Comparison of high-field and low-field magnetic resonance images of cadaver limbs of horses.

Abstract

Eleven limbs taken postmortem from 10 lame horses were examined by MRI in a low-field 0.27T system designed for standing horses and a high-field 1.5T system used to examine anaesthetised horses. Nine limbs were examined in the foot/pastern region and two in the fetlock region, and the results were compared with gross pathological examinations and histological examinations of selected tissues. The appearance of normal tissues was similar between the two systems, but the anatomical arrangement of the structures was different due to differences in positioning, and a magic angle artefact was observed at different sites in some imaging sequences. Articular cartilage could be differentiated into two articular surfaces in most joints in the high-field images but could generally be separated only at the joint margins in the low-field images. Abnormalities of tendon, ligament and bone detected by gross examination were detected by both forms of MRI, but some details were clearer on the high-field images. Articular cartilage found to be normal on pathological examination was also classified as normal on MRI, but lesions in articular cartilage detected on pathological examination were identified only by high-field MRI. An abnormality was detected on MRI of all the limbs that had abnormal navicular flexor fibrocartilage on pathological examination.

Full text

Page 1

September 5, 2009 | the VETERINARY RECORD
PAPERS
Comparison of high-field and low-field
magnetic resonance images of cadaver
limbs of horses
R. C. Murray, T. S. Mair, C. E. Sherlock, A. S. Blunden
R. C. Murray, MA, VetMB,
MS, DipECVS, DipACVS, PhD,
MRCVS,
A. S. Blunden, BVetMed, PhD,
MRCPath, MRCVS,
Animal Health Trust, Lanwades Park,
Kentford, Newmarket CB8 7UU
T. S. Mair, BVSc, PhD, DipECEIM,
EIM, DESTS, MRCVS,
C. E. Sherlock, BVetMed, MRCVS,
Bell Equine Veterinary Clinic,
Mereworth, Maidstone, Kent
ME18 5GS
Correspondence to Dr Mair, e-mail:
tim.mair@btinternet.com
Veterinary Record (2009) 165, 281-288
Eleven limbs taken postmortem from 10 lame horses
were examined by MRI in a low-field 0·27T system
designed for standing horses and a high-field 1·5T
system used to examine anaesthetised horses. Nine
limbs were examined in the foot/pastern region
and two in the fetlock region, and the results were
compared with gross pathological examinations and
histological examinations of selected tissues. The
appearance of normal tissues was similar between
the two systems, but the anatomical arrangement
of the structures was different due to differences in
positioning, and a magic angle artefact was observed
at different sites in some imaging sequences.
Articular cartilage could be differentiated into two
articular surfaces in most joints in the high-field
images but could generally be separated only at the
joint margins in the low-field images. Abnormalities
of tendon, ligament and bone detected by gross
examination were detected by both forms of MRI, but
some details were clearer on the high-field images.
Articular cartilage found to be normal on pathological
examination was also classified as normal on MRI, but
lesions in articular cartilage detected on pathological
examination were identified only by high-field MRI.
An abnormality was detected on MRI of all the limbs
that had abnormal navicular flexor fibrocartilage on
pathological examination.
NINETY-FIVE per cent of forelimb lameness in horses can be attrib-
uted to pain in the lower limb (Adams 1957, Ross 2003). Conventional
imaging techniques have limitations in the evaluation of this region,
and MRI is increasingly being used to diagnose the causes of lameness.
High-field MRI has been demonstrated to be an excellent tool for
the detection of abnormalities in the feet of horses (Schramme and
others 2001, Dyson and others 2003a, b, 2004, 2005, Schneider and
others 2003, Dyson and Murray 2004, Kristoffersen and others 2004,
Zubrod and others 2004, Boado and others 2005, Murray and Mair 2005,
Blunden and others 2006a, b, Murray and others 2006a, b), and can be
used to evaluate the limb up to and including the carpus and tarsus.
However, although most published work has been done using high-field
MRI systems designed for human beings, the expense and size of high-
field units limit their use to specialist centres. A standing low-field MRI
system has been available since 2002, and is potentially affordable and
practical for use in clinical practice (Mair and others 2003, 2005). Initial
studies using the low-field unit have revealed a similar range of lesions as
those revealed by high-field scanners (Mair and others 2003, Kinns and
Mair 2005, Mair and Kinns 2005, Sherlock and others 2007).
Many factors can affect the quality of MR images, and it is impor-
tant to understand the characteristics, benefits and limitations of each
system for the interpretation of the images of horses’ limbs and for the
diagnosis of lesions that result in lameness. One of these factors is the
field strength of the magnets. High-field magnets have a better signal-
to-noise ratio and should produce images with better resolution, which
could improve the identification of small and low-contrast lesions
compared with low-field magnets. Several studies have compared the
results of high-field MRI with the gross and histopathological changes
observed in the limbs of lame horses (Schramme and others 2005,
Zubrod and others 2005, Murray and others 2006a, b, Dyson and others
2008), but no similar comparisons of low-field imaging with pathology,
or of high-field with low-field MRI of the same limbs, have been pub-
lished. This paper describes the characteristics of the images acquired
from the same cadaver limbs of 10 lame horses by using a high-field
and a low-field MRI system and compares the images for the detection
of common lesions.
Materials and methods
Eleven limbs were examined from 10 randomly selected lame horses that
had been euthanased for clinical reasons (Table 1). The lameness had
been localised within 24 hours before the horses were euthanased, and
these regions were examined by MRI. The limbs were stored frozen at
–20°C in sealed plastic bags until they were examined.
Image acquisition
The limbs were thawed for 24 hours before they were imaged. The low-
field images were acquired with a 0·27 T standing MRI unit (Hallmarq
Veterinary Imaging). Gradient echo, fast spin echo and short tau inver-
sion recovery (STIR) sequences were acquired in three planes: sagittal,
dorsal and transverse (Table 2). A radiofrequency solenoid coil was used

Page 2

the VETERINARY RECORD | September 5, 2009
PAPERS
for image acquisition; to image the foot and fetlock, modified solenoid
coils shaped to fit the region were used. During the examinations the
limbs were stabilised in a position to simulate that in a standing horse,
using a wooden block and foam padding. To image the foot and pastern
regions, sagittal plane images were aligned perpendicular to the dor-
sal aspect and parallel to the medial and lateral aspects of the second
phalanx. Dorsal images were aligned parallel to the dorsal aspect of the
second and distal phalanges. Transverse images were aligned perpen-
dicular to the deep digital flexor tendon in the pastern region and in
the region distal to the navicular bone; at the level of the navicular
bone, transverse images were aligned perpendicular to the centre of the
palmar border of the navicular bone. The limbs were re-frozen before
being examined by high-field imaging.
The high-field images were acquired with a 1·5 T Signa Echospeed
system (GE), using gradient echo and STIR sequences (Table 3). During
image acquisition, the limbs were placed in the centre of a human
extremity radiofrequency coil, as though in right lateral recumbency in
a live horse, with the distal interphalangeal joint at an angle of between
170 and 190°C. Sagittal, dorsal and transverse images were aligned, as
in the studies with the low-field system.
Pathological examination
Each limb was examined thoroughly by a pathologist unaware of the MRI
findings, using a standard dissection protocol (Blunden and others 2006a,
b), and the gross findings and abnormalities were recorded (Table 3) and
a photographic record obtained. Samples from tissues with a suspected
abnormality were prepared routinely, stained with haematoxylin and
eosin, and examined microscopi-
cally as described by Blunden and
others (2006a, b).
Image analysis
An offline workstation (Ultra
10; Sun microsystems) with dedi-
cated software (GE Advantage
Windows 3.1) was used to analyse
the high-field images. The inte-
grated image analysis features of
the Hallmarq MRI scanner soft-
ware running on a Windows XP
workstation were used to analyse
the low-field images. The defini-
tion and appearance of differ-
ent structures and tissues were
described for each type of image.
The images from each limb were
analysed by two trained analysts
(both blinded to the results of the
pathological examination and one
blinded to each horse’s history) for
abnormalities, using a standard
image-reading protocol to include
the assessment of each tissue in a
repeatable order for signal inten-
sity and homogeneity, clarity of
margins and relationships with
other tissues.
Data analysis
A descriptive comparison of the
appearance of the tissues on the
high-field and low-field images
was made to assist in understand-
ing the relative appearance of the
anatomical structures on the two
MRI systems. A description of the
pathological features of the abnor-
malities detected on the images
was made to assist in determining
the usefulness of each system for showing specific features. The findings
on MRI were compared with the results of the pathological examina-
tions to suggest which abnormalities were most clearly defined by each
imaging technique.
Results
Positioning and plane of image acquisition
The appearance of some anatomical structures varied with the plane of
the image and the position of the limb, due to the standing, extended
position of the low-field images compared with the recumbent, flexed
position of the high-field images. The structures affected included the
deep digital flexor tendon (DDFT) and the collateral ligaments of the
distal interphalangeal joint.
Appearance of images
In both the high-field and low-field images, the definition of tissue
edges on both T2- and T2*-weighted images was less marked than on
T1-weighted images. Some details of the tissues were less clear on low-
field images, especially at the tissue margins, and there were some subtle
differences in signal intensity within them. These were probably related
to poorer resolution of the images, thicker slices and a difference in fil-
tering level leading to smoothing of the images, and were not considered
to be explained by differences in positioning.
Artefacts
Magic angle effect On T1-weighted high-field images, the magic angle
effect led to increased signal intensity within the DDFT distal to the
TABLE 1: Characteristics of 10 horses and the pathological findings observed in the limbs that were studied
Horse
Age (years),
breed, sexLimb
Region of
interestSummary of pathological findings
1 13, Haflinger, G LFFootEvidence of penetrating wound with a necrotic tract from the lateral heel bulb to
the navicular bursa; severe necrosis of DDFT with rupture/shedding of tendon
lateral>medial; navicular bone fibrocartilage abnormal, especially midline to lateral;
arthropathy of distal aspect of middle phalanx
Complete rupture of medial collateral ligament and partial rupture of LCL, and
rupture of medial joint capsule; articular cartilage damage on distal third metacarpal
and proximal phalanx especially on lateral aspect; damage to distal sesamoidean
ligaments
Dorsal soft tissue damage and disruption with periosteal irregularity and thickening
overlying bone damage and necrosis; common digital extensor tendon and DSIL
abnormal
Foot/pastern Lateral condylar fracture of middle phalanx with partial rupture of the lateral
collateral ligament and complete rupture of the DSIL; loss of articular cartilage on
the mid to lateral aspect of the articular surface of the distal phalanx
FootEmaciated horse with history of malabsorption; evidence of medullary fat atrophy/
necrosis, interstitial oedema and capillary infiltration; medial palmar arthrosis of
distal aspect of middle phalanx
Foot Full-depth cartilage defect and underlying subchondral bone irregularity on dorsal
aspect of the articular surface of the distal phalanx approximately 5 mm in diameter;
partial thickness cartilage erosion on the distal aspect of the middle phalanx
FootNavicular bone fibrocartilage damage; distal navicular bone irregular; DSIL has
severe damage including enthesiophytes at the origin and insertion; DDFT has
pitting, fibrillation and altered structure of the dorsal surface at its insertion
Fetlock Open skin wound on the plantarolateral aspect of the fetlock region with thickening
and haemorrhagic/oedematous soft tissues; lateral proximal sesamoid bone
abnormal with abaxial loss of cortical bone; medulla, especially abaxially, has
extensive bone loss with abaxial loss of cortical bone; medulla, especially abaxially,
has extensive bone loss with osteonecrosis and oedema, fat necrosis, fibroplasia
and spiculated bone fragments; bone abnormal at origin of lateral oblique distal
sesamoidean ligament
FootSevere navicular bone fibrocartilage and bone defect on midline in the distal third
at site of adhesion to the DDFT; severe dorsal DDFT pathology; continued distal
abnormality but easier to separate fibrinous/fibrous tissues between DDFT and
DSIL; DSIL abnormal; marked synovial proliferation in navicular bursa; articular
cartilage erosion on lateral condyle of distal aspect of middle phalanx
FootNavicular bone fibrocartilage abnormal; dorsal DDFT abnormal between navicular
bone and insertion; CSL thickened and abnormal on palmar aspect; DSIL abnormal
Foot Navicular bone fibrocartilage abnormal; dorsal DDFT abnormal at navicular bone
and bursa; DSIL abnormal including enthesiophytes and the origin
25, Warmblood
cross, F
RFFetlock
3 2, Pony, GLFFoot
4 4, Hanoverian
cross, G
RH
5 2, Warmblood, GLF
6 8, Warmblood, GLF
76, Warmblood, G RF
8 7, Thoroughbred, FRH
9 7, Thoroughbred, G RF
105, Hanoverian, G RF
LF
CSL Collateral sesamoidean ligament, DDFT Deep digital flexor tendon, DSIL Distal sesamoidean impar ligament, F Female, G Gelding,
LCL Lateral collateral ligament, LF Left forelimb, RF Right forelimb, RH Right hindlimb

Page 3

September 5, 2009 | the VETERINARY RECORD
PAPERS
navicular bone, in the distal collateral sesamoidean ligament (CSL)
and the distal sesamoidean impar ligament (DSIL) compared with the
T1-weighted low-field images, where the structures had lower signal
intensity. Some T1- and T2*-weighted images from the low-field
system demonstrated asymmetry in signal intensity within the collateral
ligaments of the distal interphalangeal joint and medial and lateral
oblique sesamoidean ligaments. Repositioning the limb or using T2 fast
spin echo scans produced images with a more symmetrical low signal
intensity in both ligaments.
Partial volume effect The partial volume effect was more obvious on
the low-field images, probably in association with the thicker slices and
wider spacing between slices on these images.
Appearance of tissues
DDFT The normal DDFT had a uniform low signal intensity on T1-,
T2- and T2*-weighted sequences. The high-field images showed some
septal separation that was not clear on the low-field images. Distal to
the navicular bone and near to the insertion there was increased signal
intensity on the high-field images, which was most marked on T1-
weighted images, but was absent on low-field images. The dorsal aspect
of the DDFT was smooth and clearly separated from navicular bursal
fluid on T2- and T2*-weighted images for both types of magnet.
Large tendon defects, core lesions or surface irregularity were clearly
visible as increased signal intensity or a defect in the continuity of the
tendon on both high- and low-field images of six limbs (Fig 1). Due to
the different image orientation, the configuration of the lesion differed
between the two systems for selected images on all the limbs, but the
overall interpretation remained
the same. Small focal lesions
(approximately 1 mm diameter)
visible on high-field images were
not detected on low-field images
of seven of the limbs. There was
a tendency for focal increases
in signal intensity on transverse
high-field images to have better
edge contrast and more varia-
tion in size between slices and a
greater maximum intensity than
on the low-field images. On low-
field images these focal changes in
intensity tended to be more grad-
ual between the edge of the visible
lesion and the adjacent tendon,
with less variation in size between
slices and a lower maximum
intensity, probably as a result of
partial volume effects. Subtle dor-
sal fibrillation at the level of the
navicular bursa, including at the
navicular bone and close to the
insertion that was detectable on
high-field images, was not clearly
visible on low-field images of two
limbs. Tendon splits were detect-
able on both high- and low-field
images of three limbs. For some
lesions, areas of poorly defined
changes in structure on low-field
images could be clearly defined by
lesion type on high-field images
(Fig 2). On T1-weighted images
of the region between the distal
navicular bone and the insertion,
the magic angle effect in the high-
field images made it more difficult
to detect tendon splits or core
lesions than on T1-weighted low-
field images. However, on T2- and T2*-weighted images these lesions
could be detected in both high- and low-field images.
Adhesions between the DDFT and the CSL, DSIL or navicular bone
were detected on both low- and high-field images of four limbs, but they
were more difficult to define in low-field images of one limb with an
adhesion at the level of the distal navicular bone. Increased signal inten-
sity on STIR images was detected in both high- and low-field images of
four lesions, but one lesion involving the dorsal aspect of the insertion
of the DDFT was easier to define on high-field images.
Ligaments The normal DSIL was easier to define on high-field
images, especially in the transverse plane, and the contrast between
the high signal intensity of the adjacent synovial fluid made the
ligament clearer on T2- and T2*- than T1-weighted images from
both systems. The DSIL on high-field T2-weighted images was a well-
defined structure of low signal intensity, but on low-field images its
appearance varied between a vague decrease in signal intensity and
a more clearly defined structure of low signal intensity. The ability
to visualise the DSIL better in cross-section (on transverse images)
on high-field images appeared to relate to a difficulty in obtaining
images at a consistent level on low-field images, probably due to the
greater slice thickness and inter-slice spacing. However, it was possible
to clearly define the DSIL in longitudinal orientation on low-field
sagittal images. A rupture of the DSIL in one limb was visible on both
high- and low-field images, as were bony irregularities at the origin or
insertion, and a marked structural abnormality of the ligament, in four.
However, in one limb, only high-field images gave a clear definition of
enthesiophyte formation. Palmar irregularity was also observed only
TABLE 2: Parameters used in pulse sequences for imaging horses’ limbs with a low-field MRI system
Pulse sequencePlane TE (ms) TR (ms) FEPE
FOV
(mm)
Slice thickness
(mm)
Interslice
spacing (mm) Imaging options
Pilot
STIR FSE
T1-weighted
3D HR
T1-weighted
GRE HR
T2*-weighted
3D HR
T2*-weighted
GRE HR
T2-weighted FSE Sagittal/dorsal/transverse
NA
Sagittal/transverse
Sagittal/dorsal/transverse
7
28
8
62
2910
23
150
340
340
120
175
130
220
175
170
7
5
2·5
NA
1
0
No phase wrap
Sagittal/dorsal/transverse8 92384 192170 3·50·7Motion correction
Sagittal/dorsal/transverse1332384192170 2·50
Sagittal/dorsal/transverse13 135384192 1703·5 0·7 Motion correction
842000340 17517051 Motion correction
FE Frequency encoding, FOV Field of view, FSE Fast spin echo, GRE Gradient echo, HR High resolution, NA Not applicable, PE Phase
encoding, STIR Short tau inversion recovery, TE Echo time, TR Repetition time
TABLE 3: Parameters used in pulse sequences for imaging horses’ limbs with a low-field MRI system
Pulse sequence* Plane TE (ms) TR (ms) FEPE
FOV
(mm)
Slice thickness
(mm)
Interslice
spacing (mm) Imaging options
3D SPGR Sagittal3·28·0 25625628031·5NPW, VBW, EDR, ZIP2,
Zip512
NPW, VBW, EDR, ZIP2,
Zip512
VBW
NPW, VBW, EDR, ZIP2,
Zip512
NPW, VBW, EDR, ZIP2,
Zip512
VBW
NPW, VBW, EDR, ZIP2,
Zip512
NPW, VBW, EDR, ZIP2,
Zip512
VBW
3D T2* GRESagittal 1·4 5·6 2562562803 1·5
STIR
3D SPGR
Sagittal
Dorsal
25·4
3·3
10,500
8·1
256
256
192
256
220
260
4
3
1·0
1·5
3D T2* GRE Dorsal1·5 5·8256 25626031·5
STIR
3D SPGR
Dorsal
Transverse
24·8
3·4
5000
8·4
256
256
192
256
260
220
4
3
1·0
1·5
3D T2* GRETransverse 1·45·8 256192 2203 1·5
STIR Transverse25·310,500 2561922204 1·0
* All 3D sequences used imaging options of variable band width (VBW), no phase wrap (NPW), extended dynamic range (EDR), zerofill
interpolation processing (ZIP 512) to reconstruct the image to a 512/512 matrix and slice ZIP2 to double the number of reconstructed
slices within the prescribed range. 2D sequences used variable band width. For the fast STIR sequence, an inversion time of 120 ms and
echo train length of 4 were used
FE Frequency encoding, FOV Field of view, PE Phase encoding, TE Echo time, TR Repetition time

Page 4

the VETERINARY RECORD | September 5, 2009
PAPERS
on high-field images of four limbs. However, soft tissue interposed
between the DSIL and the DDFT was detected easily by both systems,
and this was present in all cases of palmar irregularity. When there
was increased signal intensity at the origin and insertion of the DSIL
on STIR images, it was clearly detectable by both systems (Figs 2, 3),
even when it was more difficult to detect structural changes in the
ligament on low-field images.
The normal CSL had a moderate signal
intensity on T2-weighted high-field images
but a low signal intensity on T2 and T2*-
weighted low-field images. On T1-weighted
high-field images the CSL had a moderate
to high signal intensity (possibly related to
the magic angle effect) and was less easy to
define than on low-field images, in which the
ligament had a lower signal intensity.
Adhesions between the CSL and the
DDFT were detected on both high- and low-
field images.
The normal DSL had better contrast with
adjacent tissues on T2 and T2*-weighted
than on T1-weighted high- and low-field
images. Although T2- and T2*-weighted
images of the oblique distal sesamoidean
ligaments had a lower signal intensity than
T1-weighted images, the ligaments had a
slightly heterogeneous moderate signal inten-
sity on both T1- and T2-weighted high-field
images. No abnormalities were observed in
the DSL of any of the limbs.
The normal collateral ligaments (CLs) of
the proximal interphalangeal and metacar-
pophalangeal joints had a similar low signal
density and appearance on high- and low-
field T1-weighted, T2- and T2*-weighted
and STIR images. The CLs of the distal inter-
phalangeal joint had a low signal intensity on
all sequences on high-field images, but there
was some asymmetry in the signal intensity
of T1- and T2*-weighted images when using
the low-field system, attributed to the magic
angle effect.
Enlargement or marked disruption of the
CL was clearly detectable on both systems.
Complete rupture of the metacarpophalangeal
joint CL appeared on both systems as a loss of
continuation of the ligament with slight wid-
ening of the remaining ligament proximal or
distal to the rupture site, probably as a result of
retraction, but with ragged edges (Fig 4). The
partial rupture of two ligaments appeared as a
partial disruption of the ligament, with local
increased signal intensity of T2-weighted and
STIR images but decreased signal intensity on
T1-weighted images, suggesting an accumula-
tion of fluid within the ligament.
Bone On T1-weighted images, normal bone
had a clearly defined cortex and medulla
in both the high- and low-field images.
However, on the T2*-weighted images,
the stronger T2 weighting on the low-field
images made it more difficult to define the
margin between the cortex and medulla
than on the high-field images, in which the
medulla had a higher signal intensity.
Marked bone disruption, cortical defects
and changes in the medullary signal intensity
were clearly evident on both systems (Fig 5).
Moderate periosteal thickening and irregularity were clearly visible on
both low- and high-field images. Low signal intensity on T1-weighted
images and an increase in signal intensity on STIR images were visible in
both systems. One 2 mm defect in the distal phalanx observed on high-
field images was not detected on low-field images, but on T2-weighted
images a distal metacarpal condylar fracture line was more clearly visible
on the low-field than on the corresponding high-field image.
FIG 1: Comparison of (a) high-field (SPGR sequence) and (b) low-field (T1 GRE sequence)
transverse images of the foot of horse 10, demonstrating a focal high lesion at the dorsal
aspect of the medial lobe of the deep digital flexor tendon in the region just proximal to the
navicular bone (arrow). Medial is to the left of the image
(a) (b)
(a)(b)
FIG 2: Comparison of (a) high-field (STIR sequence) and (b) low-field (STIR sequence) sagittal
images of the foot of horse 7 with abnormalities in the deep digital flexor tendon (DDFT)
and distal sesamoidean impar ligament (DSIL). A focal area of increased signal is visible in
the DDFT close to its insertion into the distal phalanx, confirmed as a pathological change
on gross examination. The signal intensity is slightly increased in the body of the DSIL, and
markedly increased at its origin on the navicular bone and its insertion into the distal phalanx,
in both the high- and low-field images. On pathological examination there was severe damage
to the DSIL, particularly at its origin and insertion
(a)(b)
FIG 3: Signal intensity is high in STIR sequences in (a) high-field and (b) low-field images of the
fetlock of horse 2 at the insertion of the cruciate distal sesamoidean ligaments (arrow)

Page 5

September 5, 2009 | the VETERINARY RECORD
PAPERS
In the navicular bone, mild endosteal irregularity of the flexor border
was easily detectable on high-field T1- and T2-weighted images of three
of the limbs. On low-field images, endosteal irregularity was more easily
detectable on T2- than T1-weighted images, but mild changes were less
clearly defined than on the high-field images for two of the three limbs.
Abnormalities of the flexor surface of the navicular bone, including an
irregularity of the flexor subchondral bone, a
flexor subchondral bone defect or an accu-
mulation of fluid palmar to the bone, were
detected on both high- and low-field systems,
but these abnormalities were more clearly
defined on the high-field images. Increased
signal intensity on STIR was clearly evident
on both high- and low-field images, particu-
larly extending as a band-shaped signal from
the insertion of the CSL to the origin of
the DSIL in seven of the limbs. Distal bor-
der fragments were visible on both systems,
and were especially obvious on dorsal/frontal
plane images (Fig 6).
In horse 5, there was no fat signal in the
medulla, giving the appearance of fat sup-
pression on all the images except the STIR
images, in which there was a high signal
intensity in the medulla, suggesting that the
marrow fat had been replaced by fluid.
Articular cartilage On both high- and low-
field images normal articular cartilage gave
a moderate signal intensity, with higher
intensity on T1- than T2-weighted images.
The cartilage surface was most clearly defined
when adjacent to the contrast of synovial
fluid. On high-field images it was possible
to separate the proximal and distal articular
surfaces in all the joints examined. However,
on low-field images the cartilage from each
articular surface was visible separately only
at the joint margins, and for most of the joint
surface the cartilage appeared as a single
continuous line (Fig 7).
Irregularities of the cartilage surface were
clearly visible on high-field images, in which
the two articular surfaces could be seen as
separate layers in six limbs, but these defects
were detected on low-field images in only two
limbs (Fig 8). When there were concurrent
changes in the subchondral bone, they were
visible on both high- and low-field images, in
three of the limbs.
Synovium, synovial fluid and periarticular
tissues Normal synovial fluid and spaces were
clearly evident on T2-weighted high-field
images, but the synovial fluid pouches of the
navicular bursa and the distal interphalangeal
joint were more difficult to visualise on low-
field images. On T2-weighted images the
joint capsule had a lower signal intensity on
low-field than on high-field images.
Chronic synovial proliferation in the
navicular bursa was visible in both high- and
low-field images, replacing the low signal
intensity of the synovial fluid on T1-weighted
images and the high signal intensity on
T2-weighted images. In two cases, thicken-
ing, disruption or rupture of the joint capsule
with changes in signal intensity were easily
detected on both systems. In one case, disrup-
tion, fluid accumulation, with increased signal intensity on T2-weighted
images, and swelling of periarticular soft tissues were observed on both
systems. In horse 5, there was an absence of the expected fat signal
in the palmar soft tissues with both MRI systems, and it was appar-
ently replaced in the digital cushion by a fluid signal, probably related
to emaciation.
(a)
(b)
FIG 4: (a) High-field (SPGR sequence) and (b) low-field (GRE sequence) images, showing
a complete rupture of the medial collateral ligament of the metacarpophalangeal joint of
horse 2 (arrow). The rupture is visible as a loss of continuity, increased signal intensity and
disruption of local anatomy in both images. Medial is to the left of the images
(a)
(b)
FIG 5: (a) High-field (SPGR sequence) and (b) low-field (T1 GRE sequence) transverse images
of the foot of horse 3, showing damage to the dorsal aspect of the middle phalanx, common
digital extensor tendon and overlying soft tissues. Both images clearly show swelling and
increased signal intensity of the dorsal soft tissues to the left of the image (arrow). There is
a large area of low signal intensity in the dorsal medulla of the middle phalanx surrounding
a focal area of high signal intensity involving the dorsal cortex; both the high- and low-field
images are oblique images, so that the size and shape of the collateral ligaments of the distal
interphalangeal joint and osseous structures are different on medial and lateral sides
(a) (b)
FIG 6: (a) High-field (SPGR sequence) and (b) low-field (T1 GRE sequence) dorsal images
showing fragmentation (arrow) of the distal border of the navicular bone of horse 7