JBR–BTR, 2005, 88: 1-6.

MAGNETIC RESONANCE IMAGING OF THE INFRAPATELLAR FAT PAD:

CORRELATION WITH PATELLAR ARTICULAR CARTILAGE ABNORMALITIES
A.C. Cañizares Pérez, B. Stallenberg1

We tried to demonstrate an association between magnetic resonance imaging findings of the Hoffa fat pad and
patellar chondropathy. Parallely, we checked the correlation between the diagnosis of patellar chondropathy on mag-
netic resonance imaging and during arthroscopy.
Our study is a retrospective review of the records of 135 patients who underwent an arthroscopy and MRI study at
our institution between October 1997 and January 2001. Magnetic resonance images of the Hoffa fat pad were inter-
preted and typewritten arthroscopy reports were recorded.
A patellar chondropathy assessed during arthroscopy was present in 64 of 135 patients. Twenty of them had abnor-
mal signal intensity in Hoffa fat pad with a sensitivity of the magnetic resonance imaging findings of 31% and a
specificity of 73%. We didn’t find any significant association in the different correlations between signal abnormali-
ties of Hoffa fat pad and patellar chondropathy. On the other hand, there was a significant association between the
results of patellar chondropathy on magnetic resonance imaging findings and during arthroscopy.
No significant association was shown between the MRI findings of Hoffa fat pad and the patellar chondropathy.

Key-word: Knee, MR.

The infrapatellar fat pad is bor-
dered by the inferior pole of the
patella superiorly, the joint capsule
and patellar tendon anteriorly, the
proximal tibia and deep infrapatellar
bursa inferiorly and the synovium-
lined joint cavity posteriorly.
Infrapatellar fat pad impingement
was first described by Hoffa in 1904
and is a poorly described condition
that may in fact be a common cause
of anterior knee pain (1). The
impingement process was initially
reported to be secondary to minor
chronic trauma witch pinches the fat
pad resulting in an acute inflamma-
tory response (2). Etiology of this
pathology is not clear. In addition to
the traumatic cause, swelling of the
retropatellar fat pad is seen in
chronic patellofemoral chondroma-
lacia and instability (3).
Recent studies reported in the
literature, have demonstrated that
MRI findings visualized in Hoffa fat
pad are associated with the clinical
entity referred to as “fat pad
impingement syndrome” (4, 5).
Nevertheless, in these studies, an
association between MRI findings
and patellar chondropathy is not
clearly demonstrated.
Knowledge of the synovial
recesses within the infrapatellar fat
pad is important when evaluating
the knee for infrapatellar fat pad
abnormalities. As the quantity of
articular fluid increases or there is
From: 1. Department of Medical Imaging, Hôpital Erasme, Université Libre de
Bruxelles, Brussels, Belgium.
Address for correspondence: Dr A.C. Cañizares Pérez, Service de Radiologie et
Imagerie Médicale, Institut Bordet, Centre des Tumeurs de l’ULB, Rue Héger-Bordet 1,
Brussels, Belgium.
an inflamed synovial tissue, two
normal synovium-lined clefts along
the posterior aspect of the fat pad
become distended or better individ-
ualized (6). One is located superiorly
in a vertical orientation; the other is
located anterior to the menisci in a
horizontal orientation.
The purpose of our study was to
demonstrate an association
between MRI findings in Hoffa fat
pad and the patellar chondropathy.
Other associations between the MRI
findings of the superior synovial
recess and the signal of the patella,
and the correlation between the
findings of patellar chondropathy
visualized on MRI and during
arthroscopy were also assessed.
Materials and methods
We reviewed retrospectively the
records of 135 patients who under-
went an arthroscopy and MRI at our
institution between October 1997
and January 2001. All patients who
underwent an arthroscopy during
the 6 months following MRI of the
knee were included in our study.
Patients with previous knee surgery
were excluded from the analysis.
Our patient population comprised
85 males and 50 females (age range
17-77 years; median age 47 years).
Parallely MRI images were interpret-
ed and typewritten arthroscopy
reports were recorded.
MRI was performed on a 0.5-T
scanner (Gyroscan T5 NT; Philips,
Best, The Netherlands) between
October 1997 and November 2000
and on a 1.5-T scanner (Gyroscan
Intera Master; Philips, Best, The
Netherlands) between November
2000 and January 2001, using dedi-
cated knee coils. The sequences
included SPIR, T2-weighted (repeti-
tion time / echo time) (1969 ms/
55 ms) (n = 125) or echo planar imag-
ing (EPI) T1-weighted (425 ms/23 ms)
(n = 10) images in the sagittal plane;
magnetization transfer contrast
(MTC) 3D gradient echo (GE) (50 ms/
18 ms) (n = 132) or fat-suppressed
3D T1-weighted (39 ms/ 9’2 ms) (n =
3) images in the axial plane for the
evaluation of the cartilage.
Two independent radiologists,
blinded to the outcome of surgery,
interpreted MRI. If there was any
discrepancy between the two radiol-
ogists, a third radiologist was con-
sulted. For the interpretation of the
MRI of the infrapatellar fat pad, it
was divided into 3 areas in the
cephalo-caudal axis. The abnormal
signal intensity was characterized by
focal high signal intensity on spec-
tral presaturation with inversion
recovery (SPIR) T2 or focal low sig-
nal intensity on EPI T1. Seven grades
were recorded based on the pres-
ence or absence of pathology and
its localization: 0) normal signal
intensity; 1) abnormal signal intensi-
ty of the proximal area; 2) abnormal
signal intensity of the medial area;
3) abnormal signal intensity of the
distal area; 4) abnormal signal inten-
sity of the proximal en medial area;
5) abnormal signal intensity of the
medial en distal area; 6) diffuse
abnormal signal intensity (Fig. 1).
2 JBR–BTR, 2005, 88 (1)

The signal of the superior articu-

lar joint recess on MRI was assessed
and graded as follows: 0) normal
aspect (concave aspect of the pos-
tero-superior portion of infrapatellar
fat pad); 1) loss of concavity of the
postero-superior portion less than
1 cm. in height; 2) loss of concavity
greater than 1 cm in height.
Signal intensity of the patella was
also assessed. It was considered
abnormal when there was high sig-
nal intensity on SPIR T2 or low sig-
nal intensity on EPI T1. The results
were graded as follows: 0) no signal
abnormality; 1) abnormal signal
intensity of the subchondral bone; 2)
abnormal signal intensity at the
inferior pole of the patella; 3) abnor-
mal signal intensity anywhere else.
We also evaluated the aspect of the
patellar cartilage on MRI. In case of
chondropathy, lesions were graded
according to a previously published
arthroscopic classification adapted
to MRI (7, 8) (Table I).
Surgical grading of patellar chon-
Fig. 1. — SPIR T2-weighted image in a sagittal plane shows dropathy was based on the classifi-
the division of the Hoffa fat pad in three areas in the cephalo- cation presented by Ficat (9)
caudal axis. (Table II). The presence of lesions
not limited to the cartilage
(osteoarthritis) or imprecise chon-
dropathy was graded as 4). We con-
sistently tried to match the grades of
Table I. — Classification of patellar chondropathy at chondropathy as recorded by the
MRI based on a previously published arthroscopic arthroscopists with the grades of
classification adapted to MRI. chondropathy adapted to MRI.
Comparisons were made using
Patellar chondropathy on MRI
the Pearson test. The differences
Stade 0 Normal cartilage were considered significant when
Stade 1 Swelling and softening of cartilage p < 0.05. The sensitivity, specificity
Stade 2 Fissures or ulcers extending < 50% of and accuracy of each association
cartilage depth were determined.
Stade 3 Fissures or ulcers extending > 50% of For the correlation between the
cartilage depth MRI results and findings during
Stade 4 Exposure of subchondral bone +/- irre- arthroscopy in relation with the
gularity of bone surface patellar cartilage, we followed two
modes of correlation: comparing
each particular stage of chondropa-
thy and considering only the pres-
ence or absence of chondropathy.
The sensitivity and specificity of the
imaging findings were calculated
with the arthroscopy as the standard
Table II. — Classification of patellar chondropathy at of reference.
arthroscopy based on the classification of Ficat.
Patellar chondropathy at arthroscopy Results

Stage 0 Normal cartilage, white, nacreous and firm Correlation between the MRI find-
at palpation ings of Hoffa fat pad and the superi-
Stage 1 Closed oedematous chondropathy. or articular joint recess on MRI and
Swollen zones, weakened and depressed the arthroscopic results
Stage 2 Open chondropathy. Presence of fissures
and fibrillations. A patellar chondropathy was
Abrasion of cartilage observed in 64 of 135 patients.
Stage 3 Open chondropathy. Ulceration of cartilage Twenty of them had abnormal sig-
leading to sclerosis nal intensity in Hoffa fat pad. This
Stage 4 Osteoarthritis represents a sensitivity of the MRI
findings of 31% and a specificity of
 MRI OF INFRAPATELLAR FAT PAD — CAÑIZARES-PÉREZ et al. 3

A B
Fig. 2. — A, SPIR T2-weighted sagittal image shows no abnormalities in Hoffa fat pad. Note that signal abnormalities of the patellar
cartilage are already visible on this plane (arrow). B, MTC 3D GE axial plane of the same patient shows deep fissures in the patellar
cartilage (arrow) with exposure of the subchondral bone (arrowhead). The patellar chondropathy was classified grade 4 both on MRI
and during arthroscopy.

Correlation between the cartilage

Table III. — Correlation between patellar chondropathy at arthroscopy and
the population with abnormal MRI findings at Hoffa fat pad and superior
recess.
Abnormal Hoffa
Absence of chondropathy 19 (49%)
Presence of chondropathy 20 (51%)
Total 39
73%. On the other hand, 12 of these
patients had a pathologic superior
articular joint recess, representing a
sensitivity of 18% and a specificity
of 88% (Fig. 2).
We correlated also the presence
or absence of patellar chondropathy
in the population with MRI abnor-
malities (Table III). Of 39 patients
with an abnormal Hoffa fat pad, only
20 patients had chondropathy, while
of 21 patients with an abnormal
superior recess, only 12 had chon-
dropathy.
For the correlation of both the
aspect of Hoffa fat pad and the supe-
rior articular joint recess with the
arthroscopic results, we created
3 subgroups: 1) patients with nor-
mal aspect of Hoffa fat pad and
superior recess; 2) patients with
abnormal signal intensity of the
proximal area of Hoffa fat pad and
Abnormal superior recess
9 (43%)
12 (57%)
21
normal aspect of the superior
recess; 3) patients with abnormal
signal intensity of the proximal area
of Hoffa fat pad and loss of concavi-
ty less than 1 cm. in height. No sig-
nificant association was found.
(Table IV).
The patellar signal intensity was
normal in 119 of 135 patients.
Sixteen patients presented a high
signal intensity of the subchondral
bone (grade 1). In 10 grade 4 chon-
dropathy was found during
arthroscopy while 15 presented
grade 4 chondropathy on MRI.
In 2 patients oval high signal inten-
sities were found in Hoffa fat pad,
compatible with cyst. Diagnoses of
an arthrosynovial cyst in one and a
meniscal cyst in the other were easi-
ly made. The former presented a high
signal intensity of the proximal and
medial area of Hoffa fat pad on MRI.
findings on MRI and during
arthroscopy for the diagnosis of
patellar chondropathy.
We found a significant association
between MRI results and arthro-
scopy findings (p < 0.001), both
comparing normal and abnormal
findings and comparing different
grades of pathology using each
approach. Of 64 patients with an
arthroscopic diagnosis of chon-
dropathy, 48 (75%) were diagnosed
as having chondropathy on MRI,
representing a sensibility of these
MRI findings of 73% and a specifici-
ty of 65% (Table V).
Discussion
The pathogenesis of Hoffa dis-
ease appears to be a recurrent
impingement of the infrapatellar fat
pad between the articular surfaces
of the knee. Initially, a minor trau-
matic episode seems to be at the
origin of Hoffa disease with acute
inflammatory response, hemorrhage
and edema in the fat pad. The
swollen fat pad is so more suscep-
tible to repetitive impingement
between the articular surfaces of the
femur and patella. Actual anatomic
and arthroscopic results support
this hypothesis (10, 11). The histo-
pathologic examination of surgical
specimens reveals changes follow-
4 JBR–BTR, 2005, 88 (1)

Table IV. — Correlation between the MRI findings of Hoffa fat pad and superior articular joint recess with the
patellar chondropathy at arthroscopy.
Normal Hoffa Hoffa grade 1* Hoffa grade 1 Total
Normal recess Normal recess Recess grade 1**
Absence of chondropathy 52 (79%) 14 (21%) 0 66
Presence of chondropathy 44 (77%) 11 (19%) 2 (3%) 57
Total 96 25 2 123

* Hoffa fat pad grade 1: anomaly of focal signal of proximal 1/3 of Hoffa fat pad.
** Superior synovial recess grade 1: loss of concavity less than 1 cm in height.

Table V. — Correlation between the patellar chondropathy on MRI and at arthroscopy.

Arthroscopy MRI Grade 0 Grade 1 Grade 2 Grade 3 Grade 4 Total
Grade 0 47 (75%) 1 (2%) 5 (8%) 1 (2%) 9 (14%) 63 (100%)
Grade 1 2 (33%) 1 (17%) 2 (33%) 1 (17%) 0 6 (100%)
Grade 2 13 (48%) 1 (4%) 6 (22%) 2 (7%) 5 (18%) 27 (100%)
Grade 3 7 (37%) 0 4 (21%) 5 (26%) 3 (16%) 19 (100%)
Grade 4 2 (10%) 1 (5%) 1 (5%) 2 10%) 14 (70%) 20 (100%)
Total 71 (53%) 4 (3%) 18 (13%) 11 (8%) 31 (23%) 135 (100%)

ing the evolution of the inflamma-
tion from the acute to the chronic
phase. The acute phase (less than
4 weeks) is characterized by a
normal response to inflammation.
The anterior knee pain is likely to
occur due to a mechanical pressure
on the fat pad. The chronic phase is
maintained by persistent traumas of
the inflamed fat pad, finally result-
ing in fat necrosis and replacement
by a firm mass of fibrous tissue. The
pain in this phase is attributed to
mechanical conflict on the fibrotic
fat pad and damage to the softer
articular cartilage by the replaced
tissue (10). Concerning the etiology,
Smillie adds that this conflict can
also, in the absence of prior injury,
be secondary to joint space narrow-
ing attributable to any cause,
rendering the fat pad more vulnera-
ble to injury (12). And, in this way,
Hoffa disease is sometimes associ-
ated with underlying patellofemoral
abnormalities as patellar instability
and patellofemoral chondromala-
cia (12, 13).
Recent studies attempted to cor-
relate the signal intensity abnorma-
lities on MRI with the clinical term of
Hoffa disease.
In a retrospective study with a
population of 42 patients presenting
with anterior or lateral knee pain,
Chung et al. demonstrated focal
high signal intensity on T2-weighted
images of Hoffa fat pad in 40 of
42 patients (4). The authors introduce
the term of “patellar tendon-lateral
femoral condyle friction syndrome”
to describe a type of friction syn-
drome resulting in inflammatory
changes of the interposed fat which
is situated between the posterior
inferolateral patellar tendon and
the lateral femoral condyle. These
signal intensities seem to be closely
related to the clinical term of Hoffa
disease. The authors of this paper
propose patellofemoral tracking as
the probable cause of these signal
intensity abnormalities on MRI, due
to an imbalance in the stabilizing
structures of the patellofemoral joint.
Patellar instability may be one of
the factors that in an acute or repeti-
tive way may be at the origin of
lesion and inflammation of Hoffa fat
pad and secondarily lead to Hoffa
disease (14). Besides, the associa-
tion between the patellar subluxa-
tion both with the chondropathy
and the patellar tendinopathy has
already been evocated (15).
Brukner et al. described a popula-
tion of patients meeting the clinical
criteria for infrapatellar fat pad
impingement (5). The authors con-
sider that the infrapatellar fat pad
impingement can be distinguished
from other knee conditions causing
anterior knee pain on the basis of
both clinical and radiological crite-
ria. Particularly, concerning the MRI,
they describe an increased signal
in the superior portion of the fat
pad suggestive of edema in 7 of
8 patients on T2-weighted images
with fat-saturation. The authors add
that there is a considerable varia-
tion in the anatomic relationship
between the fat pad and the patella,
particularly in hyperextension.
McNally et al. identified a variant
of patellar tendinopathy which is
particularly associated with patel-
lar subluxation (15). This variant
is characterized by inflammatory
changes in the lateral peritendinous
structures, close to the origin of the
patellar tendon. The inflammatory
changes lie between the femoral
condyle and the patellar tendon.
They note an increase of the Q angle
(the extensor mechanism is position-
ed more laterally) and the proximal
patellar tendon abuts the lateral
femoral condyle.
In our study, we make a differ-
ence on MRI between high signal
intensity on SPIR T2 in Hoffa fat pad
and the loss of concavity of the
postero-superior aspect of Hoffa fat
pad. This loss of concavity of the
postero-superior aspect of Hoffa
may be due to both distension of the
superior joint recess by effusion or
by a synovial thickness reaction by
synovial irritation. The process origi-
nates in the synovium, distends the
synovium but produces also a dis-
tortion and replacement of the knee
fat pad (16). Consequently, we think
that the abnormalities of intracapsu-
lar fat pads appear to be a specific
indicator of synovial proliferation in
joints in which effusions are pre-
sent (6). Stoller et al. affirm that
although the synovium cannot be
imaged in early synovitis, a corru-
gated surface along Hoffa fat pad on
MRI can be interpreted as a sign of
 MRI OF INFRAPATELLAR FAT PAD — CAÑIZARES-PÉREZ et al. 5

A B
Fig. 3. — A, SPIR T2-weighted sagittal image shows a loss of concavity of the postero-superior portion of Hoffa fat pad (arrow)
assessed as pathological superior joint recess grade 1. B, MTC 3D GE axial plane of the same patient shows a patellar chondropathy
(arrows) assessed grade 2 on MRI and 0 during arthroscopy.

synovial irritation (e.g. hemophilia,
rheumatoid arthritis) (3). This irregu-
lar infrapatellar fat pad sign should
be distinguished by the signal inten-
sity abnormalities on MRI in Hoffa
disease.
Our study has several limitations.
First, the retrospective character can
have influenced the interpretation of
the arthroscopic data based on the
surgical report. Thirty-one patients
were classified grade 4 following
the arthroscopic classification for
the cartilage (imprecise chondro-
pathy) (Fig. 3).
Of 71 patients without patel-
lar chondropathy at arthroscopy
(grade 0), the MRI agrees with these
findings in only 47 patients and
disagrees in 24 patients (presence
of a patellar chondropathy). These
results can indicate a lack of either
evaluation or information of the
orthopedist over the aspect of patel-
lar cartilage during arthoscopy.
Selection bias is a second limita-
tion in our study: only patients with
a potential surgical problem were
enrolled in the analysis; patients
with clear symptoms of Hoffa dis-
ease who underwent a conservative
treatment were not included in the
analysis. Third, signal intensity
abnormalities on MRI in Hoffa fat
pad are probably modulated by the
phase of the disease and by the type
of MRI sequences used. Our popula-
tion was heterogeneous in these
aspects. No differentiation between
our patients was performed based
on these criteria. Concerning the
MRI sequences, we know e.g. that
fibrous tissue, which is the essential
component of the chronic phase in
Hoffa disease, shows low signal
intensity on T1- and T2-weighted
sequences. Because SPIR T2 was
most frequently performed in our
patients for the evaluation of the
Hoffa fat pad, we ran the risk to
under evaluate the presence of
fibrous tissue. Finally, we attempted
to correlate two classifications for
the evaluation of the cartilage on
MRI and on arthroscopy. Both clas-
sifications are well known but not
easily superposable, making any
correlation hazardous.
In our article, we tried to link
Hoffa’s fat pad disease and patellar
chondropathy. A significant associa-
tion will still have to be shown in
other studies. In previous articles,
Chung, McNally and Apostolaki pos-
tulated an association both between
Hoffa’s fat pad disease and patellar
instability and between patellar
chondropathy and patellar instabili-
ty (4, 14, 15).
Conclusion
No significant association was
shown between the MRI findings of
the Hoffa fat pad and the patellar
chondropathy. Nevertheless, we
found a significant association
between the cartilage findings on
MRI and during arthroscopy for the
diagnosis of patellar chondropathy.
References
1. Hoffa A.: Influence of adipose tissue
with regard to the pathology of the
knee joint. JAMA, 1904, 43: 795-796.
2. Ogilvie-Harris D.J., Giddens J.:
Hoffa’s disease: arthroscopic resec-
tion of the infrapatellar fat pad.
Arthroscopy, 1994, 10: 184-187.
3. Stoller D.W., Cannon W.D.,
Anderson L.J.: The knee. In.:
Magnetic resonance imaging in
Orthopaedics and sports medicine.
Edited by Stoller D.W., Cds. Printed
by Lippincott, Philadelphia, 1997,
p 390.
4. Chung C., Skaf A., Roger B.,
Campos J., Stump X., Resnick D.:
Patellar tendon-lateral femoral
condyle friction syndrome: MR imag-
ing in 42 patients. Skeletal Radiol,
2001, 30: 694-697.
5. Brukner P.D., McConnell J.,
Bergman A.G., Beaulieu C.F.,
Matheson G.O.: Infrapatellar fat pad
impingement. Correlation between
clinical and MR findings. ISMRM
Abstracts, 1998: 241.
6. Schweitzer M.E., Falk A., Berthoty D.,
Mitchell M., Resnick D.: Knee effu-
sion: normal distribution of fluid.
AJR, 1992, 159: 361-363.
7. Noyes F.R., Stabler C.L.: A system for
grading articular cartilage lesions at
arthroscopy. Am J Sports Med, 1989,
17: 505-513.
6 JBR–BTR, 2005, 88 (1)
8. Vande Berg B.C., Lecouvet F.E.,
Poilvache P., et al.: Assessment of
Knee Cartilage in Cadavers with
Dual-Detector Spiral CT Arthro-
graphy and MR Imaging. Radiology,
2002, 222: 430-436
9. Ficat C.: Les contusions du cartilage
articulaire. Etude experimental. Rev
Chir Orthop, 1976, 62: 493-500.
10. Metheny J.A., Mayor M.B.: Hoffa’s
disease: chronic impingement of the
infrapatellar fat pad. Am J Knee
Surg, 1988, 1: 134-139.
11. Magi M., Branca A., Bucca C.,
Langerame V.: Hoffa disease. Ital J
Orthop Traumatol, 1991, 17: 211-216.
12. Smillie I.S.: Lesions of the infrapatel-
lar fat pad and synovial fringes :
Hoffa’s disease. Acta Orthop Scand,
1963, 33: 371-377.
13. Chernye S.: Disorders of the knee. In:
Dee, et al., eds. Principles of
orthopaedic practice, vol. 2 New York:
Mc Graw-Hill, 1989, p 1283.
14. Apostolaki V., Cassar-Pullicino P.,
Tyrrell P.N.M., McCall I.W.: MRI
appearances of the infrapatellar fat
pad in occult traumatic patellar dis-
location. Clin Radiol, 1999, 54: 743-
747
15. McNally E.: Imaging assessment of
anterior knee pain and patellar mal-
tracking. Skeletal Radiol, 2001, 30:
484-495.
16. Jacobson J., Lenchik, L., Ruhoy M.,
Schweitzer M., Resnick D.: MR imag-
ing of the infrapatellar fat pad of
Hoffa. RadioGraphics, 1997, 17: 675-
691.