ____ ~_ b
et al. 1987). BucKwALTER et al. (l993) studied the
impact of narrow windowing of the menisci in 48
Fig. 9.3a, b. MR arthrogram following partial medial meniscec-
patients using ROC curve analysis and they could not tomy. Sagittal and coronal Tl-weighted images following the
prove that this photographing technique improves intra-articular injection of gadolinium saline solution show a
the detection of meniscal tears. normal medial meniscal stump
The Menisci
9.2.4
Field Strength
A technical aspect that has undergone extensive dis-
cussion is related to the effect of field strength for
the detection of meniscallesions. RUWE et al. (1991)
suggested that the accuracy of diagnoses based on
MR images may depend on the field strength of
the magnet. BARNETT (1993) evaluated the efficacy
of a O.5-T magnet in aseries of 118 patients with
arthroscopic correlation, and compared his results
with published series in the literature using 1.5-T
magnets. For te ars of the medial meniscus, the sensi-
tivity, specificity, and accuracy of MR at 0.5 T were
93%,90%, and 92%, respectively. For tears of the lat-
eral meniscus the sensitivity, specificity, and accura-
cy were 81 %,97%, and 93%, respectively. The author
concluded that magnetic field strength is not a sig-
nificant determinant of diagnostic reliability of MR
assessment of internal derangement of the knee.
Several published articles comparing the diagnostic
performance oflow-field (0.2 T) versus high-field mag-
nets (1.5 T) (COTTEN et al. 2000; PARIZEL et al. 1995;
KERSTIN-SOMMERHOFF et al. 1995; FRANKLYN et al.
1997; SHELLOCK et al.1998; RIEHL et al.1997) reported
that dedicated extremity magnets operating at 0.2 T are
comparable in sensitivity, specificity, and accuracy for
the diagnosis of meniscallesions. On the other hand,
other authors (FISHER et al.1991; KINNUNEN et al.1994;
WENNOGLE et al. 1998) have reported decreased accu-
racy of low-field magnets for meniscal tears as com-
pared to high-field systems. COTTEN et al. (2000) ana-
lyzed the discrepancies of the different publications
and concluded that the different results are related to
different performances of the low-field magnets, and
differences in the technical parameters used. The same
authors evaluated the two field strengths in aseries
of 219 patients using similar pulse sequences. Ninety
patients in their series underwent surgical arthroscop-
ic examinations. They concluded that the results were
comparable for both field strengths.
9.3
Normal Anatomy
The menisci of the knee are two semilunar, C-shaped
fibrocartilaginous disks that sit on the peripheral
margins of the essentially Hat tibial plateau (Fig. 9.4).
The upper surfaces of both menisci are concave and
articulate with the convex femoral condyles. Both
menisci measure approximately 4-7 mm in height at
133
Fig. 9.4. Schematic drawing of the normal menisci
the periphery and 1 mm or less at the free edge. The
peripheral one-third contains neurovascular struc-
tures, whereas the remaining two-thirds is strictly
fibrocartilaginous. The vascularized peripheral por-
tion of the meniscus has been termed the "red zone:'
while the designation "white zone" refers to the fibro-
cartilaginous, avascular area (POEHLING et al. 1990;
NEWMAN et al. 1993; HAVEN et al. 1989). These zones
can be differentiated histologically but are not obvi-
ous at visual inspection (HAUGER et al. 2000).
The lateral meniscus has the same width through-
out, approximately 7-10 mm. It is shaped like a
three-quarter circle, with its anterior and posterior
horns attached to the tibia immediately in front of
and behind the intercondylar eminen ce. The anteri-
or attachment of the lateral meniscus has a striate
appearance and its fibers are in continuity with the
more anterior fibers of the anterior cruciate liga-
ment. The central attachment sites are narrower in
width. The peripheral margin of the lateral menis-
cus is attached to the capsule except posterolaterally,
where the popliteus tendon cross es it, and more pos-
teriorly and centrally ne ar the central attachment
site, where the capsule does not extend anteriorly
into the joint.
The medial meniscus is shaped more like a half
circle. Although its posterior horn is attached to the
tibia just posterior to the posterior lateral meniscal
attachment, its anterior horn attaches far anteriorly,
approximately 10-14 mm anterior to the anterior
horn of the lateral meniscus. The width of the medial
meniscus gradually tapers from posterior to anteri-
or, in contrast to the lateral meniscus. The periph-
 134
eral margin of the medial meniscus is more firmly
attached to the joint capsule and to the tibial plateau
itself, the latter via the coronary ligament. As with the
lateral side, the capsule does not extend anteriorly
into the joint near the posterior central attachment
site.
The transverse ligament connects the anterior
horns of both menisci. Its thickness varies from 1 to
4 mm. It arises at the most anterior-superior portion
of the lateral meniscus and crosses in front of the
tibial attachment of the anterior cruciate ligament,
merging with the superior portion of the posterior
aspect of the anterior horn of the medial meniscus.
The meniscofemoralligaments are inconstant, 3-
to 4-mm fibrous bands arising from the posterior
horn of the lateral meniscus and attaching to the lat-
QX!l
a '
J. Beltran and S. Shankman
eral aspect of the medial femoral condyle. The liga-
ments of Humphry and Wrisberg arise together, the
former coursing anterior to the posterior cruciate lig-
ament (PCL) and the latter posterior to the PCL.
9.3.1
Normal MR Imaging Appearance:
Variations and Pitfalls
Sagittal images best show the anterior and posterior
horns of the medial and lateral menisci (Fig.9.5),
while coronal images best show the meniscal bodies
(Fig. 9.6). Thin-section axial images allow good visu-
alization of the menisci, especially their free edges
(Fig.9.7).
_ -'1::._.... f
Fig.9.5a-i. Normal menisci. Serial contiguous sagittal proton
density images, medial to lateral, showing normal medial and
lateral menisci
 The Menisci
Fig.9.6a-g. Normal menisci. Serial contiguous coronal 2D
gradient echo images, anterior to posterior, showing normal
medial and lateral menisci
Sagittal sections show the posterior horn of the
medial meniscus as an isosceles triangle, the sides
nearly twice as long as the base. The anterior horn of
the medial meniscus is about one-half the width of
the posterior horn, appearing more as an equilateral
triangle. It may vary in its appearance, sometimes
showing an almost rounded configuration. It sits on
the extreme edge of the anterior tibia, the transverse
ligament joining it more superiorly and somewhat
posteriorly. This junction may create an "arrowhead"
appearance, pointing posteriorly. The lower portion
represents the anterior horn attachment site, and
the upper portion represents the transverse ligament
junction (Fig. 9.8).
On the lateral side, the transverse ligament attach-
es to the most anterior-superior aspect of the ante-
135
g
rior horn; it may appear as a round dot anterior to
the meniscus and should not be confused with a tear
(Fig. 9.8) (HERMAN and BELTRAN 1988). The anterior
horn of the lateral meniscus is in fact an uncommon
site for a meniscal tear.
The anterior horn of the lateral meniscus mayaiso
demonstrate a striated or speckled increased signal
intensity pattern on proton density images. This is
another normal variation, which may represent dense
collagenous fibers of the anterior cruciate ligament
intertwined with the fibrocartilage of the meniscus
itself (Fig. 9.9). SHANKMAN et al. (1997) demonstrat-
ed this speckled pattern of the anterior horn of the
lateral meniscus in 22 MR imaging studies of the
knee in patients in whom arthroscopic surgery per-
formed for other lesions showed no evidence of a tear
136 J. Beltran and S. Shankman

Fig.9.7. Normal menisci. Axial 2D GRE image through the

joint line demonstrating the lateral meniscus, with good visu-
alization of its free edge (arrows)

Fig.9.8a-h. Normal transverse ligament. Serial contiguous

sagittal proton density images, lateral to medial, showing the
course of the transverse ligament (arrows), connecting the h
anterior horns of both menisci
The Menisci 137

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-

Fig.9.9. Normal tibial attachment of the anterior horn of the

lateral meniscus. Sagittal proton density fat -saturated image Fig.9.10a-d. Ligament of Wrisberg. Serial contiguous sagittal
showing speckled increased signal at the anterior horn of the proton density images, lateral to medial, showing the course of
lateral meniscus near its central attachment site (arrow), not the ligament of Wrisberg (arrows), including its origin at the
to be confused with a tear superior margin of the posterior horn of the lateral meniscus
 138
Fig. 9.11. Meniscal tear pitfall. Far lateral sagittal proton density
image showing a horizontalline of increased signal at the lateral
meniscus (arrows), which should not be confused with a tear
Fig. 9.12. Normal popliteus tendon. Coronal 2D gradient echo
image showing the popliteus tendon (arrow) coursing through
the posterior horn of the lateral meniscus
b
Fig. 9.13a, b. Normal popliteus tendon. Sagittal proton den-
sity images showing the popliteus tendon (arrows) coursing
through the posterior horn of the lateral meniscus. Also evi-
dent is the transverse ligament anteriorly
J. Beltran and S. Shankman
in nature. This is seen on both the sagittal and the
coronal images as medium signal intensity on Tl-
weighted and spin density images, and as high signal
intensity on T2-weighted images.
More anteriorly, coronal images show the anterior
horn of the lateral meniscus as a bandlike structure.
The anterior horn of the medial meniscus is quite
small and extends more anteriorly than the lateral.
The avascular portions of the normal menisci (the
white zone) and the transverse and meniscofemoral
ligaments all appear dark, without signal on all pulse
sequences. CHAN et al. (1998) suggested that the
peripheral vascular zone (red zone) is of medium
to high signal intensity on T2-weighted images
(Fig. 9.14). This is most prominent at the posterior
horn of the medial meniscus, and is best seen on sag-
ittal images. This is an important consideration since
tears involving the red zone are more likely to heal
than those involving the white zone. Recent inves-
tigations by HAUGER et al. (2000) cast a shadow of
doubt over the feasibility of distinguishing the two
meniscal zones with MR imaging. These authors per-
formed conventional and Gd-DTPA-enhanced MR
examinations in cadaver knees in 18 patients. They
concluded that the wedge-shaped low signal inten-
sity structure seen on MR images represents the
entire meniscus, and that the intravenous injection
of contrast material was not useful in distinguishing
between the two zones.
With age, degeneration leads to indistinct focal
areas of medium to high signal intensity change in the
avascular portion of the menisci, which must not be
Fig.9.14. Red-white zone. Sagittal Tl fat-saturated image
through the posterior horn of the medial meniscus dem-
onstrating the peripheral portion of the meniscus, which is
hyperintense (red zone; straight arrow) in relation to the
fibrous central portion (black zone; curved arrow)
The Menisci
confused with a meniscal tear. Another potential pitfall
is due to the presence of a vacuum phenomenon that
mimics a meniscal tear (SHOGRY and POPE 1991).
Normal anatomical meniscal variants are rare. The
most frequent is the discoid meniscus, discussed
below. Another variant is the so-called meniscal
ftounce. Sagittal images obtained through the body
of the lateral or medial menisci occasionally dem on-
strate a folded configuration or S-shaped fold, termed
the meniscal ftounce by Yu et al. (1997) (Fig. 9.15).
This fold occurs in the absence of a tear and it does
not increase the prevalence of meniscal tears.
Fig. 9.15. Meniscal flounce. Sagittal proton density image dem-
onstrating an undulating portion of the body of the meniscus
(arrow)
9.4
Meniscal Trauma
Suspected meniscal te ars are one of the most fre-
quent indications for musculoskeletal MR imaging.
Most health care professionals dealing with patients
presenting clinically with knee pain in whom the
possibility of a meniscal tear is considered, agree that
MR imaging provides a fast and accurate preopera-
tive assessment of the knee. Technical developments
in the last decade in both high- and low-field sys-
tems have significantly improved the ability of MR
to detect meniscal injuries. These improvements
include better surface coil design, increased gradient
139
strength, and continuously improving pulse sequence
design. As research in the field of treatment of
meniscal injuries progresses, MR imaging faces new
diagnostic challenges including accurate distinction
between degenerative and non degenerative meniscal
tears, postoperative evaluation of the meniscus, espe-
cially in those patients treated with advanced tech-
niques such as meniscal repair and meniscal trans-
plant, assessment of the added information provided
bythe intra -articular or intravenous administration of
Gd-DTPA,and the evaluation of associated injury,e.g.,
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
highly dependent on the amount of residual meniscal
tissue rather than its biomechanical integrity (NOBLE
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
and HARGREAVES 1979). The most important stress
factors contributing to the development of a menis-
cal tear include differences in the friction coefficient
between the superior and inferior articular surfaces
of the meniscus, and the nature of the biomechanical
forces acting in normal joint motion. The combina-
tion of these factors leads to the development of tears
of the microstructure of the fibrocartilage, which
in turn causes molecular changes and chondrocyte
depletion (NOBLE and HAMBLEN 1975; SEEDHOLM
1979; SEEDHOLM and HARGREAVES 1979).
 140
9.4.1
Significance of Signal Alterations
CRUES et al. (1986, 1987, 1990) and MINK et al. (1988)
studied in depth the relationship between the micro-
structural changes of the meniscus undergoing a te ar
and the signal intensity changes as reftected on MR
imaging using different pulse sequences. Tears of the
collagen fibers of the meniscal fibrocartilage provide
a milieu for water protons to become more mobile,
thus increasing the effective T2 relaxation time of the
tissue, which in turn is manifested by an increase in
the signal intensity on short TE images (Tl, gradient
echo and proton density-weighted images). At this
stage, although the T2 is increased, it is still too
short to be detected on long TE images (T2-weighted
images). With further damage of the collagen fibers,
water is absorbed by the macromolecules, further
increasing the signal on short TE images. It is rec-
ognized that short TE imaging has a sensitivity great-
er than 90% in detecting meniscal tears, while using
long TE imaging the sensitivity decreases to less than
30% (CRUES et al. 1986, 1987, 1990; MINK et al. 1988).
Normal Grade I
J. Beltran and S. Shankman
With further derangement of the collagen struc-
ture of the meniscus and development of partial or
complete tears, voids are created and water is attract-
ed to the voids. This in turn increases the T2 to
thousands of milliseconds, and then the T2-weight-
ed images can demonstrate the signal changes corre-
sponding to the tear.
Classification of the signal intensity changes of
the meniscal tissue was first reported by LOTYSH
et al. (1986). These authors graded the meniscal
signal using morphological criteria (Fig. 9.16). Grade
I signal refers to a rounded or irregular area of high
signal within the substance of the meniscus, not
extending to the articular surface. Histological corre-
lation in these cases demonstrates early degenera-
tive changes. Grade II signal corresponds to a linear
area of high signal, not extending to the articular
surface and pathologically corresponding to more
severe degenerative changes. Both grade I and grade
II signal changes rarely correlate with arthroscopi-
cally demonstrable te ars. Grade III signal alterations
extend to the articular surface, appearing well mar-
ginated, linear in configuration and high in signal
Fig. 9.16a-d. Grading
system for meniscal
signal. a Diagram of the
grading system of signal
intensity changes in the
meniscus. b Grade I
signal. c Grade 11 signal.
d Grade III signal
-"--_--I d
The Menisci
intensity. These are the distinctive features of a tear,
with very high correlation with arthroscopy. The cor-
relation between the histological findings and signal
changes was validated by a study of 12 knee spec-
imens from autopsies or above-knee amputations
published by STOLLER et al. (1987).
CRUES et al. (1986, 1987) evaluated the relation-
ship between signal changes in the menisci and the
presence of tears demonstrated arthroscopically in a
series of 277 menisci in 144 knees, with arthroscopy
or arthrotomy correlation. They found that 89% of
menisci exhibiting grade I or 11 changes were normal
at surgery, and that 94% of menisci with grade III MR
signal changes had tears. They concluded that MR
imaging can separate surgically significant from non-
significant meniscallesions, and that it is useful for
preoperative screening of suspected meniscal tears.
In a very early study, REICHER et al. (1986) evaluated
the accuracy of MR imaging in a group of 49 patients,
correlating with subsequent arthroscopy. In their
grading system they added a grade IV, representing
a gross distortion of the normal shape, truncation,
or a large focus or line of increased signal within
the meniscus. About 80% of meniscal abnormalities
graded III or IV had tears. The false-positive find-
ings involved the posterior horns of the menisci, the
site of most false-negative arthroscopic examinations
(QUINN and BROWN 1991). The negative predictive
value of MR imaging was 100%.
DE SMET et al. (1994) analyzed the cause of errors
of MR imaging in meniscal tears in aseries of 400
MR examinations performed for suspected meniscal
te ars in which the accuracy was 90%. In this group
they found 70 cases in which the MR diagnosis did
not agree with the surgical findings. They found that
40% were unavoidable errors (false-positive or false-
negative diagnoses that could not be avoided even
in retrospective examination), 39% were related to
equivocal MR imaging findings and 21 % were due to
interpretation errors. In this series, 6% of the menis-
cal tears could not be identified, even retrospective-
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%
of the 400 knees evaluated.
KAPLAN et al. (1991) evaluated the significance
of high signal in the meniscus that does not clearly
extend to the surface (grades land 11) in aseries of
142 consecutive patients undergoing MR examina-
tions of the knee. The prevalence was 14% (20 cases).
of these, 13 showed no evidence of tear at arthros-
copy or arthrotomy and one showed internal degen-
eration on histological examination. They concluded
141
that when MR scans show grade I or 11 signal chang-
es, a meniscal tear is unlikely.
In another study, DILLON et al. (1990) evaluated
the clinical significance of grade 11 meniscal abnor-
malities in aseries of 365 patients of whom 44 under-
went arthroscopic evaluation. These authors subdi-
vided grade 11 meniscal abnormalities into grade
IIA (linear abnormal signal not contacting with the
meniscal surface), grade IIB (abnormal signal con-
tacting with the surface seen only in one image) and
grade IIC (extensive wedge-shaped signal abnormal-
ity not in contact with the surface, the "meniscus
within a meniscus" appearance). Arthroscopy dem-
onstrated 3% te ars in group IIA, 0% in group IIB and
50% in group IIC. They concluded that meniscal tears
should be diagnosed only if contact is seen in more
that one image and that many group IIC lesions may
have tears. DE SMET et al. (1993a) also found that if
the signal abnormality within a meniscus is contact-
ing one articular surface, a te ar is usually present, but
if the extension into the articular surface is seen in
only one section, a tear is less likely.
DILLON et al. (1991) also performed a prospective
study of grade 11 meniscal abnormalities without
associated lesion of the anterior cruciate ligament,
to ascertain whether these lesions progress to com-
plete tears. The study group consisted of 27 menisci.
On follow-up over 3 years, six decreased in size, 18
remained unchanged and two lesions disappeared.
The authors concluded that most grade 11 lesions are
stable.
In a further attempt to establish the value of the
grading system for meniscal lesions, KORNICK et
al. (1990) reviewed the MR imaging knee examina-
tions of 64 asymptomatic volunteers in the 2nd to
8th decades of life, and analyzed these for meniscal
abnormalities. They found grades I, II and III in all
decades, with a prevalence of 25% as early as the 2nd
decade, and increasing sharply with age. They con-
cluded that there is a baseline prevalence of meniscal
signal in the asymptomatic population.
9.4.2
Terminology and Signs of a Tear
Although the grading system for meniscal signal
abnormalities has been helpful in correlating the MR
imaging patterns with histological abnormalities, it
is often found in clinical practice that interobserver
and intraobserver variability is a common problem
with any type of classification scheme. Furthermore,
the report of the presence of a grade I or 11 "lesion"
142
may compel the surgeon to perform unnecessary sur-
gery. For this reason, a more practieal approach to
reporting meniscal tears has evolved during the last
decade, utilizing terms that correlate closely with the
arthroscopie and morphologie description of tears.
This terminology includes horizontal or oblique tear,
radial tear, parrot beak tear, bucket-handle tear, ver-
tical flap tear, horizontal flap tear, meniscocapsular
separation, and peripheral tear (Table 9.2).
Table 9.2. Specific MR signs of meniscal tear
Type oftear Finding Sign
Horizontal Oblique or horizontal
line extending to
meniscal surface
Radial Verticalline
Meniscal subluxation
without OA
Parrot beak Irregular margin or se-
parated small fragment
Bucket -handle Displaced fragment in Double peL sign
top of tibial spine
Displaced frag- Flipped meniscus
ment anteriorly sign
Flap tear Displaced fragment in Absent
different locations bow tie sign
Peripheral tear Verticalline in
periphery of meniscus
Meniscocapsular Verticalline at menisco-
separation capsular junction
The most common tear encountered on MR imag-
ing examinations of the knee is the nondisplaced
oblique or horizontal tear involving most frequently
the posterior horn of the medial meniscus, and
extending to its inferior artieular surface (Fig.9.17).
Less frequently these types of tears involve the poste-
rior lateral meniscus.
Vertieal tears are produced when the femoral con-
dyle compresses the meniscus, usually superiorly to
inferiorly, posteriorly to anteriorly, and from without
inward. These tears may extend along the radius of
the meniscus or be localized to one segment, usually
the posterior horn; they are often referred to as radial
tears (Fig. 9.18). Clinieally, radial tears can result from
trauma or be related to degeneration of the menis-
cus. Not infrequently these tears are a component of
a more complex meniscal tear. Radial tears can be
subclassified into partial or full thiekness based on
their peripheral extension towards the meniscocap-
sular junction.
A complete or full-thiekness radial tear results in
two fragments of meniscus, each attached only to
the tibia at one end. The loss of meniscal integrity
J. Beltran and S. Shankman
Fig.9.17. Typical meniscal tear. Sagittal proton density image
showing increased linear signal extending into the inferior
articular surface of the posterior horn of the medial meniscus
(grade III signal), consistent with a typical, arthroscopically
detectable meniscal tear
decreases the ability of the meniscus to distribute the
radially oriented forces originating in the meniscus
during normal weight bearing. These forces are dis-
tributed though the meniscus by circumferentially
oriented collagen fibers, which are transected in
radial tears. In addition, peripheral subluxation of
the meniscal fragments occurs owing to the loss
of attachments of the meniscus at both ends of
the radial tear. These factors lead to degenerative
artieular changes of the corresponding compartment
(TuCKMAN et al. 1994). Unfortunately, radial te ars
do not improve following resection of the involved
meniscal segment since this type of surgery does not
provide arepair of the circumferentially oriented col-
lagen fibers (HENNING et al. 1991; METCALF 1991).
MR images obtained perpendieular to the orienta-
tion of a radial tear will easily demonstrate the lesion,
but if the orientation of the imaging plane is in the
same direction as the radial tear, an irregular area
of abnormal signal intensity may be seen (Fig.9.18).
This should not be confused with an area of internal
meniscal degeneration. If a full-thiekness tear is pres-
ent, a section through the same plane of the tear will
demonstrate an absent meniscus. Thin-section axial
images very often provide good visualization of the
radial tears (Fig.9.18).
Full-thickness radial tears located ne ar the tibial
insertions of the meniscus may be very difficult to
detect with MR imaging. The meniscus seen only in
one side of the tear may be seen slightly displaced
from the center of the joint and the rest of the menis-
cus may look entirely normal (TUCKMAN et al. 1994).
In the region of the posterior horn of the medial
meniscus, this diagnosis may be easier to make if
one is aware that meniscal tissue should be present
immediately medial to the tibial attachment of the
posterior cruciate ligament. Absence of meniscus in
this specific location is strongly suggestive of radial
 The Menisci
a c
Fig. 9.18a-d. Radial meniscal tear. a Diagrammatic representa-
tion of an incomplete radial tear of the medial meniscus. b
Coronal2D gradient echo image showing a radial tear through
the posterior horn of the medial meniscus (arrow), near its
tibial attachment. c Axial 2D gradient echo demonstrating
the radial tear in the posterior horn of the medial meniscus
(arrow). d Sagittal proton density image obtained through the
radial tear showing increased signal within the posterior horn
of the medial meniscus simulating an area of meniscal degen-
eration or intrasubstance tear
tear. In the lateral meniscus this assessment is more
difficult. The posterior horn of the lateral meniscus
should cover the most medial portion of the lateral
tibial plateau. Failure to do so is also suggestive
of a radial tear in this location. A secondary sign
of a radial tear is the presence of subluxation of
the meniscus, in the absence of osteoarthritis, as
described by TUCKMAN et al. (1994).
"Parrot beak" tears may be considered a variation
of a radial tear in which a portion of the meniscus
adjacent to the tear becomes slightly displaced and its
shape resembles the beak of a parrot (Fig.9.19).
They are best seen on coronal or thin-section axial
images. On sagittal views, the "bow tie" configuration
of the lateral meniscus may be disrupted centraHy
(STOLLER et al. 1987; DE SMET et al. 1993a, b).
143
d
An extensive vertieal or oblique tear with central
displacement of the free edge is referred to as a buck-
et-handle tear. The handle represents the displaced
free margin towards the intercondylar notch, and
the peripheral non-displaced fragment represents the
bucket (Fig. 9.20). The tear normally starts in the pos-
terior horn of the meniscus and extends anteriorly
in a longitudinal fashion towards the anterior horn
(SINGSON et al. 1991). Compressive forces by the fem-
oral condyle produce central displacement of the
fragment of meniscus. The tear may be asymmetrie
and spare the anterior horn.
Bucket-handle te ars are clinicaHy significant
because they produce mechanical locking or diffi-
culty in fuHy extending the knee joint. The medial
meniscus is more frequently involved. Bucket-handle
 144 J. Beltran and S. Shankman

Fig. 9.19a, b. Parrot beak tear. a Diagrammatic representation

of a parrot beak te ar of the medial meniscus. b Axial2D gradi- b
ent echo image in a patient following partial medial menis-
cectomy. Note the irregularity of the free margin related to the
surgery and the presence of a parrot beak tear of the body of
the meniscus (arrow)

tears are seen more often in young adults. The inci-

dence of this type of tear has been reported to be in
the range of 9%-24% in large series (POEHLING et
al. 1990; SHAKESPEARE and RIGBY 1983). In arecent
series published by HELMS et al. (1998), there were 33
bucket-handle tears in 350 arthroscopic surgical pro-
cedures (9% incidence).
The MR imaging diagnosis ofbucket-handle tears c
may be overlooked if one is not aware of specific
signs, which are weIl described in the literature. The
most reliable sign is the finding of the displaced frag- Fig.9.20a-c. Bucket -handle tear. a Diagrammatic representa-
ment of meniscus in the intercondylar notch, which tion of a bucket-handle tear with the displaced fragment in
is better seen on coronal images (SINGSON et al. 1991; the top of the tibial spine. b Sagittal proton density image
HELMS et al. 1998; WEISS et al. 1991; HARAMATI et al. through the midline demonstrating the displaced fragment of
a bucket-handle tear, parallel to the posterior cruciate liga-
1993). In aseries of 39 cases published by WRIG HT et ment: the "double PCL" sign (arrow). c Coronal gradient echo
al. (1995), the displaced fragment was found in image showing the displaced fragment adjacent to the tibial
84% of the cases. This sign is associated with a spine (arrow)
 The Menisci 145

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
attaehment of the flap, then a free meniseal frag-
ment oeeurs. MR imaging ean demonstrate the mor-
phology of the vertieal flap tears and the loeation of
the fragment. No series are available in the literature
assessing the aeeuraey of MR imaging in the diagno-
sis of this type of tear.
Another type of flap tear oeeurs when a horizontal
tear produees a fragment of meniseal tissue that may
beeome displaeed superiorly or inferiorly (Fig. 9.22)
(RUFF et al. 1998). These tears are less frequent than
the bueket-handle tears but in one series aeeounted
for 19% of symptomatie meniseal injuries (DANDY
and ]ACKSON 1975). They usually oeeur in the medial
meniseus. Superior displaeement of the flap is more
frequent than inferior displaeement. In one series
of 25 displaeed meniseal injuries, RUFF et al. (l998)
Fig.9.22a, b. Flap meniscal tear. a Diagrammatic representa-
tion of a fiap tear of the medial meniscus. The fragment is
displaced superiorly. b Coronal proton density fat-saturated
image demonstrating a vertically displaced fragment of the
medial meniscus (arrow), adjacent to the PCL
J. Beltran and S. Shankman
found two superiorly displaeed flap tears. In a retro-
speetive evaluation of 236 MR imaging examinations
showing displaeed meniseal tears, LECAS et al. (2000)
found 11 inferiorly displaeed flap tears. The fragment
lay inferomedial to the tibial plateau and deep in rela-
tion to the medial eollateralligament.
Meniseoeapsular separation is defined as a detaeh-
ment of the peripheral portion of the meniseus from
its eapsular attaehment. These lesions are uneom-
mon. They are treated nonoperatively if they are iso-
lated, or a meniseal repair is done if there is an asso-
ciated ligamentous injury (PRICE and ALLEN 1978;
STONE 1979; HAMBERG et al. 1983; STRAND et al.
1985). The results of this eonservative management
are good beeause the tear oeeurs in the red zone of
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).
Fig.9.23. Peripheral meniscal tear. Sagittal proton density
image showing vertical increased signal at the periphery of
the posterior horn of the medial meniscus, consistent with a
peripheral tear (arrow)
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
the meniseal eyst into the surrounding soft tissues
(TYSON et al. 1995; LEKTRAKUL et al. 1999). They are
more often seen in association with lateral meniseal
tears, but medial meniseal tears mayaiso develop
 The Menisci
parameniscal cysts. Their size varies from a few milli-
meters to several centimeters and they produce pain.
Large cysts are palpable at the level of the joint line.
The meniscal tear associated with the cyst is gen er-
ally horizontal in configuration and extends to the
periphery of the meniscus (Fig.9.24). Occasionally
they may be found adjacent to the posterior cruciate
ligament, simulating a posterior cruciate ganglion
cyst (LEKTRAKUL et al. 1999).
Meniscal ossicles are rare lesions found in young
individuals (MARTINOLl et al. 2000). Their origin is
controversial. Some researchers believe that they are
vestigial structures but others propose a post-trau-
matic etiology (BERG 1991; RICHMOND and SARNO
1988). In general they are asymptomatic and they are
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
of the ossicle.
b be oriented in different directions (complex tears)
Fig. 9.24a, b. Meniscal cyst. a Coronal, posterior 2D gradient
echo image shows a large cyst communicating with a horizon-
tal tear of the medial meniscus at the medial joint Hne, typical
of a meniscal cyst (arrow). b Axial 2D gradient echo image
shows the posteromedial extension of the cyst (arrows)
147
A discoid meniscus refers to a meniscus, almost
always the lateral one, that is not C-shaped but rather
disk-like in configuration, covering most of the tibial
plateau to varying degrees. It is prone to tearing and
is usually seen in children and adolescents. They may
be asymptomatic and incidentally noted. Although
seen in youngsters, it is believed that they are more
developmental than congenital in that the fetal menis-
cus never assurnes such a shape.
The diagnostic criteria of discoid meniscus on MR
imaging include continuity of meniscal tissue with
the anterior and posterior horns in three or more
5-mm contiguous sagittal sections and/or a trans-
verse width greater than 14 mm (ARAKI et al. 1994;
SILVERMAN et al. 1989) (Fig. 9.25).
a
Fig. 9.25a-c. Tom discoid meniscus. Serial contiguous sagittal
proton density images show thick meniscal tissue at the lateral
joint compartment with extensive tear, consistent with atom
discoid lateral meniscus
Discoid meniscus may undergo tears, which can
including the meniscocapsular junction (Ryu et al.
1998). Most frequently the meniscal te ars are periph-
eral or peripheral with a horizontal component. In a
series of 77 patients, published by Ryu et al. (1998),
the positive predictive value of MR imaging for the
148 J. Beltran and S. Shankman

diagnosis of meniscal tears in discoid menisci was

low. This was probably due to the high incidence of
multiple tears and also the tendency for a discoid
meniscus to undergo degeneration owing to its
abnormal structure (differentiation between degen-
eration and te ar may be difficult).

a
9.6
Postoperative Meniscus

The MR imaging appearance of the postoperative

meniscus depends on the nature of the surgical inter-
vention. After partial meniscectomy, it can be quite
difficult to evaluate tearing of the meniscal remnant
or incomplete meniscal resection. The arthroscopi- b
cally normal meniscal stump varies in its MR imag-
ing appearance. It may demonstrate a smooth con-
tour with homogeneous signal intensity. Standard
MR imaging criteria for meniscal tears are useful
only in this situation. Other such "normal" stumps
may demonstrate an irregular surface contour with
inhomogeneous signal intensity. In fact, such a post-
operative appearance represents anormal meniscal c
remnant most of the time. In this setting, however,
it becomes very difficult to delineate true pathology
(Fig. 9.26) (SMITH and TOTTY 1990).
Peripheral meniscal tears are often treated conser-
vatively or with arthroscopic repair (LYNCH et al. 1983;
MILLER 1988; DEHAVEN 1985; HAMBERG et al. 1983;
d
GRAF et al. 1987; ARNOCZKY et al. 1985). In either
situation it has been shown that persistent increased
signal intensity at the periphery of the meniscus Fig.9.26a-d. Retear following partial meniscectomy. Sagittal
proton density (a, b) and coronal2D gradient echo (c,d) show
remains unchanged from that seen on preoperative increased signal at the lateral meniscus,following partial men-
studies in patients who have become asymptomatic iscectomy in an asymptomatic patient
and presumably healed. The presence of such signal
therefore should not be misinterpreted as meniscal
retear. In this regard, some cases offalse-positive MR
imaging examination may in fact represent older inju- be diagnostic. Fluid, like contrast, may insinuate itself
ries that have healed by the time arthroscopic exami- within a true te ar (ApPLEGATE et al. 1993).
nation is performed (DEUTSCH et al. 1990). In recent years, meniscal transplantation using
MR arthrography of the knee is recommended in allografts has become an accepted technique for
those cases where there has been prior meniscecto- treatment of patients who have undergone prior total
my or arthroscopic repair of a peripheral tear. Dilut- or subtotal meniscectomy with progressive loss of
ed Gd-DTPA can be injected directly into the knee meniscal tissue owing to repeated trauma and of
joint or intravenously with imaging performed after patients with early osteoarthritis confined to one
10-15 min of moderate exercise. Contrast will insin- compartment of the femorotibial joint and normal
uate itself into true tears and will not extend into knee alignment (POTTER et al. 1996). MR imaging has
areas ofhealing (Fig. 9.27). Tl-weighted imaging with been found helpful in the preoperative evaluation of
fat suppression is recommended following the intra- allografts and adjacent articular cartilage, to proper-
articular injection of a Gd-DTPA saline solution. If a ly identify those patients who are at risk for failure of
joint effusion is present, a routine MR of the knee may their transplanted meniscus.
The Menisci
9.7
Accuracy
Multiple studies have been published assessing the
accuracy of MR imaging, using arthroscopy as the
"gold standard" (CRUES et al. 1986; REICHER et al.
1986; DE SMET et al. 1993a, b, 1994; QUINN et al. 1992).
In general it is agreed that MR imaging is a very accu-
rate technique for preoperative evaluation of suspect-
ed meniscallesions. Published analyses have report-
ed better sensitivity of MR imaging for detection of
tears involving the medial meniscus, as compared
with the lateral meniscus (DE SMET et al. 1993a;
RAUNEST et al. 1991). One of the main impacts of MR
imaging of the knee has been its cost effectiveness
when used as a screening tool prior to arthroscopy,
thus preventing unnecessary surgery (SUAREZ-ALM-
AZOR et al. 1999; ELVENES et al. 2000; BUI-MANSFIELD
et al. 1997).
Early publications in the orthopedic literature had
questioned the accuracy of MR imaging of the knee
(FISHER et al. 1991; GLASHOW et al. 1989; JACKSON et
al. 1988). However, after more than 15 years of wide-
spread use, MR imaging is today well accepted as
the technique of choice for screening knees with sus-
pected internal derangement. In the past contro-
versies have arisen when comparing MR imaging
and arthroscopy. Several factors have contributed
to such controversy. These include technically diffi-
cult arthroscopies, delays between MR imaging and
arthroscopy, and ambiguities in the wording of both
the referralletter and the radiological report. It is rec-
149
_ _ _ _-li b
Fig.9.27a-c. MR Arthrogram. Retear following partial menis-
cectomy. Sagittal (a) and coronal (b, c) Tl-weighted images
following the intra-articular injection of gadolinium saline
solution show contrast within the medial meniscus stump fol-
lowing partial meniscectomy, consistent with retear (arrows)
ognized that arthroscopic detection of tears involv-
ing the posterior horn of the medial meniscus may
be quite difficult to detect arthroscopically, especially
when they extend to the inferior articular surface of
the meniscus (QUINN et al. 1992; MACKENZIE et al.
1995). These factors should be considered when eval-
uating the diagnostic performance ofboth MR imag-
ing and arthroscopy.
References
Applegate GR, Flannigan BD, Tolin BS et al (1993) MR diagno-
sis of recurrent tears in the knee: value of intraarticular
contrast material. AJR 161 :821-825
Araki Y, Yamamoto H, Nakamura H et al (1994) MR diagnosis of
discoid lateral menisci of the knee. Eur J RadioI18:92-95
Arnoczky SP, Warren RF (1983) The mierovaseulature of the
meniseus and its response to injury: an experimental study
in the dog.Am J Sports Med 11:131-141
Arnoezky SP, Warren RF, Kaplan N (1985) Meniseal remodel-
ing following partial meniseeetomy: an experimental study
in the dog. J Arthrose 1:247-252
Arnoezky SP, Warren RF, Spivak JM (1988) Meniseal repair
using an exogenous fibrin clot. J Bone Joint Surg [Am]
70:1209-1219
Barnett MJ (1993) MR diagnosis of internal derangements
of the knee: effect of field strength on efficacy. AJR
161:115-118
Berg EE (1991) The meniseal ossicle: the eonsequenee of a
meniseal avulsion. Arthroseopy 7:241-243
Boden SD, Davis DO, Dina TS et al (1992) A prospeetive and
blinded abnormal findings in asymptomatic subjeets. Clin
Orthop 282:177-185
150
Buckwalter KA, Braunstein EM, Janizek DB et al (1993) MR
imaging of meniscal tears: narrow versus conventional
window width photography. Radiology 187:827-830
Bui-Mansfield LT, Youngberg RA, Warme W et al (1997) Poten-
tial cost savings of MR imaging obtained before arthros-
copy of the knee: evaluation of 50 consecutive patients. AJR
169:913-918
Chan PS, Kneeland JB, Gannon FH et al (1998) Identification
of the vascular and avascular zones of the human meniscus
using magnetie resonance imaging: correlation with histol-
ogy. Arthroscopy 14:820-823
Cheung LP, King CPL, Hollett MD et al (1997) Meniscal tears
of the knee: accuracy of detection with fast spin-echo MR
imaging and arthroscopie correlation in 293 patients. Radi-
ology 203:508-512
Cotten A, Delfaut E, Demondion X et al (2000) MR imaging
of the knee at 0.2 and 1.5 T: correlation with surgery. AJR
174:1093-1097
Crues JB, Stoller DW (1993) The menisci. In: Mink Jet al (eds)
Magnetie resonance imaging of the knee. Raven, New York,
pp 91-140
Crues JV, Mink J, Levy TL et al (1986) Meniscal tears of the
knee. AJR 47:293-30
Crues JV, Mink J, Levy TL et al (1987) Meniscal tears of the
knee: accuracy of MR imaging. Radiology 164:445-448
Crues JV, Ryu R, Morgan FW (1990) Meniscal pathology:
the expanding role of magnetie resonance imaging. Clin
Orthop 252:80-87
Czervionke LF, Czervionke JM, Daniels DL et al (1988)
Characteristie features of MR truncation artefacts. AJR
151:1219-1228
Dandy DJ (1990) The arthroscopie anatomy of symptomatie
meniscallesions. J Bone Joint Surg [Br] 72:628-633
Dandy DJ, Jackson RW (1975) Meniscectomy and chondro-
malacia of the femoral condyle. J Bone Joint Surg [Am]
57:1116-1119
DeHaven KE (1985) Rationale for meniscus repair or excision.
Clin Sports Med 4:267-273
DeSmet AA, Norris MA, Yandow DR et al (1993a) MR diagnosis
of meniscal tears of the knee: importance of high signal in
the meniscus that extends to the surface.AJR 161:101-107
DeSmet AA, Norris MA, Yandow DR et al (1993b) Diagnosis
of meniscal tears of the knee with MR imaging: effect of
ob server variation and sampIe size on sensitivity and speci-
ficity. AJR 160:555-559
DeSmet AA, Tuite MJ, Norris MA et al (1994) MR diagnosis
of meniscal tears: analysis of causes of errors. AJR
163:1419-1423
Deutsch AL, Mink JH, Fox JM et al (1990) Peripheral
meniscal tears: MR findings after conservative treatment or
arthroscopie repair. Radiology 176:484-488
Dillon EH, Pope CF, Lynch K (1990) The clinieal significance
of stage 2 meniscal abnormalities on magnetic resonance
knee images. Magn Reson Imaging 8:411-415
Dillon EH, Pope CF, Jokl P et al (1991) Follow-up of grade
2 meniscal abnormalities in the stable knee. Radiology
181:849-852
Disler DG, Kattapuram SV, Chew FS et al (1993) Meniscal tears
of the knee: preliminary comparison of three-dimensional
MR reconstruction with two-dimensional MR imaging and
arthroscopy. AJR 160:343-345
Elvenes J,Jerome CP,Reikeras 0 et al (2000) Magnetie resonance
imaging as a screening procedure to avoid arthroscopy for
J. Beltran and S. Shankman
meniscal tears.Arch Orthop Trauma Surg 120(112):14-16
Eriekson SJ, Cox IH, Hyde JS et al (1991) Effect of tendon ori-
entation on MR imaging signal intensity: a manifestation
of the "magie angle" phenomenon. Radiology 181:389-392
Eriekson SJ, Prost TW, Timins ME (1993) The "magie angle"
effect: background physies and clinieal relevance. Radiol-
ogy 188:23-25
Escobedo EM, Hunter JC, Zink-Brody GC et al (1996) Useful-
ness of turbo spin-echo MR imaging in the evaluation of
meniscal tears: comparison with a conventional spin-echo
sequence. AJR 167:1223-1227
Eustace S, Hentzen P, Adams J et al (1999) Comparison of con-
ventional and turbo spin-echo Tl-weighted MR imaging in
acute knee trauma. AJR 172:1393-1395
Ferrer-Roca 0, Vilalta C (1980) Lesion of the meniscus.
Part 1. Macroscopie and histologie findings. Clin Orthop
146:301-307
Fischer SP, Fox JM, DelPizzo W et al (1991) Accuracy of diag-
noses from magnetie resonance imaging of the knee. J Bone
Joint Surg [Am] 73:2-10
Franklyn PD, Lemon RA, Barden HS (1997) Accuracy of imag-
ing the menisci on an in-office, magnetie resonance imag-
ing extremity system. Am J Sports Med 25:383-388
Fullerton G, Cameron I, Ord V (1985) Orientation of tendons
in the magnetie field and its effect on T2 relaxation times.
Radiology 155:433-435
Gillquist J, Messner K (1993) Long-term results of meniscal
repair. Sports Med Arthrosc Rev 1: 159-163
Glashow JL, Katz R, Schneider M et al (1989) Double-blind
assessment of the value of magnetie resonance imaging in
the diagnosis of anterior cruciate and meniscal lesions. J
Bone Joint Surg [Am] 71:113-119
Graf B, Docter T, Clancy W Jr (1987) Arthroscopie meniscal
repair. Clin Sports Med 6:525-535
Haggar AM, Froelieh JW, Hearshen et al (1988) Meniscal abnor-
malities of the knee: 3DFT fast-scan grass MR imaging.AJR
150:1341-1344
Hamberg P,Gillquist J, Lysholm J (1983) Suture of new and
old peripheral meniscus tears. J Bone Joint Surg [Am]
65:193-197
Haramati N, Staron RB, Rubin S et al (1993) The flipped menis-
cus sign. Skeletal RadioI22:273-277
Hauger 0, Frank LR, Boutin RD et al (2000) Characterization
of the "red zone" of knee meniscus: MR imaging and histo-
logic correlation. Radiology 217:193-200
Haven KE,Gillquist J (1989) Non-operative treatment of menis-
cal tears. J Bone Joint Surg [Am] 71:811-822
Helms CA, Laorr A, Cannon DW (1998) The absent bow tie
sign in bucket-handle tears of the menisci in the knee. AJR
170:57-61
Hendler RC (1984) Arthroscopie meniscal repair; surgieal tech-
nique. Clin Orthop 190:163-169
Henning CE, Lynch MA (1985) Current concepts of meniscal
function and pathology. Clin Sports Med 4:259-265
Henning CE, Yearout KM, Vequist SW et al (1991) Use of the
fascia sheath coverage and exogenous fibrin clot in the
treatment of complex meniscal tears. Am J Sports Med
19:626-631
Herman LJ, Beltran J (1988) Pitfalls in MR imaging of the knee.
Radiology 167:775-781
Jackson JP (1968) Degenerative changes in the knee after men-
iscectomy. Br Med J 2:525-527
Jackson DW, Jennings LD, Maywoon RM et al (1988) Magnetic
The Menisci
resonance imaging of the knee. Am J Sports Med 16:29-38
Kaplan PA, Nelson NL, Garvin KL et al (1991) MR of the knee:
the significance of high signal in the meniscus that does
not clearly extend to the surface. AJR 156:333-336
Kersting-Sommerhoff B, Gerhardt P et al (1995) MR of the
knee joint: first results of a comparison of 0.2 T specialized
system and 1.5 T high field strength magnet. Rofo Fortschr
Geb Rontgenstr Neuen Bildgeb Verfahr 162:390-395
Kinnunen J, Bondestam S, Kivioja A et al (1994) Diagnostie
performance of low field MR imaging in acute knee inju-
ries. Magn Reson Imaging 12:1155-1160
Korniek J, Trefelner E,McCarthy Set al (1990) Meniscal abnor-
malities in the asymptomatie population at MR imaging.
Radiology 177:463-465
Lal NR, Jamadar DA, Doi K, Newman JS et al (2000) Evaluation
of bone contusions with fat -saturated fast spin -echo pro-
ton-density magnetie resonance imaging. Can Assoc Radiol
J 51(3):182-185
Lecas LK, Helms CA, Kosarek FJ et al (2000) Inferiorly dis-
placed flap tears of the medial meniscus: MR appearance
and clinieal significance.AJR 174:161-164
Lektrakul N, Skaf A, Heh L et al (1999) Pericruciate meniscal
cysts arising tears of the posterior horn of the medial
meniscus: MR imaging features that simulate posterior cru-
ciate ganglion cysts.AJR 172:1575-1579
Listerud J, Einstein S, Outwater E et al (1992) First principles
of fast spin echo. Magn Reson Q 8: 199-244
Lotysch M, Mink J, Crues JV et al (1986) Magnetie resonance
imaging in the detection of meniscal injuries. Magn Reson
Imaging 4:185
Lufkin RB, Pusey E, Stark DD et al (1986) Boundary artefact due
to truncation errors in MR imaging. AJR 147:1283-1287
Lynch MA, Henning C, Glick KR Jr (1983) Knee joint surface
changes, long-term follow-up meniscus te ar treatment
in stable anterior cruciate ligament reconstructions. Clin
Orthop 172:148-153
Mackenzie R, Keene GS, Lomas DJ et al (1995) Errors at knee
magnetie resonance imaging: true or false? Br J Radiol
68(814):1045-1051
Martinoli C, Bianchi S, Spadola L et al (2000) Multimodality
imaging assessment of meniscal ossicle. Skeletal Radiol
29:481-484
Marty-Delfaut E, Beltran J (1999) Fat saturation: MRI tech-
niques and pitfalls. Radiographies 19:373-382
McGinty JB, Guess LF, Marvin RA (I977) Partial or total
meniscectomy: a comparative analysis. J Bone Joint Surg
59:763-766
Melki PS, Mulkern RV, Schwartz RB et al (1991) Comparing
the FAISE method with conventional dual-echo sequences.
J Magn Reson Imaging 1:319-326
Metcalf RW (1991) Arthroscopie meniscal surgery. In: McGinty
JB (ed) Operative arthroscopy. Raven, New York,
pp 227-236
Miller DB (1988) Arthroscopie meniscus repair. Am J Sports
Med 16:315-320
Mink JH, Levy T, Crues JV (1988) Tears of the anterior cruciate
ligament and menisci of the knee: MR imaging evaluation.
Radiology 167:769-774
Newman AP, Daniels AU, Burks RT (1993) Principles and deci-
sion making in meniscal surgery. Arthroscopy 9:33-51
Noble J, Hamblen DL (1975) The pathology of the degenerate
meniscus lesion. J Bone Joint Surg [Br] 57:180-186
Parizel PM, Dijkstra HAJ, Geenen PJ et al (1995) Low-field
151
versus high-field MR imaging of the knee: a comparison of
signal behavior and diagnostie performance. Eur J Radiol
19:132-138
Peterfy CG, Janzen DL, Tirman PFJ et al (1994) «Magie-angle»
phenomenon: a cause of increased signal in the normal
lateral meniscus on short-TE MR images of the knee. AJR
163:149-154
Poehling GG, Ruch DS, Chabon SJ (1990) The landscape of
meniscal injuries. Clin Sports Med 9:539-549
Potter HG, Rodeo SA, Wickiewiez TL et al (1996) MR
imaging of meniscal allografts: correlation with clinieal
and arthroscopie outcomes. Radiology 198:509-514
Priee CT, Allen WC (1978) Ligament repair in the knee with
preservation of the meniscus. J. Bone Joint Surg [Am]
60:61-65
Quinn SF, Brown TF (1991) Meniscal tears diagnosed with
MR imaging versus arthroscopy: how reliable a standard is
arthroscopy? Radiology 181:843-847
Quinn SF, Brown TR, Szumowski (1992) Menisci of the knee:
radial MR imaging correlated with arthroscopy in 259
patients. Radiology 185:577-580
Raunest J, Ob erle K, LoehnertJ et al (1991) The clinieal value of
magnetie resonance imaging in the evaluation of meniscal
disorders. J Bone Joint Surg [Am] 73:11-16
Reieher MA, Hartzman S, Duckwiler GR et al (1986) Meniscal
injuries: detection using MR imaging. Radiology
159:753-757
Riehmond JC, Sarno RC (1988) Posttrauma intracapsular
bone fragments; association with meniscal tear. AJR
150:159-160
Riel KA, Reinish M, Kersting-Sommerhoff B et al (1997) 0.2
Tesla magnetie resonance imaging of internaliesions of the
knee joint: a prospective arthroscopieally controlled clini-
cal study. Knee Surg Sports Traumatol Arthrosc 7:37-41
Rubenstein JD, Kim JK, Morava-Protzner I et al (1993) Effects
of collagen orientation on MR imaging characteristies of
bovine cartilage. Radiology 188:219-226
Rubin DA, Kneeland JB, Listerud J et al (1994) MR diagnosis of
meniscal tears of the knee: value of fast spin-echo vs. con-
ventional spin-echo pulse sequences. AJR 162:1131-1135
Rubin DA, Britton CA, Towers JD et al (1996) Are MR imaging
signs of menisco-capsular separation valid? Radiology
201:829-836
Ruff C, Weingardt JP, Russ PD et al (1998) MR imaging patterns
of displaced meniscus injuries of the knee. AJR 170:63-67
Ruwe PA, Wright J, Randall RL et al (1991) Can MR
imaging effectively replace diagnostie arthroscopy? Radiol-
ogy 181:335-339
Ryu KN, Kim IS, Kim EJ, Ahn JW et al (1998) MR imaging of
tears of discoid lateral menisci.AJR 171:963-967
Schnarkowski P, Tirman PFJ, Fuchigami KD et al (1995) Menis-
cal ossicle: radiographie and MR imaging findings. Radiol-
ogy 196:47-50
Seedholm BB (1979) Transmission of the load in the knee joint
with special reference to the role of the menisci. Part I.
Anatomy, analysis, and apparatus. Eng Med 8:207-219
Seedholm BB, Hargreaves DJ (1979) Transmission of the load
in the knee with special reference to the role of the menisci.
Part 11. Experimental results, discussion and conclusion.
Eng Med 8:220-228
Shakespeare D, Rigby H (1983) The bucket-handle tear of the
meniscus. J Bone Joint Surg [Br] 65B:383-387
Shankman S, Beltran J, Melamed E et al (1997) Anterior horn of
152
the lateral meniscus: another potential pitfall in MR imag-
ing of the knee. Radiology 204:181-184
Shellock FG, Stone K, Crues JV III (1998) Assessment of
the knee using an extremity MR imaging system: diagnos-
tie finding compared to arthroscopy (abstract). Radiology
209(P):192
Shogry MEC, Pope TL (1991) Vacuum phenomenon simulating
meniscal or cartilaginous injury of the knee at MR imag-
ing. Radiology 180:513-515
Silverman JM, Mink JH, Deutsch AL (1989) Discoid menisci
of the knee: MR imaging appearance. Radiology
173:351-354
Singson RD. Feldman F, Staron R et al (1991) MR imaging of
displaced bucket-handle tear of the medial meniscus. AJR
156:121-124
Smillie IS (1970a) Clinieal features of internal derangements
relative to the menisci. In: Injuries of the knee joint, 4th edn.
Livingstone, London, pp 70-97
Smillie IS (1970b) Surgieal pathology of the menisci. In:
Injuries of the knee joint, 4th edn. Livingstone, London,
pp 39-69
Smith DK, Totty WG (1990) The knee after partial meniscec-
tomy: MR imaging features. Radiology 176:141-144
Stoller DW (1989) Magnetie resonance imaging in orthopae-
dies and rheumatology. Lippincott, Philadelphia, p 99
Stoller DW, Martin C, Crues JV et al (1987) Meniscal
tears: pathologie correlation with MR imaging. Radiology
163:731-735
Stone RG (1979) Peripheral detachment of the menisci of the
knee. Orthop Clin North Am 10:643-657
Strand T, Engesaeter LB, Molster AO (1985) Meniscus repair in
knee ligament injuries.Acta Orthop Scand 56:130-132
Suarez-Almazor ME, Kaul P, Kendall CJ et al (1999) The cost-
effectiveness of magnetie resonance imaging for patients
with internal derangement of the knee. Int J Technol Assess
Health Care 15:392-405
Trattnig S, Huber M, Breitenseher MJ et al (1998) Imaging
artieular cartilage defects with 3D fat-suppressed echo
J. Beltran and S. Shankman
planar imaging: comparison with conventional3D fat-sup-
pressed gradient echo sequence and correlation with his-
tology. J Comput Assist Tomogr 22:8-14
Tuckman GA, Miller WJ, Remo JW et al (1994) Radial tears of
the menisci: MR findings.AJR 163:395-400
Turner DA, Rapoport MI, Erwin WD et al (1991) Truncation
artefact: a potential pitfall in MR imaging of the menisci of
the knee. Radiology 179:629-633
Tyson LL, Daughters TC Jr, Ryu RKN et al (1995) MR imaging
appearances of meniscal cysts. Skeletal Radiol 24:421-424
Vahey TN, Bennett HT, Arrington LG, Shelbourne KD, Ng J
(1990) MR imaging of the knee: pseudotear of the lateral
meniscus caused by the meniscofemoral ligament. AJR
154:11237-1239
Vahlensieck M, Sommer T, Textor J et al (1998) Indirect
MR arthrography: techniques and applications. Eur Radiol
8:232-235
Warren RF (1985) Arthroscopie meniscus repair. Arthroscopy
1:170-172
Weiss KL, Morehouse HT, Levy IM (1991) Sagittal MR imag-
ing of the knee: a low-signal band parallel to the posterior
cruciate ligament caused by a displaced bucket-handle tear.
AJR 156:117-119
Wennogle CF, Chang PS (1998) Diagnostie accuracy of a
low field strength extremity-only MR scanner in internal
derangement of the knee: correlation with arthroscopy and
1.5 Tesla MR imaging (abstract). Radiology 209(P):192
White LM, Schweitzer ME, Johnson WJ et al (1996) The role
of T2 weighted fast-spin-echo imaging in the diagnosis of
meniscal tears. J Magn Reson Imaging 6:874-877
Wood MR, Henkelman RM (1985) Truncation artefacts in mag-
netie resonance imaging. Magn Reson Med 2:517-526
Wright DH, DeSmet AA, Norris M (1995) Bucket-handle
tears of the medial and lateral menisci of the knee: value
of MR imaging in detecting displaced fragments. AJR
165:621-625
Yu JS, Cosgarea AJ, Kaeding CC, Wilson D (1997) Meniscal
flounce MR imaging. Radiology 203:513-515
10 The Cruciate and Collateral Ligaments
CHARLES P. Ho

CONTENTS nance (MR) imaging, the range and resolution of

imaging of the knee was relatively limited. Radi-
10.1 Introduction 153
ography remains an important and excellent initial
10.2 Anterior Cruciate Ligament 154
10.2.1 MR Imaging Technique 154 screening examination of the osseous structures
10.2.2 The Normal ACL 154 about the knee. However, its soft tissue contrast and
10.2.3 The Injured ACL 154 resolution are very limited, and radiographie find-
10.3 Posterior Cruciate Ligament 158 ings in ligamentous derangement generally are lack-
10.4 Cruciate Ganglia 160
ing or limited. Cortieal avulsion fragments at lig-
10.5 Medial Collateral Ligament 160
10.6 Lateral Collateral Ligament 160 ament attachments may be discerned as evidence
10.7 Conclusion 162 of complete ligament detachment. For example, the
References 163 anterior tibial eminence/spine may be avulsed with
the tibial attachment of the distal anterior cruciate
ligament (ACL). The osseous avulsion fragment or
heterotopic ossification along the superior medial
10.1 femoral condyle in Pellegrini-Stieda changes may be
Introduction seen in more chronic proximal tears of the medial
collateralligament complex. The smalllateral proxi-
The ligamentous structures about the knee provide mal tibial avulsion fragment of the detached distal
critical support and stability for knee function in ath- lateral capsular ligament, termed the Segond frac-
letic and occupational endeavors, as well as the activi- ture (GOLDMAN et al. 1998), has been an important
ties of daily living. The major cruciate and collateral finding on anteroposterior radiographs as evidence
ligaments as well as the secondary supporting capsu- of lateral capsule avulsion, but also because of the
lar and other supporting ligaments may be important very high association with complete te ars of the
causes of knee dysfunction both acutely and chron- ACL. However, these radiographie findings may be
ically when injured. Clinieal evaluation alone may subtle (DELZELL et al. 1996), and cortieal avulsions
be incomplete and difficult, particularly in the more constitute only a small subset of the ligament inju-
acute setting, when substantial edema and hemor- ries that occur about the knee. Arthrography, while
rhage, muscle spasm, or patient guarding may render accurate in evaluation of the menisci, may be of lim-
physical examination very difficult. Also, the full ited use in evaluation of the ligamentous structures
range of co-existing or associated injuries may not be whieh are extrasynovial or extracapsular. Ultrasound
evident on physieal examination but may be impor- is a very focused and directed examination with lim-
tant to the treating physician in recommending and ited penetration, and comprehensive examination of
planning optimal treatment for the patient as well as all of the ligamentous supporting structures about
in counseling the patient on long-term prognosis. the knee may be difficult.
Diagnostie imaging then becomes a valuable part With the evolution of MR imaging, comprehensive
of the comprehensive examination of the injured evaluation of the cruciate and collateral ligaments
knee. Prior to the development of magnetic reso- may now be performed (RUBIN et al. 1998) as part
of a standard complete MR imaging examination of
the knee. The extent/degree and precise location of
ligament injuries as well as other associated injuries
C.P. Ho, MD, PhD
National Orthopaedic Imaging Associates, Sand Hili Imaging
about the knee may be determined and reported
Center, 2882 Sand Hili Road, Suite 118, Menlo Park, CA 94025, accurately, to assist the treating physician in planning
USA management.

A.M. Davies et al. (eds.), Imaging of the Knee

© Springer-Verlag Berlin Heidelberg 2003
154 c. P. Ho

10.2 generate signal on MR images. Ligaments should be

Anterior Cruciate Ligament smooth, well-defined structures with little or no signal
on all types of MR images. The functional bands of the
The ACL provides important support against antero- ACL may be resolved separately because of the pres-
posterior translational as well as rotational forces on ence of intervening high-signal fat tissue and may
the knee. It originates at its proximal attachment along be evaluated individually on a routine and reliable
the posterosuperior lateral aspect of the intercondylar basis. The entire length and distal tibial attachment
notch and courses anteroinferiorly and slightly medi- of the ACL are best evaluated on the sagittal images
ally to the distal attachment at the anterior tibial emi- (Fig.1O.1a) described above. However, the proximal
nence. The ligament is intracapsular but extrasynovi- femoral attachment of the ACL is not demonstrated
al, surrounded by a synovial sleeve, and consists of two reliablyon the sagittal images because of partial volume
major bands separated by fat tissue (LEE et al. 1996). averaging of the proximal ACL with the cortieal margin
Compared with the posterolateral band, the anterome-
dial band originates more superiorly at the femoral
attachment and attaches more anteriorly and medial-
ly at the tibial eminence. Functionally, the posterolat-
eral band is stressedltaut first in extension, while the
anteromedial band is tightened first in flexion. The lig-
ament has been described as isometrie, maintaining
an overall straight appearance and course, although
the individual bands may tighten and relax separately.

10.2.1
MR Imaging Technique

The course of the ACL is slightly oblique relative to

the sagittal plane of the knee, being angled by a few
degrees off of the sagittal plane as seen in both the
transverse and the coronal plane of the knee. The sag-
ittal images of the knee may be optimized for evalua-
tion of the ACL as well as the remainder of the knee a
by angling the images slightly relative to the true sagit-
tal plane of the knee along the direction of the ACL,
planned on the basis of both transverse and coronal
scout images (Ho et al. 1999). The entire length of the
ACL may then be demonstrated reliably on a single
angled sagittal image. As this is only slightly angled
off of the true sagittal plane of the knee, these oblique
sagittal images may then be extended throughout the
remainder of the knee for full evaluation of the knee
in the sagittal plane, rather than being limited to a spe-
cialized separate sequence for evaluation of the ACL
in addition to the standard true sagittal images for the
remainder of the knee.

10.2.2
The Normal ACL
The dense organized collagen bundles of ligaments
with sparse cellular elements have very few water mol-
ecules and relatively mobile hydrogen proton nuclei to
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
proximal femoral attachment of the ACL (arrows) is best seen
on transverse images. Proton density (long TR, short TE)
turbo spin echo (TSE) images
The Cruciate and Collateral Ligaments 155

of the lateral femoral eondyle at the femoral attaeh-

ment. The proximal ACL and femoral attaehment are
best evaluated on the transverse images (Fig.lO.1b),
whieh show the ligament substanee in relatively short
diameter eross-seetion on sequential images through-
out its length and extending into the femoral attaeh-
ment (Ho et al.1999; ROYCHOWDHURY et al.1997). The
eoronal images provide another view ofboth the prox-
imal and distal ligament attaehments to supplement
and eorroborate findings in the other planes.

10.2.3
The Injured ACL

In aeute injury, high-signal edema and hemorrhage

distort the morphology and signal of the ligament.
Intrasubstanee/interstitial sprain results in promi-
nent thiekening of the ligament with poorly defined
inereased signal and often irregular and indistinet
margins (Fig.1O.2). The ligament may show undula-
tion and bowing. The sprain as weH as more foeal
partial tearing may involve primarily either of the
two funetional bands, with the posterolateral band
involved more in hyperextension injuries and the
anteromedial band involved more in flexion type
injuries (Fig. 10.3). Complete tears of the ACL may
reveal diserete defeets fiHed with intervening hemor-
rhage at MR imaging.
Fig.lO.2. Severe intrasubstance sprain of the ACL. Coronal
proton density image showing diffuse thickening with poorly
defined increased signal within the ligament, with indistinct
margins (arrow)
Fig.lO.3. Sprain involving primarily the posterolateral band of
the ACL. Sagittal proton density image showing an indistinct and
edematous posterolateral band (open arrow) while the antero-
medial band (arrow) remains weil defined with low signal
The precise loeation and extent of the ACL injury
may influenee the clinieal management, from nonop-
erative to operative proeedures, and should be evalu-
ated and reported precisely and aeeurately. Proximal
tears (Fig.10A) or detaehments may be amenable to
primary repair or reattaehment. Mid-substanee tears
(Fig.10.5) generaHy do not respond weH to repair
beeause of the relative hypovaseularity and stresses
in this region, and usuaHy require eomplete reeon-
struetion in the aetive individual. Distal ACL failure
may involve avulsion of the anterior tibial eminenee
with the distal ligament attaehment (Fig.10.6). If
nondisplaeed, these avulsions may heal in the ana-
tomie or near-anatomie position and may not require
operative intervention. However, elevation, rotation,
or fragmentation of the tibial eminenee with the ACL
may heal with a clinieaHy insufficientldeficient liga-
ment if not operatively redueed to a more anatomie
position. The tibial avulsions may at times be rel-
atively subtle on eonventional radiographs, whieh
often also underestimate the degree of elevation or
rotation of the avulsed fragment. MR imaging may
aeeurately reveal this extent of injury, as weH as the
extent of damage to the ligament substanee.
In ehronie ACL tears, the hemorrhage and edema
are no longer present to point to the ligament injury,
and evaluation depends on reeognition of abnormal
156
Fig.l0.4. Proximal eomplete tear of the ACL.
Sagittal T2-weighted image revealing a dis-
erete defeet in proximal third of the ligament,
filled with high-signal fluid and hemorrhage
(arrow)
Fig. 10.5. Mid-substanee eomplete tear of the
ACL (arrow). Sagittal proton density image
morphology of the ACL. The ligament may appear
ir regular with low signal scar obliterating the normal
fat plane between the functional bands, with undu-
lation and bowing. Scarring and obliteration of the
normal surrounding synovial tissues mayaiso be
C.P.Ho
Fig.lO.6. Distal avulsion of the ACL. Sagittal T2-weighted
image demonstrating a 1.5-em avulsion fragment of the ante-
rior tibial eminenee (arrow) at the ACL attaehment, with about
I-ern elevation of the fragment. Note also the minimally dis-
plaeed fraeture of the posterior tibial eminenee (open arrow)
and the posterior eapsule tearing (curved arrow)
seen, with possible scarring to the adjacent posteri-
or cruciate ligament. Chronic tears mayaiso result in
marked attenuation and resorption of the ligament,
with only small irregular proximal and distal rem-
nants or entire absence of the ACL on the images
(Fig. 10.7). It should be no ted, however, that failure to
identify the ACL on MR imaging may rarely be due to
congenital absence of the ligaments. Severe chronic
intrasubstance scarring and/or degeneration of the
ligament may result in the appearance of a diffusely
thickened ligament with amorphous and inhomoge-
neous appearing relatively increased signal that may
be difficult to differentiate from more acute intrasu-
bstance sprain or partial tearing. The MR imaging
findings are clearly abnormal in either case, but the
clinical setting is invaluable in interpreting the abnor-
mal findings.
A number of secondary MR imaging findings in
ACL tears have been described. These include accen-
tuated buckling or bowing of the posterior cruciate
ligament and anterior translation of the lateral tibial
plateau relative to the lateral femoral condyle with
uncovering of the lateral meniscus posterior horn.
Osseous contusion or impaction injury patterns may
also be characteristic (KAPLAN et al. 1992; Ho et al.
1999) and are best seen on fat suppression T2-weight-
ed or STIR images in the acute setting. These contu-
The Cruciate and Collateral Ligaments 157

Fig.l0.7. Chronic ACL tear. Sagittal proton density image

showing attenuated distal ACL remnant (arrow) bowed poste-
riorly in the intercondylar notch
Fig.l0.8. Characteristic osseous contusion pattern of the
anterolateral pivot shift mechanism of ACL tear. Sagittal fat
suppression inversion recovery (STIR) image revealing high-
signal bone contusion and edema (arrows) of the posterior
lateral tibial condyle and anterior lateral femoral condyle

sion patterns depend on the mechanism of the ACL

injury. In the instance of an anterior internal rotation
pivot shift of the tibia relative to the femur with ACL
tearing, the anterolaterally luxated tibia may impact
against the distal femur in the luxated position, with
impaction of the posterolateral tibial plateau/condyle
against the anterolateral femoral condyle. The knee
generally has reduced by the time the MR imaging
examination is performed, but the resulting osseous
contusion or impaction injuries (Fig 10.8) are very
characteristic of this mechanism of injury and dem-
onstrate dramaticaIly the amount of the anterolater-
al pivot shift at the time of the ACL tear. The impac-
tion injury of the anterolateral femoral condyle in
the region of the sulcus terminalis has also been
described as the deep sulcus sign on conventional
lateral radiographs as weIl as on MR imaging. In
hyperextension tearing of the ACL, matching osseous
contusions of the anterolateral femoral condyle and
tibial plateau and/or the anteromedial femoral con-
dyle and tibial plateau (Fig.l0.9) may be seen. The
secondary findings are not constant, and the ACL
itself should be evaluated directly and completely
on the images (HA et al. 1998; Ho et al. 1999). The
Fig.l0.9. Osseous contusion pattern of the hyperextension
secondary findings, when present, may be described mechanism of ACL tear. Sagittal STIR image demonstrating
as corroborative of the ACL tear (BRANDSER et al. matching high -signal bone contusions (arrows) of the anterior
1996). medial femoral condyle and anterior medial tibial condyle
 158
10.3
Posterior Cruciate Ligament
The posterior cruciate ligament (PCL) originates at
its proximal attachment at the medial femoral con-
dyle along the anterior superior medial aspect of the
intercondylar notch, and shows a smooth posterior
bowed course to its distal attachment along the pos-
terior tibial eminence in the typically imaged supine
extension position (Fig.lO.lO). Like the ACL, it con-
sists of two functional bands. The posteromedial
band is tightened/stressed first in extension, while the
anterolateral band is stressed first in flexion. Unlike
a b
c images
c. P. Ho
with the ACL, the bands of the PCL typically are
not distinguished as separate bands at MR imaging,
as no macroscopic intervening fat tissue and signal
are normally seen. Rather, the ligament is seen as a
homogeneously low/black signal smooth ligament.
Because of its relative thickness, the entire length as
well as the proximal and distal attachments of the
PCL generally may be seen on one or more adjacent
sagittal images even when the sagittal images are
angled slightly obliquely along the orientation of the
ACL as described above. The short diameter cross-
section of the PCL should also be evaluated proximal-
lyon the coronal images and distally on the transverse
Fig.l0.l0a-c. Intact peL. a Sagittal image showing the entire
length of the peL (arrow) from the proximal femoral to the
distal tibial attachment as a smooth well-defined ligament
with posterior bowing in the imaged extension position. b The
proximal portion of the peL (arrow) is best evaluated in short
diameter cross-section on the coronal image. c The distal por-
tion of the peL (arrow) is optimally seen in short diameter
cross-section on the transverse image. Proton density TSE
The Cruciate and Collateral Ligaments 159

images for assessment of the extent of intrasubstance

partial tearing or degeneration (Fig. 10. lOb, c).
As with the ACL, intrasubstance or interstitial
sprain in the acute setting results in thickening of the
ligament substance with poorly defined, increased-
signal hemorrhage and edema and irregular and pos-
sibly indistinct ligament margins (Figs.lO.ll, 10.12).
Complete tears with focal high-signal hemorrhage
and fluid between the torn ligament ends may occur,
typically in the mid to proximal portions, and should
be evaluated and reported precisely and accurately.
Distal PCL failure may be seen as avulsion fracture of
the posterior tibial eminen ce at the PCL attachment
(Fig. 10.6). These injuries, clinically indistinguishable
from PCL substance tears, may be relatively subtle on
conventional radiographs. MR imaging may be very
helpful and accurate in determining the size, commi-
nution, displacement/elevation, and rotation of the
avulsion fragments, as well as the extent of associ- a
ated ligament substance injury. All of this informa-
tion may be very important in determining a treat-
ment plan and should be reported accurately.
Posterior cruciate ligament tears are not as
common as ACL tears, and secondary findings may
not be as characteristic or well described. Osseous
contusion patterns, when present, again may be help-
ful in indicating the mechanism of injury. Direct ante-
rior blows such as in a dash-board injury may dem-

Fig.lO.12a, b. Sprain of the posteromedial band of the PCL.

a Sagittal image showing prominent intrasubstance edema
(arrow) of the mid to distal portions of the PCL. b Trans-
verse image of short diameter cross-section of the distal PCL
revealing sprain involving primarily the posteromedial band
(arrow); the anterolateral band white arrow) remains weil
defined with low signal

onstrate osseous contusion/impaction injury against

Fig.lO.ll. Severe intrasubstance sprain of the PCL. Sagittal
proton density image demonstrating extensive intrasubstance,
poorly defined high-signal edema and hemorrhage with irreg-
ular and indistinct margins (arrows)
the anterior proximal tibia in conjunction with the
PCL tear. Hyperextension injuries may show match-
ing osseous contusions of the anterior femoral con-
dyles and tibial plateaus similar to those seen in and
often in conjunction with ACL tears.
160
10.4
Cruciate Ganglia
Intra-articular ganglia in the knee usually derive
from cruciate ligaments, most commonly the ACL
(TYRRELL et al. 2000; KIM et al. 2001). They have
the typical MR appearance of soft tissue ganglia, i.e.,
they are homogeneously hyperintense on T2-weight-
ed images and slightly hypo- or isointense to skeletal
muscle onTl-weighted images. Cysts from the ACL
have a fusiform shape as they are intimately related
to the fibers of the ligaments.
10.5
Medial Collateral Ligament
The medial collateralligament (MCL) attaches prox-
imally at the superior aspect of the medial femoral
condyle or epicondyle, and extends inferiorly to its
distal attachment at the proximal medial tibia, gen-
erally about 5-6 cm below the medial joint line. The
ligament complex consists of superficial medial or
tibial collateral smooth and well-defined ligament,
with deep capsular components (including medial
meniscofemoral and meniscotibial coronary liga-
ments) separated by relatively high-signal fat tissue.
A posterior oblique ligament is also present. The
entire length of the MCL from the femoral to the
tibial attachment is best seen on coronal images.
However, the axial images are also very helpful in
demonstrating the short diameter cross-section of
the MCL on sequential images along its entire length,
and are vital for evaluating the extent of partial tears
or intrasubstance sprain.
Trauma with a valgus force component tending to
open the medial joint line may result in MCL injury.
Intrasubstance sprain or partial tearing is seen as lig-
ament thickening with poorly defined intrasubstance
as weIl as overlying soft tissue/subcutaneous edema
and hemorrhage, with irregular and indistinct mar-
gins. Severe intrasubstance tearing may result in elon-
gation and undulation or stretching (plastic defor-
mity) of the ligament without a discrete tear defect,
although the ligament may be completely functional-
ly incompetent. Discrete tears commonly may be seen
proximally at the femoral attachment. Distal tears or
detachments are less frequent and may be more dif-
ficult to evaluate, as the distal tibial attachment may
extend below the edge of the field of view in a high-
resolution small field of view knee examination. In
this instance, distal tears may be suspected on the
C.P.Ho
basis of undulation and proximal retraction of the
visualized portion of the MCL.
MCL sprain or te ars often do weIl with nonop-
erative management (Fig.l0.13). Even in complete
tears, the ligament may heal with scar tissue filling
the defects and bridging the tom segments. The scar
tissue is not biomechanically as strong or weIl orga-
nized as the original ligament tissue, and gene rally
results in a ligament that shows the expected lowl
black signal but appears thickened as evidence of the
previous or chronic ligament injury.
Occasionally, however, operative intervention may
be necessary, particularly when complete tears result
in trapping of a tom ligament end in the medial joint
line (Fig. 10.14), preventing healing of the ligament
in the anatomie position necessary to restore liga-
ment function. Also, complete tears may be accompa-
nied by associated displacement or extrusion of the
medial meniscus, and operative treatment again may
be needed. MR imaging is invaluable for demonstrat-
ing and reporting these situations, which may not be
clear at physical examination.
With tearing of the MCL, secondary medial sup-
porting structures may also be injured and should be
evaluated (SCHWElTZER et al. 1995). Sprain or tearing
of the medial retinaculum and tearing of the medial
patellofemoralligament may occur, particularly with
lateral patellar dislocation, and are best evaluated on
the transverse images. Posteromedial corner struc-
tures, including the posteromedial capsule, the distal
semimembranosus, and the proximal medial gastroc-
nemius, should also be evaluated on the transverse
and sagittal images for possible sprain or strain.
10.6
Lateral Collateral Ligament
The lateral collateralligament (LCL) or primarily the
fibulocollateral ligament attaches proximally at the
superolateral aspect of the lateral femoral condyle or
epicondyle, and extends obliquely posteriorly, inferi-
orly, and laterally to its conjoined attachment with
the distal biceps femoris tendon at the fibular head.
Because of this oblique course, the ligament is not
generally seen along its entire length on a single cor-
onal image (unless the knee is slightly flexed and
the coronal images have been angled to the axis of
the tibia rather than the femur, or unless the tibia is
anterolaterally luxated relative to the femur, such as
in an ACL-deficient knee). The LCL must be followed
on sequential coronal images from anterior superior
 The Cruciate and Collateral Ligaments
b
Fig.lO.13a, b. Healing of an MCL tear. a Severe tearing of the
MCL with prominent high-signal edema and poorly defined
margins (arrow). b Repeat examination 6 weeks later shows
healing with thiek, well-defined, low-signal scar (arrow). Coro-
nal proton density images
to posterior inferior along its oblique course. Trans-
verse images are vital for evaluating the ligament in
short diameter cross-section for the extent of sprain
or partial tearing, and to follow the ligament continu-
ity on sequential transverse images from the lateral
femoral condyle proximally to the fibular head dis-
tally.
With varus force such as a direct medial blow, the
LCL may be injured. Intrasubstance sprain or partial
tearing pro duces a thickened and often undulating
or stretched ligament, with inhomogeneous high-sig-
nal hemorrhage or edema within and about the liga-
ment, which can be evaluated on the coronal images
161
Fig.lO.14. Coronal proton density image showing complete tear
of the MCL with trapping of the proximal superficial MCL end
(arrow) at the medial joint line deep to the tom meniscofem-
oral coronary ligament (open arrow). This trapping prevents
adequate anatomie healing and requires operative treatment
and confirmed on the transverse images. Complete
te ars may result in discrete defects with more focal
high-signal fluid or hemorrhage between the torn
ligament ends, often with the discrete tear in the
proximal or mid portions. Distal LCL failure may
involve discrete ligament tears or avulsion fractures
of the fibular head at the conjoined ligament/biceps
femoris tendon attachment (Figs. 10.15, 10.16). The
precise location of the tear, the size of the defect and
the location of the proximal and distal ligament ends
should be evaluated and reported accurately to assist
the treating physician in planning the incision site
and procedure when operative intervention is indi-
cated.
With LCL injury, other lateral supporting struc-
tures may also be injured and should be evaluated
carefully. The lateral capsular ligament may be torn
or avulsed (such as the Segond fracture) and is best
evaluated on coronal and transverse images. Tearing
of the lateral retinaculum should be evaluated on
transverse images. Associated strain/tearing or avul-
sion of the distal biceps femoris tendon should also
be evaluated on coronal and transverse images. Pos-
terolateral corner structures mayaiso be injured
(MILLER et al. 1997), as is often the case with
the ACL-associated injuries discussed above, and
should be evaluated on transverse and sagittal images
(Fig. 10.17). Osseous contusions/impaction injuries
on the medial aspect of the knee may occur in con-
162 C.P.Ho

Fig.l0.15. Coronal proton density image showing subacute a

complete distal avulsion of the LCL from the fibular head,
with a small avulsed osseous fragment (arrow) proximally
retracted with ab out a I-ern gap. Intervening poorly defined
scar is demonstrated

Fig.l0.16a, b. Complete distal avulsion tear of the LCL and

biceps femoris conjoined insertion on the fibular head. a Coro-
nal image showing a 1.5-cm defect between the proximally
retracted ligament/tendon ends (arrow) and the fibular head.
b Transverse image at the superior aspect of the fibular head
demonstrating the irregularity and a small cortical defect
(arrow) remaining at the fibular head. Extensive surrounding
high-signal hemorrhage and disruption of overlying superfi-
dal fasda are demonstrated. T2-weighted images b

junction with the varus force on the lateral struc-

tures, should be evaluated on coronal and transverse
images, and are best seen on fat suppression STIR or
T2-weighted images in the acute setting.

10.7
Conclusion

The cruciate and collateral ligaments as well as

Fig.l0.17. Sagittal proton density image revealing disruption
of posterolateral corner stabilizers, including complete tear of
the popliteus myotendinous junction with an undulating and
proximally retracted tendon end (arrow)
the other supporting structures about the knee are
vital to knee stability and function. Imaging may be
extremely helpful to the treating physician in evalu-
ating the fuH spectrum of cruciate and coIlaterallig-
ament injury as weIl as associated injuries in both
acute and chronic settings, when physical examina-
tion may not be sufficient. MR imaging has revolu-
The Cruciate and Collateral Ligaments
tionized the capabilities and role of diagnostic imag-
ing in the evaluation of the ligamentous structures
about the knee, as these ligamentous structures may
now be directly imaged and accurately and precisely
evaluated. Comprehensive examination of the liga-
mentous structures as weH as the remainder of the
knee at MR imaging is best achieved by using images
in aH three of the orthogonal planes of the knee, as
each plane is important and best for seeing specific
portions of these structures.
References
Brandser EA, Riley MA, Berbaum KS et al (1996) MR imaging
of anterior cruciate ligament injury: independent value of
primary and secondary signs. AJR 167:121
Delzell PB, Schils JP, Recht MP (1996) Subtle fractures ab out
the knee. Innocuous-appearing yet indicative of significant
internal derangement. AJR 167:699
Goldman AB, Pavlov H, Rubenstein D (1998) The Segond frac-
ture of the proximal tibia: a small avulsion that reflects
major ligamentous damage. AJR 151: 1163-1167
Ha TPT, Li KCP, Beaulieu CF et al (1998) Anterior cruciate liga-
ment injury: fast spin-echo MR imaging with arthroscopic
correlation in 217 examinations.AJR 170:1215
163
Ho CP, Marks PH, Steadman JR (1999) MR imaging of knee
anterior cruciate ligament and associated injuries in skiers.
MRI Clin North Am 7: 117-l30
Kaplan PA, Walker CW, Kilcoyne RF et al (1992) Occult frac-
ture patterns of the knee associated with anterior cruciate
ligament tears: assessment with MR imaging. Radiology
183:835-838
Kim MG, Kim BH, Choi JA et al. (2001) Intra-articular ganglion
cysts of the knee. Eur Radiolll :834-840
Lee SH, Petersilge CA, Trudell DJ et al (1996) Extrasynovial
spaces of the cruciate ligaments: anatomy, MR imaging and
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
imaging. AJR 169:1641
Roychowdhury S, Fitzgerald SW, Sonin AH et al (1997) Using
MR imaging to diagnose partial tears of the anterior cruci-
ate ligament: value ofaxial images.AJR 168:1487
Rubin DA, Kettering JM, Towers JD et al (1998) MR imaging
of knees having isolated and combined ligament injuries.
AJR 170:1207
Schweitzer ME, Tran D,DeelyDM et al (1995) Medial collateral
ligament injuries: evaluation of multiple signs, prevalence
and location of associated bone bruises, and assessment
with MR imaging. Radiology 194:825-829
Tyrrell PNM, Cassar-Pullicino VN, McCall IW (2000) Intra-
articular ganglion cysts of the cruciate ligaments. Eur
Radioll0:1233-1238
11 The Postoperative Knee 1:
Menisci, Cruciate Ligaments, Cartilage
VICTOR N. CASSAR-PULLICINO and s. N. J. ROBERTS
CONTENTS
11.1 Introduction 165
11.2 Surgical Overview 165
11.2.1 Instability 165
11.2.2 Meniscal Tears 170
11.2.3 Chondral and Osteochondral Damage 171
11.3 Imaging Perspective 172
11.3.1 Cruciate Ligament Reconstruction 172
11.3.2 The Postoperative Meniscus 183
11.3.3 Articular Cartilage Assessment 188
11.3.4 Extra-articular Structures 193
11.4 Conclusion 195
References 195
11.1
Introduction
Accurate interpretation of the postoperative status
of the knee cannot be achieved without a working
knowledge of the utilised surgical procedures and
techniques, as weH as familiarity with the expected
imaging appearances of the structures that have been
repaired, resected or replaced. A thorough under-
standing of these two areas is therefore a prerequi-
site to the identification and correct interpretation of
recurrent or new problems in the postoperative knee.
This chapter will focus primarily on the postopera-
tive assessment of the knee following surgery to the
menisci, cruciate ligaments and articular cartilage.
V. N. CASSAR-PULLlCINO, MD
Consultant Radiologist, Department of Diagnostic Imaging,
The Robert Jones and Agnes Hunt Orthopaedic and District
Hospital, Oswestry, Shropshire, SYlO 7AG, UK
S. N. J. ROBERTS, MA, FRCS (Orth)
Consultant Orthopaedic and Sports Injury Surgeon, The
Robert Jones and Agnes Hunt Orthopaedic and District Hos-
pital, Oswestry, Shropshire, SY10 7AG, UK
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
11.2
SurgicalOverview
From an orthopaedic surgeon's viewpoint, the clin-
ical presentation of knee problems may be divided
into three categories: pain, too much movement
and too Httle movement. The commonest intra-artic-
ular causes of pain after an acute injury has settled
are chondral and osteochondral damage along with
damage to the menisci. Too much (or the wrong kind
of) movement is caused by an inadequacy of either
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-
ment" may either be an ever-present stiffness in the
knee caused by either intra-articular or extra-articu-
lar pathology, or an intermittent catching or locking
which is likely to be caused either by pateHofemoral
subluxation or by a mobile intra-articular fragment
such as a loose body or a meniscal tear.
11.2.1
Instability
The knee is stabilised by four major ligaments: the
medial and lateral collaterals, and the anterior and
posterior cruciates. There are important secondary
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
ligaments often heal weH without surgery, and the
functional outcome is better without than with sur-
gical repair or reconstruction. This has been shown
most clearly in the case of isolated tears of the medial
coHateralligament. The intra-articular cruciates heal
less weH, but undoubtedly a low-energy"peel off" of
the ACL's femoral origin may leave a viable stump
166
which is able to re-attach itself to an area adjacent to
its anatomical origin or to the posterior cruciate liga-
ment (PCL), allowing it to provide a degree of func-
tion in anteroposterior stabilisation.
Partly for this reason, but also because of the
varying capacity for individuals to rehabilitate their
dynamic stabilisers with proprioceptive and neuro-
muscular training, not all tom ACLs will lead to sig-
nificant functional deficit. A number of attempts have
been made to devise scoring systems in an attempt to
predict which individuals will manage well with reha-
bilitation, and who will be unable to cope with the level
of activity that they demand without ligament recon-
struction. However, universal agreement on these sys-
tems has not been achieved. They have been based on
the amount and level of activity demanded, but there is
a very poor correlation between the degree of antero-
posterior laxity and the likelihood of reconstruction
being necessary. In the United Kingdom at present, it
would appear reasonable to suggest that reconstruc-
tion of the ACL in most, if not all patients should only
follow failure of an adequate attempt at non-operative
management. The exception to this would be if there
is a bony avulsion, since in this situation near anatom-
ical repair as opposed to reconstruction can be per-
formed.
If adequate function cannot be achieved in the
knee following rehabilitation, surgical reconstruction
using a graft must be performed. The indication for
surgical reconstruction of the ACL is therefore insta-
bility. If there is significant pain aside from giving
way episodes, then an alternative diagnosis should
be sought, i.e. meniscal or chondral pathology. Con-
versely, radiological evidence of degenerative change
or chronological age is not, of itself, a contra-indica-
tion to operative stabilisation of an unstable knee if it
is the instability which is the predominant complaint.
V. N. Cassar-Pullicino and S. N. J. Roberts
When the decision to operate has been made, the
surgeon has a number of choices to make regarding
the type of graft, the positioning of the graft and its
fixation to bone at either end.
11.2.1.1
Choice of Graft
Various prosthetic grafts that were popular in the
1980s have very largely been abandoned, particularly
in young patients, as the failure rate in the medium
term was unacceptably high. They may still have a
role in the older low-demand patient. Allografts, such
as patellar tendon or Achilles tendon, have the advan-
tage of not causing any donor site morbidity, but have
been shown to incorporate more slowly. They there-
fore either delay an individual's return to sport or fail
if a slower rehabilitation is not adhered to.
The common choices of autograft are the patellar
tendon (BPTB), hamstring STG tendons (semi-ten-
dinosus and gracilis) and,less commonly, the quad-
riceps tendon. The central one-third of the patellar
tendon is most commonly used (although the medial
third has also been recommended). This produces a
tendon approximately 10 mm in diameter and allows
the surgeon to harvest a bone block from both patella
and tibial tuberosity in continuity with this, making
an overall BPTB graft 10 or 11 cm in length (Fig.11.1).
The patellar tendon is broader proximally than dis-
tally, so if a harvest is taken parallel with the ten-
don's fibres, the bone block proximallywill be slightly
larger than that distally, often producing a lO-mm-
diameter cylindrical graft from the patella and a
9-mm bone block from the tibia. This is convenient
surgically since it allows the graft to be inverted and
the smaller block passed through the lO-mm-diam-
eter tibial tunnel into the 9-mm-diameter femoral
Fig. 11.1. Per-operative
BPTB graft showing the
tendon attached to bone
blocks at either end
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
tunnel more easily. It also recreates to some extent
the anatomy of the slightly convergent natural ACL,
which has a broader tibial than femoral footprint.
During the operation, a small amount of bone graft
is commonly taken from the tibia to fill the patellar
donor site in an attempt to minimise postoperative
anterior knee pain.
Hamstring tendons are harvested through a short
incision over their insertion into the tibia. The sar-
torius fascia is split and elevated, revealing the graci-
lis and semi-tendinosus tendons. Both the STG ten-
dons are taken in their entirety using a tendon strip-
per. They are then folded in half, giving a four-strand
graft with an ultimate strength of over 4,000 N,
approximately twice that of the torn cruciate which
it is used to replace. Despite being 30% stronger, the
diameter of this graft is typically rather smaller than
the patellar tendon. This is because it has a more
circular cross-section than the rather Hat patellar
tendon, and therefore contains more collagen. Reha-
bilitation allows recovery of function of the harvest-
ed tendon within a year. Imaging reveals formation of
a neo-tendon along the line of the harvested tendon,
but with a rather more proximal insertion. In the case
of the patellar tendon, repair of the defect is so good
that revision reconstructions have been performed
re-harvesting the same central third as was used for
the primary surgery. This is not currently recom-
mended owing to the inferior quality of the scar
tissue.
Fig. 11.2a, b. Anterior cruciate ligament reconstruction using the "over the top" technique
167
The graft is chosen according to the surgeon's
training and preference, there being very little differ-
ence in published outcomes. There is probably more
morbidity associated with a patellar tendon graft (in
particular anterior knee pain), but the advantage of
bone to bone fixation may outweigh this disadvan-
tage in selected patients.
11.2.1.2
Positioning of the Graft
The optimum position of the tibial attachment site of
the neo-ACL is in the centre of the tibial plateau, and
a tunnel is drilled with a jig from the anteromedial
tibial cortex, entering the knee joint at this point. A
blind-ended femoral tunnel is drilled either through
the tibial tunnel or using the medial arthroscopy
portal. The starting point for this femoral tunnel is
critical to the success of the procedure, and the most
common surgical error contributing to its failure.
The entry point of the tunnel needs to be as far poste-
rior as possible in the roof of the intercondylar notch,
allowing a minimal 2 mm of femoral cortex behind
it to prevent the tunnel blowing out into the popli-
teal fossa. The tunnel cannot be too far back without
"blowing out", and some surgeons even pass the graft
right "over the top" - outside the femur - without
drilling a tunnel at all. This accepts a position which
is a little too far back, but ensures that the error of
placing it too far forward cannot be made (Fig.l1.2).
 168
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
from the edges of the intercondylar notch) is con-
troversial. In cases of long-standing ACL deficiency,
there is often osteophytic encroachment of the space,
leaving inadequate room for graft positioning. This
is not the case in early reconstructions, but surgeons
usually find it helpful to remove a little bone from
the medial border of the lateral femoral condyle to
ensure that there is no graft impingement, and also
to aid visualisation of the femoral tunnel position-
ing. The rate of ACL rupture has been correlated with
the measured notch width (probably as a result of
the correlation between a congenitally narrow notch
and a congenitally small and weak ACL), and some
patients will require a degree of notchplasty to allow
the passage of an adequate graft.
11.2.1.3
Fixation Devices
After tensioning the BPTB graft, it is typically fixed
using an interference screwwhich may be either metal-
lic or bio-absorbable polymer - usually polylactic acid
(Fig. 11.3). These biodegradable screws do not create
any artefact, allowing optimal assessment of the graft,
its tunnels and fixation status. As the bio-absorbable
screws are not easily visible radiographically, they are
best assessed by computed tomography (CT) and mag-
netic resonance imaging (MRI). They have been shown
to be safe and effective for fixation of bone blocks
during ACL reconstruction, with no reported adverse
V. N. Cassar-Pullicino and S. N. J. Roberts
clinical responses during the degradation process
(LAJTAI et al. 1999; WARDEN et al. 1999). The screw is
passed alongside the bone block, filling the tunnel and
compressing the block against the side wall (Fig.ll.4).
There are a number of alternative fixation tech-
niques more commonly employed when a hamstring
STG graft is used. If a hamstring tendon is used, it
is placed in exactly the same mann er as described
above, but a number of techniques are used for fixa-
tion since many surgeons feel that interference screw
fixation is less satisfactory in the absence of a block
of bone attached to either end of the graft.
Pre-operative imaging in these cases excludes an
anatomical abnormality and secondary pathology. A
postoperative AP and lateral radiographs of the knee
confirms a satisfactory position of the graft and its
fixation (Fig. 11.5).
11.2.1.4
Posterior Cruciate Ligament
The PCL is three tim es stronger than the ACL and
perhaps 8-10 tim es less commonly ruptured. There
is even more debate as to which PCL ruptures require
reconstruction than is the case for the ACL, but it is
undoubtedly true that some patients with a truly iso-
lated PCL rupture manage to continue with little or
no functional deficit even in professional sport. It has
increasingly been recognised that combined injuries,
particularly of the PCL with the posterolateral corner
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
these cases.
Broadly, the surgical technique for PCL reconstruc-
tion is similar to that for ACL reconstruction, although
the tibial tunnel needs to enter the knee joint in the
popliteal fossa, and the popliteal artery is at signifi-
cant risk from this. The tibial tunnel is drilled directly
towards the popliteal artery and exits the tibia within
a few millimetres of it. If this technique is used, the
graft needs to be passed through the tibial tunnel and
forwards over the posterior tibial plateau - a right
angle known as the "killer curve" - to allow it to head
forward across the knee joint to its insertion into the
medial femoral condyle. In order to avoid this acute
angulation, it has been suggested that an "in -lay" tech-
nique should be used which involves a surgical expo-
 170
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
on the posterior tibial cortex. This reduces the angle
through which the graft must bend to enter the knee
joint (BERG FELD et al. 2001). There is debate as to the
optimum location of the femoral insertion into the
medial femoral condyle, and considerable interest in
the use of two femoral tunnels, with part of the graft
passed into each, in an attempt to recreate the mul-
tiple bundles of the natural PCL.
11.2.2
Meniscal Tears
The "C" -shaped fibrocartilaginous meniscus converts
a nearly Hat tibial plateau into something of a socket
to accommodate the convex femoral condyle. Com-
pressive forces are converted into a "hoop-stress" so
that most of the resisted force in the meniscus is cir-
cumferential rather than radial. Most of the collagen
in the meniscus runs circumferentially rather than
radially, confirming this, and it is for this reason that
longitudinal splits of the meniscal body such as those
which produce bucket-handle tears may be repaired
successfully. Radial tears, which inevitably cut across
most of the stress-bearing collagen fibres, do not
have a high rate of healing. In the adult or even ado-
V. N. Cassar-Pullicino and S. N. J. Roberts
lescent human meniscus only the peripheral third of
the meniscus has a blood supply and it is therefore
only tears in this region which are usually considered
to be repairable. The vascularity of either side of the
tear can be assessed arthroscopically without tour-
niquet (Fig. 11.6), and typically tears with a blood
supply on at least one side (described as "red on
white") have a high rate ofhealing (over 85%).
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
in (with a knot on the intra-articular surface of the
meniscus) or passing from the inside out through
curved needles with sutures tied over the capsule.
These latter inside-out techniques require small skin
incisions to enable safe knot placement, exduding
neurovascular entrapment in the suture. More recent-
ly"all inside" techniques have been developed using
bio-absorbable darts, arrows (barbed devices passed
across the tear to hold the edges together), staples
and cannulated screws. These avoid both the skin
incision and the risk of damage to neurovascular
structures and appear to have success rates similar to
the more invasive techniques.
It has been noted that the highest success rates
have been achieved with meniscal repair in conjunc-
tion with cruciate ligament reconstruction, and it has
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
Red
on white
been suggested that this is because of the haemar-
throsis which occurs postoperatively. This may pro-
vide healing factors to the poorlyvascularised healing
zone. Some surgeons have also attempted to improve
the local environment by inserting and suturing a
fibrin clot across the repair, but this is now not com-
monlyused.
When meniscal repair is impossible, the arthros-
copist needs to carry out a partial meniscectomy.
Loss in function of the meniscus varies with the
amount of meniscus removed, and so if repair is not
considered suitable, or proves unsuccessful, the sur-
geon tries to leave as much functional meniscus as
possible. There is a correlation between the amount
of meniscus removed and the risk of osteoarthritis
in the medium term. The amount removed is usually
recorded as a proportion of the radial width of the
meniscus.
11.2.3
Chondral and Osteochondral Damage
Articular cartilage has no blood supply, no nerve
supply, no lymphatic drainage and a very low meta-
bolie rate. It has little or no capacity to regenerate.
Symptomatic defects in the articular surface are
therefore quite difficult to treat and a number of tech-
niques have been used.
Cartilage repair procedures have been utilised with
varying degrees of success over the years. The pro-
171
White
on white Fig. 11.6. A diagram out-
lining the blood supply of
the meniscus and the "red"
and "white" zones
cedures have included abrasion arthroplasty, autol-
ogous chondrocyte transplant, micro-fracture tech-
nique, articular cartilage transplant, fresh osteo-
chondral allografts and transfer of osteochondral
autograft plugs. Penetration of the subchondral bone
(sometimes known as the Pridie procedure) has been
used for over 50 years in an attempt to allow granu-
lation tissue, and subsequently a fibrocartilaginous
scar, to form in the chondral defect. This may be done
either by drilling or by the very similar micro-frac-
ture technique, which involves driving a pick through
the subchondral bone in an attempt to produce a
tunnel with a rough as opposed to a smooth margin,
allowing better clot adherence. No attempt is made to
reconstruct the hyaline cartilage with this technique,
and it is known that the fibrocartilaginous scar which
is produced is not as durable as the articular cartilage
it replaces, and that it lacks cohesive attachment to
the underlying bone.
Osteochondral autografts (OATS) became popular
in the early 1990s as a me ans of filling symptomatic
defects. The technique involves harvesting a cylindri-
cal plug from a supposedly less important part of the
knee's articular surface. A plug is taken perpendicu-
lar to the articular surface including articular carti-
lage, subchondral bone and a length of subchondral
cancellous bone. Various manufacturers now pro-
duce sets of instruments which allow for plugs of an
appropriate length to be taken, varying in diameter
from 4 mm upwards. A slightly under-sized cylinder
is drilled in the recipient site and the graft is press-
172
fitted. If multiple plugs are used, they are combined
alongside each other, making a "mosaic" (hence the
term "mosaicplasty"), and careful selection and tes-
sellation of multiple sizes may allow the triangular
defects between the plugs to be minimised. This tech-
nique mayaIso be employed using larger allograft
donors (Fig. 11.7).
Since the mid 1980s, autologous chondrocyte cul-
ture techniques have been developed whereby tissue
engineering hyaline cartilage may be introduced into a
defect. This was pioneered in Sweden and is known as
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
the lateral supracondylar ridge. This is minced, digest-
ed and cultured in the laboratory for approximately 3
weeks before the second stage procedure is performed.
The second procedure involves an open arthrotomy
and, after debridement of the defect margins, a peri-
osteal patch usually taken from the anterior tibia is
shaped and sutured over the defect to the surround-
ing cartilage like the skin over ablister. A watertight
seal is attempted and often supplemented using fibrin
glue before the suspension of cultured chondrocyte is
injected under the patch. Excellent long-term results
have been reported in a cohort up to 10 years, but there
are few results from randomised trials.
Fig.l1.7. Coronal Tl-weighted image of the patella following
mosaicplasty
V. N. Cassar-Pullicino and S. N. J. Roberts
Considerable efforts have been made to produce
a non-invasive outcome assessment of the quality
and quantity of regenerated hyaline cartilage in the
defect, particularly using MRI sequences. Although
little is known of the structural and biological organ-
isation of the repair tissue, recent biopsy studies
indicate that within 12 months ACI successfully pro-
duces replacement cartilage tissue consisting of a
deeper hyaline cartilage-like zone, with an upper
zone which is more fibrocartilaginous in nature
(RICHARDSON et al. 1999).
11.3
Imaging Perspective
Although the imaging options do include convention-
al radiography and CT, the assessment relies heavily
on MRI. In view of this it is crucial that one under-
stands the biological processes that take place under
normal and pathological conditions in the three areas
identified: cruciate ligament reconstruction, menis-
cal resection or repair and surgery to the articular
cartilage. Although the ideal objective is to use MRI
as the method of choice for assessing the problem-
atic postoperative knee in preference to arthroscopy,
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
Repair of intra-substance tears of the cruciate liga-
ments has consistently shown unsatisfactory results,
with a tendency for repaired ACL tears to heal inade-
quately or to re-tear. With the exception ofbony avul-
sion of the cruciate ligaments, which is treated by sur-
gical fixation, the mainstay for surgical treatment for
cruciate ligament tears is nowadays ligament recon-
struction. Irrespective of the graft material that is
used, three different maturation phases are recog-
nised by MRI in the successful outcome of a neo-
cruciate ligament - peri-ligamentous proliferation,
intraligamentous proliferation and definitive healing
with incorporation (BELLELLI et al. 1999). Initially
 The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 173

the graft material merely acts as a scaffold, wh ich
allows it to become invaded by synovium, resulting in
neo-vascularisation and eventually neo-ligamentisa-
tion (Fig. 11.8). In stage 1 (1-3 months postopera-
tively), proliferating soft tissue is seen around the
graft, which retains its typical tendinous low signal
MR characteristics. In stage 2 (3-9 months postop-
eratively), the graft becomes progressively hyperin-
tense, while in stage 3 (12 months postoperatively),
complete ligamentisation usually occurs (Fig. 11.9).
In PCL reconstructions the ligamentisation process
as identified by MRI is similar, but it takes much
longer, requiring an average of 24 months. MR stud-
ies of tunnel placement in PCL reconstruction have
shown that improper femoral tunnel location and
clinical signs of instability are statistically significant.
The tibial tunnel is usually well placed, but proper
location of the femoral tunnel seems to be more criti-
cal owing to the lack of specific anatomicallandmarks
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
studies, although interesting, are not required in the
successfully trans plan ted ligament, and should be
reserved for instances of suspected graft impinge-
ment where the internal changes within the liga-
ment are seen to enhance. There is no requirement
for routine MRI of the reconstructed cruciate liga-
ment as elinical assessment with conventional radio-
graphs is adequate. When the postoperative course
or surgical outcome is deemed unsatisfactory, either
by the patient or by the surgeon, MRI is required
(Figs. 11.1 0, 11.1l). There is often a low correlation
between the surgeon's evaluation and the patient's
perception of knee stability and function.
Knee stiffness and residual knee instability are
the two most common indications for postoperative
imaging in the early postoperative period, usually
within the first 6 months and uncommonly up to 12
months. Knee stiffness, which is usually heralded by
the failure to achieve full knee extension, requires the
exelusion of roof impingement of the graft, arthrofi-
brosis and the "cyelops" lesion. Residual instability in
the first 6 months is usually the result of poor surgical
technique, a failure of graft incorporation or a tear
of the graft, which can result from aggressive reha-
bilitation. Failure which occurs more than 1 year after
surgery is usually due to trauma, which occurs after
resumption of full activity. Traumatic re-injury has
been reported in about 5%-10% of the athletic popu-
1ation (JOHNSON et al. 1996). These patients usually
present with a new haemarthrosis following an injury
and elinically have an increased laxity. It needs to
be stressed, however, that during the first year after
reconstruction, the graft strength and stiffness are
30%-50% of normal and excessive loads can lead
to plastic deformation and elongation (CLANCY and
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
V. N. Cassar-Pullicino and S. N. J. Roberts
SMITH 1991; DREZ etal.I99l}.Insuchinstances,apart
from the assessment of the cruciate graft integrity, a
further search needs to be made to exelude associ-
ated meniscal, chondral and osteochondral injury. In
cases where there is elearly elinical evidence of resid-
ual instability, imaging will be required to assess
the postoperative bony and soft tissue anatomy in
preparation for arevision of a failed cruciate liga-
ment reconstruction. Although MRI is on ce again
the mainstay of this investigation, CT mayaiso be
required in the full assessment of the bone contours
and tunnels.
11.3.1.1
Imaging Technique
Conventional radiographs of the postoperative knee
are obtained before MR assessment to provide infor-
mation concerning the location of the femoral and
tibial tunnels, the method of fixation, the type of graft
material used, exelusion of radi opaque loose bodies
and localisation of the patella. The mainstay of the MR
investigation is the acquisition of a full assessment
of the knee to inelude the articular surfaces, menisci
and other ligaments, along with detailed assessment
of the neo-ligament. Conventional spin echo or fast
spin echo techniques are commonly used to evaluate
the knee postoperatively. Gradient echo images are
not ideally suited owing to the sensitivity to metal
artefacts, which are quite common in this type of
operation. Sagittal proton density-weighted (TR-TE,
2600/16) and T2-weighted turbo spin echo (TSE,
2600/98) images, along with coronal Tl-weighted spin
echo sequence (500/12), coronal fat -suppressed proton
density-weighted (TRTE 2600/16) and coronal T2-
 The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
b ative infection and abscess
weighted turbo spin echo (TSE 2600-98) sequences
are obtained, along with an axial T2-weighted fast
spin echo sequence. Although probably not required
routinely, oblique coronal and sagittal images can also
be obtained along the course of the graft using one of
the previous images as the scout, which allow visuali-
sation of the entire graft in its intra-articular portion
on a single image (Fig. 11.12).
The process of successful biological incorporation
of autografts, allografts and synthetie grafts includes
graft necrosis, revascularisation, cellular repopula-
tion, collagen deposition and matrix remodelling
(CORSETTI and JACKSON 1996; ARNOCZKY et al. 1994).
Assessment of this "ligamentisation" process can in
part be done using intravenous gadolinium-DTPA to
assess the degree of enhancement of the graft and
the peri-ligamentous tissues. Gadolinium-DTPA is
175
Fig. 11.11. Tl- (a) and
T2-weighted (b) sagittal
sequences showing the rare
complication of postoper-
formation
not routinely used, however, and is probably of lim-
ited clinical value except in cases of suspected graft
impingement. Recently, contrast-enhanced sequen-
tial MRI studies, performed over a 2-year period in
successful ACL reconstructions, have shown a 9%
quantitative enhancement 2 weeks postoperatively,
an average 50% enhancement between 12 and 52
weeks postoperativelY and an average 65% enhance-
ment at 1 year, with a reduction to 25% enhancement
after 76 weeks (VOGL et al. 2001). Low-field "niehe"
magnets and more recently "open" magnets allow
the dynamic assessment of the knee and intra-artieu-
lar structures during joint function. Where available,
these can add a further dimension to the status of the
neo-ligament and in particular, help establish graft
failure or impingement. Although MR arthrography
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
the neo-ligament optimally, it is in the assessment of
the meniscus and articular cartilage that this tech-
nique is of most use.
11.3.1.2
Successful Neo-ligament
Conventional radiographs in the immediate postop-
erative period after ACL reconstruction confirm the
near-anatomical placement of the tibial and femoral
tunnels, the correct placement and orientation of the
fixation devices utilised and the absence of significant
bony debris within the joint, and identify abnormali-
ties related to the harvest site with particular refer-
ence to the patella. Ideal femoral tunnel placement is
as posterior in the notch as possible without violation
of the posterior cortical wall (blow-out). Placement
of the tibial tunnel is now known to have a profound
effect on the postoperative result of ACL reconstruc-
tion and ideally it should be positioned in the postero-
medial portion of the original footprint of the ACL.
The tunnel should be seen as parallel, but also posteri-
or to Blumensaat's line when the knee is in full exten-
sion, so as to eliminate the potential for graft impinge-
ment. Conventional radiography can diagnose several
potential complications of the arthroscopic autolo-
gous bone-patellar tendon-bone (BPTB) technique
used for ACL reconstructions. These include patellar
fracture, migration of the bone plug and hardware
failure (GRAF and UHR 1988; MANASTER et al. 1988).
Furthermore, if the lateral radiograph shows that the
intra-articular opening of the tibial tunnel is anterior
to the roof of the notch (Blumensaat's line), graft
impingement between the graft and the anterior edge
of the intercondylar notch can be expected.
Over the last 10 years or so there has been a grad-
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
rior physeal scar, as seen on the sagittal MR sequence
(MANASTER et al. 1988). On the coronal MR images
the femoral tunnel should be at the 11 o'dock posi-
tion in the right knee and the 1 0' dock position in the
left knee. The position of this femoral tunnel is crit-
ical in ensuring near isometry, which is a prerequi-
site for the maintenance of a constant length and ten-
sion of the graft throughout the range of flexion and
extension of the knee. Significant deviation from this
femoral tunnellocation will prevent isometry of the
graft, rendering the knee eventually unstable. The
tibial tunnel on the coronal MR images should open
on the intercondylar eminen ce of the tibia in the mid-
line.
Artefacts at the time of performance of MRI are
to be expected. The presence of metal interference
screws within the femoral and tibial tunnels will gen-
erate a varying degree of artefact, which predudes
complete assessment of the tunnels. There is usually,
however, no interference in the assessment of the
intra-articular component of the graft. Non-metal
fixation devices allow better assessment of fixation
and the graft within the tunnels. It is not unusual to
have microscopic fragments of metal within the joint
and these will give rise to a varying degree of MR
susceptibilityartefacts (Fig.l1.13). The gradient echo
sequences are more vulnerable to this type of arte-
fact. MR sequences obtained on more modern scan-
ners utilising shorter echo times can give rise to focal
areas of increased signal intensity on the proton den-
Fig. 11.13a, b. Metal artefacts following removal of interference screws at the
time of revision ACL reconstruction from the tibial tunnel. Note the metal
a fragments which cause significant artefact in the coronal MR images
V. N. Cassar-Pullicino and S. N. J. Roberts
sity- or Tl-weighted images within the stable asymp-
tomatic cruciate reconstructions (SCHICK et al. 1995).
It is probably unlikely that the increased signal seen
in some instances in unimpinged stable asymptomat-
ic ACL grafts is due to the so-called magic angle phe-
nomen on that can occur in tendons and ligaments
(ERICKSON et al. 1993).
Posterior cruciate ligament reconstructions are
carried out less often than ACL reconstructions. The
reconstruction is indicated particularly in athletes
and in those who demonstrate symptomatic instabil-
ity. The tom PCL is also ideally reconstructed using a
patellar tendon graft, although this is not always POS-
sible in view of the increased length that is required
and technical difficulties in a PCL reconstruction. On
MRI the intact graft has been shown to be uniformly
oflow signalintensity on Tl-weighted images (MUNK
et al. 1992), and the stable neo-ligament, unlike the
ACL, remains similar in appearance to the patellar
tendon after surgery, exhibiting low signal on all MR
pulse sequences (TUITE and DE SMET 1996). The fem-
oral tunnel should open just anterior to the femoral
insertion of the native PCL while the tibial tunnel
should open in the joint in the mid-line within the
small depression which lies just posterior to and
below the articular surface (MANASTER et al. 1988).
This is a developing field in achallenging area of
reconstructive surgery of the knee and recently use of
double bundles has been recommended (RACE and
AMIS 1998).
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
11.3.1.3
Problematic Neo-ligament
Increasingly it has become apparent that the postop-
erative rehabilitation programme employed follow-
ing cruciate ligament reconstruction can determine
the likelihood of ongoing symptoms, which range
from stiffness to recurrent instability. The stiffness
can arise either from arthrofibrosis, which can be dif-
fuse or localised anteriorly via production of the so-
called cyelops lesion, or from graft impingement pro-
duced by an imperfectly placed graft. More aggressive
and accelerated rehabilitation protocols have gained
wider acceptance in the belief that they improve the
functional result and significantly decrease the preva-
lence of postoperative stiffness. The downside of this
aggressive rehabilitation technique is the increased
risk of injury to the graft at a time when it is merely
acting as a scaffold and when it is partially revascu-
larised. The BPTB graft at harvest is 50% stronger
than the native ACL, but the revascularisation pro-
cess weakens it so that at 1 month it is only 40% as
strong as it was originally (COUPENS et al. 1992).
11.3.1.3.1
Arthrofibrosis
Arthrofibrosis can be diffuse in nature intra-articu-
larly. It is probably a result of significant haemarthro-
sis, either before the cruciate ligament reconstruction
or after surgery, and causes increased stiffness of the
joint. In its localised form anterior to the tibial inser-
tion of the neo-ligament, the arthrofibrosis takes the
configuration of a focal fibrous nodule placed inferi-
orly and anteriorly to the intercondylar notch, lying
behind Hoffa's fat pad. In this location it acts as a
mechanical block as it is trapped between the femur
and the tibia anteriorly, preventing terminal exten-
sion of the knee, which is one of the major causes of
morbidity following ACL reconstruction. Following
intravenous enhancement, the nodule has been seen
to enhance owing to its fibrovascular consistency. It
is thought to arise at the site of the stump of the tom
ACL, probably as a result of the debris produced by
drilling the tibial tunnel. The term "cyelops" is given
to its diagnostic appearance at arthroscopy, which
also allows it to be resected. It is, however, important
to realise that one of the causes put forward for the
cyelops lesion is graft impingement, which me ans
that careful search for associated impingement of the
graft needs to be done if this lesion is demonstrated
by MRI (RECHT et al. 1995). On Tl- and proton densi-
ty-weighted MRI, the cyelops lesion appears as a focal
179
nodular lesion of low to intermediate signal intensity,
while on T2-weighted images it exhibits a heteroge-
neous but predominantly low signal intensity usually
outlined by free joint fluid around it.
As previously identified, problems following liga-
ment reconstruction can arise in the early postop-
erative period (the first 6 months) or towards the
end of the first year. Those occurring early are more
often than not indicative of failure caused by imper-
fect surgical technique. Indeed, errors in surgical
technique are the most common cause of graft fail-
ure in patients presenting with recurrent instability
after ACL reconstruction. Surgicalltechnical reasons
for graft failure inelude graft impingement, improp-
er tensioning of the graft, non-anatomical tunnel
placement and inadequate fixation of the graft in
the osseous tunnels. Apart from these causes of graft
failure, consideration also needs to be given to the
graft material used, failure of adequate graft lig-
amentisation, and failure caused by trauma itself,
which usually is a problem after resumption of full
knee activity.
11.3.1.3.2
Grah Impingement
Graft impingement presents elinically with stiffness,
particularly loss of terminal extension, along with
pain. The impingement can occur in the sagittal plane
when the ACL graft impacts the intercondylar roof,
or it can be evident in the coronal plane when the
ACL neo-ligament impacts on the side walls of the
notch. The latter is a less common problem, particu-
larly because it is often detected at the time of surgery
by the arthroscopist, who also performs a prophylac-
tic notchplasty. In contrast, roof impingement is not
easy to detect at the time of surgery by the arthros-
cop ist because this tends to occur in the last 5-10°
of extension. Roof impingement is most commonly
caused by the tibial tunnel's location, having been
placed too anteriorly (HOWELL and CLARK 1992;
CASSAR-PULLICINO et al. 1994) (Fig. 11.14). The
entrance of the tibial tunnel on the tibial articular
surface should be placed completely posterior to the
intersection of the intercondylar roof onto the tibial
surface in the fully extended lateral view of the knee,
as seen both radiographically and on sagittal MR
images. The anterior wall of the opening of the tibial
tunnel in particular needs to be posterior to Blumen-
saat's line. When the tibial tunnel is either partially
or completely anterior to the slope of the intercondy-
lar roof with the knee in extension, impingement of
the graft occurs. On emerging from the tibial open-
 180 V. N. Cassar-Pullicino and S. N. J. Roberts

Fig.ll.I4a,b. Sagittal Tl-weighted images oftwo different patients showing

failure of the ACL graft due to an anteriorly located tibial tunnel. Note the
intersection of the intercondylar roof with the path of the tibial tunnel
b

ing, the graft is seen to sharply angulate as it drapes 11.3.1.3.3

around the antero-inferior margin of the roof of the
intercondylar notch, or it may be posteriorly bowed
(Fig. 11.15). Signal abnormalities are to be expected
in the impinged graft by 3 months postoperatively.
Tl- and proton density-weighted images demon-
strate increased signal intensity within the distal
two-thirds of the graft at the site of impingement
(HOWELL et al. 1991a). Furthermore, when MRI is
done either less than 3 months or more than 12
months after surgery, impingement should be sus-
pected if there is increased signal on T2-weighted
images at the point of contact (HOWELL et al. 1991a).
After the appropriate correction of tibial tunnel posi-
tion or notchplasty, the MR signal returns to normal
by 12 weeks postoperatively. At the time of diagnos-
ing graft impingement on MRI, one also needs to
evaluate the knee for associated conditions such as
rupture of the graft (partial or complete) and associ-
ated cyelops lesion.
Graft Rupture
Apart from near-anatomical positioning of the tibial
and femoral tunnels, prevention of graft impinge-
ment can also be guaranteed byan adequate notch-
plasty, which in some instances is routinely done at
the time of insertion of the graft. It is not uncom-
mon to have stenosis of the intercondylar notch bony
architecture with the presence of osteophytes, which
are thought to develop after failure of the original
ACL. These in turn, unless removed, will promote
impingement and secondary graft failure from repet-
itive trauma and damage to the neo-ligament. Graft
impingement at the intercondylar roof is not only
associated with the development of a cyelops lesion,
but can also compromise the biological incorpora-
tion of the graft. Over-indulgent notchplasty, howev-
er, will give rise to problems regarding the anatomi-
callocation of the femoral tunnel, which can lead to
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 181

Fig. 11.15a, b. An anteriorly located tibial tunnel promotes impingement of

the ligament, which appears draped over the posterior aspect of the inter-
condylar roof

incongruence at the patellofemoral and tibiofibular
articulations (JOHNSON 1993).
Biological failure presents as recurrent instability
in a patient without a history of a new episode of
trauma, and in whom there is no evidence of a
surgica1!technical error in the placement of the graft.
Biological failure results when the ligamentisation
process is inadequate and as a result the graft scaf-
fold is easily elongated and rendered biomechanical-
ly unsound and prone to damage. A number of fac-
tors affect the ligamentisation of the graft, most of
which cannot be assessed by MRI; these include type
of graft, biomechanical loads during rehabilitation,
fixation and host response. MRI in graft rupture
shows an absence of the expected continuity of the
intact graft fibres. Apart from discontinuity, absence
of the ligament within the osseous tunnels and
marked bowing of the ligament are further signs of
graft failure. These signs are useful within the first
3 months of surgery, or after maturation has taken
place 1 year after surgery. On T2-weighted images,
increased signal intensity iso-intense to fluid is pres-
ent within the intra-articular portion of the graft
(MUNK et al. 1992; YAMATO and YAMAGISHI 1992;
TUITE and DE SMET 1996); this is the most specific
sign of an acute complete tear of the graft (Fig.l1.16).
As has been previously described, the T2-weighted
sequence is very important in differentiating the
revascularisation between the 3- and the 12-month
postoperative period as the intact and unimpinged
stable neo-ligament can have increased signal on the
proton density-weighted and short TE MR images.
Partial tears of the neo-ligament can also occur: they
are highlighted by areas of increased fluid-like signal
on the T2-weighted images, affecting only part of the
graft substance with residual intact continuous fibres
still present. MAYWOOD et al. in 1993 correcdy iden-
tified such partial tears of the neo-ligament using
oblique sagittal T2-weighted images along the plane
of the graft. However, CHEUNG et al. and YAMATO
and YAMAGISHI in 1992 demonstrated a wavy con-
tour with peripheral high T2 signal in asymptomatic
stable grafts within the first year after surgery, which
could have been misinterpreted as partial graft tears
byMRI.
Apart from biological, mechanical and structural
failure of the graft, joint laxity can also arise from
improper placement of the osseous tunnels, improp-
er sizing of the graft length and loosening or fracture
of the fixation with particular reference to the bone
plug of the BPTB autograft. These possibilities can
also be assessed adequately by MRI.
In summary, at any postoperative stage, a graft
exhibiting continuity of structure and a homoge-
neous low signal intensity from the tibial tunnel
across the intra-articular component to the femoral
tunnel is indicative of an intact graft. After the first
year the graft is also expected to have matured and
should exhibit primarily low signal. In the interven-
ing period, particularly between the third and ninth
 182
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
may be seen as anormal finding owing to the revas-
cularisation and ligamentisation process with an
intact unimpinged graft. However, increased signal
intensity demonstrated on Tl- and proton density-
weighted sequences in such instances can be indica-
tive of disrupted grafts as well as intact, but impinged
grafts. It is for this reason that the integrity of the
ACL graft cannot be determined on the basis of
simply the proton density- or Tl-weighted images.
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
a graft rupture (JANSSON et al. 2001).
11.3.7.3.4
Bone Tunnels
Near-anatomical placement of the tibial and femoral
tunnels is an essential prerequisite to the success of
the graft, as has been previously described. In addi-
tion, the impact of inadequate surgical technique and
tunnel placement on graft impingement and failure
has also been stressed. Enlargement of the bone tun-
nels, however, can also be identified after ACL repair,
both radiographically and at MRI (Figs. 11.17, 11.18).
Either or both of the bone tunnels can be seen to
enlarge; this is primarily witnessed after an STG
tendon autograft, but is occasionally seen following
BPTB autograft. This may in part be due to the genera-
tion of more peri-ligamentous tissue response in the
four-strand hamstring autograft, promoting enlarge-
ment of the tunnels, or the different biomechanics that
result from different fixation of the hamstring auto-
graft as compared with the BPTB graft. In addition,
V. N. Cassar-Pullicino and S. N. J. Roberts
peri-ligamentous cysts can form within the tunnels,
presumably arising from invagination of synovium,
with synovial fluid within the bonytunnels (MURAKA-
MI et al.1999). Although unusual, occasionally one also
identifies areas of bone marrow oedema around the
enlarged bony tunnel; this may signify a true osteo-
lytic process, possibly secondary to stress shielding
at this site. Demonstration of this abnormal tunnel
by MRI is an important consideration when planning
revision surgery (FAHEY and INDELICATO 1994).
11.3.2
The Postoperative Meniscus
Meniscal tears can be treated entirely conservatively,
or by surgical conservative techniques which include
partial meniscectomy (SCHIMMER et al. 1998), menis-
cal repair with suture (VAN TROMMEL et al. 1998) or
meniscal repair with bio-absorbable arrows (WHIT-
MAN and DIDUCH 1998). The goal of treatment is to
restore the meniscus to a near-anatomical morpho-
logical structure with restoration of its stability. As
a general principle, meniscal preservation is currently
the primary goal when treating a meniscal tear. If the
tears are stable and asymptomatic and the arthroscop-
ic surgeon considers them unlikely to propagate, then
it is justifiable to leave them alone, allowing them to
heal spontaneously. In symptomatic tears that com-
promise meniscal function, ideally meniscal repair
should be performed using sutures or absorbable
implants carried out by various arthroscopic tech-
niques (CASSIDY and SHAFFER 1981; DEHAvEN 1985;
HANKS et al.1991). Ifthe meniscal tear at arthroscopy
is seen to be non-repairable, a partial meniscectomy
 The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 183

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
is carried OUt. Only the unstable meniscal fragment
is removed and in particular, the torn portion of the
meniscus that is protruding into the joint is excised
completely. The surgeon attempts to shave, fashion
and trim the residual meniscal rim from anterior
to posterior in an attempt to make it as perfectly
smooth as possible to reduce the risk of problems
postoperatively. More recently, meniscal transplanta-
tion and meniscal replacement, e.g. using collagen
meniscus implant, have been introduced as options
in instances of significant complex meniscal tears
(POTTER et al. 1996).
Over the last decade, as arthroscopic refinements
and concepts of meniscal tear treatment have evolved,
there has been a gradual increase in the knowledge of
the expected MR appearances after meniscal resection
and repair. Familiaritywith these expected MR appear-
ances is essential before embarking on an assessment
of the postoperative meniscus as they underlie the cor-
rect interpretation of the MR appearances (Fig.11.19).
Initially, conventional MRI was used in the assess-
ment of the postoperative meniscus. Gradually,howev-
a 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
V. N. Cassar-Pullicino and S. N. J. Roberts
er, it has been replaced by MR arthrography, as weIl as
dynamic assessment of the knee to enhance the accu-
racy rate by diminishing the degree of uncertainty in
distinguishing the expected postoperative appearanc-
es from pathological meniscal states.
There are two fundamental MR criteria for diag-
nosing a meniscal tear in the unoperated knee: dem-
onstration of an increased internal signal on a short
TE image unequivocally contacting an articular sur-
face of the meniscus, and an abnormal meniscal
shape. However, although these two criteria, if pres-
ent, will have an accuracy of over 90%, they cannot
be applied with the same degree of confidence in the
postoperative state. Following meniscal debridement,
rasping, partial meniscectomy or meniscal repair,
conventional MRI of the postoperative meniscus has
been shown to be unreliable (EGGLI et al. 1995;
BRONSTEIN et al. 1992; TSAI et al. 1992), with accu-
racies ranging between 38% and 80%. This is in
part because abnormal meniscal morphology is to be
expected after a previous tear which has been treat-
ed by partial meniscectomy. Furthermore, when a
b
_ ____ _ _ d
 The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
partial meniscectomy is performed, the hyperintense
intra-substance (grade I or II) signal intensity that
was present within the meniscus pre-operatively can
be converted into a seemingly grade III signal inten-
sity simulating a meniscal tear (Figs. 11.20, 11.21).
It is more likely than not that, following injury,
a meniscus will never return to its normal pre-inju-
ry signal intensity on Tl- or proton density-weight-
ed images (ARNOCZKY et al. 1994; MUELLNER et al.
1999). This means that even after a successful menis-
cal repair, the healed meniscus may still show a
hyperintense signal within its substance at the site of
the repaired tissue, which will appear as an expected
Signal Conversion
I. Pre-op
11. Post-op
___ ___ All OlWthlt2001
~_~_~
Fig.l1.20. Diagrammatic representation to explain the appear-
ances that can occur after partial meniscectomy, producing
meniscal "conversion"
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
185
finding on the short TE sequences (Tl and proton
density) and can be confused with a meniscal tear.
Access to the pre-operative MRI, the arthroscopic
findings and surgical operative details, use of MR
arthrography, and dynamic assessment using an
open MR magnet all help in reducing the uncertain-
ty, enabling correct interpretation of postoperative
meniscal findings.
11.3.2.1
Post-meniscectomy Assessment
Recurrent symptoms following meniscal surgery can
be evaluated by conventional MRI, conventional
arthrography, MR arthrography or arthroscopy. As
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.
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
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
very valuable information at the time of reporting
the MR images, it is rarely available in clinical prac-
tice. When available, based on the work carried out by
MINK et al. (1993), DEUTSCH et al. (1992), and SMITH
 b

Fig. 11.22. Meniscal remnants following partial meniscectomy of the pos-

a
terior horn (a) and the anterior horn (h)

and TOTTY (1990), there are some reasonable guide-
lines that one can follow. When less than 25% of the
meniscus has been resected, it can be evaluated using
the MRI criteria employed in the pre-operative state.
In these studies virtually 90% of the menisci exhib-
iting grade III signal were shown by arthroscopy to
have suffered a re-tear. When the partial meniscec-
tomy is more extensive (more than 30%), the stan-
dard MRI criteria used in the unoperated meniscus
become significantly less accurate. DEUTSCH et al.
(1992) divided the more extensive meniscectomy
patients (more than 30%) into two groups, depend-
ing on whether they had had a resection of less than
or more than 75% of the meniscus. In the group with
less than 75% meniscal resection, neither the pres-
ence of internal grade III signal contacting the menis-
cal surface nor abnormalities of meniscal s~ape were
as accurate in predicting a recurrent meniscal tear as
in the unoperated state, with the generation of a sig-
nificant number of false-positive and false-negative
diagnoses of recurrent tears. The authors coined the
term "signal conversion", whereby an apparent grade
III signal can be generated following arthroscopic
resection in a meniscus that demonstrated a grade
I or II signal abnormality originally. In view of this
signal conversion, the authors proposed two criteria
which can be used to distinguish a re-tear from a
stable postoperative meniscus. In the first place, dem-
onstration of a displaced meniscal fragment is to be
regarded as a highly specific but insensitive finding
in patients with recurrent meniscal tear. Secondly, if
the grade III signal abnormality seen on the short
TE images also corresponds to an area of high signal
on the T2-weighted images, then it can be reason-
ably assumed that a re-tear is present (Fig. 11.23).
Although the demonstration of this linear high T2
signal within the meniscus is probably the most
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
recurrent tear on conventional MR images, one needs
to remember that it is only present in 33% of tears in
non-operated menisci (MINK et al. 1998). In a com-
parative assessment with arthrography and conven-
tional MRI, FARLEY et al. (1991) showed that the
T2 demonstration of a meniscal tear was the most
common finding, while ApPLEGATE et al. (1993) used
this criterion in post-meniscectomy patients who had
a joint effusion, and reported a 56% sensitivity and
90% specificity for diagnosis of re-tom menisci. In
the group of patients in whom more than 75% of
the meniscus has been removed, conventional MRI
has an accuracy rate of 85% with arthroscopy. The
specificity for a re-tear of a meniscus when abnormal
signal iso-intense with fluid is seen on T2-weighted
images dearly suggests that there is a valuable role
for MR arthrography in the assessment of the post-
operative meniscus following partial meniscectomy.
ApPLEGATE et al. (1993) performed MR arthrograms
in their postoperative patients and increased the sen-
sitivity for meniscal re-tears from 56% to 90%, while
maintaining a specificity of dose to 90%.
11.3.2.2
Post-meniscal Repair
The potential for healing of meniscal te ars exists
particularly within the vascularised outer third of
the meniscus. Small stable peripheral tears iden-
tified, for example, at the time of ACL reconstruc-
tion are treated completely conservatively and they
heal spontaneously. There have been marked refine-
ments in the per arthroscopic surgical techniques
to repair peripherallarge unstable tears. As in the
post-meniscectomy state, assessment by conven-
tional MRI of the repaired meniscus can be diffi-
cult. Morphologically the meniscus is more likely
to appear normal, but the healed te ar will still pro-
duce a grade III signal appearance on the short TE
images. Second-Iook arthroscopy has shown that the
meniscus has healed and is indeed stable (DEUTSCH
et al. 1990) in these circumstances. FARLEY et al.
(1991) also showed a grade III signal abnormality
on proton density- and Tl-weighted images in 27
of 29 cases of repaired menisci, conduding that the
presence of signal contacting the articular surface
of the meniscus is not a reliable predictor for recur-
rent or residual meniscal tear. Demonstration of a
high signal iso-intense with fluid on T2-weighted
images predicted a recurrent meniscal tear, arepair
that had not healed or a new te ar with a sensitivity
of 60% and a specificity of 92% (FARLEY et al. 1991).
Clearly if a grade III signal abnormality is seen on
187
the short TE images at a site which is known to
have been normal on the pre-operative MRI, then it
can be safely conduded that this represents a new
meniscal tear. Similar results and condusions were
reached by ApPLEGATE et al. in 1993, but on ce again
these authors highlighted the important role of MR
arthrography in distinguishing a healed meniscal
repair from one that has re-tom. In a comparative
study utilising conventional MRI and conventional
arthrography in evaluating meniscal repair, VAN
TROMMEL et al. in 1998 reported a small series of
patients in whom the MR technique was more accu-
rate than arthrography in discriminating partial or
complete healing. These authors used fast spin echo
fat-suppressed sequences to evaluate the meniscal
repair, but only eight cases had arthroscopic confir-
mation of their findings. Like FARLEY et al. in 1991,
VAN TROMMEL et al. (1998) stressed the importance
of a line of high T2 signal intensity iso-intense to
fluid within the meniscus as indicating incomplete
healing of a meniscal repair. In a somewhat larger
series of 20 patients, LIM et al. (1999) carried out
a retrospective study which compared five specific
findings on conventional MRI with second-Iook
arthroscopy in patients evaluated for possible repeat
tears of the postoperative meniscus. The authors
found that the two signs with the highest correlation
were (a) signal extending to the meniscal surface
on a proton density-weighted image and (b) fluid
intensity signal on a T2-weighted image extending
into the meniscus. As found in previous studies,
postoperative abnormalities in meniscal morpholo-
gy were of little value in predicting recurrent tears.
11.3.2.3
MR Arthrography
Studies evaluating the postoperative meniscus by
conventional MRI have shown the high specificity of
a grade III signal within the meniscus which demon-
strates signal intensity that is iso-intense with fluid
on the T2-weighted sequence. It follows, therefore,
that direct MR arthrography utilising dilute gado-
linium DTPA (Gd-DTPA) with saline in a ratio of
approximately 1:250 will increase the accuracy of
MRI. Recently, there has been an ever-increasing
interest in developing the role of indirect MR
arthrography following the intravenous injection
of Gd-DTPA in assessing the postoperative knee.
ApPLEGATE et al. (1993) showed that 66% of postop-
erative patients presenting with a new knee pain
had a recurrent meniscal te ar confirmed by arthros-
copy. The greatest value of MR arthrography lies in
188
imaging of the postoperative meniscus which dem-
onstrates a grade III abnormality on a Tl- or proton
density-weighted sequence that is not also seen as
high signal on the T2-weighted images. This is par-
ticularly relevant in patients who have had a menis-
cal repair and in those who have had a partial men-
iscectomy of between 25% and 75% of the meniscus.
In such cases, MR arthrography can replace arthros-
copy as a me ans of ascertaining whether there is
evidence of a meniscal re-tear. In Applegate's series,
conventional MRI was compared with MR arthrog-
raphy and the accuracy of the latter was significantly
better (88% vs 66%). This study showed that both
techniques were about 90% accurate in instances of
minimal meniscal resection, whereas in instances
where a greater amount of the meniscus had been
resected, MR arthrography was significantly more
accurate. In the presence of a small meniscal rem-
nant, the accuracy of MR arthrography was elose to
100%. More recently SCIULLI et al. (1999) compared
conventional MRI, MR arthrography using gadolin-
ium-based contrast medium, conventional arthrog-
raphy, and MRI after the conventional arthrogram
in 33 patients who had had meniscal surgery. The
authors coneluded that MR arthrography utilising
dilute gadolinium provided the most accurate diag-
nosis of recurrent meniscal tears when compared
with the arthroscopic results obtained in 12 of their
33 patients.
In summary, depiction of a displaced meniscal
fragment always indicates the presence of a new te ar
irrespective of whether the patient has had partial
meniscectomy or meniscal repair. In patients with
partial meniscectomy, MR arthrography can confirm
conversion of signal abnormalities within the menis-
cus when there is no evidence of any contrast extend-
ing into the abnormality. In addition, by showing
contrast extending into the substance of the menis-
cus, it can exhibit re-tears or new tears within the
meniscus. Following meniscal repair, it can be con-
fidently coneluded that the tear has healed when no
contrast agent enters the repair site at MR arthrogra-
phy. In the repaired meniscus, MR arthrography can
accurately demonstrate partial or complete re-tears
along with new tears. MR arthrography is of greatest
use in assessing postoperative menisci where signifi-
cant resections have taken place. It provides a higher
signal to noise ratio, higher resolution and greater
contrast between fluid and the other soft tissue
components. The presence of sharp abrupt contour
changes and depiction of free meniscal fragments
should always suggest the presence of a re-tear.
V. N. Cassar-Pullicino and S. N. J. Roberts
11.3.3
Articular Cartilage Assessment
11.3.3.1
Osteoarthritis
Postoperative knee symptoms, and in particular pain,
may be due to underlying chondral lesions, osteo-
chondral defects or frank osteoarthritis. In part, con-
comitant injury to the articular surface at the time of
cruciate ligament or meniscal damage may be the
initiating factor which then progresses despite suc-
cessful cruciate or meniscal surgery. In the evaluation
of patients following cruciate ligament and meniscal
surgery, it is therefore imperative that the chondral
and osteochondral surfaces are fully assessed as
potential causes of underlying pathology. Delay
between injury and corrective surgery can also play
an important role in the development of chondral
injury. Only 30% of patients with an ACL tear have
symptomatic instability (MINK and DEUTSCH 1989)
and yet virtually all of them will develop premature
chondral damage leading to osteoarthritis if the knee
remains unstable (KANNUS and JARVINEN 1987).
However, what is not known is whether this is dictat-
ed by the concomitant presence of meniscal damage
and surgery, which of its own accord predisposes to
articular cartilage degeneration. Before the advent
of arthroscopy, meniscal tears were treated by com-
plete meniscectomy, which was associated with a high
incidence of premature or accelerated osteoarthritis
(ALLEN et al. 1984; Roos et al. 1998) (Fig. 11.24). The
long-term results following partial meniscectomy sug-
gest a reduced incidence of osteoarthritis (BOLANO
and GRANA 1993; COVALL and WASILEWSKI 1992;
RANGGER et al. 1995). Although a significant risk of
osteoarthritis exists following rem oval of any part of
the meniscus, and meniscal preservation is currently
the primary goal in the treatment of meniscal tear,
multiple studies have shown that the incidence and
severity of osteoarthritis after meniscectomy is direct-
ly proportional to the amount of meniscal tissue
resected (KITZlGER and DELEE 1990) (Fig. 11.25).
Indeed, KITZlGER and DELEE demonstrated a direct
correlation between the amount of meniscus removed
and the degree of radiographic osteoarthritic changes.
Prior to this, FAIRBANK (1948) had shown radiograph -
ic evidence of significant osteoarthritis in 60% of
patients following total meniscectomy. These changes
do not need MRI for diagnosis, but there is no doubt
that MRI is more likely to demonstrate changes within
the articular cartilage, not appreciated on the postop-
 The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
Fig. 11.24. AP film (al and T2- (bl and Tl-
weighted (cl coronal images following bilat-
eral complete meniscectomy. Note the early
marginal osteophyte formation
erative radiographs. Chondral pathology on the artic-
ular surface adjacent to a damaged or operated menis-
cus was found to be the cause of recurrent knee symp-
toms in 40% of patients after partial meniscectomy
(DANDY and JACKSON 1975). Loosely termed "chon-
dromalacia", this is seen arthroscopically and also on
MRI as thinning of the cartilage with fissures, in com-
bination with focal defects within the hyaline cartilage
which are best seen using T2 or fat-suppressed gradi-
ent re-call echo techniques or MR arthrography. The
incidence of radiographically detected osteoarthritis
and recurrent knee symptoms is much lower following
meniscal tear repair than after treatment by total or
partial meniscectomy (JOHNSON and BEALLE 1999).
189
c
11.3.3.2
Autologous Chondrocyte
and Osteochondral Transplantation
Superficial lesions of the hyaline cartilage surface
detected at arthroscopy are usually treated by debride-
ment in an effort to prevent enlargement. Postopera-
tive MRI in such instances will demonstrate the areas
of cartilage thinning in a focal rather than diffuse
location. The underlying subchondral bone should be
normal, although occasionally there are early chang-
es of sclerosis at this site. When the chondral defects
are seen to involve the underlying subchondral bone,
treatment is usuaUy by abrasion arthroplasty or drill-
 190
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
resultant blood clot will stimulate the development
of a fibrocartilage repair. TUITE and DE SMET (1996)
reported that the MR appearance of this fibrocar-
tilage exhibits a higher signal on the T2-weighted
images when compared with the meniscal fibrocarti-
lage. More recently, cartilage repair procedures have
been developed and are being constantly refined. The
two most commonly employed are autologous chon-
drocyte transplantation and osteochondral autograft
transplantation (OATS, mosaicplasty). Different sur-
gical techniques are employed in achieving these
repair procedures, placing unique demands on the
choice of MR pulse sequences postoperatively, and
creating problems in the correct interpretation of the
postoperative MR images. The optimum sequences
for depiction of contour and intrinsic cartilage detail
should be utilised. In addition, in some instances MR
arthrography can provide better detail of the articu-
V. N. Cassar-Pullicino and S. N. J. Roberts
lar cartilage surface, while dynamic intravenous stud-
ies can help to monitor the results of treatment in
instances of autogenous osteochondral transplanta-
tion.As surgical techniques become more refined and
acceptable, it is likely that there will be a growing
demand for non-invasive methods of monitoring the
results of treatment. However, it remains to be seen
whether MR in its monitoring role does affect the
outcome of these therapeutic procedures, or whether
it is useful in directing the treatment of overgrowth
of the transplanted cartilage before this becomes
symptomatic.
MR studies following autologous chondrocyte
transplantation show a variable degree of artefact on
the surface of the graft, which interferes with accurate
assessment of the thickness and surface contour of
the repair tissue (Fig. 11.26). Fat-suppressed proton
density-weighted fast sequences are very useful in
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
ACI graft should include an estimation of the volume
of graft that fills the original defect, surface contour
restoration, status of subchondral bone and neigh-
bouring bone marrow oedema (Fig. 11.27). Sequen-
tial MRI studies allow the identification of partial
and complete graft delamination with or without the
use of MR arthrography (Fig. 11.28) (ALPARSLAN et
al. 2001). MRI also reliably detects graft hypertrophy.
This can occur by extension into the intercondylar
notch, or when overgrowth of the graft occurs, usu-
ally at the edges with the native cartilage, particu-
lady in the non-weight-bearing surfaces. The rele-
vance and interpretation of signal changes within the
repaired cartilage remains a topic of active research.
At present MRI is unable to determine accurately the
tissue type (s) within the graft.
Osteochondral autogenous graft transfers have
recently been evaluated using conventional radiog-
raphy, CT and MRI with dynamic Gd-DTPA intrave-
nous enhancement (VERSTRAETE et al. 2000). In addi-
tion, the postoperative MR appearances have also
been correlated with the postoperative clinical status
(SANDERS et al. 2001). In about 50% of instances radi-
ography detects the donor and recipient sites, while
CT and MRI virtually show all the grafts cleady at the
recipient sites. The size of the graft, its relationship
with the subchondral bone and cartilage contour can
be evaluated. Subchondral alignment occurs in less
than 40% of grafts, while cartilage surface contour
alignment is perfect in 80% of cases. As the osteo-
chondral "plug" is taken from a different part of
the joint surface it usually has a different thickness
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
Fig. 1l.28a, b. Partial delamination at the site of previous successful autologous chondrocyte transplantation
emphasis is to ensure cartilage contour continuity, and
for this reason it is not surprising that there is often a
discrepancy in the alignment of the subchondral bone
of the plug with the recipient area (Fig. 11.29). CT is
better at showing the impaction status of the osseous
component of the plug, its incorporation and the
formation of peripherally located resorption cysts.
MRI shows bone marrow oedema around the plug in
the first few days, which increases between 3 and 9
months following surgery and then gradually disap-
pears. The transplanted osteochondral plug retains
its normal fatty marrow signal intensity on MR stud-
ies done within 2 weeks, but at 4 weeks heteroge-
neous signal is noted, highlighted by decreased Tl
and increased T2 appearances. At the 6th week post-
operatively, the plugs show a uniformly decreased
signal intensity on Tl-weighted sequences and a uni-
formly increased T2 signal intensity. At 5 months
after surgery the plugs on ce again appear heteroge-
neous as there is a gradual return of the normal fatty
marrow signal, usually commencing within the cen-
tral portion of the plug. Successfully transplanted
plugs demonstrate normal fatty marrow signal 1 year
after the surgery (Fig. 11.30). Dynamic Gd-DTPA
enhancement of the marrow surrounding the graft
is intense starting at 2 weeks following surgery, and
gradually decreases with time so that by 6-9 months
postoperatively there is minimal to no enhancement
in the surrounding marrow. This enhancement pat-
tern is thought to parallel the revascularisation of the
plug, which is initially avascular and demonstrates
V. N. Cassar-Pullicino and S. N. J. Roberts
early revascularisation at 4 weeks with improvement
at 6 weeks. The development of oedema within the
graft along with this enhancement pattern and even-
tual return of normal fatty marrow signal at 1 year
indicates normal graft incorporation. After an initial
revascularisation phase the graft undergoes a phase
of resorption which in turn is followed by an incor-
poration phase highlighted by osseous remodelling
and repair (EINHORN 1995).
In 20% of OATS cases, graft protuberance or
depression ranging between 1 and 2 mm can be
expected. This does not appear to change with time
on subsequent MR investigations, nor does it appear
to have any bearing on the short-term clinical out-
come.Arthroscopic studies in the follow-up of osteo-
chondral plug transfer have reported a fibrocartilage-
like tissue which fills the gaps between the transplant -
ed osteochondral plugs, improving the congruity of
the joint surface (HANGODY et al. 1997). Short-term
follow-up MR studies and CT arthrography,however,
have not confirmed this finding, and it may be a phe-
nomen onthat develops with time. The osteochondral
surface at the donor sites appears irregular on MRI,
with very litde alteration on follow-up examinations.
There is usually no or very poor reconstitution of
bone here, with at best only a thin fibrocartilaginous
tissue within the crater of the defects. Donor site
bone marrow oedema is generally less conspicuous
than graft site oedema. A variable degree of micro-
metal artefact can be identified in some instances at
both the donor and the recipient site.
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage 193

Fig. 11.29a-c. Tl-weighted sagittal sequences done over a l-year period

after autologous osteochondral plug transfer. Note the changes within the
bone plug with the normal fatty marrow returning at 1 year

11.3.4 sequences and are readily identified, especially in

Extra-articular Structures
Ultra sound, MRI and to a lesser extent CT are usually
the best modalities for assessment of the postopera-
tive status of the extra-articular tissues.
11.3.4.1
Scar and Metal Artefact
Extra-articular soft tissue changes occur universally
irrespective of whether the surgical procedure has
been carried out extra-articularly or intra-articu-
larly using arthroscopic techniques. Inevitably, MRI
detects a varying degree of postoperative fibrosis
and the presence of metal micro-fragments at sites
or tracks of surgical instrumentation. In combi-
nation, the scar tissue and metal artefact produce
a very low signal on both Tl- and T2-weighted
gradient echo T2-weighted sequences within the
soft tissues at the sites of arthroscopic portals and
in Hoffa's fat pad.
11.3.4.2
Supporting Structures
Surgery to the collateralligaments and patellar reti-
nacula results in the formation of scar tissue, leading
to loss of the normal anatomical detail of the involved
structures, which appear thickened on MR studies.
Primary repair of the medial supporting structures
following acute patellar dislocation pro duces a thick-
ened continuous low-signal band, which incorporates
the three main medial ligaments. Conversely, in lat-
eral release ptocedures carried out for patellar mal-
tracking and excessive patellar tilt, a defect is to be
expected within the lateral retinaculum.
 194 V. N. Cassar-Pullicino and S. N. J. Roberts

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
11.3.4.3
DonorSites
Alterations oeeur within the soft tissues where har-
vesting for ACL repair has taken plaee. Donor site
regeneration of the semi-tendinosus and gracilis
tendons following their transeetion has been shown
to oeeur clinieally and objeetively. MRI has shown
that the tendons do re-grow, inserting as a diffuse
fan-shaped attaehment into the medial popliteal
fascia, and they are also thought to resurne their
funetion (CROSS et al. 1992). Following BPTB ACL
reeonstruetions, the immediate post-harvest sono-
graphie and MR appearanee of the residual tendon
is abnormal. Adefeet at the site of rem oval of the
eentral one-third of the tendon is easily diseern-
ible, especially sonographically. The tendon dem-
onstrates diffuse thickening with inereased Tl and
T2 MR signals for up to 12 months. In the asymp-
tomatie healed state the tendons remain thiekened
but the low MR signal eharaeteristies of the normal
tendon should return. Sonographically, the abnor-
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,
both of whieh require ultrasound or MRI for diag-
nosis.
Studies utilising ultrasound (WILEY et al. 1997),
MRI and histology of tendon biopsies at the pre-
viously harvested site within the patellar tendon
have shown that at 2 years the defeet of the eentral
one-third of the tendon beeomes indistinguishable
from the rest of the normal tendon (NIXON et al.
1995; ADRIANI et al. 1995; KARTUS et al. 2000).
In about 25% of eases there is persistenee of the
tendon defeet (LIU et al. 1996; BERNICKER et al.
1998). In addition, histologieal assessment at the
bone insertion points shows an absenee of the
normal fibroeartilage zone at the end of the repaired
tendon. These morphological changes have also
been shown in animal studies (PROCTOR et al.
1997), wh ich indicate that the repaired portion of
the tendon does not have the same ultra-struetural
and biomeehanical properties as the normal tendon
by the second year of the postoperative period. In
the most recent study of 19 patients using ultra-
sound, MRI and histology, the authors eoncluded
that reharvesting of the healed patellar tendon at
least up to 2 years after the primary harvest is not
recommended for revising ACL reeonstructions as
this may be biomeehanically unsound (KARTUS et
al. 2000).
195
11.3.4.4
Patelia/Pateliofemoral Joint
Patellar maltraeking, assessed by axial imaging (CT
or MRI) in static and kinematic studies, is not uncom-
mon postoperatively. Residual lateralisation of the
patella often persists despite primary repair of the
medial soft tissue struetures for acute patellar dislo-
eation, or after a generous lateral retinaeular release
for patellar maltracking. Less often the patella can
drift medially postoperatively as a result of an over-
zealous lateral release, which can be clinically signifi-
cant. Patellar maltraeking can also be seen after BPTB
harvesting, but it may have existed prior to the sur-
gery. It is eommon to see a low-Iying patella (patella
baja) in the first 6 months after a BPTB procedure,
with areturn to the normal position by 12 months
(MINK et al. 1993). Transversely orientated stress frac-
tures have been reported and rarely complete avul-
sive fractures of the patella can also eomplicate the
harvesting procedure if a large bone plug has signifi-
cantly weakened the patellar bone strength.
11.4
Conclusion
This chapter has particularlyfocussed on cruciate lig-
ament surgery and assessment following such repair,
meniscal surgery, and the assessment of articular car-
tilage disorders and their treatment. There are other
intra-articular eauses of ongoing symptoms, includ-
ing loose bodies, osteoneerosis of the femoral and
tibial condyles and reflex sympathetie dystrophy, all
of which have been eovered elsewhere in the book.
References
Adriani E, Mariani PP, Maresea G, Santori N (1995) Healing
of the patellar tendon after harvesting of its mid-third for
anterior eruciate ligament reeonstruetion and evolution of
the unclosed donor site defeet. Knee Surg Sports Traumatol
Arthrose 3:l38-143
Allen PR, Denham RA, Swan AV (1984) Late degenerative
ehanges after meniseeetomy. Faetors affeeting the knee
after operation. J Bone Joint Surg Br 66:666-671
Alparslan L, Minas T, Winalski CS (2001) Magnetic resonanee
imaging of autologous ehondroeyte implantation. Semin
Ultrasound CT MR 22:341-351
Applegate GR, Flannigan BD, Tolin BS, Fox JM, Dei Pizzo
W (1993) MR diagnosis of reeurrent tears in the knee:
value of intraarticular eont material. AJR Am J Roentgenol
16:821-825
196
Arnoczky SP, Cooper TG, Stadelmaier DM, Hannafin JA (1994)
Magnetie resonance signals in healing menisci: an experi-
mental study in dogs. Arthroscopy 10:552-557
Bellelli A, Adriani E, Margheritini F, Camillieri G, Della Rocca
C, Mariani pp (1999) Synovial healing in reconstructed
cruciate ligaments. Our personal experience compared in
single interventions and combined reconstruction (in Ital-
ian). Radiol Med (Torino) 98:454-461
Bergfeld JA, McAliister DR, Parker RD, Valdevit AD, Kambie
HE (2001) A biomechanieal comparison of posterior cruci-
ate ligament reconstruction techniques. Am J Sports Med
29:129-136
Bernieker JP, Haddad JL, Lintner DM, DiLiberti TC, Bocell JR
(1998) Patellar tendon defect during the first year after ante-
rior cruciate ligament reconstruction: appearance on serial
magnetie resonance imaging. Arthroscopy 14:804-809
Bolano LE, Grana WA (1993) Isolated arthroscopie partial
meniscectomy. Functional radiographie evaluation at five
years. Am J Sports Med 21:432-437
Bronstein R, Kirk P, Hurley J (1992) The usefulness of MRI
in evaluating menisci after meniscus repair. Orthopaedies
15:149-152
Cassar-Pullicino VN, McCail IW, Strover AE (1994) MRI of
the knee following prosthetie cruciate ligament reconstruc-
tion. Clin RadioI49:89-99
Cassidy RE, Shaffer AJ (1981) Repair of peripheral meniscus
tears. A preliminary report. Am J Sports Med 9:209-214
Cheung Y, Magee TH, Rosenberg ZS, Rose DJ (1992) MRI of
anterior cruciate ligament reconstruction. J Comput Assist
Tomgr 16:134-137
ClancyWG Jr, Smith L (1991) Arthroscopie anterior and poste-
rior cruciate ligament reconstruction technique. Ann Chir
GynaecoI80:141-148
Corsetti JR, Jackson DW (1996) Failure of anterior cruciate
ligament reconstruction: the biologie basis. Clin Orthop
325:42-49
Coupens SD, Yates CK, Sheldon C, Ward C (1992) Magnetie
resonance imaging evaluation of the patellar tendon after
use central one-third for anterior cruciate ligament recon-
struction. Am J Sports Med 20:332-335
Covall DJ, Wasilewski SA (1992) Roentgenographie changes
after arthroscopie meniscectomy: five year follow-up in
patients more than 45 years old. Arthroscopy 8:242-246
Cross MJ, Roger G, Kujawa P,Anderson IF (1992) Regeneration
of the semitendinosus and gracilis tendons following their
transection for repair of the anterior cruciate ligament. Am
J Sports Med 20:221-223
Dandy DJ, Jackson RW (1975) Meniscectomy and chondro-
malacia of the femoral condyle. J Bone Joint Surg Am
57:1116-1119
DeHaven KE (1985) Meniscus repair in the athlete. Clin Orthop
198:31-35
Deutsch AL, Mink JH, Fox JM, Arnoczky SP, Rothman BJ,
Stoller DW, Cannon WD Jr (1990) Peripheral meniscal tears:
MR findings after conservative treatment or arthroscopie
repair. Radiology 176:484-488
Deutsch AL, Mink JH, Fox JM, Friedman MJ, Howell SM (1992)
The postoperative knee. Magn Reson Q 8:23-54
Drez DJ Jr, DeLee J, Holden JP, Arnoczky S, Noyes FR,
Roberts TS (1991) Anterior cruciate ligament reconstruc-
tion using bone-patellar tendon-bone allografts. A biologi-
cal and biomechanieal evaluation in goats. Am J Sports Med
19:256-263
V. N. Cassar-Pullicino and S. N. J. Roberts
Eggli S, Wegmuller H, Kosina J, Huckell C, Jakob RP (1995)
Long-term results of arthroscopie meniscal repair.An anal-
ysis of isolated tears. Am J Sports Med 23:715-720
Einhorn TA (1995) Enhancement of fracture-healing. J Bone
Joint Surg Am 77:940-956
Eriekson SJ, Prost RW, Timins ME (1993) The "magie angle"
effect: background physies and clinieal relevance. Radiol-
ogy 188:23-25
Fahey M, Indelieato PA (1994) Bone tunnel enlargement after
anterior cruciate ligament replacement. Am J Sports Med
22:410-414
Fairbank TJ (1948) Knee joint changes after meniscectomy. J
Bone Joint Surg Br 30:664-670
Farley TE, Howell SM, Love KF, Wolfe RD, Neumann CH
(1991) Meniscal tears: MR and arthrographie findings after
arthroscopie repair. Radiology 180:517-522
Graf B, Uhr F (1988) Complications of intra-artieular anterior
cruciate reconstruction. Clin Sports Med 7:835-848
Hangody L, Kish G, Karpati Z, Szerb I, Udvarhelyi I (1997)
Arthroscopie autogenous osteochondral mosaieplasty for
the treatment of femoral condylar artieular defects. A pre-
liminary report. Knee Surg Sports Traumatol Arthrosc
5:262-267
Hanks GA, Gause TM, Sebastianelli WJ, O'Donneli CS, Kalenak
A (1991) Repair of peripheral meniscal tears: open versus
arthroscopie technique. Arthroscopy 7:72-77
Harter RA, Osternig LR, Singer KM, James SL, Larson RL, Jones
DC (1988) Long-term evaluation of knee stability and func-
tion following surgieal reconstruction for anterior cruciate
ligament insufficiency. Am J Sports Med 16:434-443
Howell SM, Clark JA (1992) Tibial tunnel placement in anterior
cruciate ligament reconstructions and graft impingement.
Clin Orthop 283:187-195
Howell SM, Berns GS, Farley TE (1991a) Unimpinged and
impinged anterior cruciate ligament grafts: MR signal
intensity measurements. Radiology 179:639-643
Howell SM, Clark JA, Blasier RD (1991b) Serial magnetie res-
onance imaging of hamstring anterior cruciate ligament
autografts during the first year of implantation. A prelimi-
nary study.Am J Sports Med 19:42-47
Howell SM, Knox KE, Farley TE, Taylor MA (1995) Revas-
cularization of a human anterior cruciate ligament graft
during the first two years of implantation. Am J Sports Med
23:42-49
Howell SM, Gittins ME, Gottlieb JE, Traina SM, Zoellner TM
(2001) The relationship between the angle of the tibial
tunnel in the coronal plane and loss of flexion and anterior
laxity after anterior cruciate ligament reconstruction. Am J
Sports Med 29:567-574
Jansson KA, Karjalainen PT, Harilainen A, Sandelin J, Soila K,
Tallroth K,Atonen HJ (2001) MRI of anterior cruciate liga-
ment repair with patellar and hamstring tendon autografts.
Skeletal RadioI30:8-14.
Johnson LL (1993) The outcome of a free autogenous semiten-
dinosus tendon graft in human anterior cruciate reconstruc-
tive surgery: a histologieal stydy. Arthroscopy 9: 131-142
Johnson DL, Bealle D (1999) Meniscal allograft transplanta-
tion. Clin Sports Med 18:93-108
Johnson DL, Either DB, Vanarthos WJ (1996) Herniation of
the patellar fat pad through the patellar tendon defect
after autologous bone-patellar tendon-bone anterior cruci-
ate ligament reconstruction. A case report. Am J Sports Med
24:201-204
The Postoperative Knee 1: Menisci, Cruciate Ligaments, Cartilage
Kannus P, Jarvinen M (1987) Conservatively treated tears of
the anterior cruciate ligament. Long-term results. J Bone
Joint Surg Am 69:1007-1012
Kartus J, Rostgard-Christensen L, Movin T, Lindahl S, Ejerhed
L, Karlsson J (2000) Evaluation of harvested and normal
patellar tendons: a reliability analysis of magnetic res-
onance imaging and ultrasonography. Knee Surg Sports
Traumatol Arthrosc 8:275-280
Kitziger KJ, DeLee JC (1990) Failed partial meniscectomy. Clin
Sports Med 9:641-660
Lajtai G, Humer K, Aitzetmuller G, Unger F, Noszian I, Orthner
E (1999) Serial magnetic resonance imaging evaluation of
a bioabsorbable interference screw and the adjacent bone.
Arthroscopy 15:481-488
Lim PS, Schweitzer ME, Bhatia M, Giuliano V, Kaneriya PP,
Senyk RM, Oliveri M, Johnson W, Amster B, Parker L (1999)
Repeat tear of postoperative meniscus: potential MR imag-
ing signs. Radiology 210:183-188
Liu SH, Hang DW, Gentili A, Finerman GA (1996) MRI and
morphology of the insertion of the patellar tendon after
graft harvesting. J Bone Joint Surg Br 78:823-826
Manaster BJ, Remley K, Newman AP,Mann FA (1988) Kneelig-
ament reconstruction: plain film analysis. AJR Am J Roent-
genoI150:337-342
Mariani PP, Adriani E, Bellelli A, Maresca G (1999) Magnetic
resonance imaging of tunnel placement in posterior cruci-
ate ligament reconstruction. Arthroscopy 15:733-740
Maywood RM, Murphy BJ, Uribe JW, Hechtman KS (1993) Eval-
uation of arthroscopic anterior cruciate ligament recon-
struction using magnetic resonance imaging. Am J Sports
Med 21:523-527
Mink JH, Deutsch AL (1989) Magnetic resonance imaging of
the knee. Clin Orthop 244:29-47
Mink JH, Reicher MA, Crues JV III, Deutsch AL (1993) MRI of
the knee, 2nd edn. Raven Press. New York
Mink JH, Levy T, Crues JV 3rd (1998) Tears of the anterior
cruciate ligament and menisci of the knee: MR imaging
evaluation. Radiology 167:769-774
Muellner T, Egkher A, Nikolic A, Funovics M, Metz V (1999)
Open meniscal repair: clinical and magnetie resonance
imaging findings after twelve years. Am J Sports Med
27:16-20
Munk PL, Vellet AD, Fowler PJ, Miniaci T, Crues JV 3rd (1992)
Magnetie resonance imaging of reconstructed knee liga-
ments. Can Assoc Radiol J 43:411-419
Murakami Y, Sumen Y, Ochi M, Fujimoto E, Adachi N, Ikuta
Y (1998) MR evaluation of human anterior cruciate liga-
ment autograft on oblique axial imaging. J Comput Assist
Tomogr 22:270-275
Murakami Y, Sumen Y, Ochi M, Fujimoto E, Deie M, Ikuta Y
(1999) Appearance of anterior cruciate ligament autografts
in their tibial bone tunnels on oblique axial MRI. Magn
Reson Imaging 17:679-687
Nixon RG, SeGall GK, Sax SL,Cain TE, Tullos HS (1995) Recon-
stitution of the patellar tendon donor site after graft har-
vest. Clin Orthop 317:162-17l
Potter HG, Rodeo SA, Wickiewicz TL, Warren RF (1996) MR
imaging of meniscal allografts: correlation with clinical
and arthroscopie outcomes. Radiology 198:509-514
Proctor CS, Jackson DW, Simon TM (1997) Characterization of
the repair tissue after removal of the central one-third of
the patellar ligament. An experimental study in goat model.
J Bone Joint Surg Am 79:997-1006
197
Race A, Amis AA (1998) PCL reconstruction. In vitro biome-
chanical comparison of "isometrie" v single and double-
bundled "anatomie" grafts. J Bone Joint Surg Br 80:173-179
Rak KM, Gillogly SD, Schaefer RA, Yakes WF, Liljedahl RR
(1991) Anterior cruciate ligament reconstruction: evalua-
tion with MR imaging. Radiology 178:553-556
Rangger C, Klestil T, Gloetzer W, Kemmler G, Benedetto KP
(1995) Osteoarthritis after arthroscopie partial meniscec-
tomy. Am J Sports Med 23:240-244
Recht MP, Piraino DW, Cohen MA, Parker RD, Bergfeld JA
(1995) Localized anterior arthrofibrosis (cyclops lesion)
after reconstruction of anterior cruciate ligament: MR
imaging findings. AJR Am J RoentgenoI165:383-385
Recht M, Bobie V, Burstein D, Disler D, Gold G, Gray M, Kramer
J, Lang P, McCauley T, Winalski C (2001) Magnetie res-
onance imaging of artieular cartilage. Clin Orthop 391
(Suppl):S379-S396
Riehardson JB, Caterson B, Evans EH, Ashton BA, Roberts S
(1999) Repair of human artieular cartilage after implan-
tation of autologous chondrocytes. J Bone Joint Surg Br
81: 1064-1068
Roos H, Lauren M, Adalberth T, Roos EM, Jonsson K, Lohman-
der LS (1998) Knee osteoarthritis after meniscectomy: prev-
alence of radiographie change after twenty-one years com-
pared with matched controls. Arthritis Rheum 41:687-693
Sanders TG, Mentzer KD, Miller MD, Morrison WB, Campbell
SE, Penrod BJ (2001) Autogenous osteochondral "plug"
transfer for the treatment of focal chondral defects: post-
operative MR appearance with clinieal correlation. Skeletal
RadioI30:570-578
Schatz JA, Potter HG, Rodeo SA, Hannafin JA, Wiekiewiez TL
(1997) MR imaging of anterior cruciate ligament recon-
struction. AJR Am J RoentgenoI169:223-228
Schiek F, Dammann F, Lutz 0, Claussen CD (1995) Adapted
techniques for clinieal MR imaging of tendons. MAGMA
3:103-107
Schimmer RC, Brulhart KB, Duff C, Glinz W (1998) Arthroscopie
partialmeniscectomy: a 12-year follow-up and two-step eval-
uation of the long-term course. Arthroscopy 14: 136-142
Sciulli RL, Boutin RD, Brown RR, Nguyen KD, Muhle C, Lektr-
akul N, Pathria MN, Pedowitz R, Resniek D (1999) Evalu-
ation of the postoperative meniscus of the knee: a study
comparing conventional arthrography, conventional MR
imaging, MR arthrography with iodinated contrast mate-
rial, and MR arthrography with gadolinium based contrast
material. Skeletal RadioI28:508-5 14
Smith DK, Totty WG (1990) The knee after partial meniscec-
tomy: MR imaging features. Radiology 176:141-144
Stockle U, Hoffmann R, Schwedke J, Lubrieh J, Vogl T, Sudkamp
NP, Haas N (1998) Anterior cruciate ligament reconstruc-
tion: the diagnostie value of MRI. Int Orthop 22:288-292
Tsai CL, Liu CN, Lim AC (1992) Meniscal repair with autog-
enous periosteum and fibrin adhesive system (in Chinese).
Zhonghua Yi Xue Za Zhi (Taipei) 49: 170-176
Tuite MJ, De Smet AA (1996) MR of the postoperative knee.
Top Magn Reson Imaging 8:2-14
Van Trommel MF, Simonian PT, Potter HG, Wiekiewiez TL
(1998) Arthroscopic meniscal repair with fibrin clot of
complete radial tears of lateral meniscus in the avascular
zone. Arthroscopy 14:360-365
Verstraete KL, Huysse WCJ, Steines D, Lang P, Declercq D,
Rousseaux M (2000) Assessment of osteochondral auto-
grafts of the knee with CT, conventional MRI, dynamie
198
contrast-enhanced and cartilage thiekness images. Sev-
enth Annual Meeting, European Society of Musculoskeletal
Radiology, Leiden
Vogl TJ, Schmitt J, Lubrieh J, Hochmuth K, Diebold T, Dei Tredi-
ci K, Sudkamp N (2001) Reconstructed anterior cruciate
ligaments using patellar tendon ligament grafts: diagnostie
value of contrast-enhanced MRI in a 2 year follow-up regi-
men. Eur Radiolll: 1450-1456
Warden WH, Friedman R, Teresi LM, Jackson DW (1999) Mag-
netie resonance imaging of bioabsorbable polylactic acid
interference screws during the first 2 years after anterior
V. N. Cassar-Pullicino and S. N. J. Roberts
cruciate ligament reconstruction.Arthroscopy 15:474-480
Whitman TL, Diduch DR (1998) Transient posterior knee pain
with the meniscal arrow. Arthroscopy 14:762-763
Wiley JP, Bray RC, Wiseman DA, Elliott PD, Ladly KO, Vale
LA (1997) Serial ultrasonographie imaging evaluation of
the patellar tendon after harvesting its central one third
for anterior cruciate ligament reconstruction. J Ultrasound
Med 16:251-255
Yamato M, Yamagishi T (1992) MRI of patellar tendon ante-
rior cruciate ligament autografts. J Comput Assist Tomogr
16:604-607
12 The Postoperative Knee 2:
Arthroplasty, Arthrodesis, Osteotomy
THOMAS H. BERQUIST

CONTENTS (Fig. 12.1) became the "gold standard" upon whieh

12.1 Introduction 199
further modifications were developed. This section
12.2 Knee Arthrop1asty 199 will discuss basies of knee arthroplasty with empha-
12.2.1 Indications/Contraindications 199 sis on imaging of patients with knee arthroplasty and
12.2.2 Component Design 199 its potential complications.
12.2.3 Component Selection 200
12.3 Postoperative Imaging and Complications 205
12.3.1 Complications 205
12.3.2 Extensor Mechanism Dysfunction 205 12.2.1
12.3.3 Loosening 206 IndicationslContraindications
12.3.4 Infection 207
12.3.5 Fractures 209 The goals of knee arthroplasty are to relieve pain,
12.3.6 Instability 209
12.3.7 Other Comp1ications 210
improve motion and function, and enhance stability
12.3.8 Deep Venous Thrombosis 211 (BERQUIST 1995; ENGH and AMMEEN 1997; INsALL et
12.4 Arthrodesis 211 al. 1989; STUART 1991).
12.5 Osteotomy 212 Knee arthroplasty is contraindicated in patients
References 214 with active infection or fusion. Diminished strength
in the extensor mechanism, neuropathie arthropathy,
and vascular insufficiency are relative contraindica-
tions. In addition, patients with excessive occupa-
12.1 tional or physical demands as well as obese patients
Introduction are considered poor candidates for knee replacement
(BERQUIST 1995; STUART 1991).
It is essential for imagers to understand the indi-
cations, surgical approaches, and potential compli-
cations of reconstructive procedures on the knee. 12.2.2
This chapter focuses on arthroplasty, osteotomy, and Component Design
arthrodesis of the knee. Imaging techniques playa
vital role in evaluating surgical results and potential Knee replacement implants are designed to reproduce
complications of these procedures. certain knee motions depending upon the constraint
of the prosthesis. Constraint is a term used to refer to
the stability of the knee offered by the configuration
or contour of the components (BERQUIST 1995). An
12.2 unconstrained prosthesis is a resurfacing prosthesis
Knee Arthroplasty that relies on intact collateral and cruciate ligaments.
Tibial components are relatively Hat and normal knee
Total knee arthroplasty arose as an alternative to motion is maintained (BERQUIST 1995; STUART 1991).
arthrodesis in the 1930s (STUART 1991). Early hinged A semieonstrained prosthesis has a concave tibial
components frequently failed due to loosening (BER- insert (Fig. 12.1). Instability is reduced by selecting
QUiST 1995). In the 1970s, the condylar design a tibial insert with appropriate thiekness to keep the
collateral ligaments at proper tension. The anterior
T. H. BERQUIST, MD, FACR cruciate ligament is sacrificed and the posterior cruci-
Diagnostic Radio1ogy, Mayo C1inic, 4500 San Pab10 Road, Jack- ate ligament is preserved (Fig.12.1a).A posterior sta-
sonville, FL 32224, USA bilized condylar prosthesis is used when the posterior

A.M. Davies et al. (eds.), Imaging of the Knee

© Springer-Verlag Berlin Heidelberg 2003
 200 Th. 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
box in the femoral component (Fig. 12.2) provide sta-
bility in this semieonstrained system (BERQUIST 1995;
COYTE et al.1999).
Constrained prostheses are used in patients with
inadequate soft tissue support or bone loss or for revi-
sion arthroplasty. Linked or nonlinked designs are
available (Fig. 12.3). Fully hinged designs (Fig. 12.3)
frequently fail.
12.2.3
Component Selection
Considerable research has gone into developing
the most appropriate alloys and plastic for joint
implants. Stainless steel, though effective in elderly
patients, has largely been abandoned in favor of tita-
nium- and cobalt-based alloys. Cobalt-chromium
or cobalt-chromium-molybdenum alloys and tita-
nium-aluminum-vanadium alloys are most com-
monlyused in conjunction with ultra-high molecu-
lar weight polyethylene artieular surfaces (BERQUI-
ST 1995; BRYAN and PETERS 1973; FRIEDMAN et al.
1993).
Component design varies for tibial, femoral, and
patellar implants. Tibial components may be sym-
metric (Fig. 12.4) or anatomie (Fig. 12.5) (medial
and lateral polyethylene inserts configured to the
size of the tibial condyle). Tibial trays may be metal
backed with varying thickness of inserts (Figs. 12.4,
12.5) or all polyethylene (Fig. 12.6). Components
may be designed for ingrowth, in which case screws
can be used for additional fixation or for cement
 The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 201

Fig.12.2. a Photograph of a posterior stabilized semicon-

strained system with a central tibial post (arrow). The box
for the central post provides a different appearance on the lat-
eral radiograph (b), indicating the design and resection of the
posterior cruciate ligament. Note the thickness of the femo-
ral component (arrowheads) and lack of pegs compared with -..:o_ _ -, b
Fig.12.1c

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
ean be used when there is bone loss (Fig. 12.8).
Systems are also available for unieompartmental
arthroplasty (ApEL et al. 1991; BERQUIST 1995;
STUART 1991).
Femoral eomponents may be porous coated to
allow bone ingrowth or they may be designed for
eement fixation. Symmetrie and anatomie designs
are available, similar to the tibial eomponents. Aug-
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
et al. 1999). There are fewer loosening problems with
femoral components than with tibial ones (BERQUIST
1995; COLLIER et al. 1991).
Patellar resurfacing may not be required. Recent
reports suggest that the results of patellar resur-
facing and retention of the patella are similar. How-
ever, future repair is still required in 10% of patients
(BARRACK et al.1997).
Patellar surface replacement is still most commonly
performed to reduce pain and prevent later complica-
tions (BARRACK et al.1997; BERQUIST 1995; MERKow
et al. 1985). Patellar components may be all polyethyl-
ene or metal backed (Fig. 12.9). Most surgeons prefer
cemented, all polyethylene inserts (see Fig.12.11)
(BAUER and SCHELS 1999; BERQUIST 1995).
Unicompartmental arthroplasty has regained some
support in recent years (BERQUIST 1995). This tech-
 The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 203

Fig. 12.8. Photograph of the Johnson and Johnson P.F.c. modu-

lar knee with augmentation sterns (1), augmentation wedges
(2), and components assembled (3)

Fig.12.7. a The porous surface of Miller-Galante prosthesis

(titanium fiber mesh) with four porous pegs which can be
used for screw fixation. (Courtesy of Zimmer, Warsaw, Indi-
ana). b AP view of the knee shows the same tibial compo- a
nent with four screws for fixation. There is poorly defined
lucency (arrows) on the right due to infection. (From BER- b
QUIST 1995)

nique is indicated as an alternative to tibial osteoto-

my or tricompartmental arthroplasty in patients with
unicompartmental osteoarthritis (BERQUIST 1995;
COVENTRY 1984). There is less bone loss and soft Fig. 12.9. Photographs of all poly-
ethylene (a) and metal-backed (b)
tissue injury with this technique (Fig.12.10). There- patellar components
fore, it may be beUer suited as an initial procedure in
younger patients.
 204 Th. H. Berquist

Fig.12.10. AP (a) and lateral (b) radiographs of a medial compartment arthroplastywith porous-coated metal-backed components

Fig.12.11. a Standing AP view of the knees in a patient with cemented How-

mediea Duracon implants. The femoral-tibial angle on the left is 5° valgus.
The right tibial tray is 90° to the tibial axis. The femoral component is 98°
to the femoral axis. b Lateral view of the knee demonstrating the positions
of the femoral and tibial components. The tibial component should be 90° to
the tibial axis as on the AP view. If the component is >90°, it is considered
extended, and if it is less than 90°, it is considered fiexed. The femoral com-
ponent should be perpendicular to the femoral axis. The patella should be
positioned so that its height (A) and patellar tendon length (B) are at about a
1: 1 ratio. c Patellar view demonstrating symmetrie positioning of the patella
c with an all-polyethylene implant
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy
12.3
Postoperative Imaging and Complications
Anteroposterior (AP) and lateral radiographs are
obtained in the immediate postoperative period to
check alignment and component position and to
exclude obvious problems such as dislocation or peri-
prosthetic fracture (BERQUIST 1995; SCHNEIDER et
al. 1984). When the patient can tolerate a complete
knee series this should be obtained to provide a base-
line study. It should include standing AP (most prefer
a full-Iength hip to ankle study), lateral, and patellar
views. We prefer to fluoroscopically position the AP
and lateral views to assure optimal visualization of
the component-bone and bone-cement interfaces.
Flexion and extension lateral views can be obtained
to grade range of motion. Patients are usually restud-
ied at 6 and 12 months. Yearly follow-up is adequate
after the first year unless symptoms dictate more fre-
quent evaluation.
Specific features should be evaluated on each radio-
graphicview. The position of the tibial andfemoral com-
ponents and the femoral-tibial angle should be assessed
on the standing AP view of the knee (Fig.12.11a). The
tibial tray should be at 90° to the tibial axis and cover
+85% of the bony condyle. If the tibial tray is >90° it is
considered in valgus and if it is <90° it is considered in
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-
nent, and patella should also be assessed on the lat-
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
extended and if <90°, flexed. The femoral axis should
be perpendicular to a line along the base of the com-
ponent (Fig. 12.11 b). Patellar height should be about
the same as the patellar tendon length. Motion can
also be assessed with flexed and extended lateral
views. Approximately 65° of flexion is required to
climb stairs and 70° for anormal gait. Descending
stairs requires 85° and rising from a sitting position,
105° of flexion (BERQUIST 1995; BURGER et al. 1991).
Lucent zones should be assessed on AP and later-
al views. Fluoroscopic positioning is most effective
to achieve optimal evaluation of the bone-cement or
metal-bone interfaces.
The patellar view should be evaluated for symme-
try, patellar tilt, and loosening (Fig. 12.l1c).
205
12.3.1
Complications
Complications of total knee arthroplasty are sum-
marized in Table 12.1. Certain complications can be
treated conservatively. Loosening, infection, osteoly-
sis, fracture, and instability require more aggressive
therapy (BAUER and SCHELS 1999; BERQUIST 1995;
BOYD et al. 1993; COYTE et al. 1999; ELLINGSEN and
RAND 1994; GILLS and MILLS 1991; KIM 1990).
Table 12.1. Complications of total knee arthroplasty
Wound healing
Infection
Extensor mechanism dysfunction
Loosening
Instability
Fractureldislocation
Deep venous thrombosis
Pulmonary emboli
Pes anserinus bursitis
Synovitis
Osteolysis
Peroneal nerve palsy
BAUER and SCHELS (1999); BERQuIsT (1995); BoYD et
al. (1993); COYTE et al. (1999); CUCKLER et al. (1991);
ELLINGSEN and RAND (1994); ENGH and AMMEEN (1997);
GILLS and MILLS (1991); IDUSUYI and MORREY (1996); KIM
(1990); SAMBATAKAKIS et al. (1991); WILSON et al. (1990)
12.3.2
Extensor Mechanism Dysfunction
Extensor mechanism dysfunction is the most
common complication of total knee arthroplasty,
accounting for up to 50% of revision procedures
(BARRACK et al. 1997; BERQUIST 1995; MERKow et al.
1985). Most complications occur in the first 2 years
after arthroplasty (BERQUIST 1995; BLOEBAUM et al.
1998; BOYD et al. 1993). Complications include soft
tissue imbalance with abnormal patellar position
and/or abutment on the femoral condyle (Fig.12.12),
dislocation, fracture, loosening (Fig. 12.13), compo-
nent wear, and failure or rupture of the patellar
tendon or quadriceps mechanism (Fig.12.l4) (BER-
QUIST 1995; BLOEBAUM et al. 1998; BOYD et al. 1993;
MERKow et al. 1985; SAMBATAKAKIS et al. 1991).
Serial radiographs are generally adequate for diagno-
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

Fig.12.14. Lateral view of the knee with posterior stabilized

revision components. There is a low-lying patella (arrow) due
to quadriceps rupture

12.3.3
Loosening

b Loosening was very common (20%-30%) with hinged

Fig.12.13. Lateral (a) and patellar (b) views demonstrate dis-
placement (loosening) of the patellar component (arrow). The
patella is subluxed laterally
components (see Fig.12.3c) (BERQUIST 1995; HUE
and FITZGERALD 1990). The development of condy-
lar components (see Fig. 12.1) and recent modifica-
tions have reduced the frequency of loosening to only
1%-2% (APEL et al.1991; BERQUIST 1995; PATEL et al.
1991; SILVERTON et al. 1996).
Patients with mechanicalloosening usually pres-
ent with pain. Factors that contribute to loosening
include ligament laxity, improper component sizing,
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 207

obesity, and tibial bone loss. Tibial component loos-

ening occurs most frequently (APEL et al. 1991; BER-
QUIST 1995; BoYD et al. 1993; SCHNEIDER et al. 1984;
SILVERTON et al. 1996).
Serial radiographs, especially if fluoroscopieally
positioned, are very useful for the detection of com-
ponent loosening (BERQUIST 1995). Lucent zones at
the metal-bone or bone-cement interfaces are easily
evaluated on properly positioned serial radiographs
(BERQUIST 1995; BOYD et al. 1993; SCHNEIDER et al.
1984). Lucent lines less than 2 mm are often iden-
tified in the peripheral zones of the tibial (-65%)
and femoral (15%-20%) components (BERQUIST
1995; BURGER et al. 1991; GILLS and MILLS 1991).
Lucent lines <2 mm whieh are incomplete do not
usually represent loosening. The progression of
lucent lines in width (>2 mm) or extent of surface
involvement is more significant (Fig. 12.15). Lucent
lines or progressive shedding ofbeads indieate loos-
ening on uncemented components (BERQUIST 1995;
a
MANASTER 1995; SILVERTON et al. 1996). Other
indieations of component loosening include cement
fracture, collapse of adjacent trabecular bone, sub-
sidence, polyethylene wear, and shift of component
position (APEL et al. 1991; BERQUIST 1995; BOYD et
al. 1993).
Radiographie features may be inconclusive. In the
symptomatie patient, further imaging may be neces-
sary. Radionuclide bone scans (technetium-99m) are of
little value in the first 8-10 months following surgery.
Bone scans may be positive in asymptomatic patients
for up to 1 year (BERQUIST 1995; MANASTER 1995).
Subtraction arthrography is not as useful in the
knee as in the hip (BERQUIST 1995). The joint space
and bursal compartments of the knee have a large b
capacity whieh decompresses the joint, making it dif-
ficult to detect contrast extending between the bone
or cement and components.

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
overall incidence at our institution is 1.2% (BER-
QUIST 1995). The incidence of infection also varies
with the type of implant used. Hinged prostheses
became infected in 19%; stabilized components in
5.1 %; resurfacing designs in 2.8%, and unicompart-
mental components in only O. 8%. The incidence of
infection with revision is also higher (9%) (BERQUIST
1995; RAND 1993; WILSON et al. 1990).
Radiographs are normal in nearly 75% of patients
with infection (BERQUIST 1995; MANASTER 1995).
However, serial radiographs may demonstrate irreg-
ular lucency about the components, signs of loosen-
ing, and soft tissue swelling (Fig. 12.16). (BAUER and
SCHELS 1999; BERQUIST 1995; MANASTER 1995).
Radionuclide imaging is useful in patients with
suspected infection (BENGSTON and KNUTSON 1991;
KIM 1990; MINIACI et al. 1989; STUART 1991).
Most prefer a dual-radionuclide technique employing
technetium-99m methylene diphosphonate (MDP)
a sions
b c
Th. H. Berquist
or sulfur colloid scans combined with indium-lU
labeled leukocytes (BERQUIST 1995; MANASTER 1995;
PALESTRO et al. 1991). The accuracy of combined
technetium and leukocyte scans approaches 96%
(Fig.12.17) (BERQUIST 1995; PALESTRO et al. 1991;
WILSON et al.1990).
Infections are most often due to Staphylococcus
aureus (64%), Pseudomonas aeruginosa (12%), Esch-
erichia co li (5%), and an aerobic organisms (6%)
(BAUER and SCHELS 1999; BENGSTON and KNUTSON
1991; BERQUIST 1995; PALESTRO et al.1991). Manage-
ment varies with patient condition and virulence of
the organism (BERQUIST 1995; CUCKLER et al. 1991;
RAND 1993; WILSON et al. 1990). Antibiotic therapy
without component rem oval may be successful if
infection is caused by low-virulence organisms when
components are not loose. With more virulent organ-
isms, component removal, debridement, and intra-
articular and intravenous antibiotics are indicated
Fig. 12.16. Standing AP (a) and lateral (b, c) radiographs in a
patient with bilateral infection. There is bone loss, tibial com-
ponent subsidence, and soft tissue swelling with joint effu-
 The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy
b placed in the joint. The component was revised after the infec-
(Fig.12.17) (BERQUIST 1995; CUCKLER et al. 1991).
Arthrodesis can also be used in younger patients,
patients with greater mobility, and those with viru-
lent infections (CUCKLER et al. 1991; ELLINGSEN and
RAND 1994; NICHOLS et al. 1991; RAND 1993).
12.3.5
Fractures
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).
Fractures may occur during component insertion. In
this setting, fractures are usually undisplaced and
they can be treated during the arthroplasty (ENGH
and AMMEEN 1997). Postoperative fractures may
result from systemic dis orders (osteoporosis) or com-
209
Fig.12.17a-c. Infected left knee arthroplasty. a AP radiograph
demonstrates lucency under the tibial tray. b Indium-lll
labeled leukocyte scan is positive (arrow). c Components were
removed and an antibiotic-impregnated methacrylate spacer
tion cleared
ponent loosening, or be due to the component
type. Fractures are more common with hinged pros-
theses (BERQUIST 1995; HUE and FITZGERALD 1990).
They are most common in the supracondylar region
(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-
sive surgical treatment of fractures is preferred over
bed rest or long periods of nonambulation (COR-
DEIRO et al. 1990; GHAZAVI et al. 1997).
12.3.6
Instability
Ligament instability has been reported in up to 13%
of patients following knee arthroplasty (BERQUIST
1995; SAMBATAKAKIS et al. 1991). Preoperative defor-
 210
b Other (omplications
Fig.12.18. AP (a) and lateral (b) radiographs of a healing
supracondylar fracture with malunion. The femoral compo-
nent enters the anterior fracture Hne
mities are often corrected and the thickness of the
polyethylene spacer is selected to achieve proper leg
length and joint stability. Postoperative instability
may be varus, valgus, anteroposterior, or rotational.
Clinical examination is usually sufficient for diagno-
sis. Radiographs with varus, valgus, anteroposterior,
and rotational stress can be performed using fiuoro-
scopic guidance (BERQUIST 1995).
Th. H. Berquist
Fig.12.19. Standing AP (a) and notch (b) views of a right frac-
tured tibial tray (arrow) due to underlying osteolysis
12.3.7
Additional complications (Table 12.1) include poly-
ethylene wear, synovitis or osteolysis related to metal
corrosion, cement, or polyethylene particles, and
peroneal nerve palsy (ApEL et al. 1991; BERQUIST
1995; CHRISTENSEN et al. 1990; GHAZAVI et al. 1997;
GROSS and LENNON 1992; WESTRICH et al. 1997).
Pes anserinus bursitis is seen more frequently with
unicompartmental arthroplasty and when there is
medial overhang of the tibial tray.
Dissociation of polyethylene from the patellar and
tibial components can occur in addition to wear. Thin
polyethylene surfaces tend to have more problems
with wear. Wear particles result in giant cell reactions
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy 211

in the synovium with resulting synovitis and osteoly-

sis (BERQUIST 1995; CHRISTENSEN et al.1990).

12.3.8
Deep Venous Thrombosis

Obesity, prolonged immobilization, prior venous dis-

ease, and hyperlipidemia increase the incidence of
deep venous thrombosis after knee arthroplasty
(BERQUIST 1995; OISHI et al. 1994). There does not
seem to be a significant difference according to
whether one knee or both are operated on at the same
time. The incidence of deep venous thrombosis is
0.4%-3% (BERQUIST 1995). However, with more fre-
quent postoperative ultrasound screening, symptom-
atic and asymptomatic thrombosis may be evident in
up to 9% of patients (OISHI et al. 1994). a

12.4
Arthrodesis

Arthrodesis of the knee is not commonly performed

today (ARROYO et al. 1997; RAND 1993). The most
common indication at our institution is as a salvage
procedure in patients with skeletal neoplasms
(ARROYO et al. 1997). Arthrodesis mayaiso be useful
in patients with failed knee arthroplasty or other
failed surgical procedures. Arthrodesis mayaiso be
considered in patients who are not candidates for
knee arthroplasty (ARROYO et al. 1997; BERQUIST
1995; GHAZAVI et al. 1997).
The best candidates for arthrodesis are younger b
patients with high functional demand, patients with
poor soft tissue coverage or support, and immuno-
compromised patients (ELLINGSEN and RAND 1994). Fig. 12.20a, b. Compression arthrodesis with extern al fixation
Arthrodesis is contraindicated in patients with bilat- and bone grafting. a AP radiograph with bone graft (arrow)
and external fixation fracture. b Standing AP radio graph with
eral knee pathology, when there is severe bone loss, or
solid fusion. Note the pin tracts (arrowheads)
when there is associated abnormality in the ipsilat-
eral hip and/or ankle (ELLINGSEN and RAND 1994).
There are multiple surgical approaches to establish
a fused pain -free knee. External fixation (compression Complications associated with arthrodesis vary
arthrodesis), intramedullary nails, and plate fixation to some extent according to the procedure used.
have been employed (Figs.12.20, 12.21) (ARROYO et For example, pin tract infections are associated with
al. 1997; ELLINGSEN and RAND 1994; GHAZAVI et al. external fixation frames (ARROYO et al. 1997). Other
1997; RAND 1993; WEINER et al. 1996). Fusion rates common complications include nonunion, delayed
with external fixation and compression technique are union, bone or implant fracture, deep infection, reflex
often only about 50% (Fig.12.20) (ELLINGSEN and sympathetic dystrophy, and peroneal nerve palsy
RAND 1994). Intramedullary nailing is becoming the (ARROYO et al. 1997; ELLINGSEN and RAND 1994;
most popular technique, with success rates of over 90% GHAZAVI et al. 1997; RAND 1993; WEINER et al. 1996).
(ARROYO et al.1997; ELLINGSEN and RAND 1994). Overall, complications may be seen in 40%-55% of
 212
a
b tion (COVENTRY 1984,1987; MINIACI et al. 1989).
Fig. 12.21a, b. Failed compression arthrodesis treated with
plate and screw fixation. AP (a) and oblique (b) views show
fixation with pin tracts (arrowheads) from previous external
fixation
rh. H. Berquist
knee arthrodeses. Intramedullary nailing reduces this
to 38% and the majority of complications are not sig-
nificant (ARROYO et al. 1997).
Imaging of complications can be accomplished
using serial radiographs. Certain cases may require
fluoroscopie positioning to assure proper orientation
to evaluate the implant and osseous structures. Pin
tract infections (sequestra in pin tract) are also more
easily evaluated using fluoroscopic positioning. Stress
views can be useful in cases of suspected nonunion.
Nonunion and peroneal nerve injury may be most
optimally imaged using MRI. However, MR images
may be degraded by metal artifact. This is most
severe with plate and screw techniques (Fig. 12.21).
Artifact is less significant with intramedullary nails
and titanium implants (BERQUIST 1995).
Deep infection can be studied with combined
radionuclide studies, as described in the previous
section (BERQUIST 1995; MANASTER 1995; PALESTRO
et al. 1991; WILSON et al. 1990).
12.5
Osteotomy
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-
Most studies demonstrate success rates of 85% fol-
lowing osteotomy (MINIACI et al. 1989). However, it is
not unusual for joint degeneration to progress. Other
complications include delayed union (Fig.12.22) or
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

deep venous thrombosis, and peroneal nerve palsy References

(BERQUIST 1995; COVENTRY 1984, 1987; KIRGIS and
ALBRECHT 1992).
Serial radiographs (Fig.12.22) are most useful for
evaluating union and progressive deformity (BERQUI-
ST 1995). Standing AP views are essential to compare
pre- and postoperative measurements. Venous throm-
bosis can be evaluated using sonography. Patients
with peroneal nerve palsy can be studied with MRI as
metal artifact is not significant.
Ape! DM, Tozzi JM, Dorr LD (1991) Clinical comparison of all
polyethylene and metal backed tibial components in total
knee arthroplasty. Clin Orthop 273:243-252
Arroyo JS, Gurvin KL, Neff JR (1997) Arthrodesis of the knee
with a modular titanium intra-medullary nail. J Bone Joint
Surg 79A:26-35
Barrack RL, Wolfe MW, Waldman DA, Milicic AM, Bertot AJ
(1997) Resurfacing of the patella in total knee arthroplasty.
J Bone Joint Surg 79A:1l21-1131
Bauer TW, Schels J (1999) The pathology of total joint arthro-
The Postoperative Knee 2: Arthroplasty, Arthrodesisi, Osteotomy
plasty. 11. Mechanisms of implant failure. Skeletal Radiol
28:483-497
Bengston S, Knutson K (1991) The infected knee arthroplasty.
A 6 year follow-up of 357 cases. Acta Orthop Scand
62:301-311
Berquist TH (1995) Imaging atlas of orthopedie appliances
and prosthesis. Raven, New York
Bloebaum RD, Backus KN, Jensen JW, Scott DF, Hofmann AA
(1998) Porous-coated metal-backed patellar components in
total knee arthroplasty. J Bone Joint Surg 80A:518-528
Boyd AD, Ewald FC, Thomas WH, Poss R, Sledge CB (1993)
Long-term complieations after total knee arthroplasty
with or without patellar resurfacing. J Bone Joint Surg
75A:674-681
Bryan RS, Peters on LFA (1973) Polycentrie total knee arthro-
plasty. Orthop Clin North Am 4:575-585
Burger RR, Basch T, Hopson CN (1991) Implant salvage in
infected total knee arthroplasty. Clin Orthop 273:105-112
Christensen OM, Christiansen TJ, Johansen T (1990) Polyeth-
ylene failure in PCA unieompartmental knee prosthesis.
Acta Orthop Scand 61:578-579
Collier JP, Mayor MB, McNamara JL, Surprenant VA, Jensen
RE (1991) Analysis of the failure of 122 polyethylene
inserts from uncemented tibial components. Clin Orthop
273:232-242
Cordeiro EN, Costa RC, Carazzato JG, Silva JD (1990) Peripros-
thetie fracture in patients with total knee arthroplasties.
Clin Orthop 252:182-189
Coventry MB (1984) Upper tibial osteotomy. Clin Orthop
182:46-52
Coventry MB (1987) Proximal tibial varus osteotomy for osteo-
arthritis of the lateral compartment of the knee. J Bone
Joint Surg 69A:32-38
Coyte PC, Hawker G, Croxford R, Wright JG (1999) Rates of
revision knee replacement in Ontario, Canada. J Bone Joint
Surg 81A:773-782
Cuckler JM, Star AM, Alavi A, Noto RB (1991) Diagnosis and
management of infected total joint arthroplasty. Orthop
Clin North Am 22:523-530
Ellingsen DE, Rand JA (1994) Intra-medullary arthrodesis of
the knee after failed total knee arthroplasty. J Bone Joint
Surg 76A:870-877
Engh GA, Ammeen DJ (1997) Perprosthetie fractures adjacent
to total knee implants. J Bone Joint Surg 79A:1100-1113
Friedman RJ, Black J, Galante JO, Jacobs JJ, Skinner HB (1993)
The current concepts in orthopedies, biomaterials, and
implant fixation. J Bone Joint Surg 75A:1086-1109
Ghazavi MT, Stockly I, Yee G, Davis A, Gross AE (1997) Recon-
struction of massive bone defects with allograft in revision
total knee arthroplasty. J Bone Joint Surg 79A:17-25
Gills ES, Mills DM (1991) Long-term follow-up evaluation of
1000 consecutive cemented total knee arthroplasties. Clin
Orthop 273:66-76
Gross TP, Lennon DW (1992) Osteolytie cyst like area associ-
ated with polyethylene and metal debris after total knee
replacement with an uncemented Vitallium prosthesis. J
Bone Joint Surg 74A:I096-1102
Hue FC, Fitzgerald RH Jr (1990) Hinged total knee arthroplas-
ty. J Bone Joint Surg 65A:513-519
215
Idusuyi OB, Morrey BF (1996) Peroneal nerve palsy after total
knee arthroplasty. J Bone Joint Surg 78A:177-184
Insali JN, Dorr LD, Scott RD, Scott WN (1989) Rationale
of the knee society clinieal grading system. Clin Orthop
248:13-14
Kim Y (1990) Incidence of deep vein thrombosis after cement-
less and cemented knee replacement. J Bone Joint Surg
72B:779-783
Kirgis A, Albrecht S (1992) Palsy of the deep peroneal
nerve after proximal tibial osteotomy. J Bone Joint Surg
74A:1180-1185
Manaster BJ (1995) Total knee arthroplasty: post-operative
radiologie findings AJR 165:899-904
Merkow RL, Soundry M, Insali JN (1985) Patellar dislocation
following total knee replacement. J Bone Joint Surg
67A:1321-1327
Miniaci A, Ballmer FT, Ballmer PM, Jakob RP (1989) Proximal
tibial osteotomy: a new fixation deviee. Clin Orthop
246:250-259
Niehols SJ, Landon GG, Tallos HS (1991) Arthrodesis with dual
plates after failed total knee arthroplasty. J Bone Joint Surg
73A: 1020-1 024
Oishi CS, Grady-Benson JC, Otis SM, Colwell CW, Walker RH
(1994) The clinieal course of distal deep venous thrombosis
after total hip and total knee arthroplasty as determined by
duplex ultra-sonography. J Bone Joint Surg 76A:1658-1663
Palestro CJ, Swyer AJ, Kim AK, Goldsmith SJ (1991) Infected
knee prosthesis: diagnosis with In-l11 leukocyte, Tc-99m
sulfur coloid, and Tc-99m MDP imaging. Radiology
179:645-648
Patel DV, Ferris BD, Aiehroth PM (1991) Radiologie study
of alignment after total knee replacement. Int Orthop
15:209-210
Rand JA (1993) Alternatives to reimplantation for salvage of
total knee arthroplasty complieated by infection. J Bone
Joint Surg 75A:282-288
Sambatakakis A, Welton TJ, Newton G (1991) Radiographie
signs of persistent soft tissue imbalance after knee replace-
ment. J Bone Joint Surg 73B:751-756
Schneider R, Abenovoli AM, Soundry M, Insall J (1984)
Failure of total condylar knee replacement. Correlation
of radiographie clinieal and surgieal findings. Radiology
152:309-315
Silverton C, Rosenberg AO, Barden RM, Sheinkop MB, Galante
JO (1996) The prosthesis-bone interface adjacent to tibial
components inserted without cement. J Bone Joint Surg
78A:340-346
Stuart MJ (1991) Indieations and patient selection. In: Morrey
BF (ed) Joint replacement arthroplasty. Churehill-Living-
stone, New York, pp 971-979
Weiner SD, Scarborough M, VanderGreend R (1996) Resection
arthrodesis of the knee with intercalory allograft. J Bone
Joint Surg 78A:185-193
Westrieh GH, Hidaka C, Winsor RE (1997) Disengagement of
a locking screw from a modular stern in revision total knee
arthroplasty. J Bone Joint Surg 79A:254-258
Wilson MG, Kelley K, Thornhill TS (1990) Infection as a com-
plieation of total knee replacement arthroplasty. J Bone
Joint Surg 72A:878-883
13 Patellar and Quadriceps Mechanism
NIELS EGUND and LEIF RYD
CONTENTS
13.1 Introduction 217
13.2 Anatomy 218
13.2.1 Soft Tissue Stabilizers 219
13.3 Biomechanical Considerations 221
13.3.1 Patellar Tracking and the Q Angle 222
13.4 Clinieal and Pathologie Considerations 223
13.4.1 The Request for Imaging - History and Physieal
Examination 223
13.5 Routine Radiography 224
13.5.1 Lateral View 224
13.5.2 Posteroanterior View 225
13.5.3 The Standing Axial View 225
13.5.4 The Supine Axial View 226
13.5.5 Radiographie Measurement of the Axis
of the Knee 227
13.6 Roentgen Stereometrie Analysis 228
13.7 CT and MR Imaging of Patellar Tracking 229
13.7.1 Measurements 229
13.7.2 The Femoral Trochlea and Patellar Tracking 230
13.7.3 Summary of Radiologieal Measures
of Patellar Tracking 233
13.8 Osteoarthrosis of the Patellofemoral Joint 235
13.8.1 Imaging of Patellofemoral Osteoarthrosis 237
13.9 Traumatie and Overuse Conditions 238
13.9.1 Patellar Dislocation 238
13.9.2 Recurrent Dislocation of the Patella 241
13.9.3 Overuse Syndromes 243
13.10 Conclusion 244
References 245
13.1
Introduction
The extensor mechanism of the knee is composed of
the quadriceps muscle group and tendon, the patel-
la, the patellar ligament, the tibial tubercle and the
patellar retinaculum. The patellofemoral articulation
N.EGUND,MD
Professor, Department of Radiology, Aarhus University Hos-
pital, Noerrebrogade 44, 8000 Aarhus C, Denmark
L.RYD,MD
Associate Professor, Department of Orthopedies, University
Hospital of Lund, 22185 Lund, Sweden
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
centralizes the entire mechanism on the anterior
surface of the femur, and during knee extension it
mediates the forces generated by the largest muscles
through the largest lever arms in the body. Minor
variations in each component of the extensor mech-
anism may affect the centralizing function of the
patellofemoral joint by altering the patellar tracking
in the sulcus of the femur and resuIt in patellofemo-
ral pain and joint degeneration. In spite of extensive
experimental and clinical research, the problems of
the patellofemoral joint remain achallenge to cli-
nicians, and without a clear understanding of why
patellar disorders occur it is not surprising that there
is no consensus on how to treat them (THOMEE et
al. 1999). Neither the widespread use of arthroscopy
nor the availability of newer diagnostic imaging tech-
niques such as ultrasound, computed tomography
(eT) and magnetic resonance (MR) imaging has so
far contributed significantly to the understanding
of disorders related to dysfunction of the extensor
mechanism or to the traditional treatment of patel-
lofemoral disorders (INSALL 1995).
Clinical assessment and imaging of the normal
and injured patellofemoral joint require knowledge
not only about the anatomy and biomechanics of the
knee and the extensor mechanism, but also about the
site and pathoanatomical characteristics of the dif-
ferent disorders in question. In each clinical situa-
tion, the (MERCHANT 1988) classification system of
patellofemoral dis orders (Table 13.1) may be useful
in identifying the different abnormalities. The system
defines the etiology and pathomechanics of patello-
femoral disorders and categorizes the diagnoses such
that relevant examination plans can be initiated and
treatment suggested. However, in the absence of evo-
lution in the classification, examination, and treat-
ment of patellofemoral disorders, it is feit by the
authors of the present chapter that some fundamen-
tals of biomechanics of the knee have been ignored
for many years and that the relation between some
scientific observations and commonly observed path-
oanatomical and surgical findings has been over-
looked (FULKERSON 1997; SCUDERI 1995).
218
Table 13.1. The Merchant dassification system of patellofemoral
disorders
I. Trauma (conditions caused by trauma
in the otherwise normal knee)
A. Acute trauma
1. Contusion
2. Fracture
a) Patella
b) Femoral trochlea
c) Proximal tibial epiphysis (tuberde)
3. Dislocation (rare in the normal knee)
4. Rupture
a) Quadrieeps tendon
b) Patellar tendon
B. Repetitive trauma (overuse syndromes)
1. Patellar tendinitis ("jumper's knee")
2. Quadrieeps tendinitis
3. Peripatellar tendinitis (e.g., anterior knee pain in the
adolescent caused by hamstring contracture)
4. Prepatellar bursitis ("housemaid's knee")
5. Apophysitis
a) Osgood-Schlatter disease
b) Sinding-Larsen-Johansson disease
C. Late effects of trauma
1. Post-traumatie chrondromalacia patellae
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
7. Acquired quadrieeps fibrosis
II. Patellofemoral dysplasia
A. Lateral patellar compression syndrome
1. Secondary chondromalacia patellae
2. Secondary patellofemoral arthritis
B. Chronie subluxation of the patella
1. Secondary chondromalacia patellae
2. Secondary patellofemoral arthritis
C. Recurrent dislocation of the patella
1. Associated fractures
a) Osteochrondral (intra-artieular)
b) Avulsion (extra-artieular)
D. Chronie dislocation of the patella
1. Developmental
2. Acquired
III. Idiopathic chondromalacia patellae
IV. Osteochondritis dissecans
A. Patella
B. Femoral trochlea
V. Synovial plieae (anatomie variants made symptomatie
by aeute or repetitive trauma)
A. Pathologie medial patellar pliea ("shelf")
B. Pathologie suprapatellar plica
C. Pathologie lateral patellar pliea
VI. Iatrogenie dis orders
A. Iatrogenie medieal patellar compression syndrome
B. Iatrogenie chronie medial subluxation of the patella
C. Iatrogenie patella infera
N. Egund and L. Ryd
It has generally been accepted that radiographie
assessment of the femorotibial joint spaces should be
performed in the standing, weight-bearing position
with the knee in semifiexion. With the patient stand-
ing on one leg, this AP or PA radiographie view also
allows assessment of varus or valgus angulations and
subluxation. Clinieal examinations and tests as well
as imaging of the patellofemoral joint are in general
performed in the non-weight-bearing supine posi-
tion; however, it is uncertain and unexplored how
valid these measurements are for performance in the
standing position, in whieh dis orders develop and
may give rise to pain. Following the monograph of
(AHLBACK 1968), the authors of the present chapter
were brought up with weight-bearing radiographie
views of the knee including the patellofemoral joint,
clinically as well as in research. Recently, during ongo-
ing research, we became aware of some major differ-
ences between weight -bearing and non-weight -bearing
imaging, whieh at present may be considered contro-
versial but will contribute to changes in the under-
standing and treatment of patellofemoral disorders.
13.2
Anatomy
The patellofemoral joint consists of the trochlea
(patellar groove, femoral sulcus) of the femur and the
patellar artieular surfaces (Fig. 13.1). The embryol-
ogy of the patella and the femoral trochlea is fascinat-
ing (LANGER 1929), with early essential characteris-
ties of the adult. Initially, the medial and lateral facets
are equal in size, but after 23 weeks' gestation the lat-
eral facet tends to predominate. Before this time the
lateral facet of the trochlea has a greater transverse
width and is anteriorly more prominent, although the
patella is positioned distal to and without contact to
the trochlea. During growth and following function,
the shape of the patella may be modified. WIBERG
(1941) described three configurations of the patellar
joint (Fig.13.2). Wiberg type 11 (Fig.13.1) has a small-
er medial facet and represents the most common
patellar anatomy. WIBERG suggested that type III may
be correlated to chondromalacia, but this has not
been confirmed. A number of different anatomieal
variations and ossification variants may be observed
(KEATS 1996).
 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

The active stabilizers of the patellofemoral joint con-

sist of the four main muscles of the quadriceps, which
fuse distally into the quadriceps tendon. At the inser-
tion into the patella, three separate layers of the
Fig. 13.3. The superficial portion of the extensor apparatus.
quadriceps tendon can be identified (Fig. 13.4). The Medially there is a fibrous attachment between the patella and
rectus femoris originates from two attachments on the iliotibial tract, and laterally an attachment to the sartorius
the ileum and inserts into the anterior top and the tendon
220 N. Egund and 1. Ryd

Fig. 13.4. Normal MR anatomy in the sagittal plane of the

extensor apparatus, showing the separate layers of the quadri-
ceps tendon (QT). The patella is attached to the tibial tuberos-
ity (rn by the patellar tendon (PT). ACL, Anterior cruciate
ligament. There is a bright signal from the patellar and femoral
cartilage on this Tl-weighted Flash 2D sequence
Fig. 13.5. The normal MR anatomy of the vastus medialis and
its attachment to the medial portion of the patella (P). The
course of the distal fibers of the muscle and its tendon is
almost horizontal. Tl-weighted sequence

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
quadriceps tendon, and the medial and lateral reti-
nacula are the passive elements of soft tissue stabi-
lization. The patellar ligament connects the patella
with the tibial tuberosity and has a length of 4-6 cm,
a width of 25-40 mm, and a thickness of 6-8 mm
Fig. 13.6. Normal MR anatomy of the patellar retinaculum on
a Tl-weighted Flash 2D sequence. The superior portion of the
medial patellofemoralligament (MPFL) has a broad fibrocar-
tilaginous attachment to the odd facet (0) of the patella and
is dorsally attached to the femoral condyle and medial collat-
eralligament (MCL) with extension to the fascia of the dorsal
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
fascia of the leg and laterally to the fascia lata. In the
deeper layers, capsular condensations form fibrous
layers that link the patella to the medial and lateral
femoral epicondyles and the anterior aspect of
the menisci (Fig. 13.49). They include the medial
and lateral patellofemoralligaments and, below, the
medial and lateral meniscotibial patellar ligaments.
In between the medial patellofemoral and the menis-
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).
sured and calculated by numerous authors, with con-
siderable divergence of quantitative values. These
values depend on various factors, including body
weight, quadriceps force, angle of flexion, and individ-
ual anatomie factors. It is beyond the scope of this
textbook to go into detail on these considerations, an
overview of which can be obtained from orthopedic
textbooks (FULKERSON 1997; SCUDERI 1995). Howev-
er, the patellofemoral joint re action force may vary

13.3
Biomechanical Considerations

The main biomechanical function of the patella is to

improve the efficiency of the quadriceps by increasing
the lever arm of the extensor mechanism. Through-
out the range of motion, the patella displaces the quad-
riceps tendon and the patellar tendon away from the
femorotibial contact point and increases the force of
extension by as much as 50% (Fig. 13.8). The degree of
tension of the extensor mechanism and the resultant
force, the patellofemoral joint reaction force, acting Fig. 13.8. The patellofemoral joint reaction force (PFjR) is the
perpendicular to the articular surfaces, has been mea- resultant vector of the quadriceps and patellar tendon force
222 N. Egund and 1. Ryd

between 0.5 times the body weight at walking and 25

times the body weight at weight lifting and at 90° of
flexion of the knee. Thus in the assessment of patello-
femoral disorders, occupational and sports activities
are one of the cornerstones of patient history, espe-
eially when dealing with overuse syndromes and late
effects of these.
The second, and from an imaging point of view
most essential, part of the biomechanics of the exten-
sor mechanism is the behavior of patellar tracking.
Clinical as weIl as radiologie al assessment of the
patellofemoral joint requires an understanding of
those normal and abnormal factors whieh influence
the patella in its ride through the femoral trochlea
during flexion and extension of the knee.

13.3.1
Patellar Tracking and the Q Angle

The Q angle is the angle formed by a line drawn from

the anterior superior iliac spine of the pelvis through
the center of the patella and a line drawn from the
patella to the center of the tibial tuberosity (Fig. 13.9).
The Q angle is a common clinical measure and is fre-
quently discussed as a reflection of the valgus angle of
the extensor mechanism, with the underlying assump-
tion that the larger the Q angle is, the larger the lat-
eral moment on the patella. The clinieal measurement
of the Q angle is most commonly performed with the
patient in the supine position with the knee in exten-
sion, although standing measurement would reflect
weight-bearing function more accurately (POST 1997).
Different publications eite different values for the
normal Q angle, from 11° to 20°, and it is slightly
greater in standing subjects (WOODLAND and FRAN-
CIS 1992). Although measurements of the Q angle are
extensively used by orthopedie surgeons clinically as
weIl as in the planning of surgieal procedures, it has
been claimed (POST 1997) that no direct correlation
with the ineidence of patellofemoral disorders has
been established by scientific criteria.
The Q angle is helpful in understanding the func-
tion of the extensor mechanism and patellar tracking.
In the literature it has generally been agreed that, in
the terminal 30° from flexion to extension, the "screw
horne" mechanism rotates the tibia outward relative to
the femur, displaeing the tibial tuberosity laterally and
increasing the Q angle (Fig. 13.9). With a lateral posi-
tion of the tibial tuberosity (large Q angle), tension on
the quadrieeps will tend to produce a lateral displace-
ment vector of the patella, a decrease in the Q angle,
which is resisted by the vastus medialis obliquus, the
Fig. 13.9. The Q angle of the knee increases with increasing
outward tibial torsion, which may also be the result of increas-
ing femoral neck anteversion. Understanding of femorotibial
rotation and the anatomy and function of the active and pas-
sive stabilizers of the patellofemoral joint (Fig. 13.3) is a per-
quisite for understanding of abnormalities of the extensor
mechanism. (SCUDERI 1995)
medial retinaculum and the lateral facet of the troch-
lea. With the knee in full extension and the quadrieeps
contracted, the patella lies proximal to the trochlea.
During the first 30° of flexion, the tibia rotates inward,
decreasing the Q angle and also the lateral vector, and
the patella is drawn into the trochlea from the lateral
side (HUNGERFORD and BARRY 1979). These con-
siderations are in accordance with and to a great
extent based upon classieal interpretations of the
axial radiographie view of the patellofemoral joint
(LAURIN et al. 1978; MERCHANT et al. 1974) and
trans axial statie and kinematie imaging with CT
(DELGADO-MARTINS 1979; Dupuy et al. 1997; MAR-
TINEZ et al. 1983a,b) and MR imaging (BROSSMANN
et al. 1993, 1994; SHELLOCK et al. 1989). All of these
investigations were performed in the supine posi-
tion and so far, their performance for the bio-
mechanics and patellar tracking in the standing
weight-bearing position has not been established
using seientific criteria.
Patellar and Quadriceps Mechanism
Roentgen stereophotogrammetric analysis, de-
scribed below, allows very accurate measurements
of motion between rigid bodies in all three planes,
including rotation (SELVIK 1974), and has been
extensively used in orthopedic and biomechanical
research. Using this technique, BLANKEVOORT et al.
(1988) demonstrated that motion patterns of the
knee are highly susceptible to small changes in the
externaiload configuration and specifically that the
"screw horne" mechanism is not an obligatory effect
of the passive joint characteristics, but a direct result
of the externailoads. This supports the observations
reported by STEIN et al. (1993), who imaged fluoro-
scopically the patellofemoral joint of healthy volun-
teers in the anteroposterior plane during walking.
They recorded a uniform pattern of medial excursion
of the patella relative to the femur, with a sudden shift
from lateral to medial, and concluded that contrary
to conventional understanding, the patella deviates
medially rather than laterally during walking.
13.4
Clinieal and Pathologie Considerations
The multifactorial etiologies of anterior knee pain
can be related to variants of the anatomy of the patel-
la and alterations in the tensile forces of the extensor
mechanism applied to the joint surfaces of the patel-
lofemoral joint, generated during the complex move-
ment of the joint. These alterations may be caused
by overuse, disuse, and injuries to bone, cartilage
and supporting soft tissue structures. Although the
classification of patellofemoral disorders (Table 13.1)
devised by MERCHANT (1988) provides a useful
framework for the clinical and imaging approach to
diagnosis and treatment, it is feIt that it fails to rec-
ognize and interpret the pathogenesis of the most
common dis orders of the patellofemoral joint.
Patellar articular cartilage presents different modes
of degeneration. Chondrosis signifies a disorder
affecting only the articular cartilage and arthrosis, a
dis order affecting all three components of the joint:
cartilage, subchondral bone, and synovial membrane.
Chondromalacia patellae is a term applied to a syn-
drome of anterior knee pain in adolescents and
young adults (ALE MAN 1928; WIBERG 1941) and the
pathoanatomical appearances are described as soft-
ening (malacia), edema, and swelling of the carti-
lage. A proposed classification of the surgical and
arthroscopic severity of chondromalacia (FULKER-
SON and SHEA 1990) is almost identical to that of
223
arthrosis: grade 1 represents softening and swelling
of the cartilage; grade 2 is cartilage breakdown (fibril-
lation) of one half inch or less; grade 3 is cartilage
breakdown of greater than one half inch; and grade 4
is erosion of cartilage down to bone (OUTERBRIDGE
1961). This classification is also useful in the grading
of joint degeneration at MR imaging (Chap. 3). The
term "chondromalacia patellae:'however,has become
controversial since it may cover a large number of dis-
orders leading to patellofemoral pain (RADIN 1979)
in which loss of patellar cartilage is not documented,
is not present, or occurs in combination with other
abnormalities such as bursitis (pre- and infrapatellar,
pes anserinus), the plica syndrome, the fat pad syn-
drome, arthrosis, synovitis, and meniscal tears.
The terms "patellar tilt-compression" and "exces-
sive lateral pressure syndrome" (ELPS) were intro-
duced by FICAT and HUNGERFORD (1977). These
conditions are characterized clinically by pain and
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-
lar manifestations occur, the site of cartilage lesions
are the same. Classically, it is stated that the medial
facet of the patella is the typical and primary site
of cartilage lesions in chondromalacia, particularly
about the ridge that separates the medial and odd
facets (WIBERG 1941; INSALL et al. 1976). The car-
tilage changes on the medial patellar facet with a
supposed deficient contact to the femoral trochlea
(HENCHE et al. 1981) have been attributed to various
etiologic factors. Several authors have pointed out
the tendency for cartilage that is out of contact with
other cartilage to undergo surface fibrillation (GOOD-
FELLOW et al. 1976) and lose an appropriate mecha-
nism of synovial fluid nutrition (FULKERSON 1983;
LAURIN et al. 1978).
13.4.1
The Request for Imaging - History and Physical
Examination
Accurate, concise clinical evaluation of the patient
with a suspected knee disorders is almost invariably
suggestive of a working diagnosis. Together with the
clinical information, this working diagnosis forms
the cornerstone for tailoring the radiological exami-
nation and, indeed, for the interpretation of images.
Therefore it is essential that the request for imaging
is seen by an experienced radiologist and that the
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

Our standard radiographie examination is sufficient

for the evaluation of most middle-aged and elderly
patients with nontraumatie knee eomplaints. The
examination eomprises the three standard radio-
graphie views - lateral and posteroanterior views of
the knee and transaxial view of the patellofemoral
joint - all obtained in the standing, weight-bearing
position. For examination in all three views, the knee Fig. 13.10. Deviee for lower leg support and patient position
is supported by a deviee (Figs. 13.10, 13.11, 13.16) for the lateral radiographie view of the knee
whieh represents a simple eomposition of two sepa-
rate deviees (EGUND 1986; EGUND and FRIDEN 1988)
and has been adapted by several departments of radi-
ology in Denmark. The background for the eonstrue-
tion of the deviee is the normal tibial plateau angle
(Figs. 13.12, 13.14) relative to the tibial erest of 14°
(±3.6°, range 7°_22°) (MOORE and HARVEY 1974) and
the normal alignment in relation to the vertieal plane
of the posterior artieular surfaee of the patella in 15°
of inelination of the lower leg, when in the standing,
weight-bearing position (EGUND 1986).

13.5.1
Lateral View

Positioning of the patient with elose eontaet between

the entire anterior surfaee of the lower leg and the
plate of the deviee (Figs. 13.10, 13.12) is a prereq-
uisite for optimizing the posteroanterior and axial
views. The rotational position of the foot should be
Fig. 13.11. Deviee for lower leg support and patient position
with 30° of knee flexion for the anteroposterior radiographie
view. The "one-leg stand" is eonfirmed by the aetive lateral
quadrieeps muscle
 Patellar and Quadrieeps Mechanism
this position also allows assessment of sagittallaxity
(Fig. 13.13) (EGUND and FRIDEN 1988; EGUND et al.
1993; FRIDEN et al. 1992, 1993). The slopes of the tibial
plateau and the artieular surfaee of the patella are
estimated on the preview sereen or developed film
(Figs. 13.12, 13.32), as well as the position of the een-
tral beam.
Fig. 13.12. Normal true lateral radiographie view of the knee
obtained in the standing position using the described deviee
(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-
ular surface of the patella, the position being less optimal for
the axial radiographie view with a vertical beam direction
a b
225
13.5.2
Posteroanterior View
The slope of the tibial plateau is adjusted to hor-
izontal either by placing a wedge-shaped eushion
between the plate of the device and the knee (inclina-
tion of the lower leg is deereased) at anterior slope
on the lateral view or by placing the foot more pos-
teriorly (inclination of the lower leg is inereased) at
dorsal slope. The rotational position of the foot
and knee flexion (25°-35°) should be as for the lateral
view. The weight-bearing, one leg standing position
(Fig. 13.11) is strenuous, and a fast exposure is
required to avoid knee extension, upon which joint
spaee narrowing may be lost. The use of the device
allows serial exposures in different degrees of flexion.
It must be mentioned that when assessing the femo-
rotibial joint spaees the slope of the medial and lat-
eral tibial articular surfaees in the sagittal plane is
rarely the same (Figs. 13.14, 13.15): the me an differ-
enee is 3° (range 0° - 5°), and in abnormal eonditions
it may reaeh 9° (EGUND et al. 2000). Therefore it is not
possible to aeeurately measure both medial and lat-
eral femorotibial minimal joint spaees (BOEGARD et
al. 1998b; BUCKLAND-WRIGHT 1994) on single AP or
PA radiographs.
13.5.3
The Standing Axial View
The teehnique of standing axial radiographie imag-
ing of the patellofemoral joint was introdueed by
AHLBAcK (1968). His deviee for support of the patella
Fig. 13.13a, b. Lateral radiographie views of a
23-year-old female with rupture of the ante-
rior cruciate ligament. Standing without load-
ing (a) and at weight-bearing (b). Normally, in
any degree of knee flexion, standing or supine,
the lowest points of the articular surfaces of
the femoral condyles (arrow) are sited at the
center (arrowhead) of the tibial eminence (a).
In weight-bearing (b) the tibia is displaced
ventrally relative to the femur, indicating joint
instability at every step of gait
 226 N. Egund and L. Ryd

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

during the examination is still widely used in Sweden,

but it has been refined and the examination technique
is more standardized (EGUND 1986). Todaywe use the
device shown in Fig. 13.16. Guided by the lateral view
(Fig.13.12), the inclination of the lower leg is adjusted
to bring the articular surface of the patella into the
vertical plane. There should be an even distribution
of the body weight on both legs and knee flexion in
our routine view is about 30°. The technique allows
axial radiographs in different degrees of knee flexion
between 20° and 100° (Fig.l3.17) without reposition-
ing the knee relative to the device or changing the
lower leg inclination (EGUND 1986).

13.5.4
The Supine Axial View

Since SETTEGAST (1921) recognized the axial view of

Fig.l3.IS. Anteroposterior radiographie view of a knee, with a
normal difference of 4° in the slope of the medial and lateral
femoral condyles in the sagittal plane. The view through the
lateral joint space is tangential to the bony contour of the tibial
condyle. Assessment of the medial joint space is hindered by
both the anterior and the posterior aspect of the tibial condyle
the patellofemoral joint,various techniques (Fig.13.18)
have been described (BRATTSTRÖM 1964; ]AROSCHY
1924; KNUTSSON 1941), but the most simple and
accepted techniques for routine and scientific purpos-
es are those described by MERCHANT et al. (1974) and
LAURIN et al. (1979) (Figs.13.18, 13.19). The Merchant
technique as originally described required 45° of
knee flexion and a specific cassette holder. By com-
parison, it was suggested that the method of Laurin
be performed in 20° of knee flexion. Techniques
with the beam directed from the ankle to the knee
Patellar and Quadrieeps Mechanism 227

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

(BRADLEY and OMINSKY 1981; LAURIN et al. 1979)

and with the patient in the sitting position should a
be abandoned for radiation protection reasons. Both
the Merchant and the Laurin technique, however,
can be used in different degrees of knee flexion.
b
In the supine position, the beams should be angled
approximately 10° relative to the axis of the lower
leg (Figs. 13.18, 13.19) in order to obtain images
tangential to the patellar articular surface; this is c
because the patella has a lower position than in the
standing position (Figs. 13.16, 13.32).
d

Fig. 13.18a-d. Different methods to obtain the axial radio-

13.5.5 graphie view of the patellofemoral joint in the supine or prone
Radiographie Measurement of the Axis position: a SETTEGAST (1921); b JAROSCHY (1924); C KNUTSSON
ofthe Knee (1941); d LAURIN (1979)

Total radiologieal evaluation of patellar tracking

requires measurement of the long axis of the lower
extremity, the HKA angle (hip-knee-ankle), and the
Q angle as described by SANFRIDSSON et al. (2001b).
These authors used the complicated QUESTOR Preci-
sion Radiography system (QPR) (SIU et al. 1991) for
standardized and reproducible measurements in the
standing position and reported a varying lack of cor-
relation between clinical measurement of the Q angle Fig. 13.19. The Merchant et al. technique to obtain the axial
radiographie view of the patellofemoral joint at 45° of knee flex-
in the supine position and radiographie measure-
ion, commonly used in the United States. The technique requires
ment in the standing position. It appears from their a specific cassette holder. With both the Merchant and the Laurin
results (SANFRIDSSON et al. 2001a) that the axis of the technique, the beam direction relative to the tibia is less than
femoral shaft is reliable in the radiographie measure- 10°, whereas with the standing technique it is 15° (±6°)
228
ment of the Q angle, and therefore a true anteroposte-
rior view of the knee in the standing position can be
used for the measurement (Figs.13.20, 13.21). The clin-
ieal significance of standing radiographie measure-
ments of the Q angle requires further investigation.
Fig. 13.20. Biplane radiography of the knee supported by
the device. A true lateral position of the femoral condyles is
obtained at fluoroscopy using tube 1 and the image intensifier
(II). A true anteroposterior view is obtained perpendicularly
from tube 2. The film is placed in a cassette holder (arrow-
head) also used for whole limb radiography
Fig.13.21. Measurement of the
Q angle on a standing true
anteroposterior radiograph of
the knee in 20° of flexion. The
tibial tuberosity was marked
by a piece of lead
N. Egund and L. Ryd
13.6
Roentgen Stereometrie Analysis
Roentgen stereometrie analysis (RSA) is a technique
by which the relative motion of two objects, over time
or before and after application of extern al or inter-
nal forces, can be assessed using rigid body kinemat-
ics. Since the skeleton consists of a number of rigid
bodies, the method is particularly suited for skeletal
radiology. It has gained increasing popularity in the
assessment of joint replacement fixation (ÖNSTEN et
al. 1993; KÄRRHOLM et al. 1994a; RYD et al. 1995) and
fracture healing (KÄRRHOLM et al. 1983; AHL et al.
1988; RAGNARSSON et al. 1992). In the knee, in addi-
tion to the above uses, RSA has been applied to the
anterior cruciate-deficient knee (KÄRRHOLM et al.
1988; FR IDEN et al. 1992) and in vitro examinations
of the patellar joint (BLANKENVOORT et al. 1988).
RSA combines the principles of stereogrammetry,
skeletal marking, and modern computing power into
a system that allows measurement of motions down
to at least 0.1 0 or 100 11m (RYD et al. 1986). The
objects of interest are marked with balls made of
tantalum. Tantalum is inert in body fluids and also
heavy (atomic weight =73), and consequently yields
distinct radiographie images. These markers should
be placed in the objects of interest as far apart as the
objects allow. A minimum of three markers in each
object is required, but the insertion of five to seven
markers in each increases the accuracy. Radiograph-
ie images are obtained in a stereo setting, i.e. two
images are obtained simultaneously. A calibration
object is also exposed, either before the patient is
examined or, preferably, simultaneously with the
patient examination. There are usually two set-ups
available: a bi-planar set-up and a convergent-ray
stereo set-up. In the former, the body part of interest,
usually an arm or a leg, is placed inside a calibration
box; two X-ray tubes at a right angle are positioned
to expose films attached to adjacent sides of the cal-
ibration box, and frontal and lateral exposures are
obtained. In the convergent ray alternative, X-ray
tubes are aimed at approximately 40 0 from one anoth-
er and films are exposed in the same plane under-
neath a calibration object, which, in turn, is situated
underneath the tabletop carrying the patient. The
two films, constituting a "stereo pair:' are subsequent-
ly digitized and the x- and y-coordinates are fed into
dedicated software to produce first a set of 3D coor-
dinates for each patient marker and finally the kine-
matie analysis of how the two "rigid bodies:' i.e.,
objects of interest (for example, the patella and the
femur) have moved relative to one another.
 Patellar and Quadriceps Mechanism
RSA provides two advantages. First the accuracy
is about one order of magnitude better than con-
ventional radiography, and second, the system allows
perfect characterization of the motion in all 6 degrees
of freedom of motion. Hence, "out-of-plane" motion
can also be analysed.
RSA has not previously been utilized in connec-
tion with patellar motion proper. BLANKENVOORT et
al. (1988) used it to studythe screw-home mechanism,
whieh is inherently involved in the complex motions of
the patella. They found that the screw-home motion is
a facultative occurrence, dependent on how the femur
and the tibia are positioned rotationally relative to
one another at an initial point in time. Similarly, KÄR-
RHOLM and co-workers have published a number of
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-
eral they have found that femorotibial rotation is less
in an artificial knee than in the normal one. The poten-
tial of RSA has not yet been fully realized with regard
to the patellar joint, although studies are in progress
in Lund on patients in whom all three bones (femur,
tibia, and patella) have been marked.
13.7
Cl and MR Imaging of Patellar lracking
Transaxial imaging of the patellofemoral joint by eT
and MR imaging allows assessment of the position
of the patella above and within the femoral trochlea
over a range of extension and mild flexion of the knee
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
229
which is inaccessible for radiographie examination
in the axial plane (Figs. l3.27, l3.28). The rationale
for viewing the patellofemoral relationship in mild
flexion is based on the fact that most patellar sublux-
ations and dislocations occur within the first 20°-30°
of flexion. Statie eT and MR imaging can be used to
obtain axial images at defined degrees of flexion with
and without contraction of the quadriceps musele
(DELGADO-MARTINS 1979; MARTINEZ et al. 1983b).
Assessment of patellar tracking by kinematie eT and
MR imaging has been extensively reported (McNAL-
LY et al. 2000; Dupuy et al. 1997; SHELLOCK et al. 1989;
BROSSMANN et al. 1993, 1994) with and without active
loading during flexion to extension, and a number of
deviees have been developed for these purposes.
13.7.1
Measurements
Vertical patellar height is important for the biome-
chanies and stability of the extensor mechanism and
patellofemoral joint. Measurements are performed
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
of patellar height from lateral standing views at more
than 20° of flexion. The most widely employed method
for assessing patellar height has been described by
INSALL and SALVARTI (197l) and is based on the
length of the patellar tendon divided by the greatest
diagonal length of the patella (Fig. l3.22a). The
normal ratio is 1.02 with an SD of 0.l3. A ratio of
_ _ __ b,c
230 N. Egund and L. Ryd

less than 0.80 is considered indicative of patella baja

and a ratio of greater than 1.20 as indicative of patella
alta. The Insall-Salvarti index may give an inaccurate
impression of patellar height relative to the femoral
condyles owing to anatomical variations in the length
of the patella and the position of the tibial tuberosity
(GRELSAMER and MEADOWS 1992); this is especially
so following transfer of the tibial tuberosity.
To assess patellar height, BLACKBURNE and PEEL
(1977) divided the perpendicular distance from the
lower margin of the patellar articular surface to the
line joining the tibial plateau by the length of the Fig. 13.23. The radiographie measurements of the patellofemo-
articular surface of the patella (Fig. 13.22b). The ral joint performed by BRATTSTRÖM (1964) obtained with high
normal ratio was 0.80 with an SD of 0.14. It is gen- radiation doses to the patient from two complieated projections
erally considered that, compared with the Insall- of the anterior and posterior aspect of the femoral condyles
Salvarti index, this method provides a more reliable
measure of patellar height relative to the femoral
trochlea, with less interobserver variability (SEIL
et al. 2000); furthermore, it has also been adopted
in orthopedic research. The method described by
CATON et al. (1982) represents a similar measure-
me nt of patellar height (Fig. 13.22c). The vertical
position of the patella has been assessed relative to a
line drawn through the distal femoral condyles and
perpendicular to the long axis of the tibia, thereby
avoiding the potential confounding variations in the
inclination of the tibial plateau (EGUND et al. 1988).
This measurement was expressed relative to body Fig. 13.24. Measurements of the sulcus angle according to
height and has a very high correlation to the Black- Brattström (with personal communieation). a, Medial border
burne and Peel index. of the bony artieular surface; b, arbitrary landmark elose to
the lateral border of the artieular surface; c, lowest point of
the intercondylar sulcus

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).
The congruence angle, the lateral patellofemoral
angle, and measurement of lateral patellar displace-
ment have all been used frequently in the assess-
ment of patellar instability by means of CT and MR
imaging (Dupuy et al. 1997; McNALLY et al. 2000;
BROSSMANN et al. 1993). Direct and oblique trans-
axial imaging of the femoral condyles by CT and
MR imaging allows assessment of the transverse
axes and lateral articular surface of the patella rela-
tive to the plane of the posterior femoral condyles
(Figs. 13.28, 13.29), which is considered a reliable
 232
reference plane (SCHUTZER et al. 1986; BROSSMANN
et al. 1993, 1994).
The position of the tibial tuberosity relative to the
sagittal plane through the sulcus of the femoral troch-
lea (Fig. 13.30) has been measured by CT and MR
imaging in individuals with normal patellar tracking
and in patients with patellar instability (ANDO et al.
1993; MUNETA et al. 1994; BEACONSFIELD et al. 1994;
IONES et al. 1995; McNALLY et al. 2000). The consis-
tency of the results indieates that a distance exceed-
ing 20 mm may be indieative of severe maltracking.
Using a metal marker on the tibial tuberosity and
Fig. 13.28. Measurement at CT of the lateral patellofemoral
angle (vO)relative to the femoral condylar plane (CP)
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)
N. Egund and L. Ryd
axial radiographie views, NAGAMINE et al. (1999)
demonstrated an abnormal lateral position of the
tibial tuberosity in patients with lateral patellofemo-
ral arthrosis in 30° of flexion, and in comparison with
CT, the abnormal lateral position of the tibial tuber-
osity remained unchanged at extension (NAGAMINE
et al. 1997).
Axial radiographie views in the standing weight-
bearing position (AHLBÄCK 1968; EGUND 1986) have
been used widely in clinieal studies in Sweden and also
in the assessment of the degenerative patellofemoral
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
and patellofemoral contact areas are different in the
standing and the supine position (Fig.13.32).AIso, the
 Patellar and Quadriceps Mechanism
Fig. 13.30a-c. Measurement of the position of the tibial tuber-
osity relative to the femoral sulcus at cr. At the workstation,
Hnes are drawn on the screen and on three different sections:
aHne f-c (a), which joins the most dorsal aspects of the tibial
condyles, and two Hnes perpendicular to f-c, one through the
tibial tuberosity (TT; b) and one through the deepest point of
the femoral sulcus (c)
examination technique used may influence femoro-
tibial rotation (Fig. 13.32) and thus the position of the
tibial tuberosity relative to the femoral trochlea.
13.7.3
Summary of Radiological Measures of Patellar
Tracking
The results reported in the literature on normal patel-
lar motion are still inconsistent, but in general the
original results and thoughts of LAURIN et al. (1978)
and MERCHANT et al. (1974) have been confirmed by
me ans of CT and MR imaging.
In the non-loaded, supine position and in asymp-
tomatic knees, at 30° ofknee flexion and with relaxed
quadriceps muscles, the artieular surfaces of the
patella are aligned with those of the femoral troch-
lea, but in some subjects there is a slight tendency
towards lateral tilting and displacement. Between
20° of flexion and complete extension there is an
233
c
increased tendency towards lateral shift and dis-
placement. Lateral displacement and tilt may both
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-
tioned lateral to the sagittal plane through the deep-
est point of the femoral trochlea at a distance of less
than 20 mm (BEACONSFIELD et al. 1994), and in most
patients with maltracking this distance is more than
20 mm (JONES et al. 1995; McNALLY et al. 2000).
In the weight-bearingposition, the axial radiograph-
ie view of normal subjects in the standing position
always shows alignment between the medial and lat-
eral joint surfaces of the patellofemoral joint (EGUND
2001), and this is also true in those with lateral tilt and
displacement when examination is performed in the
supine position (Figs. 13.25, 13.31). During walking the
excursion of the patella is medial rather than lateral
(STEIN et al. 1993). Between flexion and full extension
when standing on both legs there is an outward rota-
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-
loaded knee (BLANKEVOORT et al.1988).
In conclusion, as previously suggested by STEIN et
al. (1993), standing examination is to be considered
the "state of the art" technique for the assessment of
patellar tracking. Visualization of the function, per-
formance and biomechanics of the knee is inadequate
when using supine imaging, including arthroscopy.
There is consequently a need to reevaluate conclu-
sions drawn without the use of views obtained in the
Patellar and Quadriceps Mechanism 235

standing position. This re evaluation should inc1ude
the present supposed etiology of exeessive lateral
pressure syndrome (FICAT and HUNGERFORD 1977)
and ehondromalacia of the medial patellar eartilage.
13.8
Osteoarthrosis of the
Patellofemoral Joint
Symptomatie knee osteoarthrosis (OA) may oeeur in
1.6%-9.4% of adults (FELSON 1998), and OA eonfined
to the patellofemoral joint or eoneomitant with fem-
orotibial OA is eommon (AHLBÄCK 1968; LEDING-
HAM et al. 1993). These figures are, however, mainly
reeorded from lateral radiographie views and the dis-
tribution within the medial and lateral patellofemo-
ral faeets has not been established by me ans ofaxial
radiographie views. Lateral patellofemoral OA is also
eommon in knees with medial femorotibial OA and
varus angulation (HARRISON et al. 1994), but the
eombination oflateral femorotibial and medial patel-
lofemoral OA is rare. Lateral patellofemoral OA may
be associated with an abnormal lateral position of
the tibial tuberosity, when examined in the supine
position by axial radiographie views and eT (NAGA-
MINE et al. 1997, 1999).
The c1inieal features in patients with patellofemoral
OA are similar to those of other patellofemoral pain
problems, but many and especially younger patients
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
lateral femorotibial joint, but there is eharaeteristie
dis ability at stair c1imbing and deseending and sudden
pain while arising from achair. The eombination of
patellofemoral and femorotibial OA is eommon, and
the lack of eorrelation between severity of symptoms
and stage of radiographie degeneration is well known.
It has to be kept in mind that any eondition involving
the hip joint with effusion/synovitis may appear with
dominating referred pain to the knee.
The method of grading radiographie change of
OA was developed by KELLGREN and LAWRENCE
(1957), but the imperfeetion of their eriteria has led
to a large number of radiographie grading seal es
for eross-seetional and longitudinal epidemiologieal
studies. We use a specifie grading system for the knee
joint (Table 13.2) whieh takes into aeeount the gener-
al radiographie features of OA as proposed by KELL-
GREN and LAWRENCE (1957), but with a more spe-
cifie measurement of joint spaee loss and attrition
as proposed by AHLBÄCK (1968) and EGUND (1986).
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

Feature Grade Definition Compartment

Femorotibial Patellofemoral
Medial Lateral Medial Lateral

Joint space narrowing 0 None D D D D

Osteophytes
Subchondral sclerosis
Sharpening of tibial spine
Chondrocalcinosis
Osteonecrosis
0.5 25% joint space narrowing D D D D
50% joint space narrowing D D D D
1.5 75% joint space narrowing D D D D
2 100% joint space narrowing D D D D
3 <5 mm attrition D D D D
4 5-10 mm attrition D D D D
0 None
Small (definite) osteophytes
2 Moderate osteophytes
3 Large osteophytes
0 Absent
1 Present
0 Absent
Present
0 Absent
1 Present
0 Absent
Present

(1957) grading system concerns the question of
whether or not marginal osteophytes (SPECTOR and
HOCHBERG 1994; BRANDT et al. 1991) represent
definitive evidence of cartilage degeneration (BoE-
GARD et al. 1998a): this is a subject which can still
unite or divide radiologists. Marginal osteophytes
appear as a sign of bony re action at sites where
articular cartilage is continuous with the synovial
membrane and periosteum (RESNICK and NIWAYA-
MA 1995). The sites of marginal osteophytes are
most commonly far away from the focal cartilagi-
nous lesion of OA. Their presence should be consid-
ered a sign of processes in the synovium and sur-
rounding soft tissue which may react to mediators
of cartilage damage and repair (EDWARDS 1998; VAN
DEN BERG et al. 1998).
Patellar and Quadrieeps Mechanism 237

13.8.1
Imaging of Patellofemoral Osteoarthrosis

Given that both isolated and combined femorotibial

and patellofemoral OA may occur, the routine radio-
graphie examination for the diagnosis of OA of the
knee joint should always comprise weight-bearing AP
and lateral views of the knee as well as trans axial views
of the patellofemoral joint (Fig.13.12). The axial radio-
graphie view obtained in the supine position may only
occasionallY demonstrate medial patellofemoral OA
(Figs. 13.35-13.37); furthermore, lateral patellofemo-
ral OA may occasionally be missed in the supine posi-
tion although it is visible with the standing technique
(EGUND 2001). Medial and lateral patellofemoral OA
is commonly associated with medial and lateral dis-
placement of the patella, respectively, and therefore Standing
imaging in the supine position may not demonstrate
the condition of the joint space of the other facet Fig.13.36. Axial radiographie views obtained in the supine and
(Fig. 13.38). It may be necessary to perform a stand- standing positions in a 51-year-old male with anterior knee
pain. A 75% reduction (stage 1.5) of the medial joint space
ing examination in different degrees ofknee flexion to
could only be demonstrated by the standing examination
visualize both patellofemoral and femorotibial OA.
A common reason for admission is the need to
perform MR imaging for the diagnosis of early OA of
the patellofemoral joint when joint spaces are normal
in all three compartments on standing radiographs.

Standing

Fig. 13.37. Axial radiographie views obtained in the supine

and standing positions in a female with medial femorotibial
arthrosis. Stage 2.0 arthrosis in the medial joint space was
visualized in the standing position only

Standing
Fig.13.35. Axial radiographie views obtained in the supine and
standing positions in a 48-year-old female with anterior knee
pain. Arthrosis suspected from the slight joint space reduction
visible only on the standing examination was confirmed at MR
imaging
In addition to anterior knee pain, many of these
patients present with swelling of the knee and effu-
sion with synovitis on contrast-enhanced MR imag-
ing (Figs. 13.33, 13.34). Our imaging protocol in these
patients comprises:
238
Fig. 13.38. Axial radiographie views of lateral patellar arthro-
sis visualized in both the supine and the standing position.
There is medial displacement of the patella between supine
and weight-bearing
1. Sagittal STIR, 512 matrix
2. Sagittal Tl, 512 matrix
3. Oblique axial Tl, 512 matrix
4. Oblique or double oblique axial proton fat saturat-
ed, 512 matrix
5. Sagittal contrast-enhanced Tl fat saturated
6. Oblique or double oblique axial contrast -enhanced
Tl fat saturated
The STIR and Tl-weighted sequences provide
an excellent overview and tissue characterization
of soft tissue, bony and cartilaginous abnormalities
(Fig. 13.33a). In addition, Tl-weighted sequences are
sensitive in detecting subchondral abnormalities,
which commonly are associated with focal cartilage
degeneration but mayaiso precede cartilage deterio-
ration, as suggested by RADIN and ROSE (1986). Con-
ventional T2-weighted sequences will not visualize
these subchondrallesions or bone marrow edema and
have little relevance in MR imaging of OA or arthritis.
Among the sequences used for cartilage imaging
for routine use, we prefer a high-resolution, proton
fat-saturated sequence, which allows grading of carti-
lage degeneration according to OUTERBRIDGE (1961)
(Figs.13.34, 13.39, 13,40). When cartilage lesions down
to bone are evident on STIR and Tl-weighted images,
the proton fat-saturated sequence is excluded. Post-
N. Egund and L. Ryd
contrast Tl-weighted sequences with fat saturation
allowan overall estimation and grading of the inflam-
matory component of OA in the synovium and its
surrounding soft tissue (OSTERGAARD et al. 1998)
as weIl as in the subchondral bone and adjacent
bone marrow (Fig. 13.33b). Cartilage defects and the
common meniscus lesions in femorotibial OA are also
weIl visualized with this postcontrast sequence.
It has been suggested that the role of processes in
the subchondral bone relative to those in the adjacent
cartilage can be further elucidated by the use of con-
trast-enhanced measurements of perfusion and per-
meability (KUHL and SCHILD 2000) followed by mea-
surements of changes in the cartilage glycosamino-
glycan concentration (BASHIR et al. 1999; BURSTEIN
et al. 2001).
13.9
Traumatic and Overuse Conditions
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
an understanding of the trauma mechanism and what
to look for, any traumatic condition is better consid-
ered a primary soft tissue injury which may involve
bone. In chronic injury and overuse syndromes, it is
less appropriate to make a distinction between soft
tissue and bone injury, considering the anatomy and
the most common sites of injury at tendon attach-
ments to bone and cartilage (Fig. 13.34).
13.9.1
Patellar Dislocation
Patellar dislocation indicates that the patella has been
completely displaced from the femoral trochlea. The
direction of displacement is almost invariably to the
lateral side. Congenital patellar dislocation (Fig.13,41)
is commonly assoeiated with neuromuscular defieien-
eies and frequently requires surgical treatment. Stand-
ing radiographs may be helpful in the planning of this
treatment (Fig. 13,42). Primary acute dislocation may
result from direct trauma to the patella, but is most
commonly due to a twisting injury on a fixed foot
with excessive outward rotation of the tibia at slight
to moderate flexion of the knee. A number of predis-
posing anatomical bone and soft tissue abnormali-
ties have been discussed, the most consistent of these
Patellar and Quadrieeps Mechanism "'20

Fig. 13.40. MR image of cartilage fis sure and adjacent bone

marrow lesions on a 3D SPGR/FLASH sequence. (Courtesy of
J. Gelineck, MD, Aarhus, Denmark)

Fig.13.41. Transaxial CT of the patellofemoral joint in 16-year-

old male with congenital patellar displacement who had previ-
ously undergone surgery on the right knee. In addition to the
Fig. 13.39. Axial radiographie views of the patellofemoral joint large outward rotation of both knees, the patient had abnor-
in a 19-year-old male with anterior knee pain and patellar mal anteversion angles of the hips
instability, obtained in the supine and standing positions.
Proton fat-saturated MR imaging demonstrated localized car-
tilage edema (arrowhead). The patient was a candidate for
lateral release until the standing radiographie view demon-
strated a normal position of the patella. Notiee edema in the
bone marrow of the patella. The patient was free of pain after
surgieal fenestration

being an abnormally high trochlear angle (BRATT-

STRÖM 1964), patella aha (INsALL et al. 1976), lateral
patellar tilt and displacement (ATKIN et al. 2000), and
generalized ligamentous laxity (RÜNOW 1983 ).Around
50% of primary acute dislocations occur in knees
without these abnormalities and the result of trauma
is most severe in these patients. Arthroscopy may Standing
reveal osteochondrallesions involving the patella and
the lateral femoral condyle in about two-thirds of
Fig. 13.42. Attempts to obtain axial radiographie views in a
patients, many with loose fragments, and tears of the
patient with congenital patellar displacement in the supine
medial patellofemoralligament are found in almost all position. There is no patellar displacement in the standing
knees at open surgical exploration (SALLAY et al. 1996; position, whieh led to an alteration in the planned surgieal
STANITSKI 1995; STANITSKI and PALETTA 1998). treatment
240 N. Egund and 1. Ryd

13.9.1.1
Imaging of the Knee with Acute Patellar
Dislocation

The radiographic examination of acute patellar dis-

placement should include AP, lateral and two oblique
views of the knee as well as an axial view of the patel-
Iofemoral joint. Almost invariably the radiographie
study will demonstrate a large effusion of the supra-
patellar bursa and commonly lipohemarthrosis, but
few intra-artieular loose osteochondral fragments
(Fig. 13.43) can be identified (STANITSKY and PAL-
ETTA 1998). The axial view may demonstrate an avul-
sion from the medial aspect of the patella (Fig. 13.43)
and an increased patellofemoral joint space with lat-
eral patellar displacement due to the effusion.
Figs.44-49
MR imaging may detect only half of the osteo-
chondral lesions seen at arthroscopy, but most of
the lesions of the medial patellofemoral ligament Fig. 13.44. Acute primary patellar dislocation. Sagittal STIR
MR sequence sectioning the center of the knee. Joint effu-
(SALLAY et al. 1996). We examine these injuries using
sion. There is a bone marrow lesion of the distal half of the
a dedicated knee coil at 20° of knee flexion and patella (white arrowhead) and a large bone marrow lesion of
our standard protocol includes sagittal and axial Tl- the femur (white arrow) at the insertion of the anterior (ACL)
weighted and STIR sequences (Fig. 13.44), and axial and posterior (PCL) cruciate ligaments

Fig. 13.45. Acute primary patellar dislocation. Transaxial MR

Fig. 13.43. Acute primary patellar dislocation. The axial radio-
graphie view in the standing position (above) demonstrates
avulsion of an osteochondral fragment from the odd facet
of the patella (white arrow) and in addition a fragment in
the lateral joint space (black arrowhead). MR imaging (below)
confirmed the osteochondral avulsion of the odd facet. For
normal anatomy, compare with Fig. 13.6
section through the femoral condyles with STIR sequence.
Joint effusion is present and there are lesions of the middle
portion of the medial patellofemoralligament (white arrow)
and at the site of the medial collateral ligament (white
arrowhead). There is a characteristie bone marrow lesion
following the contour of the lateral femoral condyle and also
a commonly seen bone marrow lesion extending in the lat-
eral direction from the insertion of the anterior cruciate lig-
ament (ACL)
Patellar and Quadriceps Mechanism 241

Fig.13.46. Acute primary patellar dislocation. Transaxial MR sec-

tion through the femoral condyles on a STIR sequence with a
matrix of 512. Joint effusion. There are minor lesions of the
medial patellofemoralligament (MPFL) at the insertion on the
patella, but at surgery there was also rupture at the attachment
on the medial femoral condyle. There is a characteristic bone Fig. 13.47. Acute primary patellar dislocation with the charac-
marrow lesion of the medial two-thirds of the patella and the lat- teristic bone marrow lesion of the lateral femoral condyle, and
eral femoral condyle. In addition, a bone marrow lesion extends also alesion of the medial femoral condyle, on a coronal MR
from the insertion of the posterior cruciate ligament (peL) STIR sequence

proton with fat saturation, all at a matrix of 512. To

date, in 41 acute primary patellar dislocations, we
have observed osteochondrallesions in about 40% of
cases at arthroscopy, but less than half of these have
been visualized at MR imaging (Egund, Wulf, Kris-
tiansen, personal communication) (Figs. 13.43, 13.48,
13.49). Partial and complete te ars of the patellofemo-
ralligament have been seen in all but two cases, and
confirmed at surgical repair (Fig. 13.45). Lesions of
the medial collateral ligament are common, as are
bone marrow lesions. The sites of the latter are char-
acteristic for patellar dislocation (Figs. 13.44-13.47)
and indicate a serious rotational injury of the fem-
orotibial joint (Figs. 13.45, 13.46), leaving this with
potential ligamentous instability. Complete tears of
the anterior or posterior cruciate ligament have not
been observed, but the common and characteristic
bone marrow lesions adjacent to the ligament inser-
tions in the femoral condyles (Figs. 13.44-13.46,
13.49) may be the result of traction injuries (Egund et
al., personal communication in preparation for pub-
Fig. 13.48. Acute primary patellar dislocation with a sagittal
lication). Acute patellar dislocation may also involve
MR STIR sequence. The typical bone marrow lesion is dem-
the fat pad of Hoffa (Fig. 13.44) (ApOSTOLAKI et al. onstrated in the lateral femoral condyle. To visualize most of
1999). Meniscus lesions are seen only occasionally. the osteochondral lesions on MR imaging, trans axial as well
as oblique coronal sectioning (shown) is necessary
242
a
b osteoarthritis mayaiso be related to patellar dis-
Fig. 13.49a, b. Acute patellar dislocation in the same patient
and with the same direction of sectioning as in Fig. 13.48. a
The joint effusion contributes to visualization of the anatomy
of the anterior portion of the medial (arrowhead) and the lat-
eral (LM) meniscus and their attachment to the patellar ret-
inaculum. The direction of sectioning may optimize assess-
ment of the anterior cruciate ligament (ACL). b Anteriorly
there is an osteochondral defect (black arrowhead) of the lat-
eral femoral condyle. Large bone marrow lesions are present
both laterally and centrally
13.9.2
Recurrent Dislocation of the Patella
Following primary acute dislocation of the patella,
most patients have to reduce their sports activities
(ATKIN et al. 2000) and will reappear with patellar
pain and instability (HAWKINS et al. 1986) or one
or more redislocations. Recurrent patellar disloca-
tion has a peak incidence at 14-15 years (range 10-30
N. Egund and L. Ryd
years) and occurs more frequently in females (FULK-
ERSON 1997). The surgical treatment of recurrent
patellar dislocation remains achallenge to ortho-
pedic surgeons and can be divided into four main
types of procedure, combinations of which may be
employed:
1. Soft tissue realignment procedures, e.g., lateral
release
2. Medial transfer of the tibial tuberosity/internal
rotational osteotomy
3. Distal transfer of the tibial tuberosity
4. Osteotomy of the lateral femoral condyle (no long-
term results)
The short-term results of these procedures are
"good" with regard to redislocation, but less prom-
ising in the long term owing to dis ability, reduced
physical activities and development of patello-
femoral osteoarthrosis (HAMPSON and HILL 1975;
MACNAB 2001; MAENPAA and LEHTO 1997a,b). In
comparison with surgery, and independent of the
surgical procedure used, conservative treatment
may yield better clinical results and reduce the
occurrence of osteoarthritis (ARNBJORNSSON et
al. 1992). Using weight-bearing radiographie tech-
niques, medial patellofemoral osteoarthritis was
recorded in 65% of 114 operated knees, followed up
for 14 years, whereas lateral patellofemoral osteoar-
thritis occurred in only three (ARNBJÖRNSSON et al.
2002). Development of mainly medial femorotibial
location (JULIUSSON and MARKHEDE 1984): based
on the use of weight-bearing radiographs, its prev-
alence 14 years following surgery was found to be
52%, with a significant correlation to medial or
lateral (3/114) displacement of the patella in the
standing position (ARNBJÖRNSSON et al. 2002). In
patients with unilateral recurrent patellar dislo-
cation, medial displacement of the patella of the
"healthy" knee has been demonstrated in the stand-
ing position; in addition, a correlation has been
observed between such recurrent dislocation and
medial patellofemoral and femorotibial osteoar-
throsis (ARNBJÖRNSSON et al. 2001 b), comparable to
findings in those with conservative treatment of
dislocation (ARNBJÖRNSSON et al. 2001a). These
observations may be in accordance with those of
SANFRIDSSON et al. (2001a), who demonstrated
increased inward rotation of the tibia in patients
with patellar dislocations.
Patellar and Quadrieeps Mechanism
13.9.2.1
Imaging of the Knee with Recurrent Patellar
Dislocation
It is feit that for the routine pre- and postoperative
radiologieal assessment of recurrent patellar disloca-
tion, there is no alternative to the standing weight-
bearing radiographie techniques. These should inc1ude
AP, lateral femorotibial and axial patellofemoral views
in 30° ofknee flexion (Figs.13.lO, 13.16). Even if there
is lateral patellar displacement in the supine posi-
tion, a normal position or medial displacement of the
patella relative to the femoral trochlea on the axial
radiographie view in the standing position (Fig. 13.50)
may serve as a contraindieation to surgieal transfer
of the tibial tuberosity and lateral release (Fig. 13.39).
Use of the Insall-Salvarti index of patellar height has
also resulted in many distal transfers of the tibial
tuberosity, whieh can be prevented by the measure
suggested by BLACKBURNE and PEEL (1977).
Fig. 13.50. Axial radiographie view of the patellofemoral joint
obtained in the standing position in a 19-year-old female
with anterior knee pain and patellar instability. There is slight
medial displacement of the patella. Would any surgeon per-
form lateral release or medial transfer of the tibial tuberosity
following this imaging appearance?
Advanced radiographie (Figs. 13.20, 13.21) and
MR imaging (Fig.13.31) in the standing position with
measurements of patellar tracking and femorotibial
rotation is still at an early stage of research, but in the
next few years may contribute to better understand-
ing and treatment of recurrent dislocation.
13.9.3
Overuse Syndromes
The term"jumper's knee" has been used to describe a
common painful condition in athletes of the quadri-
243
ceps mechanism of the bone-ligament junction at the
lower patellar pole and less comu;\QnJy at the upper
patellar pole and tibial tuberosity. The condition is
related to sudden and repetitive extension of the knee
such as occurs when running, jumping and kieking.
The correlation between histopathology, ultrasonog-
raphy and MR imaging has been well described
(KHAN et al. 1996, 1997) and indieates that the com-
monly used term "tendinitis" may be inappropriate
and should be replaced by"tendinosis." Histopathol-
ogy may demonstrate fiber failure, mucoid degenera-
tion, and fibrinoid necrosis. Complete quadrieeps or
patellar ligament rupture is rare (SONIN et al. 1995).
In the skeletally immature athlete a focal fragmenta-
tion or acute avulsion may occur at the bony inser-
tion sites of the patellar ligament (Sinding-Larsen-
Johansson disease) or the tibial tuberosity (Osgood-
Schlatter disease) (Figs. 13.51, 13.52, 13.53).
The primary imaging procedure of patients with
a c1inieal history and symptoms of patellar tendino-
sis should be sonography (Fig. 13.54). Tendon abnor-
malities at sonography may not, however, be cor-
related with symptoms (COOK et al. 1997, 2000).
Patellar tendinosis is not uncommonly an accidental
finding at MR imaging in patients with anterior knee
pain and has characteristie appearances (KHAN et al.
1996). The STIR, T2 fat-saturated and Tl gadolini-
um-enhanced fat-saturated sequences may all dem-
onstrate fiber discontinuity and adjacent abnormali-
ties of the fat pad (Fig. 13.55). The magie angle phe-
nomenon may influence the assessment of patellar
Fig.13.51. Lateral radiographie view of avulsion from the tibial
tuberosity (arrow) in a 10-year-old girl with anterior knee pain
244
Fig. 13.52. Ultrasonography of the anterior aspect of the tibial
tuberosity (TT) and patellar ligament (PL) in the normal knee
(Zeft) of the same patient as in Fig. 13.51. The avulsion with its
osseous defect is visualized (right) with loss of normal struc-
tures of the patellar ligament and Doppler enhancement distal
to the avulsion
Fig. 13.53. Sagittal MR STIR sequence of the same knee as in
Figs. 13.51 and 13.52. The defect of the patellar ligament at
the site of the tibial tuberosity is visualized but the sequence
does not contribute any additional information compared
with Figs. 13.51 and 13.52
N. Egund and L. Ryd
Fig.13.54. Jumper's knee. Ultrasonographywith sagittal direc-
tion of sectioning of the anterior aspect of the attachment of
the patellar ligament (PL) to the patella. There is disruption of
fibers of the patellar ligament (arrowhead) and infiammatory
changes are present in the fat pad ofHoffa (arrows). (Courtesy
of L. Bolvig, MD, Aarhus, Denmark)
Fig. 13.55. Jumper's knee. Sagittal MR STIR sequence dem-
onstrating a bulging infiammatory lesion of the fat pad of
Hoffa (arrow) and minor lesions of the patellar ligament at its
attachment to the patella
Patellar and Quadrieeps Mechanism
ligament abnormalities (KARANTANAs et al. 2001).
Both MR imaging and bone scintigraphy (KAHN and
WILSON 1987) may demonstrate bony involvement
of the patella.
13.10
Conclusion
For many years the understanding and treatment of
patellofemoral dis orders has been influenced by con-
ventional radiography and advanced imaging as well
as clinical examination obtained in the supine posi-
tion. We feel that the weight-bearing examination of
the patellofemoral joint may contribute not only to
more accurate diagnostics but also to the understand -
ing of why anterior knee pain occurs, as requested
by INsALL (1995). The weight-bearing techniques of
imaging do not immediately open up a new world
of treatments for patellofemoral instability, but they
may prevent patients from undergoing the tradition-
al surgical treatment, which from a "standing" radi-
ologist's point of view commonly appears contrain-
dicated. Hopefully colleagues worldwide will soon
regard the standing examination techniques with the
same enthusiasm as has been, and still is, shown for
examination in the supine position.
The role of seronegative arthritis and spondylar-
thropathy with possible involvement of entheses in
anterior knee pain has not been explored, but should
always be kept in mind when imaging these patients.
References
Ahl T, Dalen N, Selvik G (1988) Mobilization after operation
of ankle fractures. Good results of early motion and weight
bearing. Acta Orthop Scand 59:302-306
Ahlbäck S (1968) Osteoarthrosis of the knee. A radiographie
investigation.Acta Radiol Diagn (Stockh) [Suppl] 72:1-72
Aleman 0 (1928) Chondromalaeia post-traumatiea patellae.
Acta Chir Scand 63:194
Ando T, Hirose H, Inoue M, Shino K, Doi T (1993) A new
method using computed tomographie scan to measure the
rectus femoris-patellar tendon Q-angle comparison with
conventional method. Clin Orthop 289:213-219
Apostolaki E, Cassar-Pullieino VN, Tyrrell PN, McCall IW
(1999) MRI appearances ofthe infrapatellar fat pad in occult
traumatie patellar dislocation. Clin RadioI54:743-747
Arnbjörnsson A, Egund N, Rydling 0, Stockerup R, Ryd L
(1992) The natural his tory of recurrent dislocation of the
patella. Long-term results of conservative and operative
treatment. J Bone Joint Surg [Br] 74:140-142
Arnbjörnsson A, Egund N, Ryd L (2002) Recurrent dislocation
245
of the patella: the recurrence and site of osteoarthritis in 114
operated knees followed for 14 years. Int Orthop (in press)
Atkin DM, Fithian DC,Marangi KS, Stone ML, Dobson BE, Men-
delsohn C (2000) Characteristies of patients with primary
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-
tion by MRI. Magn Reson Med 41:857-865
Beaconsfield T, Pintore E, Maffulli N, Petri GJ (1994) Radiolog-
ieal measurements in patellofemoral disorders. A review.
Clin Orthop 308:18-28
Blackburne JS, Peel TE (1977) A new method of measuring
patellar height. J Bone Joint Surg [Br] 59:241-242
Blankevoort L, Huiskes R, de Lange A (1988) The envelope of
passive knee joint motion. J Biomech 21:705-720
Boegard T, Rudling 0, Petersson IF, Jonsson K (1998a) Correla-
tion between radiographieally diagnosed osteophytes and
magnetie resonance detected cartilage defects in the patel-
lofemoral joint. Ann Rheum Dis 57:395-400
Boegard T, Rudling 0, Petersson IF, Jonsson K (1998b) Joint-
space width in the weight-bearing radiogram of the tibio-
femoral joint. Should the patient stand on one leg or two?
Acta RadioI39:32-35
Boegard T, Rudling 0, Petersson IF, Sanfridsson J, Saxne T,
Svensson B, Jonsson K (1998c) Joint -space width in the axial
view of the patello-femoral joint. Definitions and compari-
son with MR imaging. Acta RadioI39:24-31
Bradley WG, Ominsky SH (1981) Mountain view of the patella.
AJR Am J Roentgenol136:53-58
Brandt KD, Fife RS, Braunstein EM, Katz B (1991) Radiograph-
ie grading of the severity of knee osteoarthritis: relation of
the Kellgren and Lawrence grade to a grade based on joint
space narrowing, and correlation with arthroscopie evi-
dence of artieular cartilage degeneration. Arthritis Rheum
34: 1381-1386
Brattström H (1964) Shape of the intercondylar groove nor-
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
active and passive knee extension: evaluation with motion-
triggered eine MR imaging. Radiology 187:205-212
Brossmann J, Muhle C, Bull CC, Schroder C, Melchert UH, Zie-
plies J, Spielmann RP, Heller M (1994) Evaluation of patel-
lar tracking in patients with suspected patellar malalign-
ment: eine MR imaging vs arthroscopy. AJR Am J Roent-
genol 162:361-367
Buckland-WrightJC (1994) Quantitative radiography of osteo-
arthritis.Ann Rheum Dis 53:268-275
Burstein D, Velyvis J, Scott KT, Stock KW, Kim YJ, Jaramillo
D, Boutin RD, Gray ML (2001) Protocol issues for delayed
Gd(DTPA)(2- )-enhanced MRI (dGEMRIC) for clinieal eval-
uation of artieular cartilage. Magn Reson Med 45:36-41
Caton J, Deschamps G, Chambat P, Lerat JL, Dejour H (1982)
Patella infera. Apropos of 128 cases. Rev Chir Orthop Repa-
ratriee Appar Mot 68:317-325
Cook JL, Khan KM, Harcourt PR, Grant M, Young DA, Bonar SF
(1997) A cross sectional study of 100 athletes with jumper's
knee managed conservatively and surgieally. The Vietorian
Institute of Sport Tendon Study Group. Br J Sports Med
31:332-336
246
Cook JL, Khan KM, Kiss ZS, Purdam CR, Griffiths L (2000)
Prospective imaging study of asymptomatie patellar tendi-
nopathy in elite junior basketball players. J Ultrasound Med
19:473-479
Delgado-Martins H (1979) A study of the position of the patel-
la using computerised tomography. J Bone Joint Surg [Br]
61:443-444
Dupuy DE, Hangen DH, Zachazewski JE, Boland AL, Palmer W
(I997) Kinematie CT of the patellofemoral joint. AJR Am J
RoentgenoI169:211-215
Edwards JCV (1998) Synovial physiology in the context of osteo-
arthritis. In: Brandt KD, Doherty M, Lohmander LS (eds)
Osteoarthritis. Oxford University Press, Oxford, pp 167-175
Egund N (1986) The axial view of the patello-femoral joint.
Description of a new radiographie method for routine use.
Acta Radiol Diagn (Stockh) 27:101-104
Egund N (2001) The radiographie axial view of the patello-
femoral joint: examination in the supine or standing posi-
tion? (abstract). Eur Radiolll:130
Egund N, Friden T (1988) Lesion of the anterior cruciate liga-
ment and sagittal disalignment of the knee in weight-bear-
ing. Acta RadioI29:559-563
Egund N, Lundin A, Wallengren NO (I988) The vertieal posi-
tion of the patella. A new radiographie method for routine
use. Acta RadioI29:555-558
Egund N, Friden T, Hjarbaek J, Lindstrand A, Stockerup R
(I993) Radiographie assessment of sagittal knee instability
in weight bearing. A study on anterior cruciate-deficient
knees. Skeletal RadioI22:177-181
Egund N, Jacobsen BW, J0rgensen A, Kristiansen SE (2000)
Angles and shape of the artieular surfaces of the tibial con-
dyles: measurement by MR imaging in the sagittal plane.
Presented at the European Society of Musculoskeletal Radi-
ology, Leiden
Egund N, Hol PK, Johnston V, Tillung T, Fosse E (2001)
Normal femorotibial rotation and patellar tracking in the
weight-bearing standing and the non-loaded supine posi-
tion: assessment and comparison by kinematic MR imag-
ing (abstract). Radiology (Suppl) 609
Felson DT (I998) Epidemiology of osteoarthritis. In: Brandt
KD,Doherty M,Lohmander LS (eds) Osteoarthritis. Oxford
University Press, Oxford, pp 13-22
Fieat R, Hungerford DS (1977) Disorders of the patello-femoral
joint. Williams & Wilkins, Baltimore
Friden T, Sommerlath K, Egund N, Gillquist J, Ryd L, Lind-
strand A (1992) Instability after anterior cruciate ligament
rupture. Measurements of sagittal laxity compared in 11
cases. Acta Orthop Scand 63:593-598
Friden T, Egund N, LindstrandA (1993) Comparison of symp-
tomatic versus nonsymptomatie patients with chronie ante-
rior cruciate ligament insufficiency. Radiographie sagittal
displacement during weightbearing. Am J Sports Med
21:389-393
Fulkerson JP (I 983) The etiology of patellofemoral pain in
young, active patients: a prospective study. Clin Orthop
179:129-133
Fulkerson JP (I997) Disorders of the patellofemoral joint. Wil-
liams and Wilkins, Baltimore
Fulkerson JP, Shea KP (I 990) Disorders of patellofemoral align-
ment. J Bone Joint Surg [Am] 72:1424-1429
Goodfellow J, Hungerford DS, Woods C (1976) Patello-femoral
joint mechanics and pathology. 2. Chondromalacia patel-
lae. J Bone Joint Surg [Br] 58:291-299
N. Egund and 1. Ryd
Grelsamer RP, Meadows S (I992) The modified Insall-Salvati
ratio for assessment of patellar height. Clin Orthop
282:170-176
Hampson WG,Hill P (1975) Late results oftransfer ofthe tibial
tubercle for recurrent dislocation of the patella. J Bone Joint
Surg [Br] 57:209-213
Harrison MM, Cooke TD, Fisher SB, Griffin MP (1994) Patterns
of knee arthrosis and patellar subluxation. Clin Orthop
309:56-63
Hawkins RJ, Bell RH, Anisette G (I986) Acute patellar disloca-
tions. The natural history. Am J Sports Med 14: 117-120
Henche HR, Kunzi HU, Morscher E (1981) The areas of contact
pressure in the patello-femoral joint. Int Orthop 4:279-281
Hungerford DS, Barry M (1979) Biomechanies of the patel-
lofemoral joint. Clin Orthop 144:9-15
Insali JN (I995) Foreword. In: Scuderi GR (ed) The patella.
Springer, Berlin Heidelberg New York
Insali JN, Salvarti E (I971) Patella position in the normal knee.
Radiology 101:101-104
Insali JN, Falvo KA, Wise DW (1976) Chondromalacia patellae.
A prospective study. J Bone Joint Surg [Am] 58:1-8
Jaroschy W (1924) Die diagnostiehe Verwertbarkeit der Patel-
laraufnahmen. Fortsehr Roentgenstr 31:781-782
Jones RB, Barlett EC, Vainright JR, Carroll RG (1995) CT deter-
mination of tibial tubercle lateralization in patients present-
ing with anterior knee pain. Skeletal RadioI24:505-509
Juliusson R, Markhede G (1984) A modified Hauser procedure
for recurrent dislocation of the patella. A long-term fol-
low-up study with special reference to osteoarthritis. Arch
Orthop Trauma Surg 103:42-46
Kahn D, Wilson MA (1987) Bone scintigraphie findings in
patellar tendonitis. J Nucl Med 28: 1768-1770
Karantanas AH, Zibis AH, Papanikolaou N (2001) Increased
signal intensity on fat -suppressed three-dimensional Tl-
weighted pulse sequences in patellar tendon: magie angle
effect? Skeletal Radio130:67 -71
Kärrholm J, Hansson LI, Svensson K (1983) Predietion of
growth pattern after ankle fractures in children. J Pediatr
Orthop 3:319-325
Kärrholm J, Selvik G, Elmqvist LG, Hansson LI (1988) Active
knee motion after cruciate ligament rupture. Stereoradiog-
raphy. Acta Orthop Scand 59: 158-164
Kärrholm J, Borssen B, Lowenhielm G, Snorrason F (1994a)
Does early mieromotion of femoral stern prostheses matter?
4-7-year stereoradiographie follow-up of 84 cemented
prostheses. J Bone Joint Surg [Br] 76:912-917
Kärrholm J, Jonsson H, Nilsson KG, Soderqvist I (1994b) Kine-
maties of successful knee prostheses during weight-bear-
ing: three-dimensional movements and positions of screw
axes in the Trieon-M and Miller-Galante designs. Knee Surg
Sports Traumatol Arthrosc 2:50-59
Keats TE (1996) Atlas of normal roentgen variants that may
simulate disease. Mosby, St Louis
Kellgren JH, Lawrence JS (1957) Radiologie assessment of
osteoarthritis.Ann Rheum Dis 16:494-501
Khan KM, Bonar F, Desmond PM, Cook JL, Young DA, Visentini
PJ, Fehrmann MW, Kiss ZS, O'Brien PA, Harcourt PR, Dowl-
ing RJ, O'Sullivan RM, Criehton KJ, Tress BM, Wark JD (1996)
Patellar tendinosis (jumper's knee): findings at histopatho-
logie examination, US, and MR imaging. Vietorian Institute
of Sport Tendon Study Group. Radiology 200:821-827
Khan KM, Cook JL, Kiss ZS, Visentini PJ, Fehrmann MW, Har-
court PR, Tress BW, Wark JD (1997) Patellar tendon ultra-
Patellar and Quadrieeps Mechanism
sonography and jumper's knee in female basketball players:
a longitudinal study. Clin J Sport Med 7:199-206
Knutsson F (1941) Über die Röntgenologie des Femoropatel-
largelenks sowie eine gute Projektion für das Kniegelenk.
Acta Radiol Diagn (Stockh) 22:371-376
Kuhl CK, Schild HH (2000) Dynamie image interpretation of
MRI of the breast. J Magn Reson Imaging 12:965-974
Langer M (1929) Über die Entwieklung des Kniegelenkes. Z
Gesamte Anat 89:83-10 1
Laurin CA, Levesque HP, Dussault R, Labelle H, Peides JP
(1978) The abnormal lateral patellofemoral angle: a diag-
nostie roentgenographie sign of recurrent patellar sublux-
ation. J Bone Joint Surg [Am] 60:55-60
Laurin CA, Dussault R, Levesque HP (1979) The tangential x-ray
investigation of the patellofemoral joint: x-ray technique,
diagnostie criteria and their interpretation. Clin Orthop
144:16-26
Ledingham J,Regan M, Jones A, Doherty M (1993) Radiograph-
ie patterns and associations of osteoarthritis of the knee in
patients referred to hospital. Ann Rheum Dis 52:520-526
Macnab J (2001) Recurrent dislocation of the patella. J Bone
Joint Surg [Am] 34:957-967
Mäenpaa H, Lehto MU (1997a) Patellar dislocation. The long-
term results of nonoperative management in 100 patients.
Am J Sports Med 25:213-217
Mäenpaa H, Lehto MU (1997b) Patellofemoral osteoarthritis
after patellar dislocation. Clin Orthop 339: 156-162
Martinez S, Korobkin M, Fondren FB, Hedlund LW, Goldner JL
(1983a) Computed tomography of the normal patellofemo-
ral joint. Invest RadioI18:249-253
Martinez S, Korobkin M, Fondren FB, Hedlund LW, Goldner JL
(1983b) Diagnosis of patellofemoral malalignment by com-
puted tomography. J Comput Assist Tomogr 7:1050-1053
McNally EG, Ostlere SJ, Pal C, Phillips A, Reid H, Dodd C (2000)
Assessment of patellar maltracking using combined statie
and dynamie MRI. Eur RadioI1O:1051-1055
Merchant AC (1988) Classification of patellofemoral disorders.
Arthroscopy 4:235-240
Merchant AC, Mercer RL,Jacobsen RH, Cool CR (1974) Roent-
genographie analysis of patellofemoral congruence. J Bone
JointSurg [Am] 56:1391-1396
Moore TM, Harvey JP (1974) Roentgenographie measurement
of tibial-plateau depression due to fracture. J Bone Joint
Surg [Am] 56:155-160
Muneta T, Yamamoto H, Ishibashi T,Asahina S, Furuya K (1994)
Computerized tomographic analysis of tibial tubercle posi-
tion in the painful female patellofemoral joint. Am J Sports
Med 22:67-71
Nagamine R, Miura H, Inoue Y, Tanaka K, Urabe K, Okamoto
Y, Nishizawa M, Iwamoto Y (1997) Malposition of the tibial
tubercle during flexion in knees with patellofemoral arthri-
tis. Skeletal Radio126:597 -601
Nagamine R, Miura H, Urabe K, Matsuda S, Chen WJ, Matsu-
nobu T, Iwamoto Y (1999) Radiologieal assessment of the
position of the tibial tuberosity by means of a marking
wire in knees with patellofemoral arthritis. Skeletal Radiol
28:27-32
Önsten I, Bengner U, Besjakov J (1993) Socket migration after
Charnley arthroplasty in rheumatoid arthritis and osteo-
arthritis. A roentgen stereophotogrammetrie study. J Bone
Joint Surg [Br] 75:677-680
0stergaard M, Stoltenberg M, Lovgreen-Nielsen P, Volck B,
Sonne-Holm S, Lorenzen I (1998) Quantification of synovi-
247
tis by MRI: correlation between dynamie and statie gado-
linium-enhanced magnetie resonance imaging and miero-
scopie and macroscopie signs of synovial inflammation.
Magn Reson Imaging 16:743-754
Outerbridge R (1961) The etiology of chondromalacia patellae.
J Bone Joint Surg [Br] 43:752-757
Post WR (1997) History and physieal examination. In: Fulker-
son JP (ed) Disorders of the patellofemoral joint. Williams
and Wilkins, Baltimore, pp 39-71
Radin EL (1979) A rational approach to the treatment ofpatel-
lofemoral pain. Clin Orthop 144: 107-109
Radin EL, Rose RM (1986) Role of subchondral bone in the
initiation and progression of cartilage damage. Clin Orthop
213:34-40
Ragnarsson JI, Eliasson P, Karrholm J, Lundstrom B (1992) The
accuracy of measurements of femoral neck fractures. Con-
ventional radiography versus roentgen stereophotogram-
metrie analysis. Acta Orthop Scand 63:152-156
Resniek D, Niwayama G (1995) Degenerative diseases. In: Resn-
iek D (ed) Diagnosis of bone and joint disorders. Saunders,
Philadelphia, pp 1263-137l
Rünow A (1983) The dislocating patella. Etiology and progno-
sis in relation to generalized joint laxity and anatomy of the
patellar artieulation. Acta Orthop Scand SuppI201:1-53
Ryd L, Lindstrand A, Rosenquist R, Selvik G (1986) Tibial
component fixation in knee arthroplasty. Clin Orthop
213:141-149
Ryd L, Albrektsson BE, Carlsson L, Dansgard F, Herberts P,
Lindstrand A, Regner L, Toksvig-Larsen S (1995) Roentgen
stereophotogrammetrie analysis as apredietor of mechani-
calloosening of knee prostheses. J Bone Joint Surg [Br]
77:377-383
Sallay PI, Poggi J, Speer KP, Garrett WE (1996) Acute disloca-
tion of the patella. A correlative pathoanatomie study. Am
J Sports Med 24:52-60
Sanfridsson J, Arnbjornsson A, Friden T, Ryd L, Svahn G, Jons-
son K (2001a) Femorotibial rotation and the Q-angle relat-
ed to the dislocating patella. Acta Radiol 42:218-224
Sanfridsson J, Ryd L, Svahn G, Friden T, Jonsson K (2001b)
Radiographie measurement of femorotibial rotation in
weight-bearing. The influence of flexion and extension in
the knee on the extensor mechanism and angles of the lower
extremity in a healthy population. Acta Radio142:207 - 217
Schutzer SF, Ramsby GR, Fulkerson JP (1986) Computed
tomographie classification of patellofemoral pain patients.
Orthop Clin North Am 17:235-248
Scuderi GR (1995) The patella. Springer, Berlin Heidelberg
NewYork
Seil R, Muller B, Georg T, Kohn D, Rupp S (2000) Reliability
and interobserver variability in radiologieal patellar height
ratios. Knee Surg Sports Traumatol Arthrosc 8:231-236
Selvik G (1974) A roentgen stereophotogrammetrie method
for the study of the kinematies of the skeletal system. Lund
University, Sweden
Settegast J (1921) Typische Röntgenbilder von normalen Men-
schen. Lehmanns Med Atlanten 5:211
Shellock FG, Mink JH, Deutsch AL, Fox JM (1989) Patellar
tracking abnormalities: clinieal experience with kinematie
MR imaging in 130 patients. Radiology 172:799-804
Siu D, Cooke TD, Broekhoven LD, Lam M, Fisher B, Saunders
G, Challis TW (1991) A standardized technique for lower
limb radiography. Practiee, applications, and error analysis.
Invest Radiol 26:7l-77
248
Sonin AH, Fitzgerald SW, Bresler ME, Kirsch MD, Hoff FL,
Friedman H (1995) MR imaging appearance of the exten-
sor mechanism of the knee: functional anatomy and injury
patterns. Radiographies 15:367-382
Spector TD, Hochberg MC (1994) Methodologieal problems
in the epidemiologieal study of osteoarthritis. Ann Rheum
Dis 53:143-146
Stanitski CL (1995) Articular hypermobility and chondral
injury in patients with acute patellar dislocation. Am J
Sports Med 23:146-150
Stanitski CL, Paletta GA (1998) Artieular cartilage injury
with acute patellar dislocation in adolescents.
Arthroscopic and radiographie correlation. Am J Sports
Med 26:52-55
Starok M, Lenchik L, Trudell D, Resniek D (1997) Normal patel-
lar retinaculum: MR and sonographie imaging with cadav-
erie correlation. AJR Am J Roentgenol168: 1493-1499
Stein LA, Endieott AN, Sampalis JS, Kaplow MA, Patel MD,
Mitchell NS (1993) Motion of the patella during walking: a
N. Egund and L. Ryd
video digital-fiuoroscopie study in healthy volunteers. AJR
Am J Roentgenol161 :617-620
Thomee R, Augustsson J, Karlsson J (1999) Patellofemoral
pain syndrome: a review of current issues. Sports Med
28:245-262
Uvehammer J, Karrholm J, Brandsson S (2000) In vivo kine-
maties of total knee arthroplasty. Concave versus poste-
rior-stabilised tibial joint surface. J Bone Joint Surg [Br1
82:499-505
van den Berg WB, van der Kraan PM, van Beuningen HM
(1998) Synovial mediators of cartilage damage and repair in
OA. In: Brandt KD, Doherty M, Lohmander LS (eds) Osteo-
arthritis. Oxford University Press, Oxford, pp 157-167
Wiberg G (1941) Roentgenographie and anatomie studies on
the femoro-patellar joint. Acta Orthop Scand 12:319-410
Woodland LH, Frands RS (1992) Parameters and comparisons
of the quadrieeps angle of college-aged men and women
in the supine and standing positions. Am J Sports Med
20:208-211
14 Infection
MOHAMED E. ABD EL BAGI, MONA S. AL SHAHED, BASSAM M. SAMMAK

CONTENTS trauma, arthropathies, haemarthrosis, infarets and

degenerative disease. Moreover the knee is the eom-
14.1 Introduction 249
14.1.1 Overview 249
monest site for bone and some soft tissue malignan-
14.1.2 Classification of Knee Infections 250 eies (FRINK et al. 1998). Aeeurate diagnosis of the
14.2 Infection Imaging Techniques 250 symptomatie knee is therefore of paramount clinieal
14.2.1 Plain Radiographs 250 importanee. Signs and symptoms of knee infeetions
14.2.2 Nuclear Medicine Scanning 251 ean be produeed by many non-infeetive eonditions;
14.2.3 Magnetie Resonance Imaging 253
14.2.4 Computed Tomography 254
sueh signs and symptoms include loeal swelling,
14.2.5 Ultrasound 254 pain, hotness, redness, limitation of movement, fever
14.3 Common Knee Infections 256 and eonstitutional symptoms. Equally, joint effusions
14.3.1 Osteomyelitis 256 may be due to trauma, including internal derange-
14.3.2 Septie Arthritis 258 ments and overuse syndromes, or non-traumatic
14.3.3 Pyogenie Myositis 259
14.3.4 Septie Bursitis 259
eauses sueh as arthritis, erystal disease, synovitis and,
14.3.5 Infective Cellulitis 260 more seriously, infeetion and malignaney. On the
14.4 Rare Bone Infections 261 other hand, infeetion ean eo-exist with or eomplicate
14.4.1 Tuberculosis 261 pre-existing diseases like infarets, trauma or arthri-
14.4.2 Brucellosis 262 tis. A thorough medieal history is a key eomponent
14.4.3 Salmonellosis 262
14.4.4 Mycetoma 262
of evaluation. Systematie physical examination and
14.4.5 Hydatid Disease 263 eomparison with the unaffeeted knee are essential.
14.5 Selected Topies 263
14.5.1 Knee Infections in Children 263
14.5.2 Knee Infection in Siekle Cell Disease 263 14.1.1
14.5.3 The Immunocompromised Patient 265
14.6 Summary 265
Overview
References 266
Early deteetion of museuloskeletal infeetions at a
stage where eomplete resolution is still possible is a
ehallenge to both clinicians and radiologists (ELGAZ-
ZAR et al. 1995). The term "eure",however,is not appli-
14.1 eable in osteomyelitis beeause it may reeur years later
Introduction if the affeeted bone is subjeeted to trauma or if the
patient's immunity is deereased (MADER et al. 1999).
The knee is a eommon site for infeetions (DIRSCHL There is no gold standard for the diagnosis of osteo-
and KINDERS 1993). It is also a eommon site for myelitis (SCHAUWECKER et al. 1990). It is therefore
vital to understand the limitations of eaeh imaging
M.E. ABD EL BAGI, MD, DMRD, FFRRCSI modality to avoid delay in diagnosis (SAMMAK et al.
Senior Consultant Radiologist (Teaching), Department of
Radiology, Riyadh Armed Forces Hospital, PO Box 7897,
1999). Likewise, laboratory investigations, blood tests
Riyadh 11159, Kingdom of Saudi Arabia and loeal fluid aspirations have their limitations even
M.S. AL SHAHED, MD, FRCR when eombined. Serial blood cultures are negative in
Department of Radiology, Riyadh Armed Forces Hospital, PO 40%-50% of patients with osteomyelitis. Cultures of
Box 7897, Riyadh 11159, Kingdom of Saudi Arabia both blood and ne edle aspirate are negative in 20%
B.M. SAMMAK, MD, FRCR
Consultant Radiologist, Department of Radiology, Riyadh
(ELGAZZAR et al. 1995). Blood eultures are frequently
Armed Forces Hospital, PO Box 7897, Riyadh 11159, Kingdom positive in aeute osteomyelitis but are usually nega-
of Saudi Arabia tive in ehronic osteomyelitis. The sensitivity of micro-

A.M. Davies et al. (eds.), Imaging of the Knee

© Springer-Verlag Berlin Heidelberg 2003
250
biological aspiration cultures has been reported to
be as low as 42% but was found to increase to 84%
when culture tests were combined with histological
core biopsies (WHITE et al. 1995). Certain laboratory
tests such as white blood count (WBC), erythrocyte
sedimentation rate (ESR) and C-reactive protein (CRP)
are helpful diagnostic and baseline tests. In one study
a WBC alone was high in only 25% of children with
osteomyelitis while the differential count was abnor-
mal in 65% (DIRSCH and KINDERS 1993). Serial CRP
measurements are a more accurate means for moni-
toring of response to treatment and predicting out-
come than ESR or radiographs (ROINE et al. 1995).
14.1.2
Classification of Knee Infections
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
trauma, general illness, malnutrition, extremes of
age, diabetes, neuropathies, sickle cell disease, sur-
gery, interventions, steroid treatment and immuno-
compromise. Understanding the pathological anato-
my of knee infections is aprerequisite for correct
diagnosis and treatment plans. Knee infection may
arise in the bone (osteomyelitis), synovium (syno-
vitis), articular structures (septic arthritis), synovial
cysts (bursitis), muscle (myositis) and subcutaneous
tissues (cellulitis or fasciitis).
Knee infections may be acute, subacute or chronic.
The chronic forms may be active or inactive. Infec-
tion may be due to (1) direct extension from a local
focus, (2) haematogenous spread, (3) direct implan-
tation of a pathogen or (4) surgery. In considering the
choice of imaging technique, it is important to know
whether infection is developing in a "violated knee",
i.e. violated by surgery or injuries, or in a "non-vio-
lated knee". Postoperative infections are dealt with in
Chap. 11. In this chapter we will concentrate on infec-
tions in the non-violated knee.
Haematogenous spread ofbone infection can arise
from urinary tract infection, pneumonia, skin boils,
otitis or endocarditis, but the source may remain
unidentified (DIRSCHL and KINDERS 1993). The
aforementioned are said to induce bacteraemia, but
transient bacteraemia is a daily event, e.g. after tooth
brushing. Staphylococci are the commonest cause of
all joint and bone infections in all groups. More than
98% of these organisms are now resistant to peni-
cillins and are hence treated by cephalosporins. In
M. E. Abd EI Bagi et al.
methicillin-resistant Staph. aureus (MRSA), cell wall
protein is resistant to all b-lactam drugs and glyco-
peptides such as vancomycin are required (BOUZA
and MUNOZ 1999). Coagulase-negative staphyloc-
coci commonly affect prosthetic joints. Anaerobes
are common in vasculopathies and diabetes. Multi-
drug-resistant tuberculosis is an emergent disease
in acquired immunodeficiency syndrome (AIDS).
In tropical countries and predisposed patients, rare
bone infections commonly involve the knee, e.g.
tuberculosis, brucellosis, salmonella and fungi (ABD
EL BAGI et al. 1999).
The radiological evaluation of knee infection is
extremely useful for treatment plans. Differentiation
between osteomyelitis, cellulitis and abscess forma-
tion is crucial in selecting treatment options such
as operative intervention, percutaneous drainage or
medical treatment (BOUTIN et al. 1998).
14.2
Infection Imaging Techniques
14.2.1
Plain Radiographs
Conventional radiographs in the frontal and lateral
projections are the first modality for the work-up of
patients suspected to have knee infections. They are
simple, readily available and cheap (Fig. 14.1). The
earliest change appreciable in knee infections is peri-
articular soft tissue swelling or displacement of the
normal fat planes. This may be apparent as early as 2
days. Development of bone hyperaemia manifests as
osteopenia after 7-10 days. The earliest bone change
is loss of the trabecular pattern. Periosteal re action
may appear next. The location of the soft tissue swell-
ing may indicate the infection harbour, e.g. swelling in
the suprapatellar recess in septic arthritis or behind
the joint in infected Baker's cyst. In acute osteomyeli-
tis,loss of bone density must reach 30%-50% before
radiographs become abnormal. Overt bone erosions
take 2-6 weeks to show on plain films. In chronic
osteomyelitis, diffuse sclerosis, sequestra, involucrum
formation and intramedullary abscess are evident on
plain radiographs (DAVID et al. 1987).
Plain radiographs often help in the interpretation
of the plentiful findings of sophisticated tests in com-
plicated cases. An important but under-emphasized
role for plain radiographs lies in indicating other
pathology or mimics. Sclerotic bone malignancies
may be confused with chronic osteomyelitis. Trau-
Infection
Fig. 14.1. Plain radiograph signs of knee infection: (1) soft
tissue swelling (short white arrows); (2) periostitis (open white
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-
tion. A Charcot's neuropathic joint may show charac-
teristie sclerosis, disorganization or periostitis.
Sinography. When a chronie discharging sinus is
present, the injection of water-soluble contrast agent
into the cutaneous opening may provide an excellent
demarcation of its extent and direction (Fig. 14.2).
14.2.2
Nuclear Medicine Scanning
14.2.2.1
Three-Phase Bone Scintiscan
Three-phase bone scintigraphy may become positive
a few hours after the onset of focal infections, but
in practiee is typically not performed until 24-48 h
from the onset of clinieal symptoms. Technetium-
99m (99ffiTc)-methylene diphosphonate (MDP) is the
principal radioisotope in most institut ions. Less com-
monly, hydroxymethylene diphosphonate (HMDP) is
used. Increased uptake in the early first-pass perfu-
sion phase and the blood pool phase at about 10 min
post injection indicates soft tissue inflammation or
251
Fig. 14.2. Sinogram demarcating the extent and direction of
the small sinus opening in the skin
infection. If such uptake is combined with increased
tracer accumulation in the delayed statie phase at 3 h,
osteomyelitis should be diagnosed (Fig. 14.3). How-
ever, if there are sizable subperiosteal pustular pock-
ets, large joint effusions or rapidly destructive pro-
cesses, cold spots will appear.
The sensitivity of 99ffiTc_MDP scintigraphy varies
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-
ware. Scintigraphy is highly sensitive but non-specif-
ic. Correlation with clinical and radiographie find-
ings is essential. Sensitivity is lower in children,
osteoporosis, vasculopathies and metabolic bone dis-
ease. Specificity is much reduced in violated knees
because the physiological effects of trauma and infec-
tion are similar. A single-photon emission computed
tomography facility increases test sensitivity for small
lesions owing to superior resolution. Other lesions,
such as arthritis, healing fractures or previously treat-
ed osteomyelitis show little abnormality in the first
two phases but may show markedly increased uptake
on delayed bone images. Furthermore, a four-phase
scintiscan incorporating a 24-h delayed image will
help to clarify the significance of any doubtful activ-
ity. On a four-phase scan at 24 h, activity of normal
252
Fig. 14.3. 99mTc_MDP bone scintigraphy showing increased
uptake in the delayed static phase due to osteomyelitis
bone disappears whereas activity in "woven" imma-
ture bone, as in osteomyelitis, is still present, giving
a high lesion to background ratio. Woven bone is
also present in fractures, tumours and degenerative
disease, where a four-phase scan is less rewarding
(PALESTRO and TORRES 1997). Some workers empha-
size comparison of quantification ratios in three-
and four-phase scans to improve bone scan specific-
ity (GuPTA and PREZIO 1988). As compared with a
three-phase scan, the four-phase scan was shown to
increase the specificity for detection of infections in
violated bone from 73% to 87% in a selected group
of patients, though this was not achieved without
a reduction in sensitivity (ELGAZZAR et al. 1995).
Other radioisotopes discussed belowwill improve the
specificity of the four-phase scan. However, if bone
scan is normal, there is no need to proceed with other
radiopharmaceuticals. Rather, the patient should be
re-evaluated for non-infective conditions. No single
radioisotope scan is useful in all situations (PALESTRO
and TORRES 1997).
14.2.2.2
Gallium-67
A gallium citrate (67 Ga) scan takes longer (24-72 h)
than a 99mTc_MDP bone scan and gives a higher radia-
tion dose. Increased uptake of 67Ga at the infection
site is attributed to in vivo labelling of serum pro-
teins, leucocyte lysosomes and endoplasmic retieu-
lum along with increased vascular permeability and
M. E. Abd EI Bagi et al.
direct bacterial uptake. 67Ga scan is useful when the
99mTc_MDP scan is non-specific. A sequential techne-
tium-gallium scan has 70% accuracy (MERKEL et al.
1985). The dual-tracer technique is useful because the
two agents have different mechanisms. It is, howev-
er,less useful in violated bone or neuropathie joints.
According to the modified Rosenthai criteria, osteo-
myelitis is diagnosed when there is incongruence or
more uptake in the 67Ga scan as compared with the
99mTc-MDP (DONOHOE 1998). Normal or relatively
less 67Ga uptake at a hot site on the 99mTc_MDP scan
excludes infection (MERKEL et al. 1985). A sequen-
tial gallium/technetium scan is equivocal for infec-
tion when the distribution of the two radiotracers is
specifically congruent and their relative intensity is
the same. 67Ga uptake alone is not very specific for
osteomyelitis as 67Ga accumulates at sites of cellulitis
as weIl. Furthermore, the resolution of a 67Ga scan
is not sufficient to differentiate between bone and
soft tissue. A gallium scan may be false-negative
in patients who have had antibiotie treatment or
patients with low-grade osteomyelitis. 67Ga scan is
also possible in primary or secondary tumours. In
children, 67Ga scan has a higher diagnostie accuracy
than conventional bone scans, but it is less specific
in the extremities (GuPTA and PREZIO 1988). 67Ga
is preferred to 99mTc scans for assessment of the
response of chronie infections to therapy.
14.2.2.3
Indium-111 Labelled Leucocytes
Imaging with indium-lI 1 B-hydroxyquinoline (oxine)
complex labelled leucocytes (lllIn-WBCs) can be pos-
itive in the first hour but is typieally performed at 4
and 24 h. This technique is 83% sensitive and 94%
specific (KOLINDOU et al. 1996), as compared to 50%
sensitivity and 78% specificity for a sequential 99mTc_
MDp/67 Ga scan (MERKEL et al. 1985). Infection sites
are usually hot, but cold spots have been reported
in 12%, usually in red marrow-rich areas (DATZ and
THORNE 1987). lllIn-WBC scan is therefore oflimited
value in red marrow-containing areas like the spine,
hips and knees. Comparison should be made with the
contralateral side.Alternatively a 99mTc-sulphur colloid
bone marrow scan is helpful. If the marrow scan
shows increased accumulation, this indieates increased
haemopoiesis. When there is increased activity on
the lllIn-WBC scan without increased activity in the
marrow image, the combined test is positive for infec-
tion. If the two tracers of WBC and marrow scan are
congruent, infection is unlikely. The accuracy of the
combined leucocyte/marrow scan ranges from 89% to
Infection
98% (PALESTRO and TORRES 1997). lllIn-WBC scan is
less sensitive for chronie osteomyelitis (60%) although
highly specific (96%). This is due to the lack of granu-
locytes in chronic inflammatorytissues (SCHAUWECK-
ER et al. 1984).
11lIn-Iabelled compounds can, however, give false-
positive results in trauma, arthritis, neuropathie joints
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
withdrawn, labelled and re-injected. This is a prob-
lem particularly in children. 11lIn-WBC test is more
expensive and not readily available, requiring special
orders. Despite these limitations, lllIn-labelled com-
pounds are probably the tracers of choiee for diagno-
sis and localization of osteomyelitis. Combined use of
lllIn-WBCs and 99mTc was found to have a high sen-
sitivity of 86% and a specificity of 94% when infec-
tion occurred in the presence of pre-existing fr ac-
ture, arthropathies, previous surgery or skin lesions
(KOLINDOU et al. 1996).
14.2.2.4
Indium-111 Labelled Polyclonallmmunoglobulin
Clinieal studies indieate that this new agent is as effi-
cacious as 11lIn-WBCs or 67Ga citrate for the evalua-
tion of focal infection. Its preparation is simpler than
that of llIIn-WBCs, evading phlebotomy and labo-
rious labelling methods. Unlike with 67Ga, there is
no gastrointestinal or bone marrow uptake. A lower
radiation dose is delivered, but lesion to background
activity ratios are relatively low. The agent is pre-
pared from pooled human serum gamma-globulin
andlabelled with lllIn viaDTPA chelation ellln-IgG).
In the first 24 h the non-specific polyclonal IgG shows
lesion uptake equal to that of specific monoclonal
IgG uptake. Thereafter, the lesion uptake of specific
monoclonal IgG exceeds that of control non-specific
IgG (NIJHOF et al. 1997). The tracer does not differ-
entiate between infection and sterile inflammation.
Indium is rarely taken up by neoplasms. False-posi-
tive results have been reported in arthritis and Char-
cot's joint.
14.2.2.5
Technetium-99m d,l-HMPAO Labelled Leucocytes
Advantages of 99mTc-Iabelled over lllIn-labelled leu-
cocytes include cost, availability, dosimetry, shorter
acquisition time and improved image quality. WBC
scan using the lipophilic complex d,l-hexamethylpro-
pylene amine oxime (HMPAO) was reported to have a
253
specificity of 93% and a sensitivity of 100% (RODDIE
et al. 1998). 99mTc-HMPAO is very reliable and is
probably the best leucocyte-Iabelling agent for the
evaluation of chronie osteomyelitis (McAFEE 1990).
Bone marrow activity is more prominent on 99mTc_
HMPAO-WBC scan than on lllIn-WBC scan; this can
give rise to false-positive results.
14.2.3
Magnetic Resonance Imaging
For the detection of osteomyelitis, magnetie reso-
nance (MR) imaging is superior to 99mTc_MDP bone
scintigraphy in terms of both sensitivity and spec-
ificity (MAZUR et al. 1995). Thus several studies
have reported sensitivities and specificities for MRI
of 92%-100% and 89%-100%, respectively (SCHAU-
WECKER et al. 1990). Similar to scintigraphy, the
MR appearance of osteomyelitis can be non-specific
because infiltrating tumours and stress fractures can
give the same appearance (BOUTIN et al. 1998).
The excellent soft tissue and bone marrow con-
trast resolution and the multiplanar capability of MR
imaging offer greater detail than computed tomogra-
phy (CT) or conventional radiographs. Early changes
in bone marrow signal on MR are a more sensitive
indicator of disease than the late-developing destruc-
tive bone lesions, whieh need further time to show
radiographieally. MR imaging becomes positive after
a few hours but a radio graph takes more than a week.
Active infections replace marrow fat with water sec-
ondary to oedema, exudate, hyperaemia and isch-
aemia. This leads to a high signal on T2-weighted
images and a low to intermediate signal on Tl-
weighted images. Short tau inversion recovery (STIR)
and fat suppression before and immediately after con-
trast enhancement improve the detection of osteo-
myelitis and soft tissue infections.
The bone marrow signal becomes heterogeneous
with the advent of chronicity. Granulation tissue
is hyperintense on T2" STIR and gradient echo
sequences (Fig. 14.4). Diffuse sclerosis can lower
the signal. When the primary signs of infection are
equivocal (low to intermediate signal on Tl weight-
ing and high signal on T2 weighting), secondary
signs of osteomyelitis will aid in the differential diag-
nosis from tumours, arthritis or neuropathie con-
ditions (MORISSON et al. 1998). Secondary signs of
infection include ulcer, cellulitis, abscess formation,
cortical sequestration and subperiosteal abscess for-
mation. Recently a penumbra sign on Tl-weighted
images was reported to be helpful in differentiating
254
Fig.14.4. This patient had a plain radiograph consistent with
chronic osteomyelitis. MR imaging demonstrated hyperin-
tense granulation tissue on gradient echo sequence. There is
breakthrough of the cortex and a soft tissue component. This
raised the possibility of a tumour. Frozen section and oper-
ative biopsy were negative for tumours or lymphoma. Para-
aortic Iymph node biopsy and culture revealed tuberculosis
(see Fig. 14.21)
subacute osteomyelitis from bone tumours, show-
ing a sensitivity of 75% and a specificity of 99%
(GREY et al. 1998). Contrast-enhanced MR imaging
is a highly sensitive technique to diagnose musculo-
skeletal infection and differentiate abscess from sur-
rounding myositis or cellulitis (HOPKINS et al. 1995)
(Fig. 14.5). Lack of contrast enhancement rules out
infection with a high degree of certainty but pres-
ence of enhancement cannot be used to exclude non-
infective inflammatory conditions entirely. Although
MR imaging is increasingly useful in diagnosis and
the planning of surgical management, it is of lim-
ited use for multifocal infection. This emphasizes
the value of scintigraphy, which allows whole-body
scanning; this capability is of particular use in chil-
dren, 7% of whom have multifocal disease.
14.2.4
Computed Tomography
Computed tomography is superior to MR imaging for
visualization of bone destruction and gas formation
(RAM et al. 1981) (Fig. 14.6). CT demonstrates bone
abnormalities much earlier than do radiographs. The
earliest sign on CT is increased attenuation values
of the medullary cavity, an abnormality undetectable
M. E. Abd EI Bagi et al.
Fig. 14.5. Post-contrast fat-suppressed Tl-weighted coronal
MR image in a girl with acute osteomyelitis of the right femur
and soft tissue involvement
on X-rays. Changes of chronic osteomyelitis are weIl
demonstrated by CT as there is no overlap, which is
a problem on radiographs. The improved soft tissue
contrast and the use of contrast enhancement are
useful for the diagnosis of bone and soft tissue infec-
tion, and particularly abscess formation. CT is espe-
cially important in patients in whom MR imaging is
contra-indicated. It is also very useful when infec-
tion is superposed on comminuted fractures (LEDER-
MANN et al. 2000) (Fig. 14.6). CT can differentiate
whether increased isotope accumulation is due to
infection or new bone formation (Fig.14.7).A further
great advantage of CT is its ability to accurately guide
bone biopsies (Fig. 14.8a).
14.2.5
Ultrasound
Ultrasound is a quick, simple and inexpensive tech-
nique. Because it does not involve ionizing radiation
and does not require sedation, it is becoming more
popular in the imaging of children. The intro duc-
tion of high-resolution prob es in the 5- to 12-MHz
range has rendered ultrasound a useful tool for
the evaluation of joint and superficial peri-articular
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

Fig. 14.7. a Child with known myositis ossificans and mito-

chondria! disease. A bone scintigram showed increased uptake
in the soft tissues. All appearances were initially thought to be
due to myositis ossificans. b CT scan identified an intramus-
a cular abscess separate from the ossification site

Fig.14.8. a CT-guided bone biopsy for an infective focus at the

lower end of left femur. b Ultrasound-guided synovia! biopsy.
Needle aspiration was negative for free fluid. Ultrasound-guided
synovia! biopsy and culture were positive for tuberculosis b
256
Musculoskeletal infections may show ultrasound
signs at as early as 1 or 2 days. Signs of knee infec-
tion include soft tissue abnormalities, cortical irregu-
larities and subperiosteal fluid collection. In addition,
ultrasound can reveal joint effusions and peri-artic-
ular abscesses or bursitis (FESSELL et al. 2000). Early
diagnosis of septic arthritis requires analysis of joint
fluid (BUREAU et al. 1999). In this context, ultrasound-
guided fluid aspiration is a quick, simple and accurate
procedure. Moreover, when there is not much fluid,
synovial biopsy can be performed under ultrasound
guidance (BUREAU et al. 1999) (Fig.14.8b). The lack of
widespread use of ultrasound for diagnosis of infec-
tions is due to lack of interest on the part of radiol-
ogists rather than limitations of the technique. The
field of view with high-resolution ultrasound is limit-
ed but suffident for most superfidal infections. Ultra-
sound has limitations in deep-seated lesions, the vio-
lated knee, purely intra-osseous processes and detec-
tion of occult or multiple sites. Anormal ultrasound
scan does not exclude the presence of infection.
14.3
Common Knee Infections
14.3.1
Osteomyelitis
The term "osteomyelitis" is used to describe infec-
tions involving bone and bone marrow. Normal bone
is usually resistant to infection unless the organism
is highly virulent, there is local tissue violation or the
general condition of the patient is defident. Because
of the lack of a gold standard test, the diagnosis of
osteomyelitis is established if two of the following
criteria are present: (1) presence of classic signs and
symptoms; (2) typical radiographie or imaging find-
ings; (3) pus on aspiration; (4) positive blood culture,
aspiration biopsy culture or histological biopsy (PEL-
TOLA and VAHVANEN 1984). We have previously sug-
gested a simplified approach to the radiological diag-
nosis of osteomyelitis (SAMMAK et al. 1999).
Osteomyelitis may be haematogenous or due to
local sepsis. Several types were described by Truetta
according to variation in the metaphyseal blood supply
in different age groups (TRUETTA 1957). The behav-
iour ofbacteria in the microcirculation ofbone is well
described (HOBO 1921). The metaphyseal capillaries
and sinusoids have poor phagocytic activity. In the
infantile form of osteomyelitis, where the metaphyseal
vessels penetrate the growth plate to supply the epiph-
M. E. Abd EI Bagi et al.
yses, infection can rapidly spread from metaphyse al
marrow with destruction of the growth plate and like-
lihood of septic arthritis. Joint effusion is present in
70% of cases and focal growth retardation may occur.
Infantile disease is areal challenge as it lacks systemic
and local clinical signs (DAVID et al. 1987). Between
the first year and puberty, in the juvenile form, vascu-
lar penetration of the growth plate no longer exists
and the infection localizes in the metaphyseal sinu-
soids. Raised pressure forces the exudate to track lat-
erally through the channels of Volkman and Havers,
spreading to perforate the cortex and elevate the thick
but loosely attached periosteum (Fig. 14.9), forming
the characteristic subperiosteal collection. Spread to
the joint space is unlikely in the juvenile form, except
in joints where the metaphyses are essentially intra-
articular, e.g. hips, elbow and shoulder. After growth
plate closure in adults and absorption of the epiphy-
seal cartilage, infection can spread from metaphyseal
marrow up to the subarticular region. Bone 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-
dal osteomyelitis due to an infected wound. Stage 3
is characterized by full thickness cortical sequestra-
tion. Finally, stage 4 entails through and through bone
involvement which may require resection with stabili-
Fig.14.9. Post -contrast axial fat -suppressed Tl-weighted image
of the lower femur of a child, showing characteristic periosteal
elevation and subperiosteal abscess formation (arrowheads)
with intense enhancement of the surrounding soft tissues and
intramedullary canal
Infection
zation (MADER et al. 1999). According to the modified
Roberts' criteria, osteomyelitis is classified according
to the exact anatomieal site (ROBERTS et al. 1982).
Type 1 refers to involvement of the central metaphy-
seal region, type 2 to eccentric metaphyseal cortieal
erosions, type 3 to eccentric diaphyseal erosion, type
4 to diffuse periostitis without erosions and type 5
to primary epiphyseal plate lesions. We consider the
lesion to be of type 6 when it involves both epiphyses
and metaphyses (Fig.14.1O).
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)
Acute osteomyelitis is commoner in children
(Fig. 14.5). Osteomyelitis tends to be subacute to
chronie in adults, in whom it is usually non-haematog-
enous and caused by trauma or local sepsis. Hae-
matogenous spread entails one organism whereas
local spread of infection in violated tissue attracts mul-
tiple organisms. Extraspinal haematogenous spread
in adults is rare except in the immunocompromised.
Bone infection may be pyogenie or non-pyogenic. We
have previously summarized the commonly isolated
bacteria as weH as the pathogens causing rare bone
infections (ABD EL BAGI et al. 1999). The commonest
isolated organism in adults is Staphylococcus aureus
(50%-75%). In non-tropieal countries, rare bone infec-
tions can be encountered in patients with any of
the three host mechanism disturbances (DIRSCHL
and KINDERS 1993). Patients with altered neutrophil
defence, such as occurs in patients with haematologi-
257
cal malignancies who are on cytotoxic treatment, have
been proven to have gram-negative bacilli,Aspergillus
and Nocardia. Those with defects in humoral immu-
nity, as in myeloma and leukaemia, tend to have encap-
sulated bacteria and Strept. pneumoniae. Patients with
ceH-mediated immunodeficiency like AIDS tend to
acquire intraceHular pathogens like mycobacteria and
Nocardia. Drug abusers have a combined deficiency of
neutrophil action and ceH-mediated immunity, and are
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.14.1I. Axial CT scan ofthe right calf showing dense"dead"
sequestra (small arrowheads) and an involucrum envelope
(long arrows) with gaps, "cloacae" (short arrows)
 258 M. E. Abd EI Bagi et al.

ingfeature in chronic osteomyelitis (Fig.14.12).A very 14.3.1.3

rare complication of osteomyelitis is malignant trans-
formation, which has been reported in no more than
0.2%-1.7% of cases (GuPTA and PREZIO 1988).
14.3.1.1
Chronic Sclerosing Osteomyelitis of Garre
Chronie sclerosing osteomyelitis of Gam~ has been
known since 1893, before X-rays were discovered.
There is diffuse sclerosis with intense periosteal prolif-
eration and cortical thiekening with no sequestration
or suppuration. This disease is particularly common
in teenagers, which may lead to confusion with scle-
rotie benign or malignant bone tumours. It affects the
diametaphyses and extends vertically,leading to fusi-
form architectural remodelling that causes encroach-
ment on the medullary canal (DAVID et al. 1987). The
shafts of the tibia and femur are the commonest sites.
14.3.1.2
Chronic Recurrent Multifocal Osteomyelitis
Chronie recurrent multifocal osteomyelitis is arecent-
ly described entity that is common in children
between 5 and 10 years and may involve the clavicle,
tibia and femur. The disease is thought to be auto-
immune. Usually there is failure to isolate pathogens
from affected areas. The disease is characterized by
multiple sites of osseous involvement, periodie exac-
erbations and remissions and failure to isolate the
pathogen from affected areas.
Brodie's Abscess
Bone abscesses that occur during the subacute or
chronic stage of haematogenous osteomyelitis are
known as Brodie's abscesses (BOUTIN et al. 1998). Bro-
die's abscess was first described in 1832. The lesion is
typieally cystic with a well-defined (geographie) scle-
rotic inner margin measuring 1-4 cm. It is commonly
eccentrie. It can be better appreciated on MR imag-
ing, particularly after intravenous gadolinium contrast
enhancement (Fig.14.13). The tibia and femur are the
commonest sites. Brodie's abscess is commonest in
young males.
14.3.2
Septic Arthritis
Septie arthritis is a medieal emergency. The knee is the
commonest large joint affected (PEREZ 1999). Infec-
tious arthritis is usually mono-articular (MADER et al.
1999). Septic arthritis can be caused by any bacterium.
The synovium has no limiting basement plate, allow-
ing easy entrance of organisms. Infective organisms
stimulate release of proteolytic enzymes, whieh cause
synovitis and cartilage damage. Staph. aureus andNeis-
seria gonorrhoeae are highly selective for the synovi-
um, as is Brucella. Predisposing factors for septie
arthritis include aging, arthridites, steroid treatment,
diabetes, intravenous drug abuse, HIV infections, cir-
rhosis, malignancy and cytotoxics. Gonococcal arthri-

Fig. 14.12. a Uncontrolled diabetic patient who presented with pathologi-

cal fracture of the upper tibia. There is focal bone erosion in the medial
margin of the medial femoral condyle (shart arrow) and periostitis (lang
arraw) due to osteomyelitis. b CT scan of the same patient demonstrated a
a
complicating intramuscular abscess
Infection 259

Fig.14.13. T2-weighted coronal image showing a geographically

well-demarcated intramedullary Brodie's abscess of the upper
third of the tibia associated with a typical elongate tunnelling Fig. 14.14. Sagittal STIR MR image sequence. Severe septic
and subcutaneous abscess. There is noticeable periosteal eleva- arthritis showing: (1) large joint effusion, (2) subchondral
tion and characteristic subperiosteal abscess (arrowheads) hyperintense erosions of the femur and tibia, (3) Baker's cyst
bursitis, posteriorly; (4) surrounding infective cellulitis with
lymphangitis, anteriorly
tis used to be common in North American females, but
its incidence has decreased markedly. Non-gonococcal
pyoarthrosis is commonly due to Pseudomonas organ-
isms. Earliest X-ray signs are peri-articular soft tissue
swelling due to effusion. Peri-articular osteoporosis,
joint space narrowing and suochondral erosions
appear later. MR imaging is useful in depicting the
extent and severity of septic arthritis (Fig. 14.14).
Ankylosis may be the end result. MRSA is common
in drug abusers. Diagnostic arthrocentesis should be
used to test for leucocyte count, culture, crystals, high
lactic acid and low or absent glucose.

14.3.3
Pyogenic Myositis

Pyogenie myositis is commonest in drug abusers,

and is usually due to Staph. aureus. Viral myositis
is common. Ultrasound is a useful tool to detect
oedema or abscess formation (Fig. 14.15), but MRI
is more accurate for subtle or diffuse inflammation
(Fig.14.16).Muscle can be involved byextension from
nearby osteomyelitis. Muscle infection may compli-
cate trauma or pathological fractures (Fig. 14.12).

14.3.4
Septic Bursitis
Bursae are synovial cysts to facilitate gliding oflocomo-
tor system components. Sepsis is responsible for 30%
Fig. 14.15. a High-resolution ultrasound showing extensive
oedema and effusion in the medial belly of the gastrocne-
mius muscle due to infective myositis. b Normal appearance
of the arrangement of the right gastrocnemius muscle fibres
on ultrasound
260 M. E. Abd EI Bagi et al.

Fig. 14.16. Coronal T2-weighted MR image showing diffuse
oedema of the right ealf muscles due to myositis
Fig. 14.17. A ehild with painful knee swelling. Sagittal T2-
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

ofbursitis. It is common in popliteal bursae, which may

or may not communicate with joints (Fig. 14.17). Dif-
ferentiation of septic from aseptic bursitis is not always
possible. Post-contrast MR imaging can detect infective
bursitis (Fig.14.18).

14.3.5
Infective Cellulitis

Skin, soft tissue and muscular infections are very

common and can progress to produce serious or
life-threatening local or systemic infections (BROOK
1999). They often occur in sites that have been com-
promised or injured by foreign bodies, trauma, isch-
aemia, malignancy or surgery. Diabetes is an impor-
tant predisposing cause. They may induce or result Fig. 14.18. Post-eontrast fat-suppressed axial Tl-weighted
sequenee showing intense enhaneement of the walls of the
from osteomyelitis or septic arthritis (Fig. 14.14).
fluid-filled popliteal bursa and the surrounding soft tissues
Staph. aureus, group A streptococci and the local (Zarge arrowheads). There is also an enhancing foeus of osteo-
flora can be the causative agents for diffuse inflam- myelitis in the medial femoral eondyle (small arrowheads)
mation involving the dermis and subcutaneous tis-
sues. Such infection can lead to osteomyelitis in adja-
cent bones. Clinical presentation varies from erysip-
elas to the more serious necrotizing fasciitis and scintigraphy, cellulitis will cause diffusely increased
gas gang rene (BOUTIN et al. 1998). There is local uptake in the perfusion and blood pool phases with-
erythema and lymphangitis (Fig. 14.14). Soft tissue out a corresponding increase in the delayed static
infections are usually apparent clinically. Imaging is phase. MR imaging will show soft tissue oedema, par-
needed to exclude abscess formation and involve- ticularly in the STIR sequences, without periosteal or
ment of the underlying bone. On three-phase bone bone lesions.
 Infection 261

14.4
Rare Bone Infections

Rare bone infections are situations where either a

small percentage of the infective organism has bone
manifestations or where the infective organism is
itself of rare occurrence (ABD EL BAGI et al. 1999).

14.4.1
Tuberculosis

It is now believed that the knee is the commonest

extraspinal site for osteoarticular tuberculosis
(HUGOSSON et al. 1996). Skeletal involvement may
precede the constitutional symptoms by a long time.
In one of our patients there was aperiod of 6 months
between the occurrence of significant knee signs and
symptoms and the development of miliary tuberculo-
sis (Fig. 14.19). Concomitant osteoarticular and pul-
monary tuberculosis occurs in 12%-50% of patients.
Knee infection with acid-fast bacilli causes a destruc-
tive caseating necrosis of bone and soft tissues, lead-
ing to cold abscess formation and sinuses. Trauma has
been associated with bone tuberculosis in 30%-50%
of cases. Radiological signs include soft tissue swell-
ing, periostitis, erosions, mild reactive sclerosis, small
sequestra and cystic lesions (Fig.14.20). The Phemis-
ter triad involves (a) peri-articular osteoporosis, (b)
peripheral osseous lesions and (c) gradual joint space
narrowing (PHEMISTER and HATCHER 1933).
Tuberculosis may mimic other infective, traumatic,
neoplastic, collagen or degenerative disease (RASOOL
et al. 1994). The patient in Fig. 4.19a and b was diag-
Fig. 14.20. Frontal radiograph of the knees showing signs
of tuberculosis in the left knee: (1) extensive peri-articular
osteoporosis; (2) numerous articular erosions; (3) subarticular
cystic lesions; (4) joint space narrowing and irregularity; (5)
shortening of the left leg. (Courtesy of Dr. Munir Madkour)
nosed as having osteochondritis dissecans on serial
plain films and MR imaging. The patient in Figs. 14.4
and 14.21 presented with knee swelling and was con-
sidered to have chronic osteomyelitis with abscess for-
mation on the basis of plain film findings. MR imag-
ing showed a homogeneously enhancing soft tissue
lesion extending subcutaneously without any case-
ation or pus collection (Fig. 14.4). This raised the
possibility of a tumour. Operative biopsy and frozen
section were negative for lymphoma, tumour and
tuberculosis. The patient had para-aortic lymphade-
nopathy and thickened small bowelloops (Fig.14.21).
CT-guided biopsy of para-aortic lymph nodes was
strongly suggestive of a tuberculous granuloma. Inoc-
ulation culture proved to be positive for 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
Fig. 14.21. CT scan of the abdomen of the patient shown in Fig.
14.4. Matted para-aortic adenopathy. CT-guided biopsy was
suggestive of tuberculosis. This was confirmed by culture
Tuberculous infection of joints may heal by anky-
losis. Limb shortening may result (Fig. 14.20). Tuber-
culosis usually affects the metaphyses, but epiphyseal
lesions can occur (Fig. 14.22). It tends to form small
sequestra, unlike pyogenic infections (Fig. 14.22).
Tuberculosis is slow to invade joints owing to the lack
of proteolytic enzymes (DAVIDSON and HOROWITX
1970). There has been a resurgence of tuberculosis
due to the HIV epidemie. It is important to note that
tuberculosis patients could have a negative Mantoux
test in 20% of cases (ROONEY et al. 1976).An enzyme-
linked immunosorbent assay (ELISA) is reported to
be highly sensitive in early but not in advanced cases
(PANDEY and TAUB 1993). Generally, tuberculosis
tends to affect the extremes of age. It is more prevalent
in the young where the disease is common and more
prevalent in the eldedy where the disease is rare.
14.4.2
Brucellosis
Peripheral bone infection is rare in brucellosis but
the knee is one of the common extraspinal sites (AL
SHAHED et al. 1994). Synovial involvement is reported
in 81 % of all peripherallesions. This is an important
differentiating point from tuberculosis, whieh tends
to spare the joint untillate. Both diseases are prev-
alent in endemie areas and may co-exist. ELISA is
more sensitive than other serologieal tests. Advances
in molecular techniques, such as polymerase chain
re action (PCR), will soon become the method of choiee
for diagnosis of brucellosis and other slow-growing
M. E. Abd EI Bagi et al.
Fig. 14.22. Sagittal T2-weighted scan showing an epiphyseal
Iesion with a small "button" sequestrum (arrowhead). Note
metaphyseal focus of the infection (arrow)
microorganisms (YOUNG 2001). Unsafe animal hus-
bandry, bad food hygiene and consumption of raw
milk are predisposing factors for the disease.
14.4.3
Salmonellosis
Salmonellosis is caused by gram-negative aerobic
bacilli. They are usually associated with siekle cell
disease (SCD), haemoglobinopathies, poor personal
hygiene, trauma, connective tissue disorders, lym-
phoma and immunosuppressive states (TU! and
CHINK 1997). Infection with Salmonella is common-
est around the knee.
14.4.4
Mycetoma
Mycetoma commonly affects the foot in sub-Saharan
Africa, India and parts of the Arabian peninsula. It
has distinct radiographie and MR findings. Radio-
logieal signs were classified by DAVIES (1958) and
MR imaging features were described by SHARIF et
al. (1991). Infection proliferates beneath the skin to
form deep abscesses with sinus tracks and tumefac-
tion leading to bizarre lesions (Fig. 14.23). The pres-
ence of coloured grains helps in the diagnosis and
classification of the disease; these grains appear as
signal-void spots on MR imaging.
Infection
Fig.14.23. Mycetoma of the knee with extensive bone destruc-
tion involving the whole of the lower third of the femur, the
upper end of the tibia and the patella
14.4.5
Hydatid Disease
Man is the secondary host of this small dog tape-
worm, infection occurring via consumption oflarvae
in sheep meat or contaminated vegetables. Liver is
infected in 75% of patients. Bone is involved in
1%-4% (MERKLE et al. 1997). Bone lesions include
the classic "soap bubble" appearance, sclerosis, peri-
ostitis and pathological fractures. Intradermal test
and immunoelectrophoresis are used to establish the
diagnosis. However, both can be false positive or false
negative in 15%-20% of cases.
14.5
Selected Topics
14.5.1
Knee Infections in Children
Perlman stated that the incidence of nearby joint
involvement in children with osteomyelitis is much
higher than is suggested in the literature (PERLMAN
et al. 2000). Knee infections are commonest in boys,
possibly the effect of repeat trauma. Osteomyelitis in
children differs from the adult pattern. It is usually
acute and haematogenous in children while it tends
to be subacute or chronic in adults. The femur and
tibia are the commonest sites. Infection is usually
263
at the metaphyses, where there is slower flow in
the venous sinusoidal plexuses, which have reduced
phagocytic cell activity. Trauma and local thrombo-
sis are predisposing factors. Macronecrosis leads to
sequestra of large devascularized fragments. Depo-
sition of periosteal and endosteal new bone forms
an involucrum envelope. Staphylococcus aureus is the
commonest pathogen. Group B streptococci are more
prevalent in neonates. Haemophilus inJluenzae is
decreasing in incidence owing to vaccination. Treat-
ment of knee infections should be star ted within 72
h if permanent sequelae are to be avoided. Because
pus is chondrolytic, septic arthritis in children is an
acute surgical emergency. Radiographs may show a
wide joint space and even dislocation can occur.
The lack of clinical manifestations in neonates
makes diagnosis difficult or delayed. Pain on move-
ment is a warning sign. Because of the thin cortex
and poor adherence of the periosteum, soft tissues are
soon invaded. The combination of plain radiographs
and high-resolution ultrasound is ideal to investigate
children suspected of osteomyelitis. MR imaging for
children causes problems of sedation availability, cost
and inability to detect multiple sites. Contrast enhance-
ment could be exaggerated by higher blood flow due
to rapid bone growth (JARAMILLO et al. 1995).
14.5.2
Knee Infection in Sickle Cell Disease
14.5.2.1
Osteomyelitis
Sickle cell disease is the most common haemoglobin
variant in the world. Compared with control subjects,
SCD patients are over a hundred times more sus-
ceptible to osteomyelitis (DAVID et al. 1987) and are
25 times more likely to acquire salmonella infection
(WANE 1997). There are three reasons for this.
First, infarcted bone in SCD is more susceptible to
infection. Secondly, infarcted bowel causes increased
mucosal penetration, leading to bacteraemia. Thirdly,
functional asplenia reduces opsonic antibody activ-
ity against bacteria.
14.5.2.2
Infection Versus Infarcts
Salmonella osteomyelitis and bone infarcts are both
common in sicklers: they may even co-exist at one or
multiple sites and there may have been previous epi-
sodes of bone infarcts. Differentiation poses a diag-
 264
nostie dilemma. Initial clinical symptoms are simi-
lar. Periosteal elevation and subperiosteal pustular
pockets are signs of infection (Fig. 14.9). The pres-
ence of sizable soft tissue swelling and intense con-
trast enhancement and breakthrough of the cortex
are features of osteomyelitis not seen in infarcts
(Fig. 14.24). On plain films, infarcts are usually scle-
rotic whereas infections are usually lytic. Infection
pro duces a thicker periosteal re action and cortical
abnormalities. Both infections and infarcts affect the
medullary space of long bones and the short tubular
bones of the hands and feet. A pseudobone within
bone appearance is commoner in infection than in
infarcts. No radiographie feature can give absolute
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
the scan is performed a week or more after the symp-
toms. Addition of 67Ga can enhance the specificity:
patients with infarct will have decreased or normal
67Ga uptake. Furthermore,addition of a 99ffiTc-sulphur
colloid scan will show photon deficiency in infarcts.
Radiocolloid photon deficiency is not a sign of osteo-
myelitis.
14.5.3
The Immunocompromised Patient
Musculoskeletal infections associated with immuno-
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
M. E. Abd EI Bagi et al.
(LEE and SARTORIS 1994). This can take the form
of cellulitis, subcutaneous abscess, pyomyositis or
osteomyelitis. Infection can be haematogenous even
in adults. Local infections are commoner in drug
abusers. HIV is blamed for the resurgence of tuber-
culosis, and the incidence of pyomyositis is rising
in these patients. The commonest pathogen is Staph.
aureus, followed by Strep. pneumoniae. Because
unusual organisms can be encountered, aspirations
and biopsies are essential for accurate diagnosis.
Examples are Histoplasma capsulatum and bacillary
angiomatosis, whieh is aggressively osteolytie. Drug
abusers ean aequire clostridial and non-clostridial
myoneerosis or gas gangrene.
14.6
Summary
The knee is a very common site for most bone
and joint infeetions. Rapid and aeeurate diagnosis is
important to avoid eomplieations in this weight-bear-
ing joint. The knee is also a eommon site for most
tumours, peripheral degenerative disease, trauma
and infarets. Differentiation of these poses a diag-
nostie ehallenge. There is no gold standard for the
diagnosis of knee infeetion as no single modality is
ideal all of the time. We reeommend an algorithm
for investigations of patients suspeeted to have knee
infeetion (Fig. 14.26).
Fig. 14.24. a Coronal Tl-weighted image of a patient
who had sickle ceH disease and presented with knee
pain, showing a hypointense epiphyseallesion (arrow)
of either an infarct or infection. b On post -contrast fat-
suppressed Tl-weighted imaging there was intense focal
enhancement as weH as synovial enhancement. This is
suggestive of infection, most Iikely salmonellosis
 Infection 265

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

H,story. Lllnleal Ind Laboralory Test•

.J.
Plain Radiographs
.J.

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

Fig. 14.26. Algorithm to be followed in cases of suspected osteomyelitis

266
References
Abd EI Bagi ME, Sammak BM, Al Shahed MS et al (1999)
Rare bone infections "excluding the spine". Eur Radiol
9:1078-1087
Al Shahed MS, Sharif HS, Haddad MC et al (1994) Imaging
features of musculoskeletal brucellosis. Radiographics
14:333-348
Boutin RD, Brossmann J, Sartoris DJ et al (1998) Update on
imaging of orthopaedic infections. Orthop Clin North Am
29(1):41-65
Bouza E,Munoz P (1999) Microorganisms responsible for osteo-
articular infections. Baillieres Clin Rheumatol13(1):21-35
Brook I (1999) Soft tissue and muscular infections (including
group A streptococci). Baillieres Clin Rheumatol
13(1 ):95-114
Bureau NJ, Chhem RK, Cardinal E (1999) Musculoskeletal infec-
tions: US manifestations. Radiographics 19:1585-1592
Datz FL, Thorne DA (1987) Cause and significance of cold
bone defects on In III labeled leucocyte imaging. J Nucl Med
28:820-823
David R, Barron BJ, Madewell JE (1987) Osteomyelitis acute
and chronic. Radiol Clin North Am 25(6):1171-1201
Davidson PT, Horowitx J (1970) Skeletal tuberculosis: a review
with patients: presentation and discussion. Am J Med
48:77-84
Davies AGM (1958) The bone changes of Madura foot. Obser-
vations in Uganda Africans. Radiology 70:841-847
Dirschi DR, Kinders A (1993) Osteomyelitis: common causes
and treatment recommendations. Drugs 45(1):29-43
Donohoe KJ (1998) Selected topics in orthopaedic nuclear
medicine. Orthop Clin North Am 29(1):85-99
Eigazzar AH, Abd EI Dayem HM, Clark JD et al (1995)
Multimodality imaging of osteomyelitis. Eur J Nucl Med
22(9):1033-1063
Espinoza LR, Berman A (1999) Soft tissue and osteoarticular
infections in HIV infected patients and other immunodefi-
cient states. Baillieres Clin Rhematol13(1):115-128
Fessell DP, Jacobson JA, Craig J et al (2000) Using sonography
to reveal and aspirate joint effusions. AJR 174:1353-1362
Frink SJ, SnearlyW, Parsons TW III (1998) Malignant tumours
about the knee.Am J Knee Surg 11(4):257-266
Grey AC, Davies AM, Mangham DC et al (1998) The 'penum-
bra sign' on Tl' weighted MR Imaging in subacute osteo-
myelitis, frequency cause and significance. Clin Rad 53(8)
587-592
Gupta NC, Prezio JA (1988) Radionuclide imaging in osteomy-
elitis. Semin Nucl Med 19(4):287-299
Hobo TZ (1921) Zur der Vitolfarbungslehere Pathogenese de
Akuten Haematogenen Osteomyelitis. Mit Beruckishtigun
Univ Imp Kioto 4:1-29
Hopkins KL, Li KC, Bergman G (1995) Gadolinium DTPA
enhanced MRI of musculoskeletal infectious processes.
Skeletal RadioI24(5):325-330
Hugosson L, Nyman R, Ny Brismar R (1996) Peripheralosteoar-
ticular and soft tissue tuberculosis. Acta RadioI37:512-516
Jaramillo D, Treves St, Kaser JR et al (1995) Osteomyelitis and
septic arthritis in children; appropriate use of imaging to
guide treatment. AJR 165(2):399-403
Kolindou A, Li Y, Ozker et al (1996) In lll WBC imaging of
osteomyelitis in patients with underlying bone scan abnor-
malities. Clin Nucl Med 21(3):183-191
Ledermann HP, Kaim A, Bongartz G et al (2000) Pitfalls and
M. E. Abd EI Bagi et al.
limitations of magnetic resonance imaging in chronic post
traumatic osteomyelitis. Eur Radioll0:1815-1823
Lee DJ,Sartoris DJ (1994) Musculoskeletal manifestations of the
human immunodeficiency virus infection: review of imag-
ing characteristics. Radiol Clin North Am 32(2):399-411
Mader JT, Shirtliff M, Calhoun JH (1999) The host and the
skeletal infection: classification and pathogenesis of acute
bacterial bone and joint sepsis. Baillieres Clin Rheumatol
13(1):1-20
Mazur JM, Ross G, Cummings RJ et al (1995) Usefulness of
magnetic resonance imaging for the diagnosis of acute
musculoskeletal infections in children. J Paediatr Orthop
15:144-147
McAfee JG (1990) What is the best method for imaging focal
bone infections? J Nucl Med 31(4):413-416
Merkel KD, Brown ML, Dewanjee MK et al (1985) Compari-
son of indium labeled leucocyte imaging with sequential
technetium gallium scanning in the diagnosis of low grade
musculoskeletal sepsis. J Bone Joint Am Surg 67:465
Merkle E, Schulte M, Vogel J (1997) Musculoskeletal involve-
ment in cystic echinococcosis - areport of eight cases and
review ofthe literature.AJR 168:1531-1534
Morrison WB, Schweitzer ME, Granville BW et al (1998) Osteo-
myelitis of the foot: relative importance of primary and
secondary MR signs. Radiology 207:625-632
Nijhof MW, Oyen WSG, Van Kemper A et al (1997) Evaluation of
infections of the locomotor system with indium lll labeled
human IgG scintigraphy. J Nucl Med 38: 1300-1305
Palestro q, Torres MA (1997) Radionuclide imaging in ortho-
paedic infection. Semin Nucl Med 27(4):334-345
Pandey J, Talib VH (1993) Laboratory diagnosis of tubercu-
losis, use of ELISA and PCR. Indian J Pathol Microbiol
35:512-518
Peltola H, Vahvanen J (1984) A comparative study of osteomy-
elitis and purulent arthritis with special reference to aetiol-
ogy and recovery. Infection 12:75
Perez LC (1999) Septic arthritis. Baillieres Clin Rheumatol
13(1):37-58
Perlman MH, Patzakis MJ, Kumar PJ et al (2000) The incidence
of joint involvement with adjacent osteomyelitis in paedi-
atric patients. J Paediatr Orthop 20(1 ):40-43
Phemister DB, Hatcher CM (1933) Correlation of pathological
and roentgenological findings in the diagnosis of tubercu-
losis. Arthritis Am J RoentgenoI29:736-7 52
Ram P, Martinez S, Korobkin M et al (1981) CT detection
of intraosseous gas. A new sign of osteomyelitis. AJR
137:721-723
Rasool MN, Govender S, Naidoo KS (1994) Cystic tuberculosis
ofbone in children. J Bone Joint Surg [Br] 76B:I13-117
Roberts JM, Drummond DS, Bread AL et al (1982) Subacute
hematogenous osteomyelitis in children, a retrospective
study. J Pediatr Orthop 2:249-254
Roddie ME,Peter AM,Osman S et al (1988) Osteomyelitis.Nucl
Med Commun 9:713-717
Roine I, Faingezicht I, Arguedas A (1995) Serial serum C-reac-
tive pro tein to monitor recovery from acute osteomyelitis
in children. Paediatr Infect Dis J 14(1):40-44
Rooney J Jr, Crocco JA, Kramer S et al (1976) Further dissec-
tions in tuberculin reaction in active tuberculosis. Am J
Med 60:517-522
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.

A.M. Davies et al. (eds.), Imaging of the Knee

© Springer-Verlag Berlin Heidelberg 2003
270 C. Wakeley and I. Watt

3. What are the probable diagnoses? In the case of the 15.2.2

knee, major causes include OA, various inflamma- Wh at 00 the Various Imaging Techniques
tory arthropathies, especially rheumatoid arthri- Teach Us?
tis, episodes of crystal shedding and infection.
15.2.2.1
Plain Radiographs

15.2 Since there is a basic need to see soft tissues as weIl

How Does One Investigate the Knee as bony structures, the principle of two views at right
Radiologically? angles still applies. The lateral radiograph is vital to
establish the presence of a joint effusion as weIl as to
15.2.1 document patellar position. An ill-defined outline of
Basic Principles the suprapatellar pouch suggests an active synovitis
(Fig.15.1), even infection.A well-defined outline sug-
Plain film radiology shows established anatomical or gests a simple joint effusion or some form of low-
morphological images at any particular moment in grade synovial mass lesion. Additional axial ("sky-
time. The bony manifestations will have evolved some line") views of the PFJ are useful when cartilage
time before the radiograph (except for acute injury). width or evidence of maltracking is sought. PA views
On the other hand, soft tissue changes are more of the knee should be carried out weight bearing, as
recent. A sense of disease evolution can be gleaned by joint space narrowing and angulation are underes-
studying both previous and follow-up images. Howev- timated in supine films (Fig. 15.2). The tibial pla-
er, functional information regarding disease processes teaux are angled, and in order to visualise the TFJ to
may be gained from perfusion based-studies includ- maximally profile the articular surfaces, the PA view
ing scintigraphy, power Doppler ultrasound and post- should also be acquired in a semiflexed position, util-
intravenous contrast medium magnetic resonance ising caudal angulation of the X-ray beam (Fig.15.3).
imaging (MRI). More rarely, computed tomography The most recently recommended view is that rem-
(eT) may be helpful. iniscent of the schuss position in skiing. However,
When considering any joint, one should remember even here accurate assessment of joint space width
to regard it as a whole organ. Thus, changes in the can only be obtained in the medial compartment.
main tissue types are inter-related, although manifesta- When overalllimb alignment needs to be assessed,
tions may predominate in one. In the case of a syno- full-Iength views of the entire lower limb can be
vial joint, three tissue types co-exist: synovium, carti- obtained using three hinged films in a cassette with
lage and bone. However, in order to make a diagnosis of
any particular disease, the localisation and distribution
of radiological signs and characteristic features within
each tissue are helpful. Nonetheless, synovium, carti-
lage and bone can only react to the various aetiological
triggers in a limited number of ways, such that changes
may be non-specific. For example, synovitis is synovitis
and can be one of dozens of causes, unless or until some
more specific sign emerges such as bony erosion.
As indicated, it is helpful in the knee to distinguish
two major compartments, the patellofemoral (PF])
and tibiofemoral (TF]) joints. OA may involve either,
perhaps with different disease subsets, although most
commonly nowadays the TFJ is involved. Other dis-
ease processes such as chondromalacia and maltrack-
ing typically involve the PFJ.
In the developing skeleton different disease pro-
ces ses may result in similar end-effects; for example,
Fig. 15.1. Septic arthritis. Lateral conventional radiographs
haemophilia, infection and juvenile arthritis result in demonstrating joint space narrowing and bone erosion (espe-
epiphyseal hypertrophy and premature growth plate cially affecting the PFJ). Note that the diffuse soft tissue swell-
fusion from the associated hyperaemia. ing from the associated synovitis is ill-defined
 Arthritis
graded screens. The length of individual bones, limb
shortening, joint angulation and the "clinical load
line" can be measured. These are of particular impor-
tance in pre-operative surgical planning.
15.2.2.2
Computed Tomography
In general, MRI has become the cross-sectional imag-
ing modality of choice owing to the ability to demon-
strate all the appropriate tissue planes. However, scout
films (scanograms) can replace full-Iength conven-
_ _ .- .- .-
._ .- . - .
_.-.- ·-0
\ '0 '
i
Fig. 15.3. Teehnique for making the 45 0 postero-anterior flex-
ion weight-bearing radiograph. [From Rosenberg TO, Paulos
LE, Parker RO, Coward OB, Seott SM (1988) The forty-five-
degree posteroanterior flexion weight-bearing radiograph of
the knee. J Bone Joint Surg Am 70:1479-1483]
271
Fig.15.2a, b. The effeet of"weight bearing". a
Supine radiographs ofboth knees. b Note the
inereased varus angulation of the left knee in
the ereet weight-bearing radiograph
tional radiographs, and also axial eT images acquired
through the hips, knees and ankles facilitate calcula-
tion of the torsional profile of individual bones, again
for pre-operative surgical planning.
15.2.2.3
Radionuclide Radiology
Skeletal scintigraphy using technetium-99m labelled
diphosphonate bone scanning agents is useful for
identifying active sites of disease, and also their dis-
tribution. Various subsets of scintigraphic patterns
have been associated with the evolution and progno-
sis of OA (see below).
15.2.2.4
Ultrasound
Ultrasound is particularly useful for distinguishing
solid from cystic masses around the knee [e.g. dis-
tended semimembranosus bursa - the so-called Bak-
er's cyst (Fig. 15.4)]. An emerging role for power Dop-
pler is evolving in the study of synovial inflamma-
tion, particularly with respect to changes related to
disease-modifying drugs.
15.2.2.5
Magnetic Resonance Imaging
The superb soft tissue contrast, multiplanar facility
and non-ionising nature of MRI make it the further
imaging modality of choice in many cases. The spa-
tial resolution and compartmental anatomical defi-
272
Fig. 15.4. Transverse ultrasound scan of the popliteal fossa
showing a large hypoechoic semimembranosus bursa. Note its
hernial neck tracking between the eccentric tendon of medial
gastrocnemius and the semimembranosus tendon (5MB)
nition demonstrate sites of abnormality. Associated
meniscal, ligamentous, bone marrow and soft tissue
abnormalities are also shown. Enhancement profiles
of synovium following intravenous contrast agents
are useful in studying response to disease-modifying
drugs. Gradient echo sequences are used routinely to
quantify hyaline cartilage volume with an increasing
degree of precision.
15.3
What Are the Arthritides That Involve
the Knee7
Without doubt, the most usual arthropathy to involve
the knee is OA. However, all the major inflammatory
arthritides involve the knee, although usually the knee
is not the presenting joint in diseases such as rheu-
matoid arthritis. Reiter's syndrome and septic arthritis
maywell involve the knee first. What follows is an over-
view of the major findings in the commoner arthriti-
des. Table 15.llists the main causes ofknee arthritis.
15.3.1
Osteoarthritis
The primary pathology of OA is unknown, but the
effects result in localised hyaline cartilage damage
C. Wakeley and I. Watt
with associated subchondral bone changes that may
result in deformity and loss of function. Mainly, the
aetiology is unknown. Thus for convenience one can
consider OA as idiopathic, or secondary when some
pre-existing cause for joint derangement can be dem-
onstrated. However, one should not be in haste to
diagnose OA. The effects of ageing must be distin-
guished from OA. As our population ages so such
changes become commoner. True, OA itself is an age-
related disorder. However, simple overall thinning or
hyaline cartilage thickness, with minor rim osteo-
phytosis, reflects joint instability, and an attempt to
"tighten up" the joint. These are normal, age-related
changes. Further, it has been pointed out that joints
become increasingly congruous with age, reducing
the natural pumping mechanism of the joint surfac-
es flexing, one upon the other, to circulate the syno-
vial fluid that provides the main nutrition of chon-
drocytes.
Idiopathic or primary DA has a gradually increas-
ing prevalence, starting usually in the fifth decade. OA
presents a spectrum of disease that usually involves
one or two of the three main knee compartments
(Fig. 15.5). OA affecting the TFJ and PFJ may be dif-
ferent subsets of disease. For example, before the
modern age PFJ was the most common, now, with an
increasingly obese population, TFJ disease predomi-
nates. Both should be distinguished from the normal
ageing process of the knee joint. Characteristic radio-
graphie features of OA include:
1. Focal joint space narrowing
2. Subchondral bone changes
3. Joint effusion
4. Osteophyte formation
5. Deformity and bone attrition
6. Separate osteochondral bodies
Table 15.1. Causes of knee arthritis
Osteoarthritis 10
20 (trauma, dysplasia, inflammation)
Inflammatory Rheumatoid arthritis
Spondyloarthropathies
(psoriasis, Reiter's syndrome, ankylosing
spondylitis, juvenile arthritis
Infection
Metabolie Gout
CPPD crystals
BCP crystals
Haemochromatosis
Miscellaneous Bleeding disorders (e.g. haemophilia)
Tumours (e.g. PVNS)
Neuropathie
Foreign body (e.g. blackthorn)
Arthritis
Fig. 15.5. Osteoarthritis. AP conventional radiograph of the
knee showing OA affecting predominantly the lateral com-
partment with joint space narrowing and marginal osteophyte
formation
Joint Space Narrowing. This should be distinguished
from the diffuse changes seen in the normal ageing
process. In OA, it is usually localised, affecting pre-
dominantly one or two compartments. The joint
space narrowing is likely to be greater than two-
thirds of the estimated joint space width. Estimating
joint space width is fraught with difficulty, however.
The space between bone ends may contain hyaline
cartilage, joint fluid or meniscus. Conventional
weight-bearing films help assess this interbone dis-
tance, but are only truly reliable in the medial knee
compartment and at the PFJ. MRI will show hyaline
cartilage excellently, but cannot be justified in most
instances.
Subchondral Changes. Subchondral sclerosis is a
response in bone at the site of hyaline cartilage
damage that may represent either a healing attempt
or trabecular failure or both! Increasingly, it would
seem that subchondral trabecular organisation is
crucial to the development and containment of OA.
The processes involved are complex and only recent-
ly receiving attention. Adjacent subchondral juxta-
articular radiolucencies or "cysts" are seen, most
likely from the ingress of synovial fluid through dam-
aged hyaline cartilage. Such lesions usually imply sig-
nificant disease, and do not occur in age-related joint
instability. Such "cysts" may stabilise, but collapse can
signal another acute attack of joint pain. OA may
be perceived to run a step-like downhill course with
such episodes marking the acute crises.
273
Joint Effusion. A joint effusion is a non-specific sign
but signifies active articular pathology. Conversely, its
absence makes aggressive OA unlikely. When a joint
effusion is present, consideration should be given
as to the cause. Could another acute pathology be
present, such as an episode of crystal shedding or
superadded inflammatory arthritis? Is the outline ill-
defined on plain film (or hypervascular on Doppler)?
If so, it is likely to be inflammatory.
Osteophytes. Osteophytes are formed by enchondral
ossification of chondrophyte, formed usually at the
periphery of the joint. These may be an attempt to
maintain joint stability and are characteristic of OA.
Other sites include the intercondylar noteh, and more
interestingly, along the articular surface. The latter,
known as a "stud" osteophyte, is thought to repre-
sent enchondral ossification of reparative chondro-
cyte proliferation through damaged articular cortex.
Osteophyte may be regarded as "good" for it repre-
sents active joint repair by chondrogenesis. Hence, the
patient is forming bone, repairing joint damage, until
the next joint crisis or progressive failure ensues.
Deformity. End-stage OA may result in gross joint
deformity from a combination of joint space loss,
subchondral collapse or bone attrition, angulation
and osteophyte formation. Deformity may be diffi-
cult to assess on simple supine views. Hence, weight-
bearing and load line images are helpful in the pre-
operative assessment of the knee.
Osteochondral Bodies. Separate, though not neces-
sarily loose, intra-articular osteochondral bodies are
often seen related to OA and are not commonly seen
in other forms of arthropathy. As indicated, they
are not necessarily loose, often being embedded in
synovium. Subsets of OA are described in relation to
crystal deposition. In association with calcium pyro-
phosphate, a form of hypertrophie OA (florid osteo-
phyte formation) particularly affecting the PFJ is
described ("pyrophosphate arthropathy"). The pres-
ence of various calcium hydroxyapatite crystals, or
basic calcium phosphates, is seen in association with
a more atrophie form of OA (with a relative absence
of osteophytes: atrophie destructive OA). The exact
relationship between these crystals and the associ-
ated arthritis is still debated. It is feIt currently that
the distinction between "crystal" and "non-crystal"
OA is largely artificial and that the crystals and bone
responses are epiphenomena, not cause and effect
(see Sect. 15.4.7). Functional information from var-
ious characteristie patterns of scintigraphic uptake
274
has been described in patients with OA that not only
permits classification of various subsets of knee OA
but also relates these changes to likely disease pro-
gression and prognosis.
OA secondary to other causes ofjoint damage is the
result of a combination of secondary reparative pro-
cesses superimposed upon pre-existing joint pathol-
ogy. The initial joint insult may be due to congenital
or developmental deformity, trauma, previous infec-
tion or pre-existing inflammatory arthritis (e.g. rheu-
matoid arthritis). The compartmental distribution of
the OA will depend on the initial joint deformity or
joint insult. Clues to the secondary nature of the OA
include:
1. Moulding abnormalities. Is the joint squared off or
are the former epiphyseal areas larger than usual,
suggesting a childhood overgrowth?
2. Symmetrical involvement. Are all three compart-
ments involved uniformly? If so, is there evidence
ofbone softening as in rheumatoid disease? Anoth-
er clue is the relative absence of osteophyte in the
face of extensive hyaline cartilage loss.
Sudden worsening of OA may be associated with
either further rapid focal hyaline cartilage loss or
increased stress. The latter includes proximal tibial
and femoral condylar insufficiency fractures and
spontaneous osteonecrosis of the knee. Character-
istically subchondral insufficiency fractures involve
predominantly the medial femoral condyle in elderly
osteoporotic women or middle-aged men. The sub-
sequent subchondral collapse causes rapid advance-
ment of OA in that compartment. Insufficiency frac-
tures may result in increased eccentric loading of
one of the compartments and resultant angulation.
This feature has been previously ascribed to avas-
cular necrosis (idiopathic medial femoral condylar
necrosis, or "SONC" - sudden onset necrosis of the
condyle). Whether or not true spontaneous avascular
necrosis occurs in the knee is debatable. Both medul-
lary and cortical avascularity do occur with system-
ic disease as with steroid therapy or alcohol abuse.
Indeed, the femoral condyles and lower femoral med-
ullary canal are typical sites (see Chap. z).
15.3.2
Inflammatory Arthritides
A large number of conditions cause an acute, or
chronic, inflammatory arthritis in the knee. Many
have no specific radiological features, simply shar-
ing the common findings of ill-defined joint swelling,
C. Wakeley and I. Watt
local reduction in bone density and diffuse hyaline
cartilage thinning. In all, increased joint fluid will be
seen on ultrasound with varying degrees of increased
blood flow on Doppler. MRI is also non-specific
unless, or until, other features emerge, such as ero-
sion of bone.
15.3.2.1
Rheumatoid Arthritis
Rheumatoid arthritis (RA) is the most common
specific inflammatory arthritis to involve the knee.
Others include the various spondyloarthritides (espe-
cially ankylosing spondylitis, psoriasis and Reiter's
syndrome).
In RA the disease process is essentially an inflam-
matory synovitis with secondary bone and articular
cartilage erosion that progresses from the marginal
"bare areas" of the joint inwards. The so-called bare
areas are the intracapsular portions of the bone at the
edge of joints that are not covered by articular carti-
lage. Characteristic features of RA include:
1. Soft tissue swelling and joint effusion
2. Marginal erosions
3. Diffuse joint space narrowing
4. Subarticular geodes
5. Demineralisation of the skeleton
Unlike OA, the changes in RA tend to involve all
three compartments of the joint diffusely, with a lack of
both subchondral sclerosis and osteophyte formation.
Tissue Swelling and Joint Effusion. Not only may the
joint effusion be pronounced, but also, underlying
synovial hypertrophy and inflammation results in an
ill-defined outline on plain films. Thus, it is difficult
to distinguish between the synovial hypertrophy and
the joint effusion. Ultrasound and MRI can reliably
distinguish between these features. The latter may be
used to assess the full extent of joint involvement and
to document therapy effects. Dynamic MRI using
intravenous contrast agents (Fig. 15.6) and the cre-
ation of enhancement profiles for synovium is a
potent means of assessing the value of disease-mod-
ifying drugs.
Approximately 30% of persons have communication
between the knee joint and one of the bursa in the
popliteal fossa. Distension of such a bursa, with a
flap valve form of one-way flow, results in a popliteal
mass, the so-called Baker's cyst. Any inflammatory
arthritis or cause of chronic joint effusion will result
in distension of the cyst, often to the point of rupture.
 Arthritis 275

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.

Joint Space Narrowing. Unlike joint space narrowing

in OA, that in RA is diffuse and involves all compart-
ments. Indeed, the major impact of RA in the knee is
diffuse hyaline cartilage loss, rather than focal bone
erosion.

Subarticular Geodes. The knee is one of the three clas-

sical sites for these large cystic lesions (Fig. 15.7), the Fig. 15.7. Rheumatoid arthritis. AP conventional radiograph
others being the superolateral aspect of the femoral of the knee showing a huge central tibial subarticular geode
276 c. Wakeley and 1. Watt

phy. Whilst little primary diagnostie importance is

attached to this "soft" clinical sign, it is at variance with
bone density changes in primary OA. The combination
of demineralisation, joint malalignment and restrieted
ranges of movement is a potent cause of insufficiency
fracture in RA. Such lesions are usually found in the
tibial plateaux and should be considered when a rapid
worsening of a patient's symptoms occurs.

15.3.2.2
Spondyloarthritis

Both psoriasis and Reiter's syndrome can affect the

knee, though far less commonlythan RA. The cardinal
features are similar to those found in the hand. Hence,
osteopenia is less likely to be demonstrated round
the knee. Indeed, density may appear increased, akin
to sclerotic medullary bone seen in phalanges. Ill-
defined new bone formation is demonstrated at the
site of the marginal erosions, although less floridly
than in the hands and feet. MRI distinguishes between
RA and spondyloarthritis. In the former, the pathol-
ogy is confined within the joint capsule initially, but in
the latter, it may commence in peri-artieular connec-
tive tissue. This may be confirmed using fat suppres-
sion techniques such as STIR.
15.3.2.3
Juvenile Arthritis
Whilst juvenile arthritis comprises aseparate subset of
disorders occurring in the paediatric population, the
disease may progress into adult life or the end results
may present as secondary adult joint disease. In chiI-
dren, the particular feature to note is the non-specific
growth changes related to hyperaemia. Most common-
ly seen are expanded epiphyses, an enlarged intercon-
dylar notch, gracile diaphyses and premature fusion of
the growth plate. As previously indicated, such changes
are non-specific and may be seen in patients with hae-
mophilia or septic arthritis (Fig. 15.8). The persistence
of these moulding changes into adult life gives a clue
as to the timing of disease in the patient's lifetime,
as weil as to the original aetiology when secondary OA
has supervened. Adolescent large joint synovitis may
presage the development of ankylosing spondylitis.
This is important since, initially, HLA antigen results
may be normal. Erosion of bone in juvenile arthritis
of the knee is even less usual than erosion with RA.
Because of the rather non-specific findings, serious
attention should be given to alternatives. These include
chronie low-grade infectious arthritis and chronic
bleeding disorders.
Fig. 15.8. Juvenile arthritis. Lateral conventional radiograph
of the knee demonstrating hypertrophy of the femoral and
tibial condyles, a coarse trabecular pattern and diffuse hyaline
cartilage thinning
Radiological assessment in juvenile arthritis of
the knee is most practieal with plain films, judicious-
ly exposed in deference to radiation burden. Ultra-
sound is extremely valuable as it may confirm syno-
vitis, joint fluid and indeed erosion without the need
for ionising radiation. As in RA, radionuclide scan-
ning will allow identification of other sites of active
disease throughout the body if this is considered clin-
ically necessary.
15.4
What Other Conditions in the Knee May
Present as Arthritis?
Obviously, as the knee is a large synovial joint, many
disease processes may be associated with a non-spe-
cific arthropathy at the knee. The major conditions to
consider include:
1. Tumours
2. Hypertrophie osteoarthropathy
3. Foreign body synovitis (e.g. blackthorn)
4. Haemophilia and other bleeding dis orders
5. Neuropathic involvement
6. Infections
7. Disorders of a possibly metabolic nature
 Arthritis
15.4.1
Tumours
15.4.1.1
Synovial Tumours
Three tumours or tumour-like diseases involve the
knee. These comprise pigmented villonodular syno-
vitis (PVNS), synovial chondromatosis and lipoma
arborescens. None is known to be a neoplasm as such
and all pursue a benign course. One malignancy is
called a synovioma, but, as will be discussed later, it
is probably not a neoplasm of synovium. Indeed, it is
striking how very rarely a malignancy, either primary
or secondary, involves a joint. The secret of the appar-
ent preservation of synovial joints from malignancy
is not known.
PVNS is an inflammatory process of synovium
resulting in a tumour-like mass of synovial tissue
causing a non-specific synovitis/arthritis. If the
tumour expands dose to the joint capsule, pressure
erosion ofbone occurs, the margins of which are sde-
rotic and well defined. MRI can assess the extent of
the PVNS and often identifies low signal within the
lesion due to haemosiderin deposits from episodes
of haemorrhage (Fig. 15.9). However, low signal in
synovium is also associated with any cause of chronic
bleeding into the joint, synovial chondromatosis and
amyloidosis, and is thus not specific. PVNS may be
focal within the joint, a dassical site being the recess
distal to Hoffa's fat pad.
a b
277
Synovial chondromatosis (osteochondromatosis)
is a common dis order, the knee, as in PVNS, being
a dassical site of presentation. The disease results
from the development of metaplastic cartilage within
synovium, of unknown initial aetiology. Enlarge-
ment and shedding of the cartilaginous foci result
in single or multiple separate chondral bodies which
may undergo enchondral ossification (hence "osteo-
chondromatosis"). The patient thus complains of
either the effect of an enlarging mass of synovium
or the mechanical effects of separate osteochondral
bodies. Bony erosion is less common than in PVNS
and synovium may appear opaque on plain film due
to calcification or ossification within the synovial
mass, rather than diffuse opacification due to iron, as
in the bleeding dis orders or PVNS.
Lipoma arborescens occurs most commonly in the
knee. It is characterised by a mass of numerous fat-
laden synovial villous projections that on conven-
tional radiographs may be evident as a fat-contain-
ing, branching mass originating in the joint. The
lesion can be demonstrated by CT or MRI, confirm-
ing the fatty nature of the tumour.
Synovioma is a highly malignant, rapidly growing
tumour with early lymph node involvement. It arises
in young adults adjacent to, but probably not from,
synovium, tendon sheaths or bursae. The knee, espe-
cially the popliteal fossa, accounts for 70% of cases.
Typically, a soft tissue mass is present, with about
one patient in five demonstrating amorphous, lesion-
al calcification. About 10% of synoviomas are asso-
Fig. 15.9a, b. PVNS.
Sagittal MRI images of
the knee. On the post-
gadolinium Tl-
weighted image (a) a
large enhancing
intermediate/high -
signal mass lesion is
seen predominantly
occupying the suprapa-
tellar pouch. The cor-
responding T2-
weighted image (b)
demonstrates the low
signal effect of the
haemosiderin deposits
within the mass
278 C. Wakeley and I. Watt

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

15.4.1.2 Foreign body re action from a penetrating wound,

Primary Bone Tumours such as a blackthorn, can initiate an aggressive
active synovitis that can be difficult to distinguish
The lower femur and upper tibia and fibula are potent from either a septie arthritis or acute inflammatory
sites for the development of primary bone tumours. It arthropathy. In the light of such a clinieal history, an
is beyond the remit of this chapter to consider these in early arthroscopy and removal of the offen ding for-
detail. However, the following should be remembered: eign body is needed. Untreated, severe hyaline carti-
l. Fibrous cortical defects: These are very common lage destruction folIows, akin to a septic arthritis.
around the adolescent knee, but are asymptomatie
save for the occasional traction lesion, or cortical
avulsion syndrome associated with traumatie enthe- 15.4.4
sis disease at the origin of a head of gastrocnemius. Bleeding Disorders
2. Osteoid osteoma: The knee joint area is a very
common site for these lesions, as weIl as pain The knee is one of the classieal joints into which
referred to the knee from one at the hip. Remem- bleeding occurs, due not only to the size of the joint
ber that an osteoid osteoma near a joint may but also to the frequency of injury. Repeated intra-
present as a non-specific synovitis and, in chil- artieular haemorrhage, whilst initially causing joint
dren, growth disturbances akin to juvenile arthri- and soft tissue dis tension from a haemarthrosis, sub-
tis. Plain films, skeletal scintigrams and CT are usu- sequently results in synovial thickening and occa-
ally adequate to make the correct diagnosis. MRI, sionally artieular erosions. Although these features
whilst showing marrow oedema, may obscure the are non-specific, the diagnosis is usually evident from
actual nidus.
3. Osteosarcoma: This, the commonest primary
malignant bone tumour, often arises near the knee.
Again, it is important to be vigilant when a young-
er person presents with knee pain to ensure that
such a tumour is not present.
4. Malignant round ceil tumours: The knee area,
especially the upper fibula, is a classieal site.
5. Giant ceil tumour: In the adult skeleton, the pres-
ence of the typieal eccentric, purely lytie lesion
should leave little doubt for differential diagnosis.

15.4.2
Hypertrophie Osteoarthropathy

Of the multiple causes of hypertrophie osteoarthrop-

athy (HOA), the commonest is carcinoma of the
bronchus. In the lower limb other causes such as
inflammatory bowel disease and chronieally infect-
ed aorto-iliac vascular grafts should be considered.
As elsewhere, the predominant feature is a periosti- Fig. 15.10. Hypertrophie osteoarthropathy. Conventional AP
tis involving the metadiaphyseal regions of the distal radiograph of the knee demonstrating a smooth periosteal reac-
femur and proximal tibia and excluding their respec- tion around both the distal femoral and proximal tibial shafts
Arthritis
the known history of haemophilia, Christmas disease
or von Willebrand's disease. Significant involvement
in childhood results in epiphyse al hypertrophy and
premature growth plate fusion from the episodes of
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
reduced synovial signal on all pulse sequences on
MRI. Recurrent haemarthrosis after joint replace-
ment may be due to arteriovenous malformation in
synovium. It is not known whether this phenome-
non arises de novo, or is an effect of surgery. The
major end-result is secondary OA. The typieal fea-
tures include diffuse hyaline cartilage loss, subchon-
dral radiolucencies, scanty osteophyte and opaque
synovium. Childhood onset is associated with the
secondary growth disturbances described above.
15.4.5
Neuropathie Involvement
Neuropathie involvement of the knee is uncommon
and usually presents as the hypertrophie variant of a
Charcot joint. The atrophie neuropathie joint is rare
in the knee. A non-specific synovitis and joint effu-
sion may be the first signs. This may be clinieally
silent, although it is worth remembering that Char-
cot's original description included painful knees. Sub-
sequent hypertrophie changes with subchondral scle-
rosis and osteophyte formation are seen, as described
Fig.15.11. Haemophilia. Lateral conventional radiograph of the
knee demonstrating the high-density radiopaque synovium,
best appreciated in the suprapatellar pouch
279
in OA. Characteristieally joint failure is associated
with gross osteochondral fragmentation of the joint
surfaces, filling the joint cavity with debris and result-
ing in end-stage collapse of the joint (Fig. 15.12).
15.4.6
Infections
An infectious arthritis is relatively common in the
knee, particularly due to atypieal organisms includ-
ing gonococcus and tuberculosis. Risk populations
include intravenous drug users and those with sex-
ually transmitted disorders. Further, the knee is a
common site of presentation for tiek-related disor-
ders, for example Lyme disease. Initial joint aspira-
tion may not be diagnostie, but persistent synovitis
results in significant hyaline cartilage destruction.
Similarly, the knee may be the presenting feature of
local bone infection, especially Brodie's abscess.Acute
septie arthritis (see also Chap. 14) in its early phase
may be indistinguishable from other acute infiam-
matory arthropathies. Septic arthritis is important,
since it is treatable and, if undiagnosed and untreat-
ed, results in joint destruction. Thus, it is important
always to consider infection as a cause of an acute
arthritis and always to consider aspiration of the
joint for bacterial mieroscopy as well as crystal analy-
sis. In the early phase of septie arthritis, there is syno-
vial thickening,joint effusion and soft tissue swelling.
On conventional radiographs, the distortion of asso-
Fig. 15.12. Neuropathie joint. AP conventional radiograph of
the knee showing gross hypertrophie destructive change and
disorganisation
280
ciated fat planes may be difficult to visualise owing
to the soft tissue oedema obscuring the normal out-
lines of fat planes. An associated hyperaemia usually
results in peri-artieular osteopenia. If left untreated,
septic arthritis will lead to cartilage destruction fol-
lowed by subchondral bone destruction and even-
tually collapse of the entire joint. The end-result is
either fibrous or bony ankylosis.
Skeletal scintigraphy is often helpful in distin-
guishing between soft tissue inflammation and bone
or joint involvement. MRI not only demonstrates the
synovial component, but also assesses the hyaline
cartilage and underlying bone marrow.
15.4.7
Metabolie Conditions
Two main subsets of disease affect the knee:
- Those associated with the deposition of crystals in
or around the joint
- Those due to systemie metabolie disease
15.4.7.1
Crystal Deposition
Three main crystals are associated with a major
inflammatory synovitis in humans. They comprise
calcium pyrophosphate dihydrate (CPPD), calcium
hydroxyapatite (also known as basie calcium phos-
phate: BCP) and monosodium urate monohydrate
(MSUM). The role of crystals in the pathogenesis of
arthritis is debated. The current view is that crystals
are really a disease marker and not simply part of
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.
CPPD. Radiologieal chondrocalcinosis is a normal,age-
related phenomenon. Approximately 40% of normal,
C. Wakeley and I. Watt
asymptomatie 80 year olds exhibit chondrocalcinosis
on knee X-rays. Patients with hyperparathyroidism or
haemochromatosis have a higher incidence of chon-
drocalcinosis than the normal population. The knee
is the classieal site for CPPD, as typically deposition
occurs in those joints with both fibro- and hyaline car-
tilage within them. CPPD deposition results in linear
or granular calcification of fibro- and hyaline carti-
lage (Fig. 15.13). In gross cases, it will also be seen in
the capsule and at entheses. Apart from being a clas-
sieal site of acute crystal shedding ("pseudogout"),
as superficiallayers of hyaline cartilage are shed into
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
and OA without CPPD crystals. Although PFJ OA has
been emphasised in association with CPPD, this may
be aseparate subset of OA as such, and the crystal-
related aetiology is again dubious.
BCP. Calcium hydroxyapatite is most frequentlydepos-
ited in peri-artieular locations (such as the supraspi-
natus tendons of the shoulder). However, it is also
obtained from joint aspirations in patients with a
form of atrophie destructive OA of the knee and
shoulder. It is important to emphasise that aspiration
of fluid from normal knee joints will yield BCP crys-
tals. They are a component of normal joint fluid. They
are in gross excess in atrophie destructive OA, how-
ever. The main radiographie features of the latter are
rapidly progressive bone destruction in the absence
of a secondary bone response. Large (often haemor-
rhagic) joint effusions containing abundant BCP crys-
tals are obtained. As with CPPD, the role of these crys-
Fig. 15.13. Chondrocalcinosis. AP conventional radiograph of
the knee demonstrating chondrocalcinosis of the menisci and
hyaline cartilage
Arthritis
Fig. 15.14. CPPD. Lateral conventional radiograph demon-
strating hypertrophie DA in association with CPPD
tals is debated. Whilst many probably represent "bone
dust" from joint disintegration, some may be newly
formed but of uncertain aetiologieal significance.
MSUM. The deposition of MSUM crystals takes place
over manyyears. It is unclear why these crystals become
shed, but when they do a severe, acute arthropathy is
caused. Acute gout does not commonly affect the knee,
the typieal joint remaining the great toe metatarsopha-
langeal joint in 80% of cases. In the more insidious sec-
ondary gout, usually associated with diuretic therapy
and renal failure, acute attacks are rare and patients
present with the chronic tophaceous form of the dis-
ease. The knee is an uncommon primary presenting
site for acute gout and hence, radiographie changes are
rarely seen. In the chronie form, peri-artieular ero-
sions may occur in association with tophaceous depos-
its. These present as well-defined peri-artieular ero-
sions with prominent overhanging margins. Intraosse-
ous lesions have also been described.
15.4.8
Systemic Metabolie Disorders
Haemochromatosis. This rare disorder of iron metabo-
lism results in iron deposition at various sites through-
out the body. In the musculoskeletal system, the small
joints of the hand are primarily involved, but even-
tually large joints (including the knee and hip) are
281
affected. More patients are surviving into later adult
life and hence large joint disease is becoming of great-
er significance, leading to joint replacement. Ten males
are affected to every female. Up to 50% of patients also
have CPPD with plain film chondrocalcinosis. Indeed,
there is argument as to whether the CPPD is responsi-
ble for the structural arthropathie changes or wheth-
er there is a true haemochromatotie arthropathy. An
intimate relationship exists between calcium, magne-
sium and iron metabolism, especially in hyaline carti-
lage. The radiographie features ofhaemochromatosis-
related arthritis are relatively non-specific, with joint
space narrowing and subchondral sclerosis. Multiple
subarticular "cystie" lesions are very suggestive of the
diagnosis. The lesions are due to separation of hya-
line cartilage from subchondral bone as the result of
straightening of the normally corrugated zone of pro-
visional calcification, the "tide mark". With hyaline
cartilage shed, the bony manifestations of OA with
osteophyte formation follow.
Hyperparathyroidism. Whilst the major manifesta-
tions of this disease present outwith the knee, a con-
stant finding is florid and premature chondrocalcino-
sis (see above). Thus, this condition should be borne
in mind when a younger patient presents with such
calcification. Two aspects also distinguish the patient
with hyperparathyroidism from that with "idiopath-
ie" CPPD. They are the rarity of acute episodes of
crystal shedding (pseudogout) and the absence of sig-
nificant structural change ("pyrophosphate arthrop-
athy"). This striking difference challenges also the
simplistie cause-effect concept of CPPD crystals as
the prime cause of pyrophosphate arthropathy.
Alkaptonuria and Ochronosis. Alkaptonuria is an
extremely rare autosomal recessive inherited meta-
bolie defect resulting in homogentisic acid deposi-
tion in articular cartilage. Radiographic features may
be similar to severe OA, with joint space narrowing,
osteophyte formation and peri -artieular calcification.
The association of these changes with severe mul-
tilevel degenerative spondylosis and discal calcifica-
tion should suggest the diagnosis, apart from obvi-
ous cartilage pigmentation clinieally.
15.5
Summary
The knee is an important joint, frequently subject to
trauma, and taking heavier loads as our populations
282
become more obese. Further, our ageing, "greying"
population rightfully demands to remain active. Yet,
one potent risk to that mobility is knee arthritis. Radi-
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
diagnosis that may be made. Most frequently, a radio-
logical report is made to a family physician, who will
never see the film and reHes solelyon the report. As in
C. Wakeley and I. Watt
all things, the radiologist is there to serve the patient
and steer clinical management in their best interest.
With a large joint like the knee, that role is crucial.
Further Reading
Resnick D (ed) (1995) Diagnosis of bone and joint dis orders,
3rd edn. Saunders, Philadelphia
16 Assessment of Knee Cartilage
SOUHIL ZAIM, ALl GUERMAZI, JOHN A. LYNCH, CHARLES G. PETERFY, HARRY K. GENANT

CONTENTS arthritis will have an increasing impact on health

care in the future.
16.1 Introduction 283
16.2 Anatomy, Function and Structure 284
Hyaline articular cartilage that lines the articular
16.3 MR Appearance 284 surfaces of the knee joint is a unique and remark-
16.3.1 Normal Appearance and Artifacts 284 able tissue. Articular cartilage sustains heavy loads
16.3.2 Spin Echo Imaging 285 of body weight and provides a low-friction surface
16.3.3 Fat-Suppressed and Fast Spin Echo Imaging 285
of unparalleled performance. In addition to support-
16.3.4 Fat-Suppressed Spoiled Gradient Echo Imaging 285
16.3.5 Magnetic Resonance Arthrography 287
ing impressive biomechanical specifications, articu-
16.4 Other Advanced MR Techniques 287 lar cartilage is a living tissue, in which mechanical
16.4.1 Magnetization Transfer 287 attrition is continuously balanced by biological syn-
16.4.2 Driven Equilibrium 287 thesis and repair. Normal cartilage can thus endure
16.4.3 Contrast-Enhanced Imaging 287
a lifetime of heavy loading without significant wear.
16.4.4 Diffusion Imaging 288
16.4.5 Sodium Magnetic Resonance Imaging 288
Preservation of the integrity of this tissue is critical
16.5 Quantitative Measurements of Cartilage 288 to the normal function of joints, and failure of the
16.5.1 Cartilage Volume and Thickness Measurement 288 biomechanical and biological properties of articular
16.5.2 Parameter Mapping 289 cartilage is widely believed to be a central element in
References 290
the pathophysiology of osteoarthritis.
Assessment of osteoarthritis and cartilage inju-
ries includes clinical, arthroscopic, and imaging
evaluations. Arthroscopy is the most common stan-
16.1 dard of reference for evaluation of cartilage abnor-
Introduction malities; however, it provides information on the
articular surface only, is invasive and expensive, and
The evaluation of articular cartilage has gained cannot be utilized for diagnosing and monitoring
increasing attention in the scientific community cartilage damage and repair. Radiography provides
as injuries and degenerative changes in the artic- information about bone changes associated with
ular cartilage are a significant cause of morbidity osteoarthritis and allows indirect assessment of car-
and impaired quality of life (YELIN and CALLAHAN tilage by means of joint space narrowing, which has
1995), with an increasing impact in our aging soci- been shown to be a poor indicator of cartilage integ-
ety. The fundamental component of osteoarthritis rity (CHAN et al. 1991). Conventional arthrography
is degradation and subsequent loss of articular car- and its variant CT-arthrography are no longer
tilage. used in the assessment of internal derangements
Osteoarthritis affects over 16 million Americans of the knee. Magnetic resonance (MR) imaging has
and is one of the leading causes of dis ability in the emerged as the imaging method of choice mainly
elderly. Given the aging of the US population, osteo- because of its excellent tissue contrast and non-
invasiveness.
The focus of this chapter will be on existing,
widely available MR imaging techniques, and on
advances in the MR imaging evaluation of articular
S ZAIM, MDj A. GUERMAZI, MDj J.A. LYNCH, PhDj
cartilage of the knee with the use of novel imaging
C.G. PETERFY, MD, PhDj H.K. GENANT, MD, Professor
Osteoporosis and Arthritis Research Group, Department of sequences and new image processing techniques
Radiology, University of California San Francisco, 350 Parnas- to assess cartilage morphology and biochemical
sus Avenue, Suite 150, San Francisco, CA 94143-1349, USA parameters.
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
284
16.2
Anatomy, Function, and Structure
Artieular cartilage consists of sparsely scattered chon-
drocytes embedded in an extracellular matrix com-
posed predominantly of collagen fibers, aggregated
proteoglycans, and water. The collagen fibers impart
tensile strength to the matrix and res ist the swelling
of the proteoglycan gel. The proteoglycans are com-
posed of core proteins to whieh are attached numer-
ous glycosaminoglycans, such as keratan sulfate and
chondroitin sulfate, which have pendant sulfate and
carboxylate groups that are ionized under physiolog-
ie conditions. These negatively charged moieties repel
each other and attract sodium ions that in turn draw
water osmotieally into the cartilage. The combination
of electrostatic repulsion and osmotie pressure pro-
duces a swelling pressure that keeps the artieular car-
tilage inflated and the collagen fibers under tension
(AKESON et al. 1995).
The structure of cartilage is highly organized and
composed of several contiguous layers (Fig.16.1). The
superficial or tangential zone is composed of tightly
packed collagen fibers that are oriented parallel to the
artieular surface, and the transitional or intermediate
zone, just below the superficial zone, consists of col-
lagen fibers with oblique orientations. The radial or
deep zone is characterized by collagen fibers that are
oriented perpendicular to the subchondral bone plate
and are anchored in the calcified zone of cartilage.
Superficial
~ne
~nsitional
zone
Upper radial
~ne
Deep radial
~ne
Calcified
~ne
Fig. 16.1. Zonal organization of hyaline cartilage
Fig. 16.2. Magie angle artifact. A foca! area of increased signal
is observed at curved area of cartilage of the femoral trochlea
(arrowhead)
S. Zaim et al.
16.3
MR Appearance
The appearance of cartilage on MR imaging reflects
its structure and composition. Contrast results mainly
from the water content, i.e. the proton density. The
high water content of artieular cartilage forms the
basis for the MR signal. However, this fundamental
MR imaging signal is modulated by a number of pro-
cesses, including TI relaxation, T2 relaxation, magne-
tization transfer and water diffusion. The thickness of
cartilage also poses achalienge as cartilage measures
onlya few millimeters at the most, the patellar carti-
lage being the thickest. Not surprisingly, the patellar
cartilage has been the focus of most in vivo studies.
16.3.1
Normal Appearance and Artifacts
The thiekness of artieular cartilage varies in the knee,
with the thiekest cartilage found at the patella and the
thinnest at the tibial plateau. At 1.5 Tesla, the Tl of car-
tilage is about 700 ms and the T2 varies between 30 ms
at the radial zone and 70 ms in the transitional zone
(DARDZINSKI et al.1997). This pattern ofT2 variation,
also known as T2 anisotropy, is the result of the orien-
tation of organized collagen fibers relative to the static
magnetic field (Ba)' or the "magie-angle" effect (ERICK-
SON et al. 1996). The magie-angle effect explains the
observed layering of cartilage related to its ultrastruc-
tural disposition when high-resolution imaging, espe-
cially in vitro imaging with special attention to the
technique, is performed (RUBENSTEIN et al. 1993; XIA
et al.1997).Additionally, the magie-angle phenomenon
is responsible for areas of increased signal observed
at curved areas of cartilage, typically at the femoral
trochlea (Fig.16.2) and femoral condyles.
 Assessment of Knee Cartilage
The laminar appearance of cartilage should be dis-
tinguished from truncation artifacts that appear in
sequences where cartilage is in high contrast, such
as the fat-suppressed spoiled gradient echo sequence
(SPGR) (ERICKSON et al. 1996). This manifests as one
or several thin bands of low signal intensity midway
through the cartilage on short-TE images. Truncation
artifacts are less common on high-resolution images,
but usuallypresent on fat-suppressed 3D SPGR images
generated with most clinical protocols (Fig. 16.3).
a Fast spin echo T2-weighted imaging increases
Fig. 16.3a, b. Truncation artifact. Sagittal fat-suppressed T2-
weighted images show a thin band of low signal intensity
midway through the cartilage (arrowhead) within the femoral
condyle (a) and the femoral trochlea (b)
285
16.3.2
Spin Echo Imaging
Conventional spin-echo imaging of cartilage com-
bines the benefits of simplicity and wide availability.
On Tl- and proton density-weighted sequences, car-
tilage appears homogeneous in signal, showing good
contrast with subchondral bone and poor contrast
with joint fluid. T2-weighted sequences capitalize on
the arthrogram-like effect provided by the adjacent
synovial fluid, thus highlighting the presence of sur-
face defects of cartilage. Intra-substance T2 abnor-
malities of cartilage can also be detected. Conven-
tional T2-weighted sequences suffer from a reduced
signal-to-noise ratio and from the poor signal of car-
tilage when long echo times are used.
16.3.3
Fat-Suppressed and Fast Spin Echo Imaging
An interesting approach to augment Tl contrast is
to suppress the fat signal and to rescale the smaller
residual Tl contrast across the image. This generates
images where cartilage is seen as a high signal intensi-
ty band in sharp contrast to adjacent low signal inten-
sity joint fluid, the suppressed fat, and bone. Fat sup-
pression also eliminates chemical shift artifacts that
can distort the cartilage-bone interface (Fig. 16.4)
(CHANDNANI et al. 1991; PETERFY et al.1994a).
the signal-to-noise ratio with reduced imaging time.
Additionally, the inherent magnetization transfer
effect (YAO et al. 1996) proper to this sequence results
in an overall improved contrast with the adjacent tis-
sues. Several studies have shown the diagnostic accu-
racy of fast spin echo imaging combined with fat
suppression in the diagnosis of cartilage lesions
of the knee (BREDELLA et al. 1999; POTTER et al.
1998). Using this type of sequence, arecent study
has shown that the presence of subchondral bone
marrow changes, especially bone marrow edema,
increases the sensitivity for detection of focal carti-
lage lesions (Figs. 16.5, 16.6) (RUBIN et al. 2000).
16.3.4
Fat-Suppressed Spoiled Gradient Echo Imaging
Augmentation of Tl contrast can also be combined
with high er through-plane spatial resolution and
higher signal-to-noise ratio by using a three-dimen-
sion gradient echo technique. This technique allows
 286 S. Zahn et al.

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
with high contrast of cartilage (Fig. 16.7). It is widely
available (3D SPGR, 3D FLASH) and has become the
most sensitive clinical technique for delineating artic-
ular cartilage lesions (DISLER et al. 1996; RECHT et
al. 1996) and measuring cartilage volume and thick-
ness (PETERFY et al. 1994a; ECKSTEIN et al.1996). This
technique requires, however, long imaging times of
approximateiy 12 min for the entire knee and is prone
to susceptibility artifacts and heterogeneity of satura-
tion of signal from fat. It does not allow analysis of
intrasubstance signal evaluation of the cartilage.
Assessment of Knee Cartilage
Fig.16.7. Normal appearance of the lateral femorotibial carti-
lage on a 3D SGPR sequence
16.3.5
Magnetic Resonance Arthrography
Magnetic resonance arthrography, through its abil-
ity to accurately delineate intra-articular structures,
additionally expands the dinical use of conventional
MR imaging in the evaluation of cartilage injury and
degeneration (VAHLENSIECK et al. 1996).
In direct MR arthrography a mixture of Gd-DTPA
and saline solution is injected into the joint. MR
arthrography has been shown to have excellent accu-
racy in the diagnosis and staging of chondral and
osteochondral injury and degeneration. Clinical use
of direct MR arthrography using a gadolinium-saline
mixture has been limited for several reasons: the
conversion of a noninvasive procedure into a mini-
mally invasive procedure, the increased cost and time
required to do MR arthrography compared with con-
ventional MR imaging, and the excellent accuracy of
conventional MR imaging in the evaluation of articu-
lar cartilage. Currently, the main role of MR arthrog-
raphy is for staging osteochondrallesions.
16.4.
Other Advanced MR Techniques
16.4.1
Magnetization Transfer
In magnetization transfer imaging a tissue and its
components are exposed to an imposed magnetic
287
field. The effect of the imposed field is modified by
the interaction and proximity of the separate compo-
nents, which comprise a local macromolecular envi-
ronment. The effect of the magnetic field on one com-
ponent of cartilage in isolation would be different
than the effect of imaging the entire tissue. Carti-
lage, because of its highly organized nature, shows
the effects of magnetization transfer. This seems to be
primarily attributable to the interaction of two com-
ponents, the collagen matrix and bulk water. Because
the intrinsic contrast produced by this effect depends
on the exchange between interacting components,
any disruption in the ordered structure would, the-
oretically, involve a change in that contrast. In its
current form, magnetization transfer results in only
moderate contrast-to-noise and its effects in abnor-
mal cartilage are not known. Its sensitivity to the
macromolecular environment may prove useful for
identifying and monitoring early changes in cartilage
(GRAY et al. 1995).
16.4.2
Driven Equilibrium
The driven equilibrium Fourier transfer sequence
pro duces image contrast that is a function of proton
density, intrinsic Tl and T2, and echo time and rep-
etition time (HARGREAVES et al. 1999). Its contrast
is well suited to articular cartilage. Synovial fluid
is high in signal intensity and articular cartilage is
intermediate in signal intensity. Bone is dark and fat
is suppressed, resulting in excellent contrast of car-
tilage from surrounding tissues. While maintaining
this contrast, this sequence also succeeds in keeping
a high signal-to-noise ratio and allows visualization
of structural elements of cartilage. Three-dimension-
al imaging capabilities have been developed for this
sequence with scan tim es that are acceptable for din-
ical applications.
16.4.3
Contrast-Enhanced Imaging
Imaging of proteoglycan concentration and its distri-
bution in cartilage can be achieved by gadolinium-
enhanced imaging. The histological and biochemical
validity of this approach has been well demonstrated
(BASHIR et al. 1997) and is based on the fact that pro-
teoglycans are lost early after injury or degeneration.
As a result, their negative charge is also lost and con-
sequently, when a negatively charged contrast agent
288
such as Gd-DTPA is administered intravenously, it
will accumulate into the degraded cartilage, no longer
being repelled by the negative charge of the proteo-
glycans. Using cartilage-nulling inversion recovery
sequences at high spatial resolutions and high field
strength, a high histological correlation of the distri-
bution of anionic Gd-DTPA with proteoglycan deple-
tion has been demonstrated (BASHIR et al. 1997;
TRATTNIG et al. 1999). The effect of the Gd-DTPA 2-
could be assessed visually and quantitatively using
mapping techniques (see below).
16.4.4
Diffusion Imaging
Water diffusion in cartilage also contributes to signal
loss on T2-weighted MR images. Because water mole-
cules that have changed positions during a portion of
the MR imaging acquisition can no longer contribute
maximally to the net signal, a resulting loss of phase
coherence is proportional to the distance traveled by
the diffusing water protons and is, therefore, worse
on long-TE images. The presence of proteoglycans,
particularly chondroitin sulfate, in normal cartilage
inhibits water diffusion and keeps this effect relative-
ly small, although with very strong gradients and
specialized phase-sensitive pulse sequences, water
diffusion can be demonstrated and quantified in
normal cartilage (BURST EIN et al. 1993). With car-
tilage degeneration and proteoglycan loss, however,
water diffusion has been shown to increase consid-
erably. Accordingly, diffusion may playa more sig-
nificant role in cartilage signal modulation in knee
osteoarthritis.
16.4.5
Sodium Magnetic Resonance Imaging
Another imaging technique that has shown pro mise
with respect to evaluation of the biochemical status
of cartilage is sodium MR imaging (INSKO et al.
1997). Similar to the use of ionic gadolinium com-
pounds, sodium imaging also allows the assessment
of the proteoglycan content of cartilage. Also based
on the charge-related loss paralleling proteoglycan
depletion, sodium imaging can depict the same
regions. Unfortunately, this technique requires the
use of sophisticated MR imaging equipment and
high-resolution imaging. In addition, it has very low
inherent signal-to-noise ratio, making it of question-
able clinical value in its current form.
s. Zaim et al.
16.5
Quantitative Measurements of Cartilage
Truly quantitative compositional data may be obtained
for biological tissue by subjecting it to bulk chemical
analysis. However, this provides no information about
the distribution of matrix components and also results
in the complete destruction of the tissue. In vivo,
however, it is possible to measure the volume of carti-
lage using the morphologic sequences and to generate
mapping based on the tissue parameters of cartilage.
16.5.1
Cartilage Volume and Thickness Measurement
Measurements of cartilage thickness and volume are
typically taken from fat-suppressed 3D spoiled GRE
images (Fig. 16.8). The earliest methods (PETERFY et
al. 1994a, b; RECHT et al. 1996; ECKSTEIN et al. 1994;
MARSHALL et al. 1995; PILCH et al. 1994) involved
region-growing techniques on a slice-by-slice basis,
with manual tracing of subtle boundaries, where car-
tilage surfaces were in contact or where overlying tis-
sues with similar intensities occurred. These meth-
ods were tedious and time consuming, requiring
many subjective decisions from experienced human
observers, especially in regions where partial volume
averaging had occurred, but can be relatively reliable
for measuring total cartilage volumes (ECKSTEIN et
al. 1994, 1998a, b; HARDY et al. 2001; TIESCHKY et al.
1997). Such techniques have been used to examine
progression in osteoarthritis (PETERFY et al. 1998),
the effects of exercise on cartilage (ECKSTEIN et al.
1998c), and the distribution of cartilage within the
knee (ECKSTEIN et al. 1998d).
Other researchers (COHEN et al. 1999; LYNCH et al.
2000; SOLLOWAY et al. 1997; STAMMBERGER et al. 1999,
2000) have developed more automated techniques that
use information on edge strengths within the image.
A technique which uses 2D deformable active shape
models has been described (COHEN et al. 1999), and
preliminary results of the use of a 3D template-based
segmentation technique have also been described. A
semiautomated technique using B-spline snakes has
been used to delineate cartilage boundaries on a slice-
by-slice basis (LYNCH et al. 2000), and a 3D model-
based approach has been used to segment many struc-
tures within the knee, including cartilage. A compari-
son of a manual and a semiautomated segmentation
technique has also been performed (STAMMBERGER
et al. 1999). All these automated or semiautomated
techniques have the potential to reduce the amount of
Assessment of Knee Cartilage
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
human interaction required for segmenting cartilage
from knee MR imaging scans, and may also provide
more reliable measures of thiekness and volume.
More recently, techniques for measuring serial
changes in localized thiekness and volume have been
developed (STAMMBERGER et al. 1999, 2000). One
technique has utilized methods for matching baseline
and follow-up cartilage surfaces (STAMMBERGER et
al. 1999), to allow measurement of localized changes
in thiekness. A different approach has been taken by
other researchers, who have utilized voxel intensity-
based image registration to match the baseline and
follow-up MR imaging data (LYNCH et al. 2001). Com-
bined with a semiautomated method for outlining the
articular cartilage (STAMMBERGER et al. 2000), this
has allowed measurement of localized changes in car-
tilage with high reproducibility and minimal human
interaction.
289
16.5.2
Parameter Mapping
The mapping of physiologie parameters with MR
imaging has emerged as a promising me ans for the
quantitative assessment of cartilage structure and
function in the normal and pathologie states. Relax-
ation times are calculated on a pixel-by-pixel basis
using a multiple echo spin-echo sequence, and the
values obtained are typically assigned a color and a
Tl or T2 map is generated (Fig. 16.9) (DARDZINSKI
et al. 1997). The technique is usually used to gener-
ate Tl maps after gadolinium administration as weIl
as for assessment of T2 distribution within the carti-
lage. This technique requires high resolution, usually
the use of a high field strength magnet or local gradi-
ent coils, and also requires significant imaging time
and postprocessing.
Fig. 16.9. Transverse T2 mapping of the
femoropatellar cartilage and color scale
(bottom right). Spatial T2 dependency of
articular cartilage is observed. Note that
the low T2 values increase progressively
from the radial to the superficial zone.
(Courtesy of Timothy J. Mosher, MD,
Pennsylvania State University College of
Medicine)
290
References
Akeson WH, Amiel DA, Gershuni DH (1995) Articular cartilage
physiology and metabolism. In: Resnick D (ed) Diagnosis
of bone and joint disorders, 3rd edn. Saunders, Philadel-
phia, pp 769-790
Bashir A, Gray ML, Boutin RD, Burstein D (1997) Glycos-
aminoglycan in articular cartilage: in vivo assessment
with delayed Gd DTPA2- -enhanced MR imaging. Radiology
205:551-558
Bredella MA, Tirman PF, Peterfy CG, Zarlingo M, FeIler JF, Bost
FW, Belzer JP, Wischer TK, Genant HK (1999) Accuracy
of T2-weighted fast spin-echo MR imaging with fat satu-
ration in detecting cartilage defects in the knee: compar-
ison with arthroscopy in 130 patients. Am J Roentgenol
172:1073-1080
Burstein D, Gray ML, Hartman AL, Gipe R, Foy BD (1993) Dif-
fusion of small solutes in cartilage as measured by nuclear
magnetic resonance (NMR) spectroscopy and imaging. J
Orthop Res 11:465-478
Chan WP, Lang P, Stevens MP, Sack K, Majumdar S, Stoller
DW, Basch C, Genant HK (1991) Osteoarthritis of the knee:
comparison of radiography, CT, and MR imaging to assess
extent and severity. Am J RoentgenoI157:799-806
Chandnani VP, Ho C, Chu P, D Trudell D, Resnick D (1991) Knee
hyaline cartilage evaluated with MR imaging: a cadaveric
study involving multiple imaging sequences and intraar-
ticular injection of gadolinium and saline solution. Radiol-
ogy 178:557-561
Cohen ZA, McCarthy DM, Kwak SD, Legrand P, Fogarasi F, Ciac-
cio EJ,Ateshian GA (1999) Knee cartilage topography, thick-
ness and contact areas from MRI: in vitro calibration, and
in vivo measurements. Osteoarthritis Cartilage 7:95-109
Dardzinski BJ, Mosher TJ, Li S, Van Slyke MA, Smith MB
(1997) Spatial variation of T2 in human cartilage. Radiol-
ogy 205:546-550
Disler DG, McCauley TR, Kelman CG, Fuchs MD, Ratner
LM, Wirth CR, Hospodar PP (1996) Fat-suppressed three-
dimensional spoiled gradient -echo MR imaging of hyaline
cartilage defects in the knee: comparison with standard MR
imaging and arthroscopy.Am J RoentgenoI167:127-132
Eckstein F, Sittek H, Milz S, Putz R, Reiser M (1994) The mor-
phology of articular cartilage assessed by magnetic reso-
nance. Surg Radiol Anat 16:429-438.
Eckstein F, Gavazzeni A, Sittek H, Haubner M, Losch A, Milz
S, Englmeier KH, Schulte E, Putz R, Reiser M (1996) Deter-
mination of knee joint cartilage thickness using three-
dimensional magnetic resonance chondro-crassometry (3D
MR-CCM). Magn Reson Med 36:256-265
Eckstein F, Schnier M, Haubner M, Priebsch J, Glaser C, Engl-
meier KH, Reiser M (1998a) Accuracy of 3D knee joint car-
tilage volume and thickness measurements with MRI. Clin
Orthop Rel Res 52:137-148
Eckstein F, Westhoff J, Sittek H, Maag KP, Haubner M, Faber
S, Engelmeier KH, Reiser M (1998b) In vivo reproducibility
of three-dimensional cartilage volume and thickness mea-
surements with MRI. Am J RoentgenoI170:593-597
Eckstein F, Tieschky M, Faber SC, Haubner M, Kolem H, Engl-
meier KH (1998c) Effect of physical exercise on cartilage
volume and thickness in vivo: an MRI study. Radiology
207:243-248
Eckstein F, Winzheimer M, Westhoff J, Schnier M, Haubner
M, Englmeier KH, Reiser M, Putz R (1998d) Quantitative
S. Zaim et al.
relationships of normal cartilage volumes of the human
knee joint: assessment by MRI. Anat Embryol (Berl)
197:383-390
Erickson SJ, Waldschmidt JG, Czervionke LF, Prost RW (1996)
Hyaline cartilage: truncation artifact as a cause of trilaminar
appearance with fat-suppressed three-dimensional spoiled
gradient-recalled sequences. Radiology 201:260-264
Gray ML, Burstein D, Lesperance LM, Gehrke L (1995) Magne-
tization transfer in cartilage and its constituent macromol-
ecules. Magn Reson Med 34:319-325
Hardy PA, Nammalwar P, Kuo S (2001) Measuring thickness of
articular cartilage from MR images. J Magn Reson Imaging
13:120-126
Hargreaves BA, Gold GE, Lang PK, Conolly SM, Pauly JM, Berg-
man G, Vandevenne J, Nishimura DG (1999) MR imaging of
articular cartilage using driven equilibrium. Magn Reson
Med 42:695-703
Insko EK, Reddy R, Leigh JS (1997) High resolution, short echo
time sodium imaging of articular cartilage. J Magn Reson
Imaging 7: 1056-1059
Lynch JA, Zaim S, Zhao J, Peterfy CG, Genant HK (2000) Car-
tilage segmentation of 3D MRI scans of the osteoarthritic
knee combining user knowledge and active contours. Proc
SPIE 3979:925-935
Lynch JA, Zaim S, Zhao J, Peterfy CG, Genant HK (2001) Auto-
mating measurement of subtle changes in articular carti-
lage from MRI of the knee by combining 3D image registra-
tion and segmentation. Proc SPIE 4322:431-439
Marshali KW, Mikulis DJ, Guthrie BM (1995) Quantification of
articular cartilage using magnetic resonance imaging and
three-dimensional reconstruction. J Orthop Res 13:814-823
Peterfy CG, Majumdar S, Lang P, van Dijke CF, Sack K, Genant
HK (1994a) MR imaging of the arthritic knee: improved
discrimination of cartilage, synovium, and effusion with
pulsed saturation transfer and fat-suppressed Tl-weighted
sequences. Radiology 191:413-419
Peterfy CG, van Dijke CF, Janzen DL, Gluer C, Namba R,
Majumdar S, Lang P, Genant HK (1994b) Quantification of
articular cartilage in the knee by pulsed saturation trans-
fer and fat-suppressed MRI: optimization and validation.
Radiology 192:485-491
Peterfy CG, White DL, Zhao J, van Dijke CF, Genant HK
(1998) Longitudinal measurement of knee articular carti-
lage volume in osteoarthritis. Arthritis Rheum 41: 1970
Pilch L, Stewart C, Gordon D, Inman R, Parsons K, Pataki I, Ste-
vens J (1994) Assessment of cartilage volume in the femoro-
tibial joint with magnetic resonance imaging and computer
reconstruction. J RheumatoI21:2307-2319
Potter HG, Linklater JM,Allen AA, Hannafin JA, Haas SB (1998)
Magnetic resonance imaging of articular cartilage in the
knee. An evaluation with use of fast-spin-echo imaging. J
Bone Joint Surg Am 80:1276-1284
Recht MP, Piraino DW, GA Paletta, JP Schils, Belhobek
GH (1996) Accuracy of fat-suppressed three-dimensional
spoiled gradient-echo FLASH MR imaging in the detection
of patellofemoral articular cartilage abnormalities. Radiol-
ogy 198:209-212
Rubenstein JD, Kim JK, Morova-Protzner I, Stanchev PL, Hen-
kelman RM (1993) Effects of collagen orientation on MR
imaging characteristics of bovine articular cartilage. Radi-
ology 188:219-226
Assessment of Knee Cartilage
Rubin DA, Harner CD, Costello JM (2000) Treatable chondral
injuries in the knee: frequency of assodated focal subchon-
dral edema.Am J RoentgenoI174:1099-1106
Solloway S, Hutchison CE, Waterton JC, Taylor CJ (1997) The
use of active shape models for making thickness measure-
ments of articular cartilage from MR images. Magn Reson
Med 36:943-952
Stammberger T, Eckstein F, Michaelis M, Englmeier KH, Reiser
M (1999) Interobserver reprodudbility of quantitative car-
tilage measurements: comparison of B-spline snakes and
manual segmentation. Magn Reson Med 17:1033-1042
Stammberger T, Hohe J, Engelmeier KH, Reiser M, Eckstein F
(2000) Elastic registration of 3D cartilage surfaces from MR
image data for detecting local changes in cartilage thick-
ness. Magn Reson Med 44:592-601
Tieschky M, Faber S, Haubner M, Kolem H, Schulte E, Englmei-
er KH, Reiser M, Eckstein F (1997) Repeatability of patel-
lar cartilage thickness patterns in the living, using a fat-
291
suppressed MRI sequence with short acquisition time and
3D data processing. J Orthop Res 15:808-813
Trattnig S, Mlynarik V, Breitenseher M, Huber M, Zembsch A,
Rand T, Imhof H (1999) MRI visualization of proteoglycan
depletion in articular cartilage via intravenous administra-
tion of Gd-DTPA. Magn Reson Imaging 17:577-583
Vahlensieck M, Peterfy CG, Wischer T, Sommer T, Lang P,
Schlippert U, Genant HK, Schild HH (1996) Indirect MR
arthrography: optimization and clinical applications. Radi-
ology 200:249-254
Xia Y, Farquhar T, Burton-Wurster N, Lust G (1997) Origin of
cartilage laminae in MRI. J Magn Reson Imaging 7:887-894
Yao L, Gentili A, Thomas A (1996) Inddental magnetization
transfer contrast in fast spin -echo imaging of cartilage. J
Magn Reson Imaging 6:180-184
Yelin E, Callahan LF (1995) The economic cost and sodal and
psychological impact of musculoskeletal conditions. National
Arthritis Data Work Groups. Arthritis Rheum 38: 1351-1362
17 Osteochondritis and Osteonecrosis
ARTHUR A. DE SMET

CONTENTS However, they differ in their likely etiology and age

distribution. Since OCD is often seen in individuals
17.1 Introduction 293
17.2 Osteochondritis Dissecans 293
who are active in sports, acute or repetitive trauma
17.2.1 Etiology 294 is likely important in its etiology. In contrast, many
17.2.2 Clinical Presentation 294 cases of osteonecrosis are related to predisposing
17.2.3 Natural History and Treatment 296 systemie causes such as cortieosteroid usage, alcohol-
17.3 Imaging of Osteochondritis Dissecans 296 ism, collagen vascular disease and marrow packing
17.3.1 Radiography of OCD of the Femoral Condyles 296
17.3.2 Radiography of Less Common Sites of OCD 298
disorders such as siekle cell anemia.
17.3.3 Radionuclide Bone Scintigraphy 298 The age distribution is different with OCD, which
17.3.4 Computed Tomography 299 characteristically occurs in adolescence, although it
17.3.5 Arthrography 299 may first present in the third or fourth decade of life.
17.3.6 MR Imaging 300 Osteonecrosis is most common in the middle-aged
17.4 Osteonecrosis of the Knee 301
17.4.1 Etiology 301
and elderly, depending on the predisposing causes.
17.4.2 Clinical Presentation and Natural History 302 In view of the significant differences between these
17.5 Imaging of Osteonecrosis 302 two conditions, theywill be discussed separatelywith
17.5.1 Radiography 302 comparison and contrast as appropriate.
17.5.2 Radionuclide Scintigraphy and Computed
Tomography 303
17.5.3 MR Imaging 304
17.6 Summary 304
References 304 17.2
Osteochondritis Dissecans

Osteochondritis dissecans is a condition in which an

17.1
Introduction
Osteochondritis dissecans (OCD) and osteonecrosis
are two of the many types of pathology that may
affect the bony structures of the knee. They are sim-
ilar in at least three ways. Both lesions occur most
commonly in the femoral condyles when they involve
the knee. In advanced lesions, both OCD and osteo-
necrosis are manifested radiographieally by collapse
of the subchondral bone plate with formation of
intra-articular loose bodies. Finally, they may have
similar histology, with necrotic bone identified within
the involved regions.
A.A. DE SMET, MD
Professor, Department of Radiology - E3/311, University of
Wisconsin Hospital and Clinics, 600 Highland Avenue, Madi-
son, WI 53792, USA
A.M. Davies et al. (eds.), Imaging of the Knee
© Springer-Verlag Berlin Heidelberg 2003
osteochondral fragment shows radiographie sepa-
ration from the underlying bone. In some cases,
the fragment is firmly connected to the underlying
bone by a fibrous, cartilaginous, or fibrocartilaginous
union. Although not as firm as a bony union, this
may be asymptomatie. However, without bony union,
there is often motion of the fragment that gives rise
to pain with use of the joint.
In 1888, KOENIG was the first to use the term
"osteochondritis dissecans" (CLANTON and DELEE
1982; SCHENCK and GOODNIGHT 1996). He believed
that OCD is caused by inflammation of the bone and
cartilage with subsequent necrosis and dissection of
the fragment from the underlying bone (SCHENCK
and GOODNIGHT 1996).
Histological evaluation has shown that these are
not inflammatory lesions, as believed by KOENIG, but
the exact etiology is unproven. Although not etiolog-
ically correct, the term "osteochondritis dissecans"
is wen established and is still widely used. Currently
294
proposed etiologies for OCD inc1ude trauma, periph-
eral avascular necrosis, and congenital abnormalities
of ossification of the epiphysis (CLANTON and DELEE
1982; GARRETT 1991).
17.2.1
Etiology
Although many etiologies have been proposed, most
investigators believe that direct and indirect trauma
plays an important role in the formation of osteo-
chondritic lesions (SCHENCK and GOODNIGHT 1996;
PAPPAS 1981;FAIRBANK 1933;MILGRAM 1978;CAHILL
et al. 1989). There is a history of prior knee trauma
in 40%-60% of patients who have OCD of the knee
(CLANTON and DELEE 1982). OCD is more common-
ly seen in athletes than in those who do not partiei-
pate in organized sports (SCHENCK and GOODNIGHT
1996). OCD also occurs two to three times more com-
monly in males than females (CLANTON and DELEE
1982; SCHENCK and GOODNIGHT 1996; LINDEN 1976).
This male predominance has been used as evidence
of the traumatic etiology in view of the greater physi-
cal activity levels of boys than girls, as was the pat-
tern until recent years. In addition, the location of
these lesions on convex surfaces suggests a shearing
or impaction injury. The most common site for OCD
in the body is on the convex joint surfaces of the femo-
ral condyles (CLANTON and DELEE 1982).
Previous studies of the histological features of OCD
in the knee have not definitively identified an etiol-
ogy. The fragments may consist of articular cartilage
alone or with variable amounts of attached viable or
necrotic subchondral bone. Some biopsy studies of
the underlying bone found normal viable bone with
reparative granulation tissue, so at least in these cases,
the osteochondritic lesions were not due to peripheral
avascular necrosis (CHIROFF and COOKE 1975). How-
ever, others have noted ischemic changes in the base
of the lesions, causing them to believe that OCD rep-
resents «••• pathological fractures through subchondral
necrotic bone.. :' (CAMPBELL and RANAWAT 1966).
Based on his extensive research, MILGRAM (1978)
conc1uded that OCD is not a form of peripheral osteo-
necrosis. He analyzed 29 partially attached and 21
detached osteochondritic fragments, and found that
the foei of increased density seen on radiographs
within an OCD can be due to: (1) attached subchon-
dral bone, (2) caleification of degenerating cartilage,
(3) new bone formation, or (4) caleification of new
surface layers of cartilage and bone. Because of the
absence of attached bone in 24 fragments, he com-
A.A.De Smet
mented, «It is difficult to conceive how avascular necro-
sis ofbone could produce speeimens without attached
necrotic subchondral bone" (MILGRAM 1978).
In addition to trauma and avascular necrosis, some
authorities feel that OCD results from an abnor-
mality of endochondral ossification (SCHENCK and
GOODNIGHT 1996).A predisposition to OCD has been
described in several families but a familial assoeiation
is rare (PAES 1989; KOZLOWSKI and MIDDLETON 1985;
CLANTON and DELEE 1982; SCHENCK and GOOD-
NIGHT 1996; MUBARAK and CARROLL 1979,1981).
17.2.2
Clinical Presentation
The most common presenting symptom of patients
with OCD is vague joint pain. If the fragment is par-
tially or completely detached, there may be symp-
toms of catching, locking or swelling with use of the
affected knee (CLANTON and DELEE 1982; SCHENCK
and GOODNIGHT 1996). OCD of the knee presents
more commonly in childhood and early adolescence
than does OCD at other sites in the body. Femoral
condylar osteochondritic lesions are two to three
times as likely to be unilateral as bilateral (CLANTON
and DELEE 1982; SCHENCK and GOODNIGHT 1996).
Osteochondritis dissecans of the knee is divided
into juvenile and adult forms based on patient age at
presentation (CLANTON and DELEE 1982; SCHENCK
and GOODNIGHT 1996). In juvenile OCD, the lesion
first becomes symptomatic prior to the c10sure of the
adjacent growth plate (Fig. 17.1). In the adult form,
the adjacent growth plate is fused at the time of clini-
cal presentation (Fig.17.2). This distinction is impor-
tant for patient management. Juvenile lesions will
often heal with conservative treatment while patients
with adult femoral condylar OCD usually have per-
sistent knee dis ability unless their lesions are treated
surgically.
Although far less common than femoral condylar
lesions, patellar OCD is not a rare entity. A survey
of the literature in 1977 found 40 recorded cases
(EDWARDS and BENTLEY 1977). Since that time there
have been two series with 13 cases in one study and
12 in the other (DESAI et al. 1987; RENU et al. 1994).
Although speeific episodes of prior trauma have not
been assoeiated with patellar OCD, all 13 patients
in one series had a history of repeated trauma in
sports (DESAI et al. 1987). The most common pre-
senting symptom in these patients is the gradual
onset of vague anterior knee pain. Patients with
patellar OCD are more commonly male than female
a

Fig. 17.1a-c. Unstable juvenile osteochondritis of the medial femoral con-

dyle in a 16-year-old boy. a Anteroposterior radiograph shows a crescentic
lucent lesion with a sclerotic border filled with a central calcified density,
which begins in the notch and extends medially onto the weight-bearing
surface. b Because of its posterior location, the lesion is not weH seen on
the tunnel view but is manifested as an area of ill-defined radiolucency and
sclerosis (black arrowheads). C Sagittal spin echo T2-weighted MR image
shows a high signal intensity line (curved white arrow) at the margins of
c the lesion, indicating an unstable lesion

a ...._ _ _ _---.:- . . _....1 b

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
SCHENCK and GOODNIGHT 1996). The sclerosis can
vary in thickness from 1 to as much as 10 mm. The
lesion itself may be entirely radiolucent (Fig. 17.3),
may have one or more small fragments within the
lesion (Fig. 17.2), or may be almost completely filled by
a solitary bony fragment (Fig. 17.1). There is usually
disruption of the subchondral bone plate at the mar-
gins of the lesion unless the lesion has healed.
In children around 10 years of age, the major dif-
ferential diagnosis is normal variation in ossifica-
tion of the epiphysis. In some children, the epiphysis
will not ossify from a single enlarging ossification
center but will have islands of ossification. As the
child develops, one of the small islands of ossification
may persist at the articular surface and may mimic
an OCD. The normal variation differs from an OCD
in that there is no marginal sclerosis at the base of
an ossification center, while an osteochondritic frag-
ment usually has a rim of reactive sclerosis.
17.3.2
Radiography of Less Common Sites of OCD
Patellofemoral osteochondritic lesions are consider-
ably less common than those in the femoral condyles.
However, they have distinctive findings, which may
be easily overlooked without proper filming. Because
they occur on the sloping surfaces of the femoral
sulcus or the patellar facets, they are best visualized
on either the lateral view of the knee or tangential
views of the patellofemoral joint, such as the Laurin
view.
The patellar lesions appear as radiolucent defects
adjacent to the articular surface with thin well-
defined borders, which may or may not have central
bony fragments (Fig.17.5). They are most commonly
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
A.A.De Smet
present on the inferior half of the patella, especially
on the median ridge and medial facet, with a minor-
ity on the lateral facet (DESAI et al. 1987; RENU et al.
1994; EDWARDS and BENTLEY 1977). No lesions have
been noted on the odd facet or on the margins of the
patella.
The anterior femoral lesions are almost always
radiolucent with variable amounts of marginal sclero-
sis ranging from minimal to extensive (Fig.17.6). The
lesions may occur anywhere in the femoral sulcus,
with sites varying from the medial edge through the
central groove onto the lateral edge of the intercondy-
lar sulcus (CAYEA et al. 1981; KURZEWEIL et al.1988).
The tibial plateau and anterior tibial spine are
rare sites of OCD in the knee (BUI-MANSFIELD et al.
2000; SAGASTIBELZA et al. 1993; TOWBIN et al. 1982).
Although not many lesions have been reported at
these two sites, the radiographie appearance is usu-
ally a well-defined radiolucent area surrounded by
variable amounts of marginal sclerosis and often
containing a central ossified fragment.
17.3.3
Radionuclide Bone Scintigraphy
Radionuclide bone scintigraphy is highly sensitive
in the detection of femoral osteochondritic lesions
(LOOMER et al. 1993; Ly and FALLAT 1993; CAHILL
and BERG 1983). The typical appearance is an intense
focus of radionuclide uptake in a femoral condyle
that extends down to the subchondral bone plate.
MESGARZADEH et al. (l987) found that the presence
of hyperemia on the blood pool phase was 100% sen-
sitive and 83% specific in identifying unstable OCD
lesions and that intense uptake on the delayed images
was 90% sensitive and 100% specific. CAHILL and
 Osteochondritis and Osteonecrosis
,
b
BERG (1983) found that radionuclide bone scintigra-
phy performed every 6 weeks was useful in predicting
the potential for healing of OCD. If serial bone scin-
tigrams show continued or increasing high levels of
uptake, then the site is undergoing constant remodel-
ing and the prognosis for healing with conservative
management is good (CAHILL and BERG 1983). Serial
radionuclide scanning has not received widespread
application because of the radiation exposure to the
patient. Most physicians prescribe reduced weight
bearing in patients with juvenile OCD of the knee
and use serial radiographs and change in symptoms
to assess healing.
17.3.4
Computed Tomography
Computed tomography (CT) will occasionally detect
alesion which is occult on radiographs and can be
used for preoperative staging by determining the
exact size of the lesion and the degree of fragmen-
299
Fig. 17.6a-c. Stable anterior lateral femoral condyle OCD. a
Axial radiograph shows a subtle radiolucency (black arrow-
head) of the lateral femoral condyle. b Sagittal fast spin echo
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
tation (HOWIE 1985; HOLLAND and DAVIES 1994;
LOOMER et al. 1993; Ly and FALLAT 1993). In my own
practice, CT is often used to assess the size of osse-
ous fragments within the lesion (Fig. 17.7). Determi-
nation of osseous fragment size allows surgical plan-
ning for the method of stabilization of the fragment.
Larger fragments are amenable to screw fixation while
smaller fragments cannot be internally fixed.
17.3.5
Arthrography
Early reports suggested that arthrography would be
a useful technique to assess the stability of the lesion
and the integrity of the overlying articular cartilage
(ALMGARD and WIKSTAD 1964). In theory, contrast
material injected into the joint should pass beneath
the fragment. Visualization of radio-opaque contrast
beneath the lesion is proof of an articular cartilage
defect and confirms an unstable lesion. However,
because these lesions commonly have granulation
 300
a
Fig. 17.7a, b. OCD of the medial femoral condyle. Coronal (a)
and sagittal (b) reformatted cr images show an entirely radio-
lucent lesion without central calcified fragments that might be
amenable to internal fixation
tissue beneath them, the amount of contrast that
passes beneath the lesion is often small and hard to
detect by radiographs. For this reason, conventional
arthrography is not very useful for assessing lesion
stability.
Computed arthrotomography is a useful modifi-
cation of conventional arthrography. With its high
resolution, cross-sectional imaging and sensitivity
to tissue contrast differences, CT after intra-articu-
lar contrast injection allows excellent definition of
the articular cartilage (HOLLAND and DAVIES 1994;
PAILLE et al. 1988). However, although CT arthrog-
raphy should be an accurate way to determine the
integrity of the articular cartilage over an OCD, few
studies have compared its accuracy with operative
findings of stability or instability.
A more recent modification of knee arthrography
has been the use of magnetic resonance (MR) imag-
A. A. De Smet
ing after intra-articular injection of gadolinium
(KRAMER et al. 1992). In one study, MR arthrogra-
phy of the knee was correct in differentiating intact
from partially or completed detached osteochondrit-
ic lesions in 24 of 25 lesions (KRAM ER et al. 1992).
Animal models have also shown that MR arthrogra-
phy is a reliable way of detecting cartilage defects and
fractures in femoral OCD (KRAMER et al. 1992).
17.3.6
MRlmaging
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
make identification of instability difficult. T2-weight-
ed imaging has been the most useful in assessment
of instability of an osteochondritic lesion as detec-
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
intensity joint fluid allows assessment of defects in
the articular cartilage.
In a study performed using an animal model for
OCD, Tl-weighted imaging after intravenous injection
of gadolinium was as helpful as T2-weighted imaging
in identifying whether a fragment was stable or unsta-
ble (ADAM et al. 1991). This same study provided his-
tological confirrnation that the interface between an
unstable lesion and the underlying femur consisted of
vascularized granulation tissue (ADAM et al. 1991).
My current protocol for assessment of femoral
OCD is fast spin echo T2-weighted imaging with fat
saturation in the sagittal and coronal planes using a
field ofview of 14 cm, matrix of 256+ 192, two excita-
tions, an echo train length of 8, a slice thickness of
3 mm, a TR of 3,000 and a TE of 80eff. This is com-
bined with routine coronal Tl-weighted and sagittal
proton density-weighted images to assess the menis-
Osteochondritis and Osteonecrosis
ci as patients with OCD are usually athletically active
and their knee pain can be due to a meniscal tear
rather than an unstable OCD lesion.
The criteria for diagnosing instability of an OCD
lesion have been well described (DE SMET et al. 1990,
1996; MESGARZADEH et al. 1987). Alesion is unstable
by MR criteria if: (l) there is a high signal intensity
line representing granulation tissue or fluid beneath
the deep portion of the lesion, (2) if there is a cyst of
more than 5 mm in width beneath the lesion, (3) if
there is a focal articular defect, or (4) if fluid is seen
passing through an articular cartilage fracture (DE
SMET et al.1996). If one or more of these is present, the
lesion can be diagnosed as being unstable (Figs. 17.1,
17.2,17.5). Using these criteria,MR imagingwas found
to be 97% sensitive and 100% specific for diagnosing
the instability of OCD lesions in 31 patients with femo-
ral condylar lesions and nine patients with talar dome
lesions (DE SMET et al. 1996). In this same study, the
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
(DE SMET et al.1996).
If the lesion itself has signal within it, but none of
the four signs are present, then the lesion should be
diagnosed by MR imaging as being stable (Fig. 17.6)
There have been only two reported cases in which a
high signal intensity line was present beneath lesions
that were surgically proven to be stable (MESGAR-
ZADEH et al. 1987; DE SMET et al. 1990). Fast spin
echo T2-weighted imaging with fat saturation should
be an even more sensitive method than convention-
al spin echo T2-weighted imaging in determining
lesion instability, but there have been no reported
studies to confirm its accuracy.
Although MR imaging is accurate in the assess-
ment of whether or not alesion is unstable, it is not
definitive for assessment of the integrity of the artic-
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.
1990). Although instability may in itself be an indi-
cation for surgery in a child whose lesion has not
responded to conservative treatment, some surgeons
wish to prove that the cartilage is fractured before
proceeding to surgery. In this scenario, MR arthrog-
raphy provides the best assessment of articular car-
tilage. One study reported an accuracy of 100% in
assessing the status of the articular cartilage over
OCD in 25 knees (KRAMER et al. 1992).
Magnetic resonance imaging can also be useful
in predicting patient outcome when an OCD is treat-
301
ed conservatively. Arecent study reported the long-
term follow-up of 14 patients treated conservatively
for 1.5-8.5 years. Both patients with stable lesions as
diagnosed by MR imaging had a good outcome while
10 of the 12 patients whose lesions were unstable by
MR imaging had poor outcomes with continued knee
disability or required surgery (DE SMET et al. 1997).
Magnetic resonance imaging also appears to be
effective in confirming the healing of surgically treat-
ed OCD. Fifty-six patients with knee OCD treated
surgically by retrograde drilling were followed post-
operatively by physical examination, radiographs,
and MR imaging (SCHNEIDER et al. 1998). MR imag-
ing was superior to radiographs for predicting heal-
ing of the lesions (SCHNEIDER et al. 1998). One case
report noted that MR imaging could be used to assess
healing of OCD even when absorbable pins had been
used for fixation of the osteochondritic fragment
(SMITH et al. 1994).
17.4
Osteonecrosis of the Knee
Osteonecrosis of the knee is similar to OCD in that its
etiology is unknown when it is a primary condition.
However, unlike OCD, where there is variable histol-
ogy, osteonecrosis by definition always shows necrosis
on histological examination. In its earliest stages, the
necrosis involves primarily bone marrow, as this is the
tissue most sensitive to ischemia (MUNK and VELLET
1993). With continued ischemia, there is death ofbone
with necrosis of osteocytes. Being relatively resistant
to ischemia, the fat within the medullary cavity is the
last tissue to undergo necrosis. If a predisposing cause
is known, the treating physician will often suspect
osteonecrosis when the patient presents with knee
pain even before any imaging is performed.
17.4.1
Etiology
Osteonecrosis may be considered to be of two types,
idiopathic and secondary. Although vascular injury
and trauma have both been proposed as causes for
idiopathic osteonecrosis of the knee, the exact etiol-
ogy is unknown (LOTKE and ECKER 1988). Those who
feel that it is caused by vascular injury believe that
there is occlusion of the microcirculation, while those
who favor a traumatic etiology speculate that sub-
chondral microfractures in osteopenic bone cause
302
bleeding with increased marrow pressure and then
ischemia (LOTKE and ECKER 1988).
Secondary osteonecrosis has many possible causes
and is the most common type of osteonecrosis of
the knee. The list of secondary causes of osteone-
crosis is extensive and includes corticosteroid usage,
marrow packing dis orders such as sickle ceH anemia
and Gaucher's disease, alcoholism, gout, systemic
lupus erythematosus, dysbaric osteonecrosis due to
undersea diving and working in caissons, femoral
neck fracture and femoral head dislocation (CRUESS
1986). Corticosteroids are used to treat many condi-
tions such as autoimmune disease and marrow-based
malignancies like lymphoma. Unfortunately, osteo-
necrosis is not infrequently a complication of either
acute or chronic moderate to high dosage corti-
costeroid treatment. In my own practice, the most
common cause of osteonecrosis is the use of cortico-
steroids to suppress transplant rejection in patients
who have had organ transplants (Fig. 17.8). Recently
osteonecrosis of the femoral condyles and tibial pla-
teau has been identified after the use of a laser for
meniscectomy (JANZEN et al. 1997).
Fig. 17.8. A 22-year-old woman on corticosteroids for a liver
transplant. Anteroposterior radiograph shows collapse of the
articular surface of the lateral femoral eondyle from osteone-
erosis
17.4.2
Clinical Presentation and Natural History
The most common form of idiopathic osteonecrosis
of the knee is spontaneous osteonecrosis (LOTKE and
ECKER 1988). This is seen in the elderly and is much
more common in the medial femoral condyle than
in the lateral femoral condyle or the tibial condyles
(LOTKE and ECKER 1988). Patients usually present with
sudden onset of severe pain while doing a minor phys-
A.A.De Smet
ical activity such as walking or climbing stairs. The
pain can be quite disabling and collapse of the articu-
lar surface is an unfortunate common sequela. Spon-
taneous osteonecrosis is three times more common in
women than in men (LOTKE and ECKER 1988).
An association has been noted between sponta-
neous osteonecrosis of the knee and meniscal tears
(NORMAN and BAKER 1978).Since manypatients with
spontaneous osteonecrosis have collapse of the artic-
ular surface, it is not certain whether the meniscal
tear has preceded the osteonecrosis or has occurred
secondary to the articular incongruity.
The most common presentation for both idiopath-
ic and secondary osteonecrosis is pain. The pain is
likely due to the increased pressure from a compart-
ment syndrome, which occurs when edema and hem-
orrhage develop within the closed space of the medul-
lary canal (CRUESS 1986). During the early stages, the
radiographs are normal and diagnosis of the cause
of the patient's pain requires either radionuclide scin-
tigraphy or MR imaging (CRUESS 1986; LOTKE and
ECKER 1988). Although the pain can be severe, there is
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,
KOPECKY et al. (1991) found osteonecrosis in 25 hips
yet only seven of these hips became symptomatic.
The MR examinations returned to normal in 6 of the
18 asymptomatic hips.
17.5
Imaging of Osteonecrosis
17.5.1
Radiography
In contrast to OCD of the knee, in which symptom-
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

graphs of patients with pain from early osteonecrosis 17.5.2

are normal (CONKLlN et al. 1983). Only as the disease Radionuclide Scintigraphy and Computed
becomes more advanced do the radiographs show Tomography
abnormalities. Just as for OCD, the most common site
for osteonecrosis of the knee is the femoral condyles In view of its low cost, widespread availability, and
(Fig.17.8). The frequencywith whieh a site is involved high sensitivity, radionuclide scintigraphy is an excel-
depends on the underlying etiology. In one study of lent way to screen for early osteonecrosis which
46 knees affected with osteonecrosis secondary to cor- is radiographically occult (CONKLlN et al. 1983;
tieosteroid usage or alcoholism, the lateral condyle D' AMBROSIA et al. 1978). In one study of 40 knees
alone was affected in 20 knees, with both condyles in 31 patients with spontaneous osteonecrosis, the
affected in 23 (SAKAI et al. 1998). In this same study, scintigram was positive in 11 knees with normal
29lesions were in the posterior third of a condyle only radiographs. An added advantage of scintigraphy is
and 14 involved both the middle and posterior thirds. the ability to survey the entire skeleton for multiple
In another study in which most of the patients had sites of osteonecrosis, such as can occur in the sec-
spontaneous osteonecrosis, almost alilesions involved ondary forms. The accuracy of scintigraphy and
the medial femoral condyle (Fig.17.9) (MOTOHASHI et MR imaging is comparable. One prospective study
al.1991). of corticosteroid-treated renal transplant patients
The earliest appearance of femoral condylar osteo- showed significant agreement between scintigraphy
necrosis is a subtle lucencywithin the condyle extend- and MR imaging for detection of femoral head osteo-
ing down to the subchondral bone plate. As the necrosis (SIDDIQUI et al. 1993). Only a few cases were
lesion matures, sclerosis begins in the margins and seen only on scintigraphy or only on MR imaging
the lesion becomes more evident radiographieally (SIDDIQUI et al. 1993).
(MARTEL and SITTERLEY 1969). If the forces of The appearance of osteonecrosis on delayed scin-
weight-bearing exceed the mechanieal strength of tigraphie imaging is characteristieally that of intense
the weakened necrotic area, there is collapse of the radionuclide uptake centered within a femoral or
articular surface and the lesion is readily evident with tibial condyle (GREYSON et al. 1982). While this pat-
fragmentation of the bone. These fragments may be tern is not specific for osteonecrosis, it may be suf-
contained within the lesion or may become osteocar- ficiently diagnostie when the clinieal scenario is typi-
tilaginous loose bodies. cal for osteonecrosis. In the middle-aged and elderly,
In contrast to OCD, where tibial lesions are rare, who are the most commonly affected by spontaneous
spontaneous osteonecrosis of the knee is not rare in osteonecrosis, localized uptake in a femoral condyle
the tibial condyles. The radiographie appearance is mayaiso be due to severe degenerative joint disease
similar to osteonecrosis in the condyles, with progres- or primary or metastatie neoplasm. Correlation with
sion from subtle radiolucency to mixed radiolucency radiographs usually allows differentiation between
and sclerosis and finally to subchondral collapse. these conditions and avascular necrosis.
304
If a three-phase bone scintigram is performed,
there is increased blood flow at the site of acute
osteonecrosis with intense uptake within the lesion
on the blood pool phase (GREYSON et al. 1982).As the
lesion matures, the hyperemia and increased radio-
nuclide activity on the blood pool phase are no longer
seen (GREYSON et al. 1982).
When the radionuclide uptake occurs in the tibial
condyles, an additional differential diagnosis in the
elderly would be a tibial insufficiency fracture. Insuf-
ficiency fractures usually have a linear pattern of
radionuclide uptake, which allows differentiation
from the globular uptake pattern of osteonecrosis.
Although the initial radiographs may be negative in
the early stages of a tibial insufficiency fracture, serial
films will show a characteristic linear pattern of scle-
rosis in the tibial condyle paralleling the adjacent
tibial plateau.
Computed tomography has litde value in the initial
diagnosis of osteonecrosis of the knee. If the lesion is
not visible by radiographs, it is only occasionally vis-
ible by CT. CT may be useful in determining the exact
amount of the bone loss for surgical planning prior to
total knee arthroplasty, as has been shown in assess-
ment of osteonecrosis of the femoral heads.
17.5.3
MRlmaging
Magnetic resonance imaging has become the defini-
tive method for assessment of occult avascular necro-
sis because it offers both high sensitivity and high
specificity.lts major limitations when compared with
radionuclide scintigraphy are its greater cost and the
inability to screen multiple areas of the body with
one examination. The cost difference between MR
imaging and scintigraphy can be reduced by the use
of rapid MR screening protocols.Although the effica-
cy of a rapid screening protocol has not been proven
for the knee, use of only coronal Tl-weighted images
has been shown to be sensitive for detection of osteo-
necrosis of the femoral heads in transplant patients
(TERVONEN et al. 1992; MULLIKEN et al. 1994). It is
likely that coronal Tl-weighted images would be as
cost-effective for the detection of osteonecrosis in the
knees as it is in the hips.
The MR appearance of osteonecrosis of the femo-
ral and tibial condyles varies depending on the acute-
ness of the lesion. Early on, the lesions may be man-
ifested by diffuse edema, which is visualized as an
ill-defined large area of low signal intensity on Tl-
weighted images and high signal intensity on T2-
A.A.De Smet
weighted images (Fig. 17.9). This is not a specific
appearance and can also be seen in osteomyelitis and
bone bruises from trauma. The clinical presentation
will usually allow easily discrimination with fever
and systemic symptoms in osteomyelitis and a his-
tory of re cent trauma for bone bruises. In these cases,
it is also important to compare the radiographs with
the MR images. Elderly patients with tibial plateau
insufficiency fractures can present clinically with
spontaneous medial pain that may resemble the pain
in spontaneous osteonecrosis. However, MR imaging
allows definite diagnosis of an insufficiency fracture,
which shows a linear pattern of abnormal signal
intensity that is different from the globular pattern of
osteonecrosis (LE GARS et al. 1996).
However, most cases of avascular necrosis have a
characteristic appearance once they reach the chronic
phase. The resistance of the fatty marrow to ischemia
results in lesions that on Tl-weighted MR images
have the central high signal intensity of fat and a
thin well-defined low signal intensity border. This
low signal intensity border may remain of low signal
intensity on T2-weighted images, reflecting sclerosis
or fibrosis, or may have a high signal intensity due
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
signal intensity on T2-weighted images.
17.6
Summary
Osteochondritis dissecans and osteonecrosis are two
interesting conditions affecting the articular surface
of the knee. Although they may have similar histolog-
ical and radiographie features, they are distincdy dif-
ferent entities. OCD characteristically affects young-
er individuals in the second and third decades of life,
while osteonecrosis often affects the middle aged and
elderly. The MR appearance is markedly different,
with a focal, ellipticallesion adjacent to the joint sur-
face in OCD and a large oval area of edema within a
femoral condyle in osteonecrosis.
References
Adam G, Buhne M, Prescher A et al (1991) Stability of osteo-
chondral fragments of the femoral condyle: magnetie res-
onance imaging with histopathologie correlation in an
animal model. Skeletal Radio120:601-606
Osteochondritis and Osteonecrosis
Aiehroth P (1971) Osteochondritis dissecans of the knee: a
clinieal survey. J Bone Joint Surg 53B(3):440-447
Almgard LE, Wikstad I (1964) Late results of surgery for osteo-
chondritis dissecans of the knee joint. Acta Chir Scand
127:588
Bui-Mansfie!d LT, Kline M, Chew FS et al (2000) Osteochondri-
tis dissecans of the tibial plafond: imaging characteristies
and a review of the literature. AJR 175:1305-1308
Cahill BR (1985) Treatment of juvenile osteochondritis disse-
cans and osteochondritis dissecans of the knee. Clin Sports
Med4:67
Cahill BR, Berg BC (1983) 99 m -Technetium phosphate com-
pound joint scintigraphy in the management of juvenile
osteochondritis dissecans of the femoral condyles. Am J
Sports Med 11(5):329-335
Cahill BR, Phillips MR, Navarro R et al (1989) The results of
conservative management of juvenile osteochondritis dis-
secans using joint scintigraphy: a prospective study. Am J
Sports Med 17:601
Campbell CJ, Ranawat CS (1966) Osteochondritis dissecans:
the question of etiology. J Trauma 6(2):201-221
Cayea PD, Pavlov H, Sherman MF et al (1981) Lucent artieular
lesion in the lateral femoral condyle: source of patellar fem-
oral pain in the athletie adolescent.AJR 137:1145-1149
Chiroff RT, Cooke CP III (1975) Osteochondritis dissecans:
a histologie and mieroradiographie analysis of surgieally
excised lesions. J Trauma 15(8):689-696
Clanton TO, DeLee JC (1982) Osteochondritis dissecans: his-
tory, pathophysiology and current treatment concepts. Clin
Orthop 167:50-64
Conklin JJ, Alderson PO, Zizie TM et al (1983) Comparison
of bone scan and radiograph sensitivity in the detection
of steroid-induced ischemie necrosis of bone. Radiology
147:221-226
Cruess RL (1986) Osteonecrosis of bone: current concepts as
to etiology and pathogenesis. Clin Orthop 208:30-39
D'Ambrosia RD, Shoji H, Riggins RS et al (1978) Scintigraphy
in the diagnosis of osteonecrosis. Clin Orthop 130:139-143
Desai SS, Pate! MR, Miehelli LJ et al (1987) Osteochondritis
dissecans of the patella. J Bone Joint Surg [Br]
69B(2):320-325
De Smet AA, Fisher DR, Graf BK et al (1990) Osteochondritis
dissecans of the knee: value of MR imaging in determin-
ing lesion stability and the presence of artieular cartilage
defects. AJR 155:549-553
De Smet AA, Ilahi OA, Graf BK et al (1996) Reassessment of
the MR criteria for stability of osteochondritis dissecans in
the knee and ankle. Skeletal RadioI25:159-163
De Smet AA, Ilahi OA, Graf BK et al (1997) Untreated osteo-
chondritis dissecans of the femoral condyles: predietion of
patient outcome using radiographie and MR findings. Skel-
etal Radiol 26:463-467
Dipaola JD, Nelson DW, Colville MR et al (1991) MR: character-
izing osteochondrallesions by magnetie resonance imag-
ing. Arthroscopy 7:101
Edwards DH, and Bentley G (1977) Osteochondritis dissecans
patellae. J Bone Joint Surg [Br] 59B(1):58-63
Engel A, Hajek P, Kramer J et al (1990) Magnetie resonance knee
arthrography. Acta Orthop Scand 61 [SuppI240]:33-43
Fairbank HAT (1933) Osteochondritis dissecans. Br J Surg
21:67-82
Garrett JC (1991) Osteochondritis dissecans. Clin Sports Med
10(3):569-593
305
Garrett JC, Kress KJ, Mudano M (1992) Osteochondritis disse-
cans of the lateral femoral condyle in the adult. Arthroscopy
8(4):474-481
Greyson ND, Lotem MM, Gross AE et al (1982) Radionuclide
evaluation of spontaneous femoral osteonecrosis. Radiol-
ogy 142:729-735
Holland P, Davies AM (1994) Computed tomographie arthrog-
raphy in the assessment of osteochondritis dissecans of the
elbow. Clin RadioI49:231-235
Howie JL (1985) Computed tomography in osteochondritis
dissecans of the patella. J Can Assoc Radiol 36: 197-199
Hughston JC, Hergenroeder PT, Courtenay BG et al (1984)
Osteochondritis dissecans of the femoral condyles. J Bone
Joint Surg 66A(9):1340-1348
Janzen DL, Kosarek FJ,Helms CA et al (1997) Osteonecrosis after
contact neodymium: yttrium aluminum gamet arthroscop-
ie laser meniscectomy. AJR 169:855-858
Kopecky KK, Braunstein EM, Brandt KD et al (1991) Apparent
avascular necrosis of the hip: appearance and spontaneous
resolution of MR findings in ranal allograft recipients.
Radiology 179:523-527
Kozlowski K,Middleton R (1985) Familial osteochondritis dis-
secans: a dysplasia of artieular cartilage? Skeletal Radiol
13:207-210
Kramer J,Stiglbauer R,Enge!A et al (1992) MR contrast arthrog-
raphy (MRA) in osteochondrosis dissecans. J Comput Assist
Tomogr 16(2):254-260
Kurzeweil PR, Zambetti GJ, Hamilton WG et al (1988) Osteo-
chondritis dissecans in the lateral patellofemoral groove.
Am J Sports Med 16(3):308-310
Le Gars LL, Orce! P, Liote F et al (1996) Osteonecrosis-like syn-
drome of the medial tibial plateau can be due to a stress
fracture. Rev Rheum 66( 6):323-330
Linden B (1976) The incidence of osteochondritis dissecans in
the condyles of the femur. Acta Orthop Scand 47:664-667
Linden B, Telhag H (1977) Osteochondritis dissecans: a his-
tologie and autoradiographic study in man. Acta Orthop
Scand 48:682-686
Loomer R, Fisher C, Lloyd-Smith R et al (1993) Osteochondral
lesions of the talus. Am J Sport Med 21: 13
Lotke PA, Ecker ML (1988) Current concepts review osteone-
crosis of the knee. J Bone Joint Surg 70A(3):470-474
Ly PN and Fallat LM (1993) Transchondral fractures of the
talus: a review of 64 surgieal cases. J Foot Ankle Surg
32:352
Martel W, Sitterley BH (1969) Roentgenologic manifestations
of osteonecrosis. AJR 106(3):509-522
McGill JJ, Demos TC, Lomasney LM et al (1995) Osteo-
chondritis dissecans: imaging modalities. Orthopedics
18(12):1180-1185
Mesgarzadeh M, Sapega AA, Bonakdarpour A et al (1987)
Osteochondritis dissecans: analysis of mechanical stability
with radiography, scintigraphy, and MR imaging. Radiol-
ogy 165:775-780
Milgram JW (1978) Radiologieal and pathologieal manifesta-
tions of osteochondritis dissecans of the distal femur. Radi-
ology 126:305-311
Motohashi M, Morii, T, Koshino T et al (1991) Clinical course
and roentgenographic changes of osteonecrosis in the
femoral condyle under conservative treatment. Pro gnosis
Medial Gonarthrosis 266:156-161
Mubarak SJ, Carroll NC (1979) Familial osteochondritis dis-
secans of the knee. Clin Orthop 140: 131-136
306
Mubarak sJ, Carroll NC (1981) Juvenile osteochondritis dis-
secans of the knee: etiology. Clin Orthop 157:200-211
Mulliken BD, Renfrew DL, Brand RA et al (1994) Prevalence of
previously undetected osteonecrosis of the femoral head in
renal transplant recipients. Radiology 192:831-834
Munk PL, Vellet AD (1993) Lesions of cartilage and bone
around the knee. Top Magn Reson Imaging 5(4):249-262
Norman A, Baker ND (1978) Spontaneous osteonecrosis of the
knee and medial meniscal tears. Radiology 129:653-656
Paes RA (1989) Familial osteochondritis dissecans. Clin Radiol
40:501-504
Paille P, Quesnel C, Baunin C et al (1988) Computed arthrog-
raphy: its role in the screening of joint diseases in pediatric
radiology. Pediatr RadioI18:386-390
Pappas AM (1981) Osteochondrosis dissecans. Clin Orthop
158:59-69
Pfeiffer WH, Gross ML, Seeger LL et al (1991) Osteochondritis
dissecans of the patella: MRI evaluation and a case report.
Clin Orthop 271:207-211
Renu JM, Bou CV, Portet RV et al (1994) Osteochondritis dis-
secans of the patella: 12 cases followed for 4 years. Acta
Orthop Scand 65(1):77-79
Sagastibelza J, Zuniga JJ, Blasco JJ et al (1993) Osteochondritis
dissecans of the anterior tibial spine. Arthroscopy
9(6):695-696
Sakai T, Sugano N, Ohzono K et al (1998) MRI evaluation of ste-
A.A.De Smet
roid- or alcohol-related osteonecrosis of the femoral con-
dyle. Acta Orthop Scand 69( 6):598-602
Schenck RC, Goodnight JM (1996) Current concepts review
osteochondritis dissecans. J Bone Joint Surg
78A(3):439-456
Schneider T, Fink B, Jerosch J et al (1998) The value of mag-
netic resonance imaging as postoperative control after
arthroscopic treatment of osteochondritis dissecans. Arch
Orthop Trauma Surg 117:235-239
Siddiqui AR, Kopecky KK, Wellman HN et al (1993) Pro-
spective study of magnetic resonance imaging and SPECT
bone scans in renal allograft recipients: evidence for a
self-limited subclinical abnormality of the hip. J Nucl Med
34:381-386
Smith DS, Sharp DC, Resendes M et al (1994) MRI of healing
osteochondritis dissecans fragment with absorbable pins. J
Comput Assist Tomogr 18(5):832-833
Tervonen 0, Mueller DM, Matteson EL et al (1992) Clinically
occult avascular necrosis of the hip: prevalence in an
asymptomatic population at risk. Radiology 182:845-847
Towbin J, Towbin R, Crawford A et al (1982) Osteochondritis
dissecans of the tibial plateau. J Bone Joint Surg [Am]
64A(5):783-784
Twyman RS,Desai K,Aichroth PM et al (1991) Osteochondritis
dissecans of the knee: a long-term study. J Bone Joint Surg
[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

"From DORFMAN and CZERNIAK (1995)

18.1 18.2
Introduction Detection

Bone sarcomas are uncommon when compared with
other malignancies, accounting for only 0.2% of all
tumours (DORFMAN and CZERNIAK 1995). Their
incidence is approximately one-tenth that of soft
tissue sarcomas (MAcK 1995) and one-sixtieth that
of either lung or breast carcinoma. The annual inci-
den ce for bone sarcomas is approximatelyO.8/1 00,000
(DORFMAN and CZERNIAK 1995). The subject is par-
ticularly pertinent when dealing with the knee, as
A.M. DAVIES, MD
Consultant Radiologist, MRI Centre, Royal Orthopaedic Hos-
pital, Bristol Road, Birmingham, B31 2AP, UK
D. VANEL,MD
Department of Radiology, Institut Gustav Roussy, 39 rue
Camille Desmoulins, Villejuif 94895, France
The majority of patients with a bone or soft tissue
tumour will present with pain and/or swelling. Alter-
natively, a pathologieal fracture may be the precipi-
tous presenting feature. Occasionally, a bone tumour,
typieally a benign lesion, can be an incidental radio-
graphie finding. Despite newer imaging techniques,
the radio graph is the preliminary and single most
important imaging investigation. Frequently, the diag-
nosis may be obvious to the trained eye, and further
imaging, if required, is then directed towards staging
the lesion. Alternatively, if an abnormality is present
on the film and the precise nature is not immediately
apparent, certain features will indieate a differential
diagnosis and other forms of imaging can then be
employed to assist in establishing a more definitive

A.M. Davies et al. (eds.), Imaging of the Knee

© Springer-Verlag Berlin Heidelberg 2003
308
radiological diagnosis. If the initial radiograph is
normal, however, with persisting and increasing
symptoms, a repeat radio graph may be indicated in
due course.
Early signs of a bone tumour, or for that matter
infection, include subtle areas of ill-defined lysis or
sclerosis, cortical destruction, periosteal new bone
formation and soft tissue swelling (Fig.18.I) (ROSEN-
BERG et al. 1995). Not surprisingly, bone lesions are
frequently missed or overlooked on the initial radio-
graph. In an audit performed at one of the authors'
institutions (A.M.D.), in approximately 20% of cases
neither the clinician nor the radiologist at the refer-
ring centre detected the bone tumour on the initial
radiographs, although evidence was present on ret-
rospective review of the films (GRIMER and SNEATH
1990). This is a bigger problem with tumours of flat
bones, such as the pelvis, than around the knee joint.
The pathological process may be well established
even in the presence of anormal radio graph. At
least 40%-50% of trabecular bone must be destroyed
before a discrete area of lucency can be discerned
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-
cal to medullary bone, the easier it will be to detect
an abnormality on the radiographs. Therefore, were
three tumours of similar size to arise around the
knee joint, it is fair to assurne that the lesion affect-
Fig. 18.1. AP radiograph of the knee in a 10-year-old boy
showing an early lytic osteosarcoma of the proximal tibial
metaphysis
A. M. Davies and D. Vanel
ing the smallest bone, the proximal fibula, would be
radiographically more conspicuous than those in the
femur or tibia. Beware, however, the patella. Although
it is a rare site, tumours of the patella are easily over-
looked as the bone is projected over the distal femur
on the anteroposterior projection.
In the presence of anormal radiograph, referred
pain needs to be considered. Hip joint pathology pre-
senting with referred pain to the knee is a well-recog-
nised entity in the child. If referred pain is suspected,
then radiographs of the pelvis and proximal femur
are indicated. On occasion, radiographically occult
lesions may be detected by bone scintigraphy and/or
magnetic resonance (MR) imaging. It is important to
stress that, owing to its high sensitivity, MR imaging
all too frequently reveals abnormalities of little or no
clinical significance. An increasing number of knee
MR scans are performed each year for a variety of
conditions. Incidental medullary abnormalities will
be revealed in the distal femoral meta-diaphysis in
a small percentage of cases which alm ost invariably
prove to be innocuous chondromas (Fig.18.2).
18.3
Diagnosis
18.3.1
Diagnosis of Bone Tumours
Once a skeletal abnormality has been detected around
the knee, the next objective of imaging is to attempt
to characterise the lesion and, in doing so, to indicate
an appropriate differential diagnosis to the referring
clinician. At this stage, important maxims that should
be appreciated include not over-treating a benign
lesion, not under-treating a malignant lesion and not
misdirecting the approach to biopsy, which might
prejudice subsequent surgical management (MOSER
and MADEWELL 1987). Before assessing the imaging,
the prudent radiologist should establish some basic
facts regarding the patient. By recognising the rele-
vance of certain clinical details an extensive differ-
ential diagnosis may be significantly reduced even
before the imaging is considered. Important factors
to be considered include:
1. Age. The age of the patient is arguably the single
most useful piece of information as it frequently
influences the differential diagnosis. Many mus-
culoskeletal neoplasms exhibit a peak incidence
at different ages. For osteosarcoma this is in the
second and third decades (Fig. 18.I). Metastases
 Tumours and Tumour-like Lesions 309

a b

Fig. l8.2a, b. Enchondroma of the distal

femoral diaphysis that was initially
an incidental finding on MR imaging.
a Lateral radiograph; b sagittal Tl-
weighted image

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

RIlUND CUlLES IONS

CORJICAl flBRauS DYSPlASIA\ EWING SARCOMA
ADAIMNTINOMA J RETICUlUM Cut SARCOMA
MVElOMA

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~~:~~-------

Fig. 18.5. A composite diagram of the sites of origin oE primary

bone tumours. (From MADEWELL et al. 1981, with permission)

Fig. 18.4. Lateral radiograph showing a peripheral chondro-

sarcoma in a patient with diaphyseal aclasis

What is the tumour doing to the bone (pattern of

destruction)? What form of periosteal re action, is
any, is present? What type of matrix mineralisation, if
any, is present?

Site in Skeleton. Most bone tumours and infections

occur around the knee and as such litde diagnostic
information can be deduced from noting the affect-
ed bone in most cases. There are exceptions. Most
tumours arising in the patella are benign. The majori-
ty of tumours involving the proximal fibula are aneu-
rysmal bone cysts (ABCs),giant ceH tumours (GCTs),
osteosarcoma and Ewing's sarcoma.

Loeation in Bone. The site of origin of a bone tumour

is an important parameter of diagnosis (Fig. 18.5)
(MADEwELL et al. 1981). It refiects the site of greatest
ceHular activity. During the adolescent growth spurt
the most active areas are the metaphyses around the
knee and in the proximal humerus. Tumours origi-
nating from marrow ceHs may occur anywhere along
the bone. Conventional osteosarcoma will tend to Fig. 18.6. AP radiograph
originate in the metaphysis or meta-diaphysis, while oE a Ewing's sarcoma of
Ewing's sarcoma tends to arise in the metaphysis or, the proximal fibular dia-
more distinctively, in the diaphysis (Fig. 18.6). In the physis. Typical features
child the differential diagnosis of alesion arising include the ill-defined
bone destruction,
within an epiphysis can be realistieally limited to lamellar periosteal reac-
chondroblastoma (Fig.18.7),epiphyseal abscess (pyo- tion and Codman angles
genie or tuberculous) and, rarely, eosinophilic granu- medially
 Tumours and Tumour-like Lesions
Fig.18.7. a AP radiograph in a 14-year-old boy with a chondroblastoma arising in the
a femoral epiphysis (arrowheads). b The matrix mineralisation is easily seen on CT
loma. Following skeletal fusion, subarticular lesions,
analogous in the adult to the epiphyseallesions,
include GCT, clear cell chondrosarcoma (rare) and
intra-osseous ganglion. Most cases of osteomyelitis
will arise within the metaphysis of a long bone, typi-
cally the tibia and femur.
It can also be helpful to identify the origin of the
tumour with respect to the transverse plane of the
bone. Is the tumour central, eccentric or cortieally
based? For example, a simple bone cyst, fibrous dys-
plasia and Ewing's sarcoma will tend to be centrally
located. Chondromyxoid fibroma and fibrous corti-
cal defect/non-ossifying fibroma (Fig. 18.8) are typi-
cally eccentric. Lesions that usually arise in an eccen-
tric position may appear central if the tumour is
particularly large or the involved bone is of a small
calibre. Therefore, most tumours arising in the proxi-
mal fibula will appear "central". There are numerous
surface lesions of bone which are related to the outer
cortex (KENAN et al. 1993a; SEEGER et al. 1998). The
majority of the malignant surface lesions of bone are
the rare forms of osteosarcoma, e.g. periosteal, high-
grade surface and parosteal osteosarcoma. Most of
the cases of parosteal osteosarcoma arise from the
posterior metaphysis of the distal femur.
Pattern of Bone Destruction. Analysis of the interface
between tumour and host bone is a good indieator of
the rate of growth of the lesion. A sharply marginat-
ed lesion usually denotes slower growth than a non-
311
marginated lesion (Fig. 18.8). The faster the growth,
the more aggressive the pattern of destruction and
the wider the zone of transition between tumour and
normal bone (Fig. 18.1). Aggressivity per se does
not conclusively indieate malignancy, but the malig-
nant tumours tend to be faster growing than their
benign counterparts. Geographie bone destruction
is the term applied to bone lesions that appear well
marginated with a thin zone of transition. The thick-
er the sclerotic border, the longer the host bone has
had to respond to the lesion and, therefore, by impli-
cation, the slower the rate of growth of the lesion. The
vast majority of bone tumours in children showing a
geographic pattern of destruction are benign, such as
simple bone cyst (SBC), ABCs, fibrous dysplasia and
enchondroma.
Moth-eaten and permeative bone destruction are
terms used to describe bone destruction in whieh
there are multiple tiny corticallucencies with an ill-
defined zone of transition. These patterns indieate the
aggressive nature of these lesions in contrast to those
with a geographie pattern. The rapid growth of these
lesions does not allow the host bone sufficient time to
react and produce a response. Typically malignancies,
including osteosarcoma, Ewing's sarcoma and neu-
roblastoma metastasis, exhibit a moth-eaten or per-
meative pattern of bone destruction (Figs. 18.1, 18.6,
18.9). Acute osteomyelitis is the "benign" condition
whieh mayaiso give a moth-eaten appearance.
 312
a b
a b
Periosteal Reaction. The periosteum is normallyradio-
lucent but will mineralise when stimulated by an adja-
cent osseous or para-osseous process. The rate o(
mineralisation is partly dependent on the age of the
patient. The younger the patient, the more rapid the
appearance of radiographie change and vice versa.
Periosteal reaction, otherwise known as periosteal
new bone formation, may occur in any condition
whieh elevates the periosteum, whether it be blood,
pus or tumour. The appearance and nature of a peri-
A. M. Davies and D. Vanel
Fig.18.8a, b. AP radiographs 27 months
apart, showing growth of fibrous corti-
cal defect/non -ossifying fibroma of the
femoral metaphysis
Fig.18.9a, b. AP and lateral radiographs
of a mixed lytic and sclerotic osteosar-
coma. There is a combined lamellated
and spiculated periosteal reaction with
Codman angles
osteal reaction is frequently valuable in narrowing
down the differential diagnosis of a bone tumour.
A "shell" is used to describe a lytie lesion with
bone expansion. The shell is the periosteal new bone
laid in response to the growing tumour. The thieker
the shell, the slower growing the lesion and vice
versa. Shells are typically found in benign lesions
such as SBC, ABC, fibrous dysplasia and chondro-
myxoid fibroma (Fig. 18.10). They mayaIso be seen
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.

Matrix. A number of tumours produce a matrix, the Fig. 18.11. Lateral

intercellular substance, that can calcify or ossify. The radiograph of a
radiodense foci should be differentiated from other eentral ehondro-
causes of calcifications such as fracture callus, sclerot- sareoma of the
femur. Typical ear-
ic response adjacent to a tumour, necrotic debris and
tilage ealcifieation
dystrophie calcification. Radiodense tumour matrix proximally. The
is either osteoid or chondroid. The exception is permeative appear-
fibrous dysplasia, where the collagenous matrix may anee distally with
be sufficiently dense to give a ground-glass appear- anterior eortical
ance. Tumour osteoid is typified by solid (sharp- destruction indi-
eates a more
edged) or cloud to ivory-like (ill-defined edge) aggressive compo-
patterns (Fig. 18.9). Tumour cartilage is variously nent of the tumour
described as stippled, flocculent, ring-and-arc and (arrow)
314
may show caleification of a similar nature. However,
the distribution is typically central in enchondroma,
while it is peripheral in medullary infarction.
18.3.2
Diagnosis of Soft Tissue Tumours
The lack of contrast resolution is a well-recognised
limitation of radiography, but the value of the exami-
nation should not be underestimated in the evalua-
tion of soft tissue masses. It will not identify the pre-
eise diagnosis in the majority of cases, but it can still
provide valuable information, e.g. on the presence of
caleification and bone involvement. The absence of
any bony abnormality in the presence of a clinically
palpable mass immediately indieates that the pathol-
ogy is of soft tissue origin, albeit with a large differ-
ential diagnosis.
The radiodensity of most soft tissue masses
approximates to that of water and is similar to that
of muscle; such masses are, therefore, only revealed
by virtue of their mass effect. In a minority of cases,
part or all of the tumour may exhibit a radiodensity
suffieiently different from that of water for it to be
visualised directly on radiographs. Lipomas, the com-
monest of all soft tissue tumours, produce a low
radiodensity between that of muscle and air. For this
reason, lipomas are typieally well demarcated from
the surrounding soft tissues and, if of suffieient size,
can be diagnosed on radiographs with moderate con-
fidence. It should be noted that low-grade liposarco-
mas may contain variable amounts of fat that will
also appear relatively radiolucent on radiography. A
low-kilovoltage technique will accentuate the differ-
ences between fat and muscle.
Increased radiodensity may be seen in the tissues
due to haemosiderin, calcification or ossification.
Haemosiderin deposition typically occurs in synovi-
al tissues exposed to repeated haemorrhage such as is
seen in pigmented villonodular synovitis. Caleifica-
tion or ossification in the soft tissues is a feature of a
large spectrum of pathologies, including congenital,
metabolic, endocrine, traumatic and parasitie infec-
tions. Primary soft tissue tumours are one of the less
common causes of caleification that the general radi-
ologist can expect to come across in his or her routine
practiee.
Ultrasound is an important technique in the ini-
tial assessment of a suspected soft tissue mass. First,
it can confidently confirm or exclude the presence of
a mass. Second, it can to a degree characterise the
lesion by distinguishing purely cystic lesions, such as
A. M. Davies and D. Vanel
ganglia, from solid tumours. Doppler ultrasound can
be employed to assess the vascularity of alesion and
ultrasound is ideally suited for image-guided biopsy.
18.3.3
CT and MR Imaging in Diagnosis
The prineipal role of computed tomography (CT)
and MR imaging in the management of the patient
with a suspected musculoskeletal tumour is in stag-
ing (see Sect. 18.4). In selected cases both techniques
can be useful in establishing a differential diagnosis.
The CT features that should be assessed are similar
to those described above when evaluating the radio-
graphs. This reflects that fact that both are radio-
graphie techniques relying on the attenuation of an
X-ray source. Cortieal breaching, soft tissue exten-
sion and faint mineralisation are all more readily
appreeiated on CT scans than on radiographs.Assess-
ment of CT attenuation values will allow distinction
between fat-containing and fluid-containing masses.
Although the physical basis of MR imaging is very
different, similar morphological information can be
easily identified. The exception is the signal voids of
fine mineralisation, whieh can be easily missed on
MR imaging. Potentially misleading MR features that
might suggest a bone sarcoma are prominent marrow
oedema and soft tissue oedema (HAYES et al. 1992).
These are, however, common with osteoid osteoma,
osteoblastoma, chondroblastoma, stress fracture and
infection. Many soft tissue sarcomas will appear well
defined on MR imaging owing to the presence of
a pseudocapsule, whereas inflammatory processes,
such as abscesses, will appear poorly defined owing
to the surrounding inflammatory exudate.
The majority of tumours will have prolonged Tl
and T2 relaxation times, thereby showing low to
intermediate signal on Tl-weighted and high signal
on T2-weighted sequences. Tl shortening, with a
high signal intensity, will be seen in fat-containing
tumours, subacute haemorrhage and gadolinium
chelate enhancement. A low signal intensity on T2-
weighted images is seen with dense mineralisation,
hypocellular/fibrous tumours, signal voids from flow-
ing blood,haemosiderin deposition,surgiealimplants
and bone cement. Fluid-fluid levels are well dem-
onstrated on both CT and MR imaging in a large
number of different musculoskeletal conditions. In
the immature skeleton with the appropriate radio-
graphie appearances fluid-fluid levels are most com-
monly seen in ABCs (Fig. 18.lOb) (DAVIES and CAS-
SAR-PULLICINO 1992).
Tumours and Tumour-like Lesions
Dynamic contrast-enhanced MR imaging has been
used to differentiate benign from malignant bone
lesions using the slope of the derived time-intensity
curves (VERSTRAETE et al. 1994). Benign bone lesions
tend to show a low slope as compared with the high
or steep slope of malignant lesions; however, there
is considerable overlap such that this technique is of
limited value in routine practice. For example, highly
vascularised or perfused lesions such as ABC, eosino-
philic granuloma, osteoid osteoma and acute osteo-
myelitis may all show slope values in the malignant
tumour range. Similarly, in the soft tissues, early myo-
sitis ossificans will show a steep slope mimicking
malignancy.
18.4
Surgical Staging
Accurate surgical staging is a fundamental requisite
of all oncological imaging. The staging system reg-
ularly used for bone and soft tissue sarcomas is
that adopted by the Musculoskeletal Tumor Society
(ENNEKING et al. 1980). This assigns one of three
grades according to the local extent of the tumour,
the presence or absence of metastases and the his-
tological grade. Clarification of the first two features
of the staging system relies entirely on imaging. The
value of a straightforward staging system, such as
this, is that it is easily applied, correlates well with
prognosis and allows valid comparison of studies
of differing treatments and treatment centres. An
alternative staging system for bone sarcomas is the
American Joint Committee on Cancer, which iden-
tifies tumour extension, i.e. whether it is confined
to bone (Tl) or extends beyond bone (T2), grade
(Gl-4), nodal involvement (N) and distant tumour
spread (M).
Determination of local tumour extent around the
knee usually relies on MR imaging. One study has
shown CT to be as good as MR imaging in staging
(PANICEK et al. 1997a) although there has been some
doubt expressed as to whether the technique and
quality of technology used in that multicentre study
were stricdy comparable (STEINBACH 1998). Howev-
er, where access to MR imaging remains limited, CT is
an adequate alternative, albeit with a significant radi-
ation burden. The MR scan should preferably be per-
formed before the biopsy as the trauma of the proce-
dure may result in haemorrhage and oedema which
can exaggerate the true extent of the tumour. The
tumour characteristics that should be assessed on
315
MR imaging for the purposes of staging a suspected
bone sarcoma are as follows:
Extent in Bane? To assess the extent of bone involve-
ment by tumour, a Tl-weighted sequence should be
performed oriented along the long axis of the bone
involved (Figs. 18.2b, 18.12a, b). This sequence is par-
ticularly sensitive to marrow changes. It is necessary
to measure the tumour extent from a recognised ana-
tomical reference point, which, for the purposes of
a bone sarcoma arising around the knee, can be the
articular cortices of the femur or tibia. A gadolini-
um-chelate should not be used at this stage as uptake
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-
sive way of achieving the same result. Many benign
as well as malignant bone tumours show a variable
degree of peritumoral oedema. These include osteoid
osteoma, chondroblastoma, GCT and osteosarcoma
(Fig. 18.13b). The oedema appears as a zone of inter-
mediate signal intensity merging imperceptibly with
the main tumour. With sarcomas it can be difficult
to distinguish where tumour ends and oedema com-
mences. Arguably, it is prudent to include all reduced
marrow signal within the measurements of the
tumour extent as malignant cells may contaminate
the oedematous area beyond the immediate confines
of the main tumour. Some researchers have suggest-
ed that it is possible on MR imaging to distinguish
between tumour tissue and peritumoral oedema
using a dynamic contrast-enhanced sequence, but it
is difficult to believe this technique would pick up
isolated nests of malignant cells. It is interesting to
speculate whether in the future diffusion-weighted
MR imaging might help distinguish tumour from
oedema. Fortunately, this is not a significant manage-
ment problem in the majority of patients with a sar-
coma arising in the distal femur or proximal tibia.
Increasing the length of a custom -made prosthesis by
several centimetres to accommodate the oedematous
zone is unlikely to affect the functional outcome.
Extent in Soft Tissue? If the tumour is confined to
bone the cortex will remain intact. Cortical bone
appears black on all MR sequences as it does not pro-
duce a signal. Cortical destruction with loss of the
black line is a frequent and characteristic finding of
bone malignancy (Fig. 18.13). Not infrequently, how-
ever, highly malignant sarcomas such as osteosarco-
ma and Ewing's sarcoma can penetrate the cortex
without frank destruction. In this situation, best
 316
Fig. 18.12a-c. Osteosarcoma of the distal femur. The
small field of view sagittal Tl-weighted MR image
(a) demonstrates the primary tumour weil but fails
to reveal the proximally located skip metastases
which are visible on the lower resolution large field
of view image (b). The skip metastases can also be
identified as discrete foci of increased activity on
bone scintigraphy (c)
demonstrated on axial images, the dark contour of
the cortex will persist with a permeative appearance
analogous to the permeative or moth-eaten pattern
on radiographs. It is convention to describe any
tumour tissue identified outside the cortex as extra-
osseous or soft tissue extension. Strictly speaking this
is often incorrect as the tumour can remain con-
fined by a largely intact periosteum. Nevertheless,
this convention persists and is usually only a source of
problems when resolving the findings of MR imaging
versus the examination of the pathological specimen.
The relatively high water content of most tumours,
both bone and soft tissue sarcomas, renders them
iso-intense and therefore indistinguishable from sur-
rounding muscles on Tl-weighted images. It is for
this reason that to assess soft tissue extension a
T2-weighted sequence, with good contrast between
tumour and muscle, is required. A disadvantage of
the widely used fast spin echo (turbo) T2-weighted
sequence is slightly reduced spatial resolution and
A. M. Davies and D. Vanel
b,c
relatively high signal of fat, which may limit the con-
trast with tumour. This problem can be overcome
with the use of fat suppression. A STIR sequence is
an alternative, albeit with a poorer signal to noise
ratio (Fig. 18.13b). The STIR sequence will also tend
to overstage the extent of the tumour owing to its
increased sensitivity to raised water content in a tissue.
As in bone, the distinction of soft tissue tumour from
perineoplastic oedema can be problematic (SHUMAN
et al. 1991). Knowledge of the compartmental anato-
my around the knee is essential when determining the
stage of the tumour (ANDERSON et al. 1999).
Joint Involvement? It is important to identify knee
joint invasion by a sarcoma because, although the
joint will usually be excised at the time of definitive
surgery, prior knowledge will prevent the surgeons
from opening the joint and thereby potentially con-
taminating the surgical field with tumour cells. MR
imaging is highly sensitive for detecting joint inva-
Tumours and Tumour-like Lesions
c weighted fast spin echo axial image (c)
sion but false positives due to subsynovial rather than
true intra-articular spread can lead to overstaging
(SCHIMA et al. 1994). This is problematic in the knee,
where anterior extra-osseous spread of a distal femo-
ral sarcoma will appear to invade the suprapatellar
pouch while in reality it is frequently displacing it.
Of significance is the fact that the absence of a joint
effusion has a high negative predictive value for joint
invasion (SCHIMA et al. 1994). The articular cartilage
is a relative barrier to tumour growth and is usually
only involved in very large or late presenting tumours.
Therefore, typical sites of joint invasion in the knee
are the meniscocapsular reflections and the intercon-
dylar notch along the cruciate ligaments (Fig.18.13c).
Transarticular spread is rare. Identification of tumour
on both sides of the knee should suggest that it arose
de novo in the joint rather than the bone.
Neurovascular Involvement? Once a sarcoma has
extended beyond the confines of the bone it will tend
317
Fig. 18.Ba-c. MR imaging of an osteosarcoma of the distal
femur. The intra-osseous extent and cortical breaching are
weH demonstrated on the coronal Tl-weighted image (a).
The extra-osseous component and peritumoral oedema are
more conspicuous on the coronal STIR image (b). Early joint
invasion is evident at the capsular junction lateraHy (straight
arrow) and the intercondylar notch (curved arrow) on the T2-
to grow in the line of least resistance. Around the knee
this is typically into the popliteal fossa. MR imaging
can demonstrate whether a tumour is close to or in
contact with a neurovascular structure in the poplite-
al fossa, but usually cannot distinguish mere contact,
adherence or early invasion (PANICEK et al. 1997b).
Fortunately, the prevalence of neurovascular involve-
ment in bone sarcomas is less than 4%, such that,
although the positive predictive value of MR imaging
for involvement is poor, the negative predictive value
is over 90% (PANICEK et al. 1997b). MR angiography
can be used to delineate the relationship of the tumour
to vessels (LANG et al. 1995; SWAN et al. 1995).
Skip Metastases and Lymph Node Involvement? Small
synchronous foci of tumour, usually osteosarcoma,
that are present within the same bone as the pri-
mary tumour, or within a bone on the other side of
an unaffected joint, are called skip metastases. Skip
metastases in osteosarcoma have been reported in
318
up to 25% of eases (ENNEKING and KAGAN 1975)
although in the author's experienee the true inci-
denee is less than 5%. The majority will be deteeted
by bone scintigraphy (Fig. 18.12e) but scintigraph-
ieally negative skip metastases have been reported
(BHAGIA et al. 1997). When staging a sareoma around
the knee, the best resolution images will undoubt-
edly be obtained utilising the knee eoil (Fig. 18.l2a).
It is prudent however, to include a single large
field of view Tl-weighted sequenee along the line
of the femur or tibia to exclude a skip metastasis
(Fig. 18.12b). In the author's experienee, for some
unknown reason most transartieular skip metastases
are identified in the proximal tibia in patients pre-
senting with a distal femoral osteosareoma. Lymph
node spread in bone sareomas is uneommon and
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
those eases of osteosareoma with mineralisation,
indieating metastatie involvement, whieh ean be
easily deteeted on radiographs or CT and will show
inereased aetivity on bone scintigraphy. A false diag-
nosis of a skip metastasis may oeeur on scintig-
raphy when inereased aetivity within an involved
lymph node is projeeted over the distal femur when
only anterior or posterior projeetions are obtained
(BEARCROFT and DAVIES 1999). Rarely, in patients
with longstanding prostheses, regional lymphade-
nopathy may oeeur owing to a foreign body reae-
tion in response to the lymphatie uptake of metal
debris (DAVIES et al. 2001). In patients treated by
amputation, post-traumatie neuromas may mimie
lymphadenopathy.
Gd-DTPA has little value in the initial staging of
bone sareomas (SEEGER et al. 1991). It may help
distinguish subsynovial spread from true joint inva-
sion (SCHIMA et al. 1994), and adynamie eontrast-
enhaneed sean ean be obtained at this stage as a base-
line study for the subsequent assessment of tumour
response to ehemotherapy (see Seet.18.5).
The principles of staging soft tissue sarcomas aris-
ing around the knee are very similar to those detailed
above for bone sareomas. For both bone and soft
tissue sareomas the exclusion or eonfirmation of pul-
monary metastases requires ehest CT. The sensitivity
of pulmonary CT has been improved with the intro-
duetion of spiral CT (GINSBERG and PANICEK 2000).
Overstaging with spiral CT is a potential hazard as up
to 70% of solitary nodules less than 5 mm in diameter
at initial presentation in ehildren with solid extratho-
racie tumours may be benign (GRAMPP et al. 2000).
A. M. Davies and D. Vanel
Bone scintigraphy is used to exclude skeletal metas-
tases. However, over 95% of scintigraphie abnormali-
ties oeeurring at the time of presentation of osteosar-
eoma at loeations distant from the primary tumour
do not represent metastatie disease (KELLER et al.
1984). It is, therefore, important to eorrelate scinti-
graphie abnormalities with radiographs of the rele-
vant area.
At the same time as the staging imaging studies
are performed, it is the usual praetiee in the author's
unit to obtain measurement radiographs of both the
affeeted and the eontralateral lower limbs to aid
the manufaeture of a eustom-made prosthesis. Also,
the bone age of the skeletally immature patients is
estimated as eertain designs of prosthesis allow for
growth, i.e. are extendable.
18.5
Imaging Follow-up
The imaging follow-up for a patient with a proven
sareoma arising around the knee ean be divided into
short-term (i.e. pre-definitive surgery) andlong-term
(i.e. post-definitive surgery). In the short-term many
patients with a sarcoma will be ente red into one of
the international adjuvant ehemotherapy trials. After
a predetermined number of eycles of ehemotherapy
and immediately before surgery, the patient is re-
staged with an MR sean of the primary tumour and a
CT sean of the ehest. This is to ensure that the stage of
the tumour has not altered and that the planned sur-
gery is still appropriate. Also, this is an opportunity
to use imaging to assess the response of the tumour
to the ehemotherapy. Histologieal response to ehe-
motherapy expressed as pereentage neerosis is one
of the most important prognostie indieators in both
osteosareoma and Ewing's sareoma. Over the years
all types of imaging have been used to estimate the
response to ehemotherapy.
Post -ehemotherapeutie radiographie and CT find-
ings do not consistently differentiate the good from
the poor responder (SHAPEERO and VANEL 2000).
For example, an inerease in tumour volume may sug-
gest a poor response but mayaiso represent haemor-
rhage seeondary to neerosis in a responsive tumour
(VAN DER WOUDE et al. 1998). Conventional angi-
ography is eonsidered too invasive a proeedure
for monitoring tumour response to ehemotherapy.
Although it ean identify over 90% of responders, it
will miss 50% of the poor responders (CARRASCO et
al. 1989). It remains to be seen whether MR angiog-
Tumours and Tumour-like Lesions
raphy can fulfil a useful role in this respect (LANG et
al. 1995).
If there is a significant extra-osseous component
to the tumour, Doppler ultrasound can be used to
monitor response (VAN DER WOUDE et al. 1995).
The technique is operator dependent, which may
affect reproducibility of results on sequential scan-
ning. Scintigraphy using technetium-99m methylene
diphosphonate, thallium-201 and gallium-6? and
fluorine-18 fluorodeoxyglucose positron emission
tomography (PET) scanning have been advocated for
the estimation of tumour response (SHAPEERO and
VANEL 2000). Inherent to all of these methods is the
limited anatomical resolution and, with PET scan-
ning, limited availability. To date these techniques are
largely reserved for research purposes.
Unenhanced MR imaging has a limited role.
Increased or unchanged tumourvolume and increased
peritumoral oedema after chemotherapy suggest
a poor histological response in osteosarcoma and
Ewing's sarcoma. Virtual obliteration of the extra-
osseous component combined with a hypo-intense
rim in Ewing's sarcoma usually indicates a good
response. It is, however, impossible to exclude small
foci of viable tumour without contrast medium. Stan-
dard contrast-enhanced MR imaging is also of lim-
ited value as viable tumour, revascularised necrotic
tissue, reactive hyperaemia etc. may all enhance. It
is for this reason that much of the work on imaging
assessment of the response of sarcoma to chemo-
therapy over the past decade has concentrated on
dynamic contrast-enhanced MR imaging.A number
of different techniques have been described but all
rely on the underlying principle that viable tumour
enhances rapidly (i.e. within seconds of the contrast
medium arriving in the adjacent artery) whereas all
other enhancing tissues take much longer. It is pos-
sible on the console of most modern scanners to
plot a time/intensity curve showing the uptake of the
contrast medium. By comparing the curve obtained
before commencement of chemotherapy with that
obtained afterwards, the tumour response can be
estimated. It should be noted that this is a time-con-
suming and costly exercise, with numerous variables
that directly influence patient management in very
few cases.
In the long-term the patients are closely moni-
tored for evidence of local recurrence (DAVIES and
VANEL 1998), metastatic disease (BEARCROFT and
DAVIES 1999) and complications of treatment. Local
recurrence of a sarcoma is almost inevitable if the
original resection margin was not wide. Recurrence
may be detected on radiographs as a soft tissue mass
319
with or without bone destruction. Locally recurrent
bone sarcoma will usually occur within the soft tis-
sues at the site of the initial surgery as the host bone
will have been excised and replaced with a prosthe-
sis. Detection on radiographs is easier if there is evi-
dence of matrix mineralisation. Recurrent tumours
with the propensity to mineralise (i.e. osteosarcoma)
will usually exhibit focal increased activity on scin-
tigraphy but it is rarely used for this purpose.
MR imaging is the technique of choice in the
detection of early recurrence when local control may
still be surgically achievable. While ultrasound does
have some attractions (CHOI et al. 1991), MR imag-
ing will still be required for preoperative evaluation
if a recurrence is identified. Depending on the pres-
ence or absence of mineralisation, most recurrences
will show a high signal intensity mass on T2-weight-
ed or STIR images (VANEL et al. 1994). Diffuse high
signal intensity is frequently seen shortly after sur-
gery or can be prolonged following radiation therapy
(RICHARDSON et al. 1996). Contrast medium may be
required to distinguish enhancing recurrent tumour
from seromas, haematomas etc. Dynamic contrast-
enhanced MR imaging can be helpful in differen-
tiating small recurrences from other postoperative
changes.
It is generally accepted that it is usually the meta-
static disease that will eventually kill the patient, and
not the primary tumour itself. It is for this reason that
follow-up imaging is concentrated on the site where
metastases are likely to occur, namely the lungs. Chest
radiographs are usually considered adequate. Serial
ehest CT scans are of doubtful value in view of the
considerable radiation dose involved. The natural
history of osteosarcoma has been modified by che-
motherapy in that up to 20% of those who develop
metastases will first do so in bone prior to there
being any evidence of pulmonary metastases. The
prognosis for a patient with osseous metastases is so
poor that serial follow-up scintigraphy is unlikely to
modify the outcome. Scintigraphy is indicated should
a patient on follow-up develop bone pain.
It should be recognised that the prolonged medical
and surgical management of a patient with a sarcoma
is not without risk of complications. Prostheses may
become loose or infected or require replacement if
a child has outgrown the extended length of a grow-
ing prosthesis (KASTE et al. 2001). Allografts may also
become infected and are prone to fracture. In the
long-term follow-up of patients who have received
radiotherapy, pain or functional impairment within
the radiation field should lead to consideration or
bone necrosis or radiation-induced sarcoma.
320
18.6
Bone Tumours
The distal femur and proximal tibia are the com-
monest sites of occurrence for most bone tumours.
Tumours can be broadly classified according to their
tissue of origin or the tissue they most closely resem-
ble, e.g. osseous, cartilaginous, fibrous, lipomatous
and an unknown or miscellaneous category.
18.6.1
Benign Bone Tumours
18.6.1.1
Osseous Tumours
Osteoma is a benign, slow-growing focus of mature
dense bone that commonly involves the frontal and
ethmoid sinuses. It rarely arises on the outer surface
of the long bon es. On the distal femur, if large, it can
resemble a parosteal osteosarcoma. If extensive and
multifocal, the sclerosing dysplasia melorheostosis
should be considered. Fod of dense compact bone
within the medulla are known as bone islands or
enostoses. Multiple bone islands clustered towards
the bone ends are diagnostie of osteopoikilosis. Occa-
sionally bone islands may grow and show increased
activity on bone sdntigraphy.
Approximately 16% of osteoid osteomas arise
around the knee. The radiographic features of a small
lucent nidus, with or without mineralisation, and sur-
rounding sclerosis are the same as can be expected
elsewhere in the skeleton. On MR imaging if a large
field of view and thiek slices are employed, the nidus
may be obscured by florid perilesional oedema. The
larger variant of osteoid osteoma, the osteoblastoma,
is uncommon around the knee.
18.6.1.2
Cartilaginous Tumours
Benign cartilage tumours of bone can be divided
into those that arise within the medulla and those
that arise from the surface of bone. Medullary car-
tilage tumours are enchondromas, chondromyxoid
fibroma (CMF) and chondroblastoma. Surface car-
tilage tumours are osteochondroma and periosteal
chondroma. It is important to stress that the pres-
ence of cartilaginous calcification is helpful in indi-
cating the tissue of origin but does not distinguish
benign from malignant cartilage tumours (Figs.
l8.2a, 18.11).
A. M. Davies and D. Vanel
Enchondroma is a benign tumour of mature hya-
line cartilage accounting for 4% of all primary bone
tumours.Approximately 15% occur around the knee,
two to three times more commonly in the distal
femur than in the proximal tibia. The typieal radio-
graphie features are a well-defined oval or rounded
lytic defect, usually central, containing a variable
amount of cartilage calcification within the metaphy-
sis or diaphysis (Fig. 18.2a). The calcifications tend
to be scattered throughout the tumour, as opposed
to the peripher al linear distribution that is found
in medullary infarcts, which also tend to occur in
the distal femoral diaphysis. Bone expansion is not
common in the femur or tibia but may be observed
in thinner long bones, such as the proximal fibula.
Enchondromatosis (Ollier's disease) is a condition
marked by multiple enchondromas involving the
metaphysis and meta-diaphysis of the long bones.
There is no hereditary or familial tendency and it is
usually classified as a bone dysplasia. A monomelic
or hemimelic distribution is common. The spectrum
of skeletal change around the knee can vary enor-
mously from tiny fod of cartilage, to linear columns
of dysplastic unmineralised cartilage, to major mod-
elling deformities resulting in marked deformity. If
there is any doubt as to the condition, radiographs
of the hands and feet will usually clinch the diag-
nosis. Malignant transformation and the distinction
of enchondroma from low-grade chondrosarcoma is
discussed in Sect. 18.6.2.2.
Chondromyxoid fibroma is a rare (five times less
common than enchondroma) benign tumour com-
posed of varying amounts ofhyaline cartilage, fibrous
and myxomatous tissue. Approximately 43% of CMFs
arise around the knee, with the proximal tibial meta-
diaphysis a more common site than the distal femur.
CMF is typically eccentric in location, with a lobulat-
ed sclerotie margin (Fig. 18.14). The overlying cortex
is thinned, expanded and may even be absent. Matrix
mineralisation is uncommon. CMF has no malignant
potential.
Chondroblastoma is a benign cartilage tumour
arising almost exclusively in epiphysis or apophysis,
with three-quarters occurring at less than 20 years of
age. Approximately 39% occur around the knee, with
an equal inddence in the distal femoral and proxi-
mal tibial epiphyses. Radiographically, chondroblas-
toma appears as a well-defined lytie lesion within
the medulla of the epiphysis, although smalliesions
around the knee may not be easily identified. Matrix
mineralisation is seen in approximately one-quarter
of cases (Fig. 18.7b). Breaching of the growth plate
with metaphyseal involvement occurs in larger lesions
Tumours and Tumour-like Lesions
Fig. 18.14. AP radiograph of a chondromyxoid fibroma of
the proximal tibia. Typieal radiographie features include the
eccentrie location, well-defined lobulated endosteal margin
and peripheral expansion
with late presentation.A florid inflammatory response
is common, similar to osteoid osteoma, being typified
by surrounding marrow oedema with or without an
associated joint effusion. Both are weIl demonstrat-
ed on MR imaging (OXTOBY and DAVIES 1996). Rare
instances of malignant change in chondroblastoma
have been reported.
Osteochondroma is the commonest benign bone
tumour. It represents a bony protuberance (exosto-
sis) covered by a hyaline cartilage cap. Approximate-
ly 44% arise around the knee, with the distal femur
being involved twice as often as the proximal tibia.
It may be single or multiple, as in the hereditary form,
diaphyseal aclasis. It can be broad based (sessile)
or narrow based (pedunculated). Osteochondromas
arise from the metaphysis and are angulated away
from the adjacent joint. An important diagnostic fea-
ture is the continuity of the host bone marrow with
the marrow of the osteochondroma. The cartilage
cap is not visible on radiographs unless it calcifies.
The cartilage cap can be readily identified with ultra-
sound, CT and MR imaging. Complications of osteo-
chondromas include fracture, pressure on adjacent
tendons, nerves and vessels, overlying bursitis and
malignant degeneration (MURPHEY et al. 2000).
321
18.6.1.3
Fibrous Tumours
The commonest benign fibrous tumours ansmg
around the knee are fibrous cortical defects, which
occur in up to 30% of the normal population in the
first and second decades. Multiple lesions are pres-
ent in less than 10% cases. Radiographs are diag-
nostic, showing a well-defined elliptical, radio lu-
cent defect confined to the cortex of the long bone
near the growth plate. In time, with further skeletal
growth, the defects heal spontaneously and either
are no Ion ger visible or leave a residual focus of end-
osteal sclerosis. Larger lesions are called non-os si-
fying fibroma; 67% of such cases arise around the
knee, with 33% in the distal femur. Radiographi-
caIly, apart from their greater size, they resemble
fibrous cortical defects, often with further trabecu-
lation (Fig. 18.8). The greater the involvement of the
bone, the greater the risk of fracture. Benign fibrous
histiocytoma resembles non-ossifying fibroma his-
tologically but radiographically tends to appear
more aggressive. It has a predilection for the sub ar-
ticular eccentrie parts of the distal femur and proxi-
mal tibia.
Desmoplastic fibroma is a rare locally aggressive
fibrous tumour with a predilection for the distal
femur and to a lesser extent, the proximal fibula.
The radiographie appearance of a lytie expansile
lesion with trabeculation makes distinction from
other bone tumours difficult on imaging alone.
18.6.1.4
Lipomatous Tumours
Intra-osseous lipoma classieally arises in the anteri-
or calcaneus. It may occur in the long bones around
the knee, most frequently the proximal tibia. Radio-
graphs show a well-defined lucency within the medul-
la, frequently containing dystrophie calcification. In
this situation it may mimic a central chondrosarco-
ma. The diagnosis can be made with confidence on
MR imaging by identifying the predominant fat com-
ponent of the lesion.
18.6.1.5
Unknown Tumours
The simple bone cyst (SBC), also known as a unieam-
eral bone cyst, is a non-neoplastie fluid-filled cavity.
It is more common in males and is usually detected
in the first two decades of life. Only 6% of cases occur
around the knee; SBCs are slightly more common
 322
in the distal femur than the other bon es. The radio-
graphie appearances are a well-defined lytie lesion,
cent rally located within the metaphysis and migrat-
ing with time into the diaphysis. A typieal, but not
pathognomonic sign, is the so-called fallen fragment.
This represents a fragment of fractured cortex that
descends to the dependent portion of the cyst.
The differential diagnosis around the knee includes
aneurysmal bone cyst (ABC), fibrous dysplasia and
non-ossifying fibroma; the last-mentioned is usually
eccentrically located.
Aneurysmal bone cyst is a non-neoplastie lesion
consisting of multiple blood-filled spaces with vary-
ing amounts of fibrous, riehly vascular connective
tissue. Three-quarters of cases present in patients
under 20 years of age. Approximately 26% of cases
arise in the bones around the knee,ABC being slight-
ly more common in the proximal tibia than in the
other bones. The radiographie features are a well-
defined multiloculated expansile lesion arising in
the metaphysis (Fig. 18.10). Occasionally ABCs may
arise in a subperiosteallocation but these are usually
in the diaphysis. If the ABC is growing particularly
fast, the expanded shell may be thinned or absent
(Fig. 18.15). ABC may be a secondary phenomenon
occurring in pre-existing bone lesions, in whieh case
there may be imaging evidence of the underlying
abnormality. Fluid-fluid levels are a typical, but not
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
A. M. Davies and D. Vanel
pathognomonic, feature of ABCs seen on CT and MR
imaging (Figs. 18.10b, 18.15b ).It is necessary for the
scan to be perpendieular to the fluid-fluid levels for
them to be visible. The differential diagnosis for an
ABC includes SBC, chondromyxoid fibroma, non-
ossifying fibroma, GCT (in older patients) and, most
important of all, telangiectatie osteosarcoma. To
the unwary, telangiectatie osteosarcoma may resem-
ble an ABC both on imaging (including fluid-fluid
levels) and histology (Fig. 18.16). Misdiagnosis can
have potentially disastrous consequences as the man-
agement of an ABC is curettage while that of a tel-
angiectatie osteosarcoma is chemotherapy and wide
surgical excision.
Giant cell tumour (GCT) is a locally aggressive
tumour representing approximately 5% of all prima-
ry bone tumours and 22% of benign bone tumours.
Approximately 50% arise around the knee, more
commonly in the distal femur. Patients are usually
between 20 and 40 years of age. The tumour is consid-
ered benign although occasionally it may be multi-
focal, metastasise to the lungs or undergo malignant
transformation. The characteristie radiographie fea-
tures are an expansile, eccentric, lytic, subarticular
lesion which is weIl defined without marginal scle-
rosis (Fig. 18.17). On the AP projection of the knee,
distal femoral GCTs may appear remote from the
articular margin but inspection of the lateral will
 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)

show the tumour extending to the anterior, more 18.6.2

proximal articular margin. The important differen-
tial diagnosis, particularly in the older age group,
would be a metastasis (Fig. 18.3). Treatment of GeT
is usually curettage with or without bone grafting
or filling of the surgical defect with bone cement.
Local recurrence occurs in 20%-30% cases. Dynam-
ic contrast-enhanced MR imaging is of value in dis-
tinguishing early intra-osseous recurrence from scar
tissue.
Malignant Bone Tumours
18.6.2.1
Osseous Tumours
Osteosarcoma is the commonest primary malignan-
cy of bone after myeloma. Three-quarters of cases
are conventional osteosarcoma, also known as high-
grade intramedullary osteosarcoma. Over half arise
324
in the bones around the knee, with the peak inci-
dence in the second decade (Fig. 18.1). The radio-
graphie appearances can vary from purely lytie to
purely sclerotie but most will show a mixed appear-
ance with permeative margins, cortieal destruction,
soft tissue extension and periosteal new bone forma-
tion (Figs.18.1, 18.9). The latter may appear lamellat-
ed or spieulated and, if interrupted, will have Codman
angles. Approximately 10% of osteosarcomas are
the telangiectatic variety, which occurs most com-
monly in the distal femur followed by the proximal
tibia. This tumour is categorised by permeative bone
destruction with bony expansion. CT and MR imag-
ing show multiple blood-filled cavities with or with-
out fluid-fluid levels. It is the expansion and fluid-
fluid levels that may cause diagnostie problems with
ABC (Fig. 18.16).
Parosteal osteosarcoma is a low-grade malignancy
and is the commonest of the surface osteosarcomas,
accounting for approximately 5% of all osteosarco-
mas. Sixty percent arise on the posterior metaphysis
of the distal femur. Radiographieally, parosteal osteo-
sarcoma appears as a dense lobulated mass attached
to the outer cortex with a thin radiolucent cleft
between part of the mass and the cortex as it wraps
around the bone (Fig. 18.18). Intramedullary exten-
sion visible on CT or MR imaging will be present in
approximately one-third of cases. Dedifferentiation
to high-grade osteosarcoma occurs in about 20% of
cases and should be suspected if there is a large soft
tissue component to the tumour.
Multieentrie osteosarcomas may be synchronous
or metachronous. Although rare, the bones around
the knee are, as with solitary osteosarcoma, the com-
monest sites involved. Secondary osteosarcoma may
be seen in Paget's disease and following radiotherapy.
The pelvis, proximal femur and proximal humerus
are more commonly involved than the knee.
18.6.2.2
Cartilaginous Tumours
Malignant cartilage tumours, chondrosarcomas, are
the third most common primary malignant tumour
of bone, after multiple myeloma and osteosarcoma.
They can be distinguished from many other prima-
ry sarcomas of bone in that they occur in late adult-
hood rather than childhood or adolescence. Like
their benign counterparts, they arise in a central
or peripherallocation. Approximately 25% of cen-
tral chondrosarcomas arise in the femur, more com-
monly proximally than distally. Patients with Ollier's
disease are at risk of sarcomatous transformation in
A. M. Davies and D. Vanel
Fig. 18.18. Lateral radiograph of a parosteal osteosarcoma of
the distal femur
5%-30% of cases (LIU et al. 1987). The radiographie
appearance of central chondrosarcoma is variable.
High-grade lesions will appear permeative with car-
tilage mineralisation, cortieal destruction and soft
tissue extension (Fig. 18.11). In particularly aggres-
sive tumours, dedifferentiation to a high-grade sar-
coma (e.g. osteosarcoma or malignant fibrous his-
tiocytoma) should be considered. Diagnostic prob-
lems are usually encountered with low-grade central
chondrosarcoma, whieh can be difficult to differen-
tiate from an enchondroma on both radiography
and histology. One study concluded that a size great-
er than 5 cm was the most reliable predictor of
chondrosarcoma and that all other morphologieal
features, such as endosteal scalloping, were of little
value (GEIRNAEDT et al. 1997). Lesions demonstrat-
ing uptake on whole-body bone scintigraphy that
is less than uptake in the anterior iliac spine are
unlikely to be malignant (MURPHEY et al. 1998).
Although dynamic contrast-enhanced MR imaging
has been claimed to be of value in predicting malig-
nancy (GEIRNAEDT et al. 2000), this is not a univer-
sally held view (FLEMMING and MURPHEY 2000).
Less than 1% of solitary osteochondromas under-
go malignant transformation to a peripheral chon-
drosarcoma. It is generally accepted that the rate of
malignant change in diaphyseal aclasis is high er, but
probably no more than 1% if one includes all asymp-
tomatic cases of diaphyseal aclasis that do not pres-
ent for medical treatment (Fig. 18.4) (VOUTSINAS and
Tumours and Tumour-like Lesions
WYNNE-DAVIES 1983). Clinical features that suggest
malignant change include pain and increasing size
following skeletal fusion. Measurement of the thick-
ness of the cartilage cap using ultrasound, CT or MR
imaging can be helpful. A cartilage cap of less than 2
cm is likely to be benign, whereas as the cap exceeds
2 cm in thickness the likelihood of chondrosarcoma
increases. Complications of osteochondromas such
as overlying bursitis and pseudo-aneurysm forma-
tion can mimic malignant change.
18.6.2.3
Fibrous Tumours
Malignant fibrous tumours consist of fibrosarcoma
and malignant fibrous histiocytoma. While these
may be distinct histologie al entities, the imaging fea-
tures are indistinguishable and they are therefore
discussed together. Approximately 50% occur in the
bones around the knee. Twenty percent arise in pre-
existing bone lesions, including Paget's disease, bone
infarction (Fig. 18.19), irradiated bone, fibrous dys-
Fig. 18.19. AP radio graph of a medullary infarct of the distal
femur. A malignant fibrous histiocytoma is developing in the
proximal portion of the infarct
325
plasia and non-ossifying fibroma. As with chondro-
sarcoma, they tend to present in a slightly older age
group than most other primary sarcomas of bone,
with a peak in the fourth decade. Typical radio-
graphie appearances are geographie bone destruc-
tion with a wide zone of transition in an eccentric
metaphyseal or meta-diaphyseallocation. They tend
not to extend to the articular margin. Cortical
destruction with soft tissue extension is common but
periosteal new bone formation is unusual and there
is no matrix mineralisation. In the older patient the
appearances can be indistinguishable from metasta-
sis or lymphoma of bone.
18.6.2.4
Unknown
Originally of unknown histogenesis, Ewing's sar-
coma and the similar peripheral neuroectodermal
tumour are now known to be linked to primitive
neuroectoderm. Together they are the second com-
monest malignancy of bone in children and adoles-
cents, after osteosarcoma. The classic site in long
bones, as illustrated in many texts, is the diaphy-
sis, but this is the site in only one-third of cases.
Approximately 60% arise in the meta-diaphysis,
with only 14% of all cases around the knee. The
radiographie features are of an aggressive lesion
with a permeative or moth-eaten pattern, cortical
destruction and periosteal new bone formation
(Fig. 18.6). Around the knee, Ewing's sarcomas may
radiographically resemble the more lytic form of
osteosarcoma.
18.6.3
Patellar Tumours
The patella is the largest sesamoid bone in the human.
Unlike the distal femur and proximal tibia, it is a rare
site for tumours (KRANSDORF et al. 1989). Despite
this, it deserves separate mention in a treatise on the
knee. Arecent review of the literature identified a
total of 158 patellar tumours (O'MARA et al. 2000).
Of these, 72.6% were benign neoplasms, 21.7% were
malignant neoplasms and the remaining 5.7% were
benign non-neoplastic conditions. The most common
benign tumours were GCT (Fig. 18.20) and chondro-
blastoma, with brown tumours the most common
non-neoplastic condition. Osteosarcoma was the
most common malignant neoplasm. Imaging of patel-
lar tumours with radiographs and MR imaging will
usually suffice.
326 A. M. Davies and D. Vanel

of the popliteal or peroneal nerves. The fusiform

mass with entering and exiting nerves can be confi-
dently diagnosed with either ultrasound or MR imag-
ing. The eccentric position of the nerve with respect
to the main mass helps to distinguish aschwannoma
from a neurofibroma (Fig. 18.21).

18.7.2
Malignant Soft Tissue Tumours

Approximately 45% of soft tissue sarcomas in adults

occur in the lower extremity (VARMA 1999), with
no particular predilection for the knee. Radiographs
will reveal calcifications in extraskeletal osteosar-
coma, extraskeletal chondrosarcoma and approxi-
mately one-third of synovial sarcomas. Peripheral
mineralisation/ossification suggests myositis ossifi-
cans rather than a sarcoma. Most soft tissue sar-
Fig. 18.20. Lateral radio graph of a GeT arising within the comas, irrespective of their tissue origin, will be
patella iso-intense to musele on unenhanced CT and het-
erogeneous but predominantly hyperintense on T2-
weighted MR imaging. Cystic and haemorrhagic
18.7 areas are common in high-grade sarcomas. Synovial
Soft Tissue Tumours sarcoma is the commonest lower extremity malig-
nancy in the child and young adult and frequently
18.7.1 arises around the knee. It may show evidence of
Benign Soft Tissue Tumours haemorrhage with a "bunch of grapes" appearance,
with or without fluid-fluid levels (JONES et al. 1993).
The majority of benign soft tissue masses arising in
relation to the knee are non-neoplastic cystic lesions
(see Sect. 18.8.1.1). Post-traumatic conditions around
the knee which may mimic a soft tissue tumour
inelude occult rectus femoris musele tears, myositis
ossificans and popliteal artery pseudo-aneurysms
secondary to mechanical irritation from an adjacent
osteochondroma. The commonest benign soft tissue
neoplasms are myxomas, lipomas, vascular tumours,
neurogenic tumours and aggressive fibromatosis. The
relative incidence of each group of these benign con-
ditions in the lower extremity, exeluding the hip, foot
and ankle, is: myxomas, 55%; lipomatous tumours,
14%; vascular tumours, 15%; neurogenic tumours,
17%; and aggressive fibromatosis, 10% (KRANSDORF
and MURPHEY 1997). The imaging features of these
conditions are weH described and do not vary depen-
dent on their anatomicallocation. A couple of entities
are briefly mentioned because of their predilection
for the knee. Juxta-articular myxoma, also known
as peri-articular myxoma, occurs most commonly
around the knee. Because of cystic change it fre- Fig.18.21. Sagittal Tl-weighted MR image of aschwannoma of
quently mimics a ganglion on imaging. Neurogenic the popliteal nerve. The nerve can be seen eccentricallyenter-
tumours around the knee are usuaHy schwannomas ing and exiting the mass
Tumours and Tumour-like Lesions
The latter is a non-specific sign as it may be seen in
the soft tissues in cavernous haemangioma, myositis
ossificans and abscess. Distinction of the different
types of soft tissue sarcoma on imaging will rarely
influence subsequent biopsy or management. Clear-
ly, it is important to differentiate a sarcoma from
other non-malignant conditions. It is prudent to
consider all deep-seated solid or semi-solid masses,
without any specific diagnostic features, as poten-
tially malignant until proved otherwise by biopsy. If
there is marked necrosis or cystic change, imaging
may help indicate the most appropriate solid part
of the sarcoma to biopsy. In low-grade liposarcoma,
MR imaging will help differentiate the lipomatous
portion of the tumour from the liposarcomatous
elements (Fig. 18.22).
18.8
Joint Tumours
In this section the term "tumour" is taken in its liter al
sense as a swelling or mass. The discussion is there-
fore not limited just to true neoplasms but includes
cystic, metaplastic and proliferative conditions which
may present with a mass arising from or adjacent to
the knee joint. Polyarticular dis orders are not includ-
ed as they rarely present with an isolated mass.
Fig.18.22. Sagittal Tl-weighted MR image of a low-grade lipo-
sarcoma arising in the popliteal fossa. Although much of the
mass shows high signal intensity corresponding to fat, the
lower signal areas distally suggest sarcoma
327
18.8.1
Benign Joint Tumours
18.8.1.1
Bursae, Ganglia and Synovia I Cysts
There are numerous normal and abnormal cystic
structures that can occur around the knee (JANZEN et
al. 1994; MORRISON and KAPLAN 2000). While ultra-
sound can detect and diagnose many of these con-
ditions, MR imaging is ideally suited to identify the
relationship to the intra- and peri-articular struc-
tures. There is considerable confusion regarding the
classification/terminology of cystic lesions around the
knee. Synovial cysts can be distinguished from gan-
glion cysts histologically by identification of a syno-
viallining in the former which is absent in the latter.
Despite this distinction, some entities have been his-
torically incorrectly named. For example, proximal
tibio-fibula ganglia arise fl"om the proximal tibio-fib-
ula joint. On histological grounds they are synovial
cysts, not ganglia. Further problems occur with mul-
tiple names for the same condition. For example, pos-
terior cysts that communicate with the knee joint
are variously known as synovial/popliteal cysts, Bak-
er's cysts or, referring more precisely to the common
anatomical site, the gastrocnemius-semimembrano-
sus bursa. Fortunately, the specific histological nature
of most of these lesions rarely influences individual
patient management. The majority will appear hypo-
intense on Tl -weighted and hyperintense on T2-
weighted MR imaging. If complicated by haemor-
rhage, the contents may appear hyperintense on Tl-
weighted images. The correct diagnosis is made by
identifying the precise anatomical relationship of the
cystic structure with respect to the other tissues.
Bursae are synovial-lined sacs that reduce fric-
tion between adjacent moving structures. They can
become inflamed due to, for example, repetitive
trauma, infection or gout, or distended due to syno-
vial proliferation (e.g. pigmented villonodular syno-
vitis). The bursae most commonly abnormal around
the knee are (MORRISON and KAPLAN 2000):
- Prepatellar bursa
- Superficial and deep intrapatellar bursae
- Pes anserine bursa
- Gastrocnemius-semimembranosusbursa(popliteall
Baker's cyst)
Ganglia may be intra- or extra-articular. Intra-
articular ganglia arise in relation to the cruciate liga-
ments or the al ar folds over the infrapatellar fat pad
(TYRRELL et al. 2000; KIM et al. 2001). The absence
328 A. M. Davies and D. Vanel

of an associated meniscal tear can help distinguish

the latter from another form of ganglion seen in the
knee, the meniscal cyst. Ganglion cysts affecting the
bones of the knee may be one of two types. First,
the intra-osseous ganglion, whieh typieally appears
as a loculated subartieular lucency, most commonly
in the proximal tibia near the insertions of the cruci-
ate ligaments. Second, the periosteal ganglion, whieh
arises most commonly on the medial metaphysis of
the proximal tibia, causing well-defined pressure ero-
sion of the outer cortex.

18.8.1.2
Pigmented Villonodular Synovitis

Pigmented villonodular synovitis (PVNS) is a benign

proliferative dis order of synovium. There are two cat-
ego ries of this condition. First, the relatively common
extra-articular form, known as giant cell tumour of Fig. 18.23. Coronal Tl-weighted MR image of diffuse intra-
tendon sheath, which ocoors almost exclusively in articular PVNS. Typical features include proliferation of"black
the hand and wrist and less commonly the foot and synovium" and degenerative joint disease
would, therefore, be rare around the knee. Second,
the intra-artieular form, termed PVNS, whieh tends
to involve the large joints, with approximately 80% tellar fat pad with a rounded soft tissue mass show-
cases involving the knee joint (DORWART et al. 1984). ing relatively low signal intensity on all sequences
Early in the disease radiographs may be normal or (Fig. 18.24) (PALUMBO et al. 1994; DELCOGLIANO et
show a joint effusion with or without soft tissue mass. al. 1998; CHOI 2000). PVNS may occasionally arise in
Multiple well-defined erosions on both sides of the a synovial-lined bursa. The typieal site in relation to
joint with relative preservation of joint space is well the knee joint is the pes anserine bursa.
recognised in PVNS but is only seen in approximately
one-quarter of cases involving the knee (DORWART et 18.8.1.3
al. 1984). This is thought to be due to the relatively Synovial (Osteo)chondromatosis
lax capsule of the knee as compared with the hip,
whieh allows the mass to extend in the line of least Synovial osteochondromatosis (also known as syno-
resistance into the popliteal fossa and suprapatellar vial chondromatosis or synovial chondrometaplasia)
pouch. The MR imaging appearances of PVNS are is due to metaplastie proliferation of multiple cartilag-
characteristic. These comprise diffuse heterogeneous inous nodules within the synovium of joints, bursae
synovial masses with predominantly low signal inten- and tendons. It is more than twice as common in men
sity on all sequences due to the deposition in the as in women and tends to present in middle age. Over
synovium of haemosiderin resulting from repetitive 50% cases involve the knee, about 10% of whieh are
intra-articular bleeds, the so-called black synovium bilateral (MILGRAM 1977). The radiographie appear-
(Fig. 18.23) (JELINEK et al. 1989; STEINBACH et al. ances depend on the chronicity of the disorder and
1989; HUGHES et al. 1995). The haemosiderin can the degree of mineralisation. Early on the radiographs
cause susceptibility artefacts whieh are most conspie- may be normal or show a synovial-based soft tissue
uous on gradient echo sequences. Similar appearanc- mass. Mineralisation can vary from faint punctate and
es may be seen in other conditions associated with curvilinear cartilage calcification to more dense calci-
repeated episodes of intra-artieular bleeding, such as fication and occasionally ossification (Fig. 18.25a). In
haemophiliac arthropathy and synovial haemangio- the late disease there is usually a degree of degen-
ma, as well as hypocellular deposition dis orders, such erative joint disease. In the past, both arthrography
as amyloid arthropathy. Most cases of PVNS of the and CT were used to establish the diagnosis, partieu-
knee show diffuse involvement of the synovium but larly in the presence of non-mineralised intra-artieu-
there is a less common localised form. This typieally lar bodies. Nowadays most patients will undergo MR
arises in the anterior knee joint,replacing the infrapa- imaging in preference to either of these techniques.
 Tumours and Tumour-like Lesions 329

Fig.18.24. Sagittal Tl-weighted (a) and axial T2*-weighted (b) MR images

of localised PVNS affecting the infrapatellar fat pad. The haemosiderin
a
deposition is more conspicuous on the gradient echo image (arrows)

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-
articular soft tissue masses of intermediate signal
intensity on Tl-weighted images and high signal inten-
sity on T2-weighted images (Fig. 18.25b) (KRAMER et
al. 1993). The extent of focal areas of signal void, best
seen on T2-weighted or STIR images, will depend on
the degree of mineralisation. In severe cases, bone ero-
sion will be present. In longer-standing cases, discrete
intra-articular bodies can be identified. This condi-
tion of primary synovial osteochondromatosis should
be distinguished from the much commoner secondary
form. In the latter there are osteocartilaginous loose
330
bodies of varying size associated with more advanced
osteoarthritis. The loose bodies in the secondary form
are assumed to originate from the debris shed by
the damaged articular cartilage. There is a rare asso-
ciation between synovial osteochondromatosis and
synovial chondrosarcoma (see Sect.18.8.2.2).
18.8.1.4
Lipoma Arborescens
Lipoma arborescens is a cause of slowly increasing
monarticular pain and swelling in the absence of a
history of trauma. The aetiology is unknown; how-
ever, it has been described in patients with rheu-
matoid arthritis, psoriasis, degenerative joint disease
and diabetes mellitus (KLOEN et al. 1998), suggesting
a reactive rather than a neoplastic process. It occurs
most commonly in the knee joint, particularly the
suprapatellar pouch (FELLER et al. 1994; MARTIN et
al. 1998), but the shoulder, hip, wrist, elbow and ankle
are also recognised sites. Although usually monartic-
ular, cases of bilateral involvement of the knees
and hips have been reported (MARTIN et al. 1998;
SUMEN et al.1998). It occurs more commonly in men
than women, usually in the fourth and fifth decades.
Patients as young as 9 years (DONNELY et al. 1994)
and as old as 90 years (LAoRR et al. 1995) have been
reported. The appearance on MR imaging is typically
that of a large joint effusion with a frond-like mass
arising from the synovium, which has fat density on
all sequences (Fig. 18.26). The mass will therefore
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
Fig. 18.26. Axial Tl-weighted MR image of lipoma arbore-
scens. Typical high signal intensity frond-like subsynovial fat
deposits in the suprapatellar pouch
A. M. Davies and D. Vanel
PVNS. Lipoma arborescens can be confidently diag-
nosed on MR imaging, and this has probably led to
the recent increase in the number of reported cases.
To give some idea of the relative incidence of this
condition, the diagnosis is made in approximately 1
in 1,400 knee MR scans performed in the unit of one
of the authors (A.M.D.).
18.8.1.5
Synovia I Haemangioma
Soft tissue haemangioma arising in the synovium
accounts for less than 1% of all haemangiomas
(RESNICK and OLIPHANT 1975). Over 65% of these
involve the knee joint, typically the suprapatellar
pouch. Approximately one-third will also have extra-
articular extension. On MR imaging a lobulated intra-
synovial mass with or without an extra-articular
component will be seen containing prominent ser-
piginous vessels (Fig. 18.27). Fatty elements, identi-
fied as high signal intensity on Tl-weighted images,
may be seen, similar to an intramuscular haemangi-
oma. Repetitive intra-articular bleeding will result in
haemosiderin deposition in the synovium similar in
appearance to PVNS and haemophiliac arthropathy.
In the author's experience the degree of haemosid-
erin deposition in synovial haemangioma tends to
be less florid than in cases of either PVNS or hae-
mophiliac arthropathy. Localised pressure erosion of
the adjacent subcortical bone is common, particular-
ly affecting the distal femoral metaphysis (Fig. 18.27)
Fig. 18.27. Axial T2-weighted MR image of a synovial haeman-
gioma. Typical features include a lobulated intrasynovial mass
with prominent vessels and pressure erosion of the distal fem-
oral metaphysis
Tumours and Tumour-like Lesions 331

(SUH et al. 1994; GREENSPAN et al. 1995; LLAUGER et

al. 1995). In the growing skeleton, chronic hyperae-
mia associated with a large synovial haemangioma
can result in localised limb enlargement.

18.8.1.6
Intracapsular/Para-articular Chondromas

Chondromas of soft tissue are uncommon lesions of

which over 80% arise in the hands and feet. Rarely
chondromas arise from the joint capsule or para-
articular tissue sites around large joints. Ninety per-
cent of cases arise at the knee, typically involving
the infrapatellar fat pad (STEINER et al. 1994). A lat-
eral radio graph of the knee will show a mass lesion
replacing the infrapatellar fat pad with a variable
amount of punctate cartilage calcification. If ossifi.-
cation is present, the lesion is more appropriately
termed a soft tissue osteochondroma (Fig. 18.28).
This condition should be distinguished from dyspla-
sia epiphysealis hemimelica (Trevor's disease), which
typically affects the ankle and/or knee and in which
there is an osteocartilaginous mass arising from the Fig. 18.28. Lateral radiograph of an intracapsular osteochon-
epiphysis (Azouz et al. 1985). droma arising in the infrapatellar fat pad

18.8.2
Malignant Joint Tumours
Malignancies arising within the knee joint, or any
synovial joint for that matter, are extremely rare.
Because of their relatively common occurrence in
general, on occasion one might expect to see metasta-
ses or lymphoma arising within a joint. A single case
of malignant change in PVNS has been reported but
this arose in the ankle joint (KAUL and UNNI 1998).
The commonest primary joint malignancies are syno-
vial sarcoma and synovial chondrosarcoma.
18.8.2.1
Synovia I Sarcoma
Less than 10% of synovial sarcomas arise in an intra-
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
More commonly the tumour arises within the soft
tissues dose to the knee, with or without secondary
joint invasion. The imaging features of the intra-
articular form mimic other synovial proliferative or
infiammatory disorders such as primary synovial
chondromatosis and gout. The bone destruction,
however, tends to be more infiltrative than erosive
and the disease shows rapid progression.
18.8.2.2
Synovia I Chondrosarcoma
Synovial chondrosarcoma is rare, with the knee joint
the commonest site. Although it may arise de novo,
most cases described in the literature are due to
malignant degeneration of synovial chondromatosis
(KENAN et al. 1993b ). There can be considerable diffi-
culty in differentiating between synovial chondrosar-
coma and synovial chondromatosis, both on imaging
and on histology.As with intra-articular synovial sar-
coma, the bone destruction in synovial chondrosar-
coma is more infiltrative than erosive. Extra-articular
extension, rapid local recurrence following surgery
and the development of metastatic disease dearly
favour the diagnosis of a sarcoma.
18.9
Tumour-like Lesions of Bone
There are a large number of disparate bone condi-
tions which can have similar imaging appearances to
tumours. What constitutes a tumour mimic depends
very much on the expertise of the individual review-
ing the imaging. The majority can be dassified as
normal variants or post -traumatic, infiammatory and
metabolie conditions. The prudent radiologist will
always have dose at hand, when reporting, one of the
standard reference texts on normal variants which
illustrates the wide spectrum of developmentallucen-
eies and irregularities of ossification that can be seen
on radiographs of the growing knee.
18.9.1
Avulsive Corticallrregularity
Avulsive cortical irregularity (also known as peri-
osteal desmoid, cortical desmoid and distal femoral
metaphyseal defect) is a tumour-like fibrous pro-
liferation of the periosteum which affects the pos-
teromedial cortex of the distal femoral metaphysis
A. M. Davies and D. Vanel
(BUFKIN 1971; BARNES and GWINN 1974). The radio-
graphie appearances are erosion of the outer cortex
with small bony spicula which can simulate an
aggressive or malignant lesion. It is seen in adoles-
cents, particularly males, and is often bilateral. There
is some debate as to whether it is a normal variant
or represents a traction enthesopathy at the sites of
attachment of the tendinous fibres of the adductor
magnus or medial head of gastrocnemius musdes.
Bone scintigraphy is typically normal, which would
be unusual for a traumatic lesion (BURROWS et al.
1982; VELCHIK et al. 1984). MR imaging, however,
can show some oedematous change in relation to the
cortical irregularity (HYMAN et al. 1995; YAMAZAKI
et al. 1995) which would not be expected with a
normal variant.
18.9.2
Stress Fractures
The fatigue form of stress fracture can cause diag-
nostic difficulties in that the healing fracture may fre-
quently be mistaken for a malignant lesion (LEVIN et
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
as a sarcoma of bone (DAVIES et al. 1988). A less
common site where similar problems may occur is
the distal femoral diaphysis (DAVIES et al. 1989).
Some indication of the in eiden ce of misdiagnosis is
indicated by two studies conducted in an orthopae-
die oncology unit. In the first, fatigue fractures of the
proximal tibia accounted for 11 % of referrals with
tibial lesions and was the second commonest final
diagnosis after osteosarcoma (DAVIES et al. 1988).
In the second study on femoral diaphyseallesions,
fatigue fractures were the second commonest diag-
nosis after Ewing's sarcoma but the commonest diag-
nosis if only skeletally immature patients were con-
sidered (DAVIES et al. 1989).
Computed tomography will show the periosteal
and endosteal new bone formation, the fracture
line (if viewed on appropriate bone windows) and,
most important of all from the point of view of
exduding malignancy, absence of a soft tissue com-
ponent (SOMER and MEURMAN 1982; YOUSEM et al.
1986; DAVIES et al. 1989). The increased attenuation
within the medulla around the fracture site due to
oedema and/or haemorrhage should not be mistak-
en for intra-osseous tumour tissue (ARRIVE et al.
1988; DAVIES et al. 1988). The value ofMR imaging in
Tumours and Tumour-like Lesions
stress fractures is questionable (TYRRELL and DAVIES
1994). If the correct diagnosis is not considered on
the radiographs, MR imaging may further confuse
the unwary radiologist by revealing an extensive area
of marrow signal change (LEE and YAO 1988; TYR-
RELL and DAVIES 1994).
18.9.3
Inflammatory Conditions
A severe monarticular presentation of an inflamma-
tory arthropathy of the knee, such as in gout, may
occasionally mimic a tumour of joint origin. A more
common problem with inflammatory disease is dis-
tinguishing between osteomyelitis and bone neo-
plasia (WILLIS and ROZENCWAIG 1996). Around the
knee, it is usually subacute osteomyelitis of either the
distal femur or, more frequently, the proximal tibia
that gives rise to problems with diagnosis. The imag-
ing features of osteomyelitis are discussed in Chap.
14. A useful MR feature of subacute osteomyelitis is
the so-called penumbra sign, which is identified on
Tl-weighted images as a discrete peripheral zone of
marginally high er signal intensity than the abscess
cavity and surrounding marrow oedema/sclerosis
and lower signal intensity than the fatty bone marrow
(GREY et al. 1998). This sign is characteristic but not
pathognomonic of subacute osteomyelitis. Another
pointer to osteomyelitis rather than tumour is early
breaching of the growth plate. Infection with abscess
formation should be included in the differential diag-
nosis of a lytic epiphyseallesion although chondro-
blastoma remains the more likely diagnosis.
18.9.4
Brown Tumours
The radiographie similarities of brown tumours of
hyperparathyroidism to other bone tumours has long
been recognised. Solitary tumours include SBC, ABC
and GCT. Multiple brown tumours may be mistaken
for osteolytic metastases, myeloma and histiocyto-
sis. Unless the underlying hyperparathyroidism is
clinically suspected, the associated osteopenia may
well be mistaken for disuse osteoporosis secondary
to pain.
333
References
Anderson MW, Temple HT, Dussault RG, Kaplan PA (1999)
Compartmental anatomy: relevance to staging and biopsy of
musculoskeletal tumors. Am J RoentgenoI173:1663-1671
Ardran GM (1951) Bone destruction not demonstrable by
radiography. Br J RadioI24:107-109
Arrive L, Sellier N, Kalifa G, Seringer R (1988) Difficultes diag-
nostiques des apopositions periostees unilamellaires symp-
tomatiques isolees. J RadioI69:351-356
Azouz EM, Slomie AM, Marton D, Rigault P, Finidori G (1985)
The variable manifestations of dysplasia epiphysealis hemi-
melica. Pediatr RadioI15:44-49
Barnes GR Jr, Gwinn JL (1974) Distal irregularities ofthe femur
simulating malignancy. Am J Roentgenol 122: 180-185
Bearcroft PWP, Davies AM (1999) Follow-up of musculoskel-
etal tumors. 2. Metastatie disease. Eur Radiol 9: 192-200
Bhagia SM, Grimer RJ, Davies AM, Mangham DC (1997) Scin-
tigraphieally negative skip metastases in osteosarcoma. Eur
RadioI7:1446-1448
Bufkin WJ (1971) The avulsive cortieal irregularity. Am J
Roentgenoll12:487 -492
Burrows PE, Greenberg ID, Reed MH (1982) The distal femo-
ral defect: Tc-99m pyrophosphate bone scan results. J Can
Assoc RadioI33:91-93
Campanacci M (1999) Bone and soft tissue tumors: clinieal fea-
tures, imaging, pathology and treatment, 2nd edn. Springer,
Vienna New York
Carrasco CH, Charnsangavej C, Raymond AK (1989) Osteosar-
coma: angiographie assessment of response to preoperative
chemotherapy. Radiology 170:839-842
Choi H, Varma DGK, Fornage BD, Kim EE, Johnston DA
(1991) Soft-tissue sarcoma: MR imaging vs sonography
for detection of local recurrence after surgery. Radiology
157:353-358
Choi NH (2000) Localized pigmented villonodular synovitis
involving the fat pad ofthe knee.Am J Knee Surg 13:117-119
Davies AM, Vanel D (1998) Follow-up of musculoskeletal
tumors.l. Local recurrence. Eur RadioI8:791-799
Davies AM, Evans N, Grimer RJ (1988) Fatigue fractures
of the proximal tibia simulating malignancy. Br J Radiol
61:903-908
Davies AM, Carter SR, Grimer RJ, Sneath RS (1989) Fatigue
fractures of the femoral diaphysis in the skeletally imma-
ture simulating malignancy. Br J RadioI62:893-896
Davies AM, Cassar-Pullicino VN (1992) The incidence and sig-
nificance of fluid-fluid levels on computed tomography of
osseous lesions. Br J RadioI65:193-198
Davies AM, Cooper SA, Mangham DC, Grimer RJ (2001) Metal-
containing lymph nodes following prosthetie replacement
of osseous malignancy: potential role of MR imaging in
characterisation. Eur Radiolll :841-844
Delcogliano A, Galli M,Menghi A, Belli P (1998) Localized pig-
mented villonodular synovitis of the knee: report of two
cases of fat pad involvement. Arthroscopy 14:527-531
Donnely LF, Bisset GS, Passo MH (1994) MRI findings of
lipoma arborescens of the knee in a child: case report. Pedi-
atr RadioI24:258-259
Dorfman HD, Czerniak B (1995) Bone cancers. Cancer
75:203-210
Dorwart RH, Genant HK, Johnston WH, Morris JM (1984) Pig-
mented villonodular synovitis of synovial joints: clinieal,
pathologie and radiologie features. AJR 143:877-885
334
Edelstyn GA, Gillespie PJ, Grebbel FS (1967) The radiologieal
demonstration of osseous metastases: experimental obser-
vations. Clin RadiolI8:158-162
Enneking WF, Kagan A (1975) The implications for skip metas-
tases in osteosarcoma. Clin Orthop 111:33-41
Enneking WF, Spanier SS, Goodman MA (1980) A system
for the surgieal staging of musculoskeletal sarcoma. Clin
Orthop 153:106-120
Feller JF, Rishi M, Hughes EC (1994) Lipoma arborescens of the
knee: MR demonstration. Am J Roentgen 163:162-164
Flemming DJ, Murphey MD (2000) Enchondroma and chon-
drosarcoma. Semin Musculoskeletal Radio14:59-71
Geirnaerdt MJA, Hermans J, Bloem JL, Kroon HM, Pope TL,
Taminiau AHM, Hogendoorn PCW (1997) Usefulness of
radiography in differentiating enchondroma from central
grade I chondrosarcoma.Am J Roentgen 169:1097-1104
Geirnaerdt MJA, Hogendoorn PCW, Bloem JL, Taminiau AHM,
van der Woude HJ (2000) Cartilaginous tumors: fast con-
trast-enhanced MR imaging. Radiology 214:539-546
Ginsberg MS, Panieek DM (2000) Subcentimetre pulmonary
nodules detected in patients with sarcoma. Sarcoma
4:63-66
Grampp S, Bankier AA, Zoubek A, Wiesbauer P, Schroth B,
Henk CB, Grois N, Mostbeck GH (2000) Spiral CT of the
lung in children with malignant extra-thoracic tumors: dis-
tribution of benign vs malignant pulmonary nodules. Eur
RadiollO:1318-1322
Greenspan A, Azouz EM, Matheus IJ, Decarie JC (1995) Synovi-
al hemangioma: imaging features in 8 histologieally proven
cases, review of the literature and differential diagnosis.
Skeletal Radio124:583-590
Grey AC, Davies AM, Mangham DC, Ritchie DA (1998) The
"penumbra sign" on Tl-weighted MR imaging in subacute
osteomyelitis: frequency, cause and significance. Clin Radiol
53:587-592
Grimer RJ, Sneath RS (1990) Diagnosing malignant bone
tumours: editorial. J. Bone Joint Surg [Br] 72B:754-756
Hayes CW, Conway WF, Sundaram M (1992) Misleading aggres-
sive MR imaging appearance of some benign musculoskel-
etalIesions. RadioGraphies 12:1119-1134
Hughes TH, Sartoris DJ, Schweitzer ME, Resniek DL (1995) Pig-
mented villonodular synovitis: MRI characteristies. Skel-
etal Radio124:7 -12
Hyman AA, Heiser WJ, Kim SE et al (1995) An excavation ofthe
distal femoral metaphysis: a magnetie resonance imaging
study: case report. J Bone Joint Surg 77A:1897-1901
Janzen DL, Peterfy CG, Forbes JR, Tirman PFJ, Genant HK
(1994) Cystic lesions around the knee joint: MR imaging
findings. Am J RoentgenolI63:155-161
Jelinek JS, Kransdorf MJ, Utz JA, Berrey BH, Thomson JD,
Heekin RD, Radowieh MS (1989) Imaging of pigmented
villonodular synovitis with emphasis on MR imaging. Am
J RoentgenolI52:337-342
Jones BC, Sundaram M, Kransdorf MJ (1993) Synovial sarco-
ma: MR imaging findings in 34 patients. Am J Roentgenol
161:827-830
Kalil RK, Unni KK (1998) Malignancy in pigmented villonodu-
lar synovitis. Skeletal Radio127:392-395
Kaste SE, Neel MN, Rao BN, Thompson VF, PraU CB (2001)
Complications oflimb-sparing procedures using endopros-
thesis replacements about the knee for paediatrie skeletal
sarcomas. Paediatr Radio131:62-71
Keller SM, Rosenbaum RC, Rosenberg SA (1984) The signifi-
A. M. Davies and D. Vanel
cance of bone scan abnormalities in patients with primary
osteogenie sarcoma. J Surg Onco126:122-129
Kenan S, Abdelwahab IF, Klein MJ, Herman G, Lewis MM
(1993a) Lesions of juxtacortieal origin (surface lesions of
bone). Skeletal Radio122:337-357
Kenan S,Abdelwahab IF, Klein MJ et al (1993b) Synovial chon-
drosarcoma secondary to synovial chondromatosis. Skel-
etal Radio122:623-626
Kim MG, Kim BH, Choi JA et al (2001) Intra-artieular ganglion
cysts of the knee: clinieal and MR imaging features. Eur
Radiol11:834-840
Kloen P, Keel SB, Chandler HP, Geiger RH, Zarins B, Rosenberg
AE (1998) Lipoma arborescens of the knee. J Bone Joint
Surg 80B:298-301
Kramer J, Recht M, Deely DM, Schweitzer M, Pathria M, Gentili
E, Greenway G, Resniek D (1993) MR appearance of idio-
pathie synovial chondromatosis. J Comput Assist Tomogr
17:772-776
Kransdorf M, Moser R, Vinh T et al (1989) Primary tumours of
the patella. Skeletal RadiolI8:365-371
Kransdorf MJ, Murphey MD (1997) Soft tissue tumors in a
large referral population: prevalence and distribution of
diagnoses by age, sex and location, In: Kransdorf MJ, Mur-
phey MD (eds) Imaging of soft tissue tumors. Saunders,
Philadelphia, pp 3-35
Krieun ME (1983) Radiographie evaluation of solitary bone
lesions. Orthop Clin North Am 14:39-64
Lang P, Grampp S, Vahlensieck M, Johnston JO, Honda G,
Rosenau W, MaUhay KK, Peterfy C, Higgins CB, Genant HK
(1995) Primary bone tumors: value of MR angiography for
preoperative planning and monitoring response to chemo-
therapy. Am J RoentgenolI65:135-142
Laorr A, Peterfy CG, Tirman PF, Rabassa AE (1995) Lipoma
arborescens of the shoulder: MR imaging findings. J Can
Assoc Radio146:311-313
Lee JK, Yao L (1988) Stress fractures: MR imaging. Radiology
196:217-220
Levin DC, Blazina ME, Levine E (1967) Fatigue fractures of
the shaft of the femur: simulation of a malignant tumor.
Radiology 89:883-885
Liu J, Hudkins PG, Swee RD et al (1987) Bone sarcomas associ-
ated with Ollier's disease. Cancer 59:1376-1385
Llauger J, Monill JM, Palmer J, Clotet M (1995) Synovial hem-
angioma of the knee: MR imaging findings in two cases.
Skeletal Radio124:579-581
Mack TM (1995) Sarcomas and other malignancies of soft
tissues, retroperitoneum, pleura, heart, mediastinum and
spleen. Cancer 75 [Suppl]:211-244
Madewell JE, Ragsdale BD, Sweet DE (1981) Radiologie and
pathologie analysis of solitary bone lesions, part 1. Internal
margins. Radiol Clin North Am 19:715-748
Martin S, Hernandez L, Romero L, Lafuente J, Poza AI, Ruiz
P, Jimeno M (1998) Diagnostie imaging of lipoma arbore-
scens. Skeletal Radio127:325-329
McKinney CD, Mills SE, Fechner RE (1992) Intraartieular syno-
vial sarcoma. Am J Surg Patholl6: 10 17-1020
Milgram JW (1977) Synovial osteochondromatosis. A his-
topathologieal study of thirty cases. J. Bone Joint Surg
59A:792-801
Morrison JL, Kaplan PA (2000) Water on the knee: cysts, bursae
and recesses. MRI Clin North Am 8:349-370
Moser RP, Madewell JE (1987) An approach to primary bone
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

A chronic recurrent multifocal osteomyelitis 258

accessory ossification centres 9 Codman's angle/triangle 8,310,324
aggressive fibromatosis 32 collateralligaments 52,53,56
aneurysmal bone cyst 310-315,322,333 computed tomography (CT) 6,10,23-40,117
anterior cruciate ligament - anatomy 30, 31
- anatomy 51-53,56,73,83,135,154,155 - arthrography 32,35,37,38
- avulsion 33 - contrast enhancement 36
- congenital absence 83 - fluoroscopy 25,26,38,39
- ganglion 32,39,160 - helical 24,25,26,32
- replacement/reconstruction 57,58,59,70,165-168, - Hounsfield units 27,28
172-183 - image quality 27
- rupture/tear 15,16,20,117,122,123,153,155-157 - infection 254,255
arthrography - interventional 25, 38, 39
- conventional arthrography 19-21,37,49,79,129,153,283, - measurements 33
299 - metal artefact 27
- CT arthrography 19,32,35,37,38,283,299 - multislice 25,26,35
- MR arthrography 19,47-50,54,56,57,132,148,175,184, "cyclops" lesion 179
187,188,300
arthrodesis 211,212
arthroplasty 199-211 D
- infection 207-209 diabetes 250,258, 260
- loosening 206, 207 diaphyseal aclasis 86,308,324
articular/hyaline cartilage 53-56,66,68,283-291
- postoperative 171, 172
- transplantation 189-192 E
- volume & thickness measurement 288, 289 effusion 67,110,249,256,273
avascular necrosis see osteonecrosis enchondroma see chondroma
avulsive cortical irregularity 332 eosinophilic granuloma 313,315
Erlenmyer flask deformity 86,87
Ewing's sarcoma 307,308,310,311,313,315,318,319,325,
B
332
Blount's disease (tibia vara) 78,79
exostosis see osteochondroma
bone marrow oedema/bone bruise 48,52,53,56,157,161
extra-abdominal desmoid see aggressive fibromatosis
Brodie's abscess 258, 259
brown tumour 325,333
bursa/bursitis 327
F
- gastrocnemius-semimembranosus (Baker's cyst) 71,72,
fabella 74,92,93
250,259,260,272,274,327
fibrous cortical defect/non-ossifying fibroma 9,278,
- prepatellar 69,218,327
310-312,321,322,325
- septic 259,260
fibrous dysplasia 81,310-312,322,325
fracture
C - distal femur 26
calcium hydroxyapatite disease 272,280 - intraarticular 32
calcium pyrophosphate deposition disease (CPPD) 272,280, - osteochondral 122
281 - Ottawa knee rule 110
chondroblastoma 310,314,320,321,325 - patella 6, 114, 115
chondroma 308-311,320,324 - Salter Harris 111-113
- intracapsular/para-articular 331 - Segond 33,122,123,153,161
chondromyxoid fibroma 310-312,320,322 - supracondylar 110
chondrosarcoma 307,310,311,313,324,325 - tibial plateau/condyle 6,33,116,117
- extraskeletal 326 - tibial spine 117, 118
338 Subject Index

G - spatial resolution 44,45,129,316

ganglion cyst 38,71,327,328 - spin echo/fast spin echo 46,47,48,51,53,54,57,130,174,
- intraosseous ganglion 311 285,316
Gaucher's disease 87,88,302,308 - STIR 48,53,56,57,59,156,235,254,260,316,317,319,323
genu - truncation artefact 54,58,131,137,285
- recurvatum 80 - volume averaging 137
- valgum 77,78 - vacuum phenomenon 58,131
- varum 77, 78, 80 malignant fibrous histiocytoma 307,325
giant cell tumour 278,310,311,322,323,325,326,333 medial collateralligament 70,117,121,160,161
gout 35,272,280,281,302 meniscofemoralligaments 134,137,138
growth arrest lines 9 meniscus 50-52,129-152
- anatomy 133-139
- bucket handle tear 50
H - computed tomography 32
haemangioma - cyst 71,146,147
- bone 313 - discoid 83,139,146-148
- soft tissue 327 - fiounce 131,139
- synovial 328,330,331 - ossicles 37,70,147
haemophilia 85,272,276,279,328,330 - postoperative 49,51,57,148,170,171,182-188
Hoffa's fat pad 67,69,179 - tear 20,37,51,70,130,132,139-148
hypophosphataemie riekets 81 metastasis 308,309,313,318,325
hypertrophie osteoarthropathy 276,278 multiple epiphyseal dysplasia 89
myeloma 308,311,324
myositis ossificans 255,315,326,326,327

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

patella 217 - 248 - measurement teehniques 14-16

- anatomy 218-221 - osteoarthritis 10-13
- alignment/traeking 34,35,57,91,92,195,222,223,230-235 - projeetions 6-10
- alta 16,91 - stereophotogrammetry 15
- baja 16 - stress views 10, 11 0
- bipartite 9,68,90 - tomography 5,6, 10, 117
- eartilage 36,49,55 rheumatoid arthritis 269,272,274-276
- ehondromalacia 68,218,223 riekets 80,89,90, 106
- eomputed tomography 34,35,229,230,231-233 rubella 84
- disloeation 10,83,90-92,124,125,218,238-243
- fraeture 5,10,121
- jumper's knee 121,218,243,244 S
- osteoarthritis 235-237 sehwannoma 326
- radiographie projeetions 7,8,91,224-228 scintigraphy 251-253
- tendon 9,57,65,66,69,118,217,218 seurvy 106
- tumours 308,325,326 septie arthritis 20,258,259,270,279
Pelligrini-Stieda lesion 121,153 siekle eell disease 263-265,302,309
periosteal desmoid see avulsive eortieal irregularity simple bone eyst/unieameral bone eyst310,311,321,322,
peroneal nerve 66, 73 333
pigmented villonodular synovitis 36,272,277,314,328-331 Sinding-Larsen-Johansson disease 9,121,218
plasmaeytoma 313 soft tissue sareoma 36
popliteal eyst 20 staging of tumours 315
popliteus tendon 57,133,137,138 stress fraetures 126,314,332,333
posterior cruciate ligament suprapatellar poueh 4,5,9,66,67
- anatomy 73,158 synovial ehondromatosis 277,328,329
- reeonstruetion 169,170,171,173,174,178 synovial ehondrosarcoma 332
- rupture/tear 53,168 synovial pliea 37,56,67
prosthesis 13, 15 synovial sareoma (synovioma) 277,278,326,331,332
pyogenie myositis 259 syphilis 84, 106

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

M. E. ABD EL BAGI, MD, DMRD, FFRRCSI S. CHAPMAN, MD

Senior Consultant Radiologist (Teaching) Consultant Paediatric Radiologist
Department of Radiology Birmingham Children's Hospital
Riyadh Armed Forces Hospital Steelhouse Lane
PO Box 7897 Birmingham, B4 6NH
Riyadh 11159 UK
Kingdom of Saudi Arabia
A. M. DAVIES, MD
M. S. AL SHAHED, MD, FRCR Consultant Radiologist
Department of Radiology MRI Centre
Riyadh Armed Forces Hospital Royal Orthopaedic Hospital
PO Box 7897 Bristol Road
Riyadh 11159 Birmingham, B31 2AP
Kingdom of Saudi Arabia UK

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

S. N. J. ROBERTS, MA, FRCS (Orth) S.ZAIM,MD

Consultant Orthopaedic and Sports Injury Surgeon
The Robert Jones and Agnes Hunt
Orthopaedic and District Hospital
Oswestry
Shropshire, SYlO 7AG
UK
Osteoporosis and Arthritis Research Group
Department of Radiology
University of California San Francisco
350 Parnassus Avenue, Suite 150
San Francisco, CA 94143-1349
USA