Musculoskeletal Imaging • Original Research

Stacy et al.
Infarct-Associated Bone Sarcomas

Musculoskeletal Imaging
Original Research
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved

Infarct-Associated Bone Sarcomas:

Multimodality Imaging Findings
Gregory Scott Stacy 1 OBJECTIVE. The objective of our study was to characterize infarct-associated bone sar-
Ryan Lo1 coma and its imaging features.
Anthony Montag2 MATERIALS AND METHODS. Our databases were searched for instances of sarco-
ma arising in association with osteonecrosis. Demographic and imaging data were recorded.
Stacy GS, Lo R, Montag A The imaging studies of 258 patients with sarcomas were reviewed to determine whether un-
derlying osteonecrosis was present. Radiographic and MRI studies of patients with bone in-
farction were reviewed to categorize the various appearances of infarction and to determine if
sarcomas tended to arise in a particular pattern. A literature review was performed.
RESULTS. Nine infarct-associated bone sarcomas were found in eight patients: seven
malignant fibrous histiocytomas (MFHs) and two osteosarcomas. All occurred in the fe-
mur or tibia; multifocal infarction was documented in all patients except one. Sarcomas were
commonly associated with a so-called “mature”-type pattern of osteonecrosis—that is, with
well-defined calcified margins. Osteolysis of infarct-associated MFHs was often overlooked
at initial presentation and was often detected only after pathologic fracture. CT and MRI re-
vealed cortical penetration in all cases; infarct margin disruption was evident, but preserva-
tion of fat within the infarct was typical. Increased radiotracer activity with relative central
photopenia was characteristic of large infarct-associated bone sarcomas on scintigraphy. All
lesions, including those treated at our institution and those found in the literature, were me-
taphyseal or diaphyseal, and although epiphyseal extension of sarcoma from a metadiaphy-
seal infarct was common, no purely epiphyseal lesions were encountered.
CONCLUSION. Radiologists must remain vigilant for this rare occurrence, especially
Keywords: malignant fibrous histiocytoma, MRI, in patients with new pain in an area of known bone infarction.
osteonecrosis, osteosarcoma, sarcoma

DOI:10.2214/AJR.14.13871
Received September 27, 2014; accepted after revision
April 1, 2015.
Based on presentations at the Society of Skeletal
Radiology 2008 annual meeting, La Quint, CA/USA,
and at the International Skeletal Society 2013 annual
meeting, Philadelphia, PA/USA.
1
Department of Radiology, University of Chicago,
5841 S Maryland Ave, MC 2026, Chicago, IL 60637.
Address correspondence to R. Lo (dr.ryan.lo@gmail.com).
2
Department of Pathology, University of Chicago,
Chicago, IL.
WEB
This is a web exclusive article.
AJR 2015; 205:W432–W441
0361–803X/15/2054–W432
© American Roentgen Ray Society
T
he term “bone infarction” is gen-
erally used to describe osteone-
crosis occurring in the metaphy-
seal or diaphyseal segments of
the long bones. Bone infarction has been at-
tributed to a variety of predisposing condi-
tions, including alcoholism, sickle cell ane-
mia, collagen vascular diseases, Gaucher
disease, decompression sickness (caisson
disease), and steroid use. In many instances,
the underlying cause of bone infarction is
unknown. Although bone infarcts can ini-
tially be painful, most are discovered inci-
dentally on imaging studies performed for
the evaluation of other abnormalities.
Bone infarcts can have a variety of appear-
ances on radiographs, ranging from subtle ill-
defined radiolucencies to the more classic ap-
pearance of well-defined shell-like sclerosis
[1]. On MRI, bone infarcts often have a char-
acteristic appearance of a “maplike” rim of
low signal intensity, with an adjoining inner
rim of high signal intensity on T2-weighted
images (i.e., the double line sign), surround-
ing a central region that follows the signal in-
tensity of fat; however, a central region of low
signal intensity may also be observed presum-
ably due to marrow fibrosis or sclerosis [2].
Although bone infarction is frequently en-
countered in radiology practices, the devel-
opment of sarcoma in association with bone
infarction is rare. In 1960, Furey et al. [3] de-
scribed two patients who developed a fibro-
sarcoma arising at the site of a bone infarct.
Since then, approximately 70 additional ex-
amples of infarct-associated sarcomas have
been reported, many of which have been clas-
sified histologically as malignant fibrous his-
tiocytoma (MFH), a relatively uncommon
primary bone tumor. Most of these infarct-
associated sarcomas in the literature are de-
scribed in case reports or small series, with

W432 AJR:205, October 2015


few collections and reviews of the reported
data. Furthermore, none of the reviews pub-
lished to date focuses collectively on the im-
aging findings seen with this phenomenon.
Our study aims to further characterize this
rare condition by providing a detailed descrip-
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
tion of the imaging features of bone sarcomas
arising in bone infarcts as seen on radiogra-
phy, CT, MRI, and skeletal scintigraphy.
Materials and Methods
This retrospective study was approved by our
institutional review board, and informed consent
was waived. Two methods were used to find in-
stances of bone sarcoma arising in association
with osteonecrosis. For the first method, a list of
422 patients with bone sarcoma who presented to
our institution from January 1, 1978 through Jan-
uary 31, 2008 was generated from the database of
our institution’s orthopedic oncology clinic. From
this list of 422 patients, we excluded patients who
had been referred to our institution after surgery
or initiation of chemotherapy elsewhere, patients
who developed a sarcoma in an area previously
treated with radiation therapy, and patients who
were treated at our institution but had missing pa-
thology reports or whose medical record other-
wise lacked sufficient documentation for a use-
ful review. These exclusions yielded a subset of
308 patients with bone sarcoma who had received
no treatment when they presented to our institu-
tion and who were subsequently treated at our in-
stitution and had accessible pathology reports. A
search of the orthopedic oncology clinic and de-
partment of pathology databases at our institu-
tion of these 308 patients was then undertaken to
find instances of sarcoma of bone arising in as-
sociation with osteonecrosis; for that search, the
terms “osteonecrosis,” “infarct,” “infarction,” and
“avascular necrosis” were used.
The second method to find cases of infarct-as-
sociated bone sarcomas involved direct inspection
of the available imaging studies from the same
30-year period. The majority of cases of sarcoma
arising in association with bone infarction report-
ed in the literature have been osteosarcoma and
MFH, with fibrosarcoma, spindle cell sarcoma,
undifferentiated sarcoma, and angiosarcoma oc-
curring less frequently. Therefore, we collected
the imaging studies of the patients with these par-
ticular sarcomas and reviewed them to determine
whether underlying osteonecrosis was present.
At least one cross-sectional imaging study, either
CT or MRI, was required for review. This review
yielded a total of 258 patients (206 patients with
osteosarcoma; 36 patients with MFH; and 16 pa-
tients with diagnoses of fibrosarcoma, spindle cell
sarcoma, undifferentiated sarcoma, or angiosar-
AJR:205, October 2015 W433
Infarct-Associated Bone Sarcomas
coma) whose imaging studies consisted of various
combinations of conventional radiographs, CT ex-
ams, MR exams, and skeletal scintigrams.
For each patient with infarct-associated sarco-
ma, the sex, age at the time of diagnosis, presence
or absence of any condition that could predispose
to osteonecrosis, and the duration of symptoms
before diagnosis were recorded. All pertinent im-
aging studies, including radiography, CT, MRI,
and skeletal scintigraphy, were reviewed by two
radiologists. All imaging studies were used to as-
sess lesion location in the skeleton and bone, le-
sion size and anatomic extent, visibility of the un-
derlying infarct, and the presence or absence of
additional bone infarcts. Other features reviewed
on radiographs included the characteristics of the
sarcoma (osteolytic vs bone-forming tumor, pres-
ence or absence of cortical destruction, periosteal
reaction, visible soft-tissue mass, and pathologic
fracture) and the characteristics of the underlying
infarct using a categorization scheme described
later in this article.
On CT, the density of the tumor relative to skel-
etal muscle and the presence or absence of matrix
mineralization were noted. When there were con-
trast-enhanced images, the degree of enhancement
and the pattern of enhancement were also noted.
On MRI, the signal intensity of the tumor rel-
ative to skeletal muscle on T1- and T2-weighted
images was recorded. When gadolinium-en-
hanced images were available, enhancement of
the tumor was graded as mild (resultant signal in-
tensity slightly greater than that of nonenhanced
tumor), moderate (resultant signal intensity slight-
ly less than that of fat), or marked (resultant signal
intensity equal to or greater than that of fat). The
homogeneity or heterogeneity of the tumor on all
pulse sequences was noted.
On skeletal scintigraphy, the activity of the tu-
mor was characterized as mild (slightly greater
than that of long bone cortex on whole-body pla-
nar images), moderate (similar to that of the lum-
bar spine), or marked (similar to concentrated ra-
diotracer in the urinary bladder). Because all of
these patients were referred to our orthopedic on-
cology clinic, the pertinent imaging studies and
reports from the referring institutions were also
reviewed. The final pathologic diagnosis (i.e., type
of sarcoma) was recorded.
Next, the imaging studies of 63 patients with
bone infarction without sarcoma detected on both
MRI and radiography within 1 month of each oth-
er were obtained. Using MRI as the criterion stan-
dard for the determination of infarction based on
its virtually pathognomonic appearance, radio-
graphs of patients with bone infarction were re-
viewed to categorize the various appearances of
bone infarction and to subsequently determine if
sarcomas tended to arise in association with a par-
ticular pattern of bone infarction. Patients with
acute complications associated with the infarction
(e.g., osteomyelitis) were excluded. Bone infarcts
were categorized as type 1 (normal-appearing on
radiography), type 2 (mottled, ill-defined radiolu-
cency without sclerosis), type 3 (lesion with poor-
ly defined or incomplete calcified margins with
density less than that of cortical bone and variable
central density), type 4 (the classic appearance of
a peripherally sclerotic lesion with well-defined
calcified margins of density visibly similar to cor-
tical bone and either a lucent or dense center), or
type 5 (mixed sclerosis and radiolucency with
haphazard coarsened trabeculae but without de-
finable margins, typically reflecting infarction of
nearly the entire bone).
Finally, the literature was reviewed for case re-
ports and series describing sarcomas arising in as-
sociation with bone infarction. The results of our
study were added to a compilation of data from
the literature.
Results
A review of the databases revealed a total
of nine infarct-associated sarcomas occur-
ring in eight patients. One patient developed
an infarct-associated osteosarcoma in his fe-
mur and subsequently developed an infarct-
associated MFH in his contralateral tibia 7
years later. This patient’s first sarcoma had
been previously described in a case report
[4], although his second sarcoma had not.
One additional patient developed an infarct-
associated osteosarcoma that had also been
described in a previous case report [4]. The
remaining six patients developed infarct-as-
sociated MFH. Of the eight patients with in-
farct-associated sarcomas, four were women,
and four were men. The average age of the pa-
tients at initial diagnosis was 55.1 years with
an SD of 16.7 (range, 31–80 years). Two of the
patients had a predisposition for bone infarc-
tion from steroid treatment of inflammatory
bowel disease. The patient who developed
two infarct-associated sarcomas had familial
hypertriglyceridemia but no other predispos-
ing factors for bone infarction. The remaining
five patients had infarcts that were deemed
idiopathic—that is, without a discernible pre-
disposition identified—on the basis of a re-
view of their clinical records. Symptom du-
ration data were available for six of the eight
patients; on average, the patients complained
of pain for 4 months before diagnosis (range,
1–8 months). Data are summarized in Ta-
ble 1. Three additional patients with a patho-
logic diagnosis of ­possible infarct-­associated
 Stacy et al.

TABLE 1: Patients With Infarct-Associated Sarcoma Treated at Our Institution

Pain Duration Histologic Additional Sites of
Age at Predisposition to Before Diagnosis of Osteonecrosis
Patient No. Diagnosis (y) Sex Infarction Workup (mo) Location of Sarcoma Sarcoma Documented
1 61 F None 8 Tibia, proximal metaphysis with Osteosarcoma Yes
epiphyseal extension
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved

2 (First sarcoma) 35 M Hypertriglyceridemia 4 Femur, proximal diaphysis Osteosarcoma Yes

2 (Second sarcoma) 42 M Hypertriglyceridemia 2 Tibia, mid distal diaphysis MFH Yes
3 46 M None NA Tibia, proximal diaphysis MFH No
4 58 M None 6 Tibia, proximal metaphysis with MFH Yes
epiphyseal extension
5 68 F Steroid therapya 4 Femur, distal metaphysis with MFH Yes
epiphyseal extension
6 62 F None 6 Femur, distal metadiaphysis MFH Yes
7 80 M None NA Femur, distal metaphysis with MFH Yes
epiphyseal extension
8 31 F Steroid therapya 1 Femur, distal metaphysis with MFH Yes
epiphyseal extension
Note—MFH = malignant fibrous histiocytoma, NA = not available.
aTreatment for inflammatory bowel disease.

s­ arcoma were revealed as a result of a search
of the pathology database; however, reexami-
nation of the histologic specimens of these pa-
tients showed tumor surrounding small foci
of dead bone rather than an interface of tumor
with an area of bone infarction. Furthermore,
the imaging studies of these patients, includ-
ing MRI, revealed no evidence of osteonecro-
sis; therefore, these patients were excluded
from our study.
Of the 206 patients seen in our orthope-
dic oncology clinic and ultimately diagnosed
with osteosarcoma, only the two aforemen-
tioned patients showed imaging evidence of
tumor associated with underlying bone in-
farction. Of the 36 patients seen in our or-
thopedic oncology clinic and ultimately di-
agnosed with MFH, only the aforementioned
seven patients showed imaging evidence of
tumor associated with underlying bone in-
farction. None of the patients diagnosed with
fibrosarcoma, spindle cell sarcoma, undiffer-
entiated sarcoma, or angiosarcoma showed
evidence of underlying bone infarction.
Radiographs were available for all pa-
tients with infarct-associated sarcoma. All
infarct-associated sarcomas occurred in ei-
ther the femur or the tibia, with five of nine
sarcomas occurring near the knee joint with
epiphyseal extension from a metaphyseal in-
farct; four of the sarcomas were diaphyseal
or metadiaphyseal in location. The underly-
ing bone infarcts were visible, to varying de-
grees, in all cases. The most common pattern
of underlying infarction associated with sar-
coma, seen in six of the eight patients, was
type 4 (Figs. 1 and 2). The type 3 pattern
was seen in one patient with sarcoma (Fig.
3), and the type 5 pattern was seen in the one
patient who developed two sarcomas.
Radiographs depicting infarct-associated
osteosarcoma revealed an ossified soft-tissue
mass arising from bone in both patients. In
one of the patients (described in a previous
case report [4]), an underlying destructive
lesion with aggressive features, including la-
mellated periosteal reaction and cortical dis-
ruption, was evident; in the other patient, the
ossified soft-tissue mass was the dominant
abnormality (Fig. 2). The estimated maxi-
mum dimensions of each osteosarcoma on
the basis of radiographs were approximate-
ly 7 and 11 cm, respectively. Radiographs
depicting infarct-associated MFHs revealed
poorly defined osteolysis with cortical thin-
ning and destruction adjacent to the bone
infarct (Figs. 1 and 3). No matrix mineral-
ization was observed that would suggest a
classic osteosarcoma or chondrosarcoma.
Despite the cortical penetration, distinct
periosteal reaction was minimal or absent
and a discrete soft-tissue mass was not vis-
ible except in one patient with a Codman tri-
angle and associated soft-tissue prominence.
Based on medical record review, only three
of the seven infarct-associated MFHs were
detected on initial radiographs, with four pa-
tients requiring follow-up radiography for
detection, three of whom presented after sus-
taining a pathologic fracture. Tumor-associ-
ated osteolysis was visible in retrospect in all
four cases but was subtle (Fig. 1). Estimated
size of MFHs (based on osteolysis) was ap-
proximately 5.4 cm on average, and the tu-
mors tended to arise eccentrically within the
bone (five of seven patients).
All patients with infarct-associated sar-
comas underwent cross-sectional imaging
(three cases, CT only; four, MRI only; two,
both CT and MRI). Total cortical penetra-
tion was noted in all cases, and a discrete
soft-tissue mass was present in eight cases
(89%). Tumor was also noted to disrupt the
infarct margin in all cases; however, the ma-
jority of the hyperdense or hypointense mar-
gin was present in all patients. Preservation
of at least some fat attenuation or fat signal
intensity within the infarct was noted in all
patients except one.
Of the five sarcomas evaluated with CT,
three were MFHs and two were osteosarco-
mas. The CT examinations of the patients
with infarct-associated osteosarcoma were
unenhanced. CT of the patient with the in-
farct-associated osteosarcoma depicted in
Figure 2 revealed an incomplete dense ring-
like calcification in the proximal tibial me-
taphysis with fat attenuation centrally, repre-
senting the infarct. There was osteosclerosis
of the bone between the superior margin of
the infarct and the tibial articular surface
compatible with osteosarcoma. There was a
pathologic fracture of the lateral tibial pla-
teau. The sarcoma extended inferiorly along
the periphery of the infarct and posteriorly

W434 AJR:205, October 2015

 Infarct-Associated Bone Sarcomas
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
A B
C D
F G
AJR:205, October 2015 W435
Fig. 1—46-year-old man (patient 3 in Table 1) with
malignant fibrous histiocytoma arising at site of bone
infarction.
A, Anteroposterior knee radiograph shows multifocal
bone infarction (black arrowheads) in proximal tibia
and distal femur with well-defined, dense sclerotic
margins (type 4 pattern). Adjacent osteolysis in
proximal tibia (white arrowheads) was not detected
on initial interpretation.
B, Anteroposterior knee radiograph obtained
4 months after A reveals progression of bone
destruction (white arrowheads) with collapse of
lateral tibial plateau. Note unchanged appearance of
bone infarct (black arrowhead).
C and D, Coronal (C) and transverse (D) T1-weighted
MR images of proximal tibia show tumor (white
arrowheads) surrounding underlying infarct (black
arrowheads), which is largely preserved with
possible exception of focal destruction along its
superior margin (arrow, C).
E, Transverse T2-weighted MR image of proximal
tibia shows hyperintense tumor (white arrowheads)
adjacent to infarct (black arrowheads).
F, Transverse T1-weighted MR image of proximal tibia
obtained after IV administration of 15 mL of gadolinium
chelate shows enhancement of tumor (white
arrowheads) adjacent to infarct (black arrowheads).
G, Frontal planar skeletal scintigram shows increased
uptake in proximal tibia with photopenic focus centrally
corresponding to infarct with surrounding tumor.
H, Photograph of bisected gross specimen shows
tumor (white asterisk) adjacent to bone infarct (black
asterisk) eroding through lateral tibia.
E
H
 Stacy et al.
through the cortex to form a soft-tissue mass;
most of this mass was ossified and, hence, hy-
perdense to skeletal muscle. CT of the other
patient with infarct-associated osteosarcoma
revealed a similar-appearing ossified soft-
tissue mass arising anteromedially from the
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
proximal femoral diaphysis with underlying
cortical destruction. Haphazard coarsened
trabeculae in the medullary space of nearly
A B
C D
Fig. 2—61-year-old woman (patient 1 in Table 1) with osteosarcoma arising at site of bone infarction.
A, Lateral knee radiograph shows ossified mass (white arrowheads) in soft tissues along proximal aspect of
posterior tibia. Note underlying bone infarction in proximal tibia with well-defined sclerosis (type 4 pattern) and
bone infarction in distal femur (black arrowheads).
B, Transverse CT image through proximal tibia confirms ossified mass (white arrowhead) and underlying bone
infarct (black arrowhead).
C, Planar skeletal scintigram shows uptake in proximal tibia corresponding to osteosarcoma.
D, Photograph of bisected gross specimen shows sarcoma (white asterisk) eroding through tibial cortex with
relative preservation of intramedullary infarct (black asterisk).
W436 AJR:205, October 2015
the entire femur indicated extensive bone in-
farction. The maximum dimensions of these
two osteosarcomas on CT were similar to the
estimated maximum dimensions on radiogra-
phy (≈ 7 and 11 cm, respectively).
Of the three CT studies used to evaluate
infarct-associated MFH, one study was un-
enhanced and the other two studies were
contrast-enhanced without any unenhanced
images. The unenhanced CT scans of one
of the lesions revealed an incomplete dense
ringlike calcification of the distal femoral
metaphysis, representing the margins of the
infarct. Centrally, the infarct was replaced
with sarcoma that extended into the distal
femoral epiphysis. The tumor was slightly
hypodense relative to skeletal muscle and
penetrated the lateral cortex, resulting in a
soft-tissue mass and a pathologic fracture.
No residual fat density was noted in the in-
farcted bone on CT, although a small amount
was seen on subsequent MRI. Coarse calcifi-
cations within the sarcoma oriented in a row
in the sagittal plane were thought to repre-
sent residua of the infarct margin. The con-
trast-enhanced CT scans of the second case
of infarct-associated MFH revealed a de-
structive lytic lesion of the anterolateral as-
pect of the proximal tibial diaphysis associ-
ated with a soft-tissue mass and pathologic
fracture. Centrally, the mass was slightly
hypodense to skeletal muscle, although
there was a thick rind of enhancement with
a density similar to nearby enhancing ves-
sels; no matrix mineralization was evident.
Within the tibia, the sarcoma was bordered
by a thick dense band of calcification repre-
senting the infarct margin. The contrast-en-
hanced CT scans of the third case of infarct-
associated MFH revealed a destructive lytic
lesion of the anterior aspect of the mid tibial
diaphysis associated with a soft-tissue mass.
The mass was diffusely and homogeneously
slightly hyperdense to skeletal muscle, pre-
sumably because of contrast enhancement,
although no unenhanced images were avail-
able; no matrix mineralization was evident.
Haphazard coarsened trabeculae and incom-
plete dense ringlike calcifications in the tibia
adjacent to the superior aspect of the sarco-
ma represented the infarct.
T1-weighted MR images were available
for five of the six infarct-associated MFH
cases. The sarcoma was of intermediate
signal intensity with T1-weighting, com-
parable to that of skeletal muscle or articu-
lar cartilage, in all cases. The sarcoma was
juxtaposed with a well-defined curvilinear
low-signal-intensity band, representing the
 Infarct-Associated Bone Sarcomas

infarct margin, with focal disruption of the
margin in all cases. Preservation of at least
some fat signal intensity within the infarct
was typical, although the amount varied
from case to case and this finding was not
present in one of the cases. Relatively ill-de-
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
fined curvilinear or nodular foci of low sig-
nal intensity were noted within the substance
of the sarcoma in four of the five cases in
which T1-weighted imaging was available,
presumably representing residua of tumor-
enveloped infarct margins; otherwise, the
with non–fat-suppressed T2-weighted imag-
es, four of the sarcomas were predominantly
of high signal intensity, similar to that of fat,
with smaller ill-defined components of inter-
mediate signal intensity approaching that of
skeletal muscle (postulated to be more cel-

signal intensity on T1-weighted images was lular regions of the tumor), as well as bet-
homogeneous. Foci of increased signal in- ter-defined low-signal-intensity components
tensity suggestive of hemorrhage were not a (postulated to be residua of calcified infarct
dominant feature of infarct-associated sarco- margins). One of the tumors was predomi-
mas but were present in one case. nantly hypointense to fat, containing equal
T2-weighted MR images were available volumes of tissue that were isointense and
for all six of the infarct-associated MFH cas- slightly hyperintense to skeletal muscle in
es that had MRI studies, with only fat-sup- addition to bandlike and nodular foci of low
pressed T2-weighted images available for signal intensity. The MRI examination with
one of the six cases. Of the five MRI studies only fat-suppressed T2-weighted images
showed a predominantly hyperintense tumor
(signal intensity approaching fluid signal in-
tensity) with nodular components similar
in signal intensity to skeletal muscle. Corti-
cal penetration was evident in all cases, and
soft-tissue masses were seen on five of the
six MRI examinations.
Four of the MRI examinations included
non–fat-suppressed T1-weighted sequences
obtained after IV gadolinium administra-
tion. In all four cases, the majority of the tu-
mor showed enhancement with a heteroge-
neous pattern. Two of the sarcomas showed
regions of enhancement that were moderate
(signal intensity slightly less than that of fat)
to marked (signal intensity equal to or great-
A B
er than that of fat), whereas the other two sar-
comas showed regions of enhancement that
were mild (slightly hyperintense to skeletal
muscle) to moderate. Small nonenhancing
elements within the sarcoma likely were a
combination of focal necrosis, relatively hy-
povascular tissue, and residua of infarct mar-
gins. On cross-sectional imaging, the aver-
age MFH size was 7.5 cm.
Seven skeletal scintigraphy examinations
(two patients with osteosarcoma, five pa-
tients with MFH) were available for review.
All infarct-associated sarcomas showed in-
tensely increased radiotracer uptake (equal
to concentrated radiotracer in the bladder)
except one MFH that showed moderate up-
C D
take (similar to the lumbar spine). An adja-
Fig. 3—31-year-old woman (patient 8 in Table 1) with history of steroid therapy for ulcerative colitis and
malignant fibrous histiocytoma (MFH) arising at site of bone infarction. cent or central area of decreased uptake cor-
A, Lateral radiograph of knee shows small focus of osteolysis and cortical irregularity (white arrowheads) along responding to the infarcted area of bone was
anterior aspect of distal femur. Note underlying bone infarct with type 3 pattern of incomplete sclerosis (black characteristic of MFHs associated with rela-
arrowheads).
B, Transverse T1-weighted MR image through distal femur shows tumor (white arrowheads) surrounding
tively large infarcts.
infarct (black arrowheads); note preservation of fat signal intensity within infarct with exception of low-signal- Based on all imaging studies provided
intensity focus laterally (arrow). and reviewed, seven of the eight patients had
C, Photograph of bisected gross specimen shows tumor (asterisk) adjacent to bone infarct eroding anterior multifocal osteonecrosis. Only one patient
cortex of femur.
D, Photograph of histologic specimen shows pleomorphic cells of MFH (black asterisk) adjacent to dead bone with infarct-associated MFH showed no ev-
(white asterisk). idence of additional sites of osteonecrosis,

AJR:205, October 2015 W437


a­ lthough the imaging studies were limited to
the area of the sarcoma with the exception of
skeletal scintigraphy. Of note, the additional
sites of osteonecrosis were typically not visi-
ble on skeletal scintigraphy except in the one
patient who developed two sarcomas.
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
All of the infarct-associated bone sarco-
mas were intermediate- or high-grade tumors
treated with chemotherapy and amputation or
wide resection of the tumor with endopros-
thetic reconstruction. Lung metastases were
present in one patient at the time of diagnosis
and developed in three other patients over the
next 3 months to 3 years after diagnosis. All
patients with metastases died within 8 months
of discovery of the metastases (average, 4
months). Patients without metastases lived 11
months to 25 years after diagnosis.
Discussion
Although its prevalence is unknown, os-
teonecrosis, or “bone death,” is a relatively
common condition that can occur second-
ary to a wide variety of predisposing fac-
tors (e.g., trauma, hemoglobinopathies such
as sickle cell disease, vasculitides such as
systemic lupus erythematosus, corticoste-
roid use, alcoholism) or can occur idiopathi-
cally [2]. For the purposes of further discus-
sion, we define the term “avascular necrosis”
as osteonecrosis that occurs in subarticular
locations and the terms “bone infarct” and
“bone infarction” as osteonecrosis that oc-
curs in the metaphysis or diaphysis of a long
bone; however, these terms are not used con-
sistently in the literature.
Avascular necrosis, particularly of the
femoral head, has been studied extensively
given that untreated disease can lead to ar-
ticular surface collapse and osteoarthritis of-
ten requiring arthroplasty. Conversely, bone
infarction has received relatively little atten-
tion in the literature.
Bone infarcts can have a variable appear-
ance on radiographs, which we have cat-
egorized into five types as described earli-
er. Although the classic type 4 pattern likely
correlates with a mature bone infarct, Munk
et al. [1] described an alternative appear-
ance of infarction consisting of “mottled,
ill-defined radiolucencies” sometimes with
“mild reactive sclerosis which was attributed
to an immature bone infarct.” This appear-
ance corresponded to the type 2 and type 3
patterns of infarction in our categorization.
Finally, when most of the bone showed ev-
idence of infarction on MRI, an additional
pattern of bone infarct—a pattern of mixed
W438 AJR:205, October 2015
Stacy et al.
sclerosis and lucency with haphazard coars-
ened trabeculae without definable margins
(type 5)—was seen; this pattern was not
described in the article by Munk and col-
leagues. Although the appearance of osteo-
necrosis on MRI is considered to be pathog-
nomonic in most cases, we acknowledge that
using MRI as the criterion standard for iden-
tifying patients with osteonecrosis has its
limitations. It stands to reason that a patient
with radiographs showing a classic pattern
of bone infarction (type 4) diagnosed confi-
dently by the radiologist might not undergo
subsequent MRI if the clinician concludes
that osteonecrosis might be the source of the
patient’s pain. Conversely, if one of the oth-
er patterns of infarction is encountered or
overlooked by the radiologist and a diagno-
sis of infarction is not confidently made, then
the clinician might be more inclined to or-
der MRI for further evaluation of symptoms,
perhaps at the urging of the radiologist. Be-
cause this potential bias was not controlled
for in our retrospective study, we did not at-
tempt to determine the true frequency of the
different patterns of infarction.
In 1960, Furey et al. [3] reported two cas-
es of fibrosarcoma arising at the site of bone
infarcts. Since then, several case reports de-
scribing a variety of sarcomas arising in as-
sociation with bone infarction, including
MFH, osteosarcoma, angiosarcoma, and ep-
ithelioid hemangiothelioma, have been pub-
lished [4–44]. In 1992, Torres and Kyriakos
[40] reported a case of osteosarcoma aris-
ing in a medullary infarct of the humerus
and also reviewed and summarized existing
case reports, including the cases described
by Furey et al. and the two cases of infarct-
associated osteosarcoma from our institution
reported in 1987 [4]. More recently, Domson
et al. [15] described their experience with the
treatment of 15 patients with infarct-associ-
ated sarcoma and also reviewed additional
cases that had been documented after the ar-
ticle by Torres and Kyriakos was published.
The series reported by Domson et al. is the
only article other than ours that describes
more than five original cases of infarct-as-
sociated sarcoma, and this article is the first
to date to focus on the imaging findings of
a series of cases and of additional cases de-
scribed in the literature.
Including our set of eight patients, a total
of 78 patients have been described in the Eng-
lish-language literature [3–42]. Two of these
patients developed two infarct-associated sar-
comas, yielding a total of 80 infarct-associat-
ed sarcomas. Data available for all 78 patients
included the age at diagnosis, sex, involved
bone or bones, and histologic diagnosis. The
average age at diagnosis was 53.2 years (SD,
13.9 years). Fifty of the 78 (64%) patients
were male. This sex discrepancy may be re-
lated to the greater likelihood of men to have
predisposing conditions, such as alcoholism,
and to work in occupations that are predis-
posing to dysbaric osteonecrosis. Fifty-five of
the 80 malignancies arose either in the distal
femur (the most common location) or in the
proximal tibia. Malignancies were also re-
ported to arise in the proximal or mid femur
(11 cases), mid or distal tibia (eight cases),
proximal humerus (four cases), distal radius
(one case), and acetabulum (one case). Over-
all, 93% of the cases occurred in the lower
extremity. The most common sarcoma histo-
logic diagnosis seen in 53 tumors was MFH
(66%), which is interesting in light of the fact
that MFH comprises only a small percentage
(< 2%) of primary bone tumors in the gen-
eral population [45]. A histologic diagnosis
of osteosarcoma was made for 13 of the tu-
mors (16%), angiosarcoma or malignant epi-
thelioid hemangioendothelioma for six of the
tumors (8%), and fibrosarcoma for five of the
tumors (6%). Leiomyosarcoma, anaplastic
sarcoma, and pleomorphic sarcoma, one case
each, were also reported as diagnoses.
Not all of the case reports and series dis-
cussed predisposing conditions for the under-
lying bone infarctions; however, the reports
describing 29 patients mentioned that no pre-
disposing factors were apparent [3–5, 8, 11,
13, 14, 18, 20, 22, 24, 27, 30, 32, 38–40, 44].
Decompression (i.e., caisson disease), steroid
therapy, and alcoholism were the most com-
monly reported predisposing factors [14, 16,
17, 21, 23, 25, 26, 29, 32, 34, 43]. There was
only one report of infarct-associated sarcoma
in a patient with sickle cell anemia [33] de-
spite the fact that this condition affects 1 in
500 African American patients. Only three
reports of infarct-associated sarcoma were in
patients with sickle cell trait [9, 21, 37], which
affects approximately 8% of African Ameri-
cans but does not typically result in the same
degree of osteonecrosis as sickle cell anemia
[46]. The explanation for this apparent para-
dox is unknown. Other underlying conditions
included hereditary bone dysplasia, dyslipid-
emia, Gaucher disease, and lupus erythema-
tosus [28, 36, 42, 43]. Of the 78 patients re-
ported (including our eight patients), at least
44 had multifocal infarction [3, 4, 9–11, 14,
17–19, 21–24, 26–29, 31–34, 36–38, 41–44],

although this number may be artificially low
given that several of the reports did not men-
tion or provide ­images that might have shown
additional sites of infarction. Pathologic frac-
ture was reported in eight cases (in addition
to our three cases) [5–7, 29, 32, 35, 40, 41].
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
The remaining cases in which the present-
ing symptom was described showed that pain
was the most common malady, with an aver-
age duration of pain before diagnosis of 3.6
months [3–12, 14–16, 18, 19, 21–24, 26–29,
31–36, 38, 40–44].
Most reports in the literature included fig-
ures depicting imaging studies of the infarct-
associated sarcoma. Radiographs of 47 in-
farct-associated sarcomas (excluding those
in our series) were presented in the articles
[3–12, 14–16, 18–29, 31–41, 43, 44]. Based
on our analysis of the photographs, 30 of the
infarcts arising adjacent to sarcomas showed
the type 4 pattern [3–8, 12, 14, 18, 20, 22,
24–27, 32, 35, 37–41, 43, 44]. The type 3
pattern was observed in four cases and the
type 5 pattern, in four cases [9, 10, 14, 15, 21,
23, 28, 33]. For the remaining eight cases,
the picture quality was judged to be subop-
timal for classification [11, 16, 19, 29, 31, 34,
36, 39]. Sarcomas presented as areas of os-
teolysis and were frequently associated with
periosteal reaction, a soft-tissue mass, or a
pathologic fracture; 4 of the 13 osteosarco-
mas presented with a mineralized soft-tissue
mass [3–12, 14–16, 18–29, 31–41, 43, 44].
Photographs of CT scans [6–8, 11, 17, 20,
23, 34, 37, 42, 44], available for 13 of the le-
sions, showed features similar to those seen
on radiographs but with better definition of the
sarcoma and underlying infarct. Photographs
of diagnostic-quality MR images [6, 15, 17,
19, 23, 28, 36, 38] were available for eight of
the lesions and showed the typical serpentine
margins of osteonecrosis with variable inter-
nal signal intensities ranging from that of fat
on T1-weighted imaging to low signal intensi-
ty. Sarcomas were clearly differentiated from
the infarcts as masslike lesions of abnormal
signal intensity extending partially or com-
pletely through the cortex.
Thirteen articles discussed the skeletal
scintigraphic findings of infarct-associated
sarcoma [6, 8, 11, 20, 23, 28, 29, 34, 38, 39,
41, 42, 44]. Based on these articles, all sar-
comas showed increased uptake. However,
figures were available in only four of the ar-
ticles: Two depicted spot images showing in-
creased uptake of the sarcoma [8, 11], one
depicted a whole-body scan with increased
uptake only in the sarcoma [38] and one de-
AJR:205, October 2015 W439
Infarct-Associated Bone Sarcomas
picted a whole-body scan with multifocal up-
take in a patient with hereditary bone dys-
plasia [28].
Sarcoma arising in association with pre-
existing osteonecrosis is rare. Despite the
relatively common occurrence of epiphyseal
avascular necrosis, all reported cases of sar-
coma arising in osteonecrosis appear to have
occurred in association with metaphyseal or
diaphyseal bone infarcts [3–44]. Epiphyse-
al extension of bone infarction, however, was
common in our series, occurring in five of
nine cases. Five of nine cases in our series
occurred in the vicinity of the knee, in keep-
ing with summative data from the literature
showing that nearly 70% of infarct-associat-
ed sarcomas arise around the knee [3–44].
All of our patients presented with pain, and
three cases were diagnosed after pathologic
fracture. Of the eight patients in our series,
only three had an identifiable cause for their
infarct: Two patients were being treated with
steroids for inflammatory bowel disease, and
one had a diagnosis of hypertriglyceride-
mia—both of which have been implicated as
risk factors for the development of osteone-
crosis. When our series of patients is includ-
ed with those patients previously described
in the literature, fewer than 40% of cases
show an identifiable cause of infarction.
Based on a review of the radiology reports
of the patients treated at our institution for
MFH arising in association with bone in-
farction, the MFHs were commonly over-
looked on initial radiographs, likely because
of a combination of the features of the sar-
coma itself (subtle radiolucency adjacent to
the infarct with or without radiographic evi-
dence of cortical penetration) and of the dis-
tracting nature of the dense infarcts, the im-
aging manifestations of which were readily
evident on the radiographs. We postulate that
because infarcts are typically central in loca-
tion, because the MFHs appear to arise along
the periphery of the infarct, and because the
avascular nature of the infarcted areas may
make them relatively resistant to tumor pen-
etration, the sarcomas may penetrate the cor-
tex earlier than primary tumors. These fac-
tors could lead to symptoms (i.e., pain from
irritation of periosteal nerves) when the tu-
mor is relatively small and therefore is more
difficult to detect radiographically.
Comparison with prior studies, if available,
is advised to evaluate for new aggressive fea-
tures in the bone surrounding the infarct. If
there is a detectable change, the next appro-
priate step is MRI; in our series, MRI read-
ily showed the tumor as an enhancing mass
of abnormal signal intensity that was at least
partially invading the infarct and was always
penetrating the cortex, and a soft-tissue mass
was present in most cases. Preservation of at
least some fat signal intensity within the in-
farct was typical and should therefore not be
misinterpreted as a sign of an adipocytic tu-
mor. However, it is theoretically possible for
a bone sarcoma to completely obliterate all
findings of the original infarct. In these cases,
detection of bone infarction on prior studies
could support the diagnosis of infarct-associ-
ated sarcoma, which typically carries a worse
prognosis than a primary bone tumor, pos-
sibly warranting more aggressive treatment
[15]. The accepted treatment of infarct-asso-
ciated sarcoma is wide surgical resection. Al-
though an optimal treatment regimen has not
been established, it appears that chemothera-
py is a beneficial adjuvant treatment that can
help increase the dismal 2-year survival rate
from 24–62% [15].
Skeletal scintigraphy could also be useful
for the evaluation because infarct-associated
sarcomas result in increased uptake on skel-
etal scintigraphy, whereas uncomplicated in-
farcts often do not. Although a central or adja-
cent area of decreased uptake corresponding
to the infarcted area of bone was seen in the
cases of infarct-associated MFH in our se-
ries, this finding should not be considered
pathognomonic of infarct-associated sarco-
ma because a similar appearance has been de-
scribed with other entities such as giant cell
tumor of bone [47]. Recently, Dua et al. [17]
presented the MRI and FDG PET/CT studies
of a patient with multifocal infarction and as-
sociated sarcoma: Only the sarcoma showed
increased uptake. The bone infarcts in our se-
ries of patients tended to be multifocal, sug-
gesting that patients with relatively extensive
infarction are more likely to develop sarcoma.
Unfortunately, the fact that there are very
little data on bone infarction limits our un-
derstanding of infarct-associated bone sarco-
mas. To date, there are no published data on
the prevalence of bone infarcts, so the percent-
age of bone infarcts that undergoes sarcoma-
tous degeneration is also unknown. The inci-
dence of bone sarcoma arising in association
with infarct is low, with percentages of 0.6% or
1% reported [14, 33]. However, the incidence
may be underestimated because there may be
cases in which the bone sarcoma obscures the
original bone infarct and the infarct is over-
looked. Furthermore, because uncomplicat-
ed bone infarcts are often asymptomatic, their

prevalence is likely underestimated. The time
interval between bone infarct and the devel-
opment of bone sarcoma is unknown because
the time of the initial bone infarction is usu-
ally unknown. One of the patients in our series
developed a sarcoma 4 years after initiation of
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
steroid therapy, but it is generally assumed that
the development of sarcoma from infarction is
a relatively slow process based on the study of
Caisson workers who developed sarcoma 17–
22 years after they had left their jobs [32]. The
fact that immature infarcts (i.e., infarcts with
little or no sclerosis) seem to develop sarcoma
less commonly than mature infarcts with more
robust sclerosis supports a long latency. The
fact that most sarcomas appear to arise in as-
sociation with bone infarcts in which a well-
defined sclerotic margin is evident radiographi-
cally (type 4) suggests that this imaging pattern
may represent a predisposing factor for sarco-
ma development; however, this theory does not
explain why epiphyseal avascular necrosis with
a similar radiographic appearance is seldom, if
ever, associated with sarcomatous transforma-
tion. Furthermore, it is possible that the appar-
ent association between the type 4 pattern of
bone infarction and sarcoma development may
simply be because this pattern of infarction is
more readily detected and diagnosed as osteo-
necrosis than other patterns of infarction.
The argument has been made that sarco-
mas may simply develop adjacent to infarcts
by chance [12]. However, if this theory were
true, we would expect to see a different dis-
tribution of sarcoma histologic diagnoses
rather than a large majority of MFHs. In fact,
some of the early diagnoses of infarct-asso-
ciated fibrosarcoma and pleomorphic sarco-
ma would probably be reclassified as infarct-
associated MFH or another type of sarcoma
today [28]. Nearly 20% of cases of MFH of
bone treated at our institution had an asso-
ciation with bone infarction. This percent-
age is a greater percentage than has been
reported elsewhere and underscores the as-
sociation. On the other hand, only 1% of the
cases of osteosarcoma treated at our institu-
tion had an association with bone infarction
despite osteosarcoma being more prevalent
than MFH of bone. However, MFH has re-
cently been reclassified by the World Health
Organization as pleomorphic undifferentiat-
ed sarcoma, and previous diagnoses of MFH
might also be revised in light of more current
morphologic, immunohistochemical, and
molecular diagnostic approaches [48, 49].
We acknowledge four limitations of our
study. First, we acknowledge the aforemen-
W440 AJR:205, October 2015
Stacy et al.
tioned use of MRI as the criterion standard
for the diagnosis of infarction, which likely
led to a biased representation of the differ-
ent patterns of bone infarction encountered
on radiographs. Second, we acknowledge the
small sample size of infarct-associated sar-
comas in our series, although this series is
the largest series presented in the radiolo-
gy literature to date. Third, we acknowledge
that the characterization of the imaging ap-
pearance of infarct-associated sarcomas in
the literature based on photographic figures
is not ideal. Fourth, we note that the histo-
logic diagnosis of MFH may be different to-
day as a result of advances in morphologic,
immunohistochemical, and molecular ap-
proaches. Nevertheless, we present a case
series and review of an interesting subset of
malignancies that warrants recognition and
understanding by radiologists.
In summary, infarct-associated sarcomas
are rare. They appear to arise in long-stand-
ing mature infarcts, particularly those that oc-
cur in the lower extremity around the knee.
Many patients who develop infarct-associ-
ated sarcoma do not have a defined risk fac-
tor for the development of osteonecrosis, but
a large proportion of patients have multifo-
cal bone infarction. Pain is the most common
presenting symptom and should alert the ra-
diologist to carefully examine the bone sur-
rounding the infarct for signs of malignancy;
these signs can be extremely subtle and ini-
tially overlooked amid the distracting appear-
ance of the sclerotic infarction. The radiolo-
gist must be vigilant for this rare occurrence
and should be suspicious when poorly defined
areas of osteolysis (or mineralization) with
cortical thinning and bone destruction arise
next to a bone infarct. When infarct-associat-
ed sarcoma is suspected, MRI is recommend-
ed to confirm the development of a sarcoma
and show the extent of bone and soft-tissue in-
vasion. Infarct-associated bone sarcomas car-
ry a poor prognosis; therefore, it is important
to identify these malignancies early and to de-
termine whether metastasis is present to prop-
erly guide therapy.
Acknowledgments
We thank the Oncology Fellows of the De-
partment of Orthopaedic Surgery at the Uni-
versity of Chicago for allowing us to use the
Orthopaedic Oncology Program’s database
and image files.
References
1. Munk PL, Helms CA, Holt RG. Immature bone
infarcts: findings on plain radiographs and MR
scans. AJR 1989; 152:547–549
2. Saini A, Saifuddin A. MRI of osteonecrosis. Clin
Radiol 2004; 59:1079–1093
3. Furey JG, Ferrer-Torrer-Torells M, Reagan JW.
Fibrosarcoma arising at the site of bone infarcts: a
report of 2 cases. J Bone Joint Surg Am 1960;
42:802–810
4. Heater K, Collins PA. Osteosarcoma in associa-
tion with infarction of bone: report of two cases.
J Bone Joint Surg Am 1987; 69:300–302
5. Abdelwahab IF, Hermann G, Lewis MM, Klein
MJ. Transformation of an idiopathic bone infarct
into malignant fibrous histiocytoma in a female: a
case report. Bull Hosp Jt Dis Orthop Inst 1988;
48:197–203
6. Abdelwahab IF, Kenan S, Klein MJ, Lewis MM.
Case report: angiosarcoma occurring in a bone
infarct. Clin Radiol 1992; 45:412–414
7. Abdelwahab IF, Klein MJ, Hermann G,
Springfield D. Angiosarcomas associated with
­
bone infarcts. Skeletal Radiol 1998; 27:546–551
8. Abrahams TG, Hull M. Case report 394: malig-
nant fibrous histiocytoma (MFH) arising in an
infarct of bone. Skeletal Radiol 1986; 15:578–583
9. Arnold WH. Hereditary bone dysplasia with sar-
comatous degeneration: study of a family. Ann
Intern Med 1973; 78:902–906
10. Bertoni F, De Santis E, Laus M, Bravo JS. Malig-
nant fibrous histiocytoma of bone. Ital J Orthop
Traumatol 1978; 4:369–378
11. Breen TF, Healy WL. Malignant fibrous histiocy-
toma arising in medullary long bone infarcts: a
case report. Orthopedics 1987; 10:1169–1173
12. Cerilli LA, Fechner RE. Angiosarcoma arising in
a bone infarct. Ann Diagn Pathol 1999; 3:370–373
13. Dahlin DC, Unni KK. Bone tumors: general
aspects and data on 8,542 cases, 4th ed.
­
­Springfield, IL: Thomas, 1986:xv, 522
14. Desai P, Perino G, Present D, Steiner GC. Sarco-
ma in association with bone infarcts: report of five
cases. Arch Pathol Lab Med 1996; 120:482–489
15. Domson GF, Shahlaee A, Reith JD, Bush CH,
Gibbs CP. Infarct-associated bone sarcomas. Clin
Orthop Relat Res 2009; 467:1820–1825
16. Dorfman HD, Norman A, Wolff H. Fibrosarcoma
complicating bone infarction in a caisson worker: a
case report. J Bone Joint Surg Am 1966; 48:528–532
17. Dua SG, Purandare N, Shah S, Sharma AR, Zade
AA, Rangarajan V. Bone infarct-associated sar-
coma detected on FDG PET/CT. Clin Nucl Med
2011; 36:218–220
18. Dunham WK, Wilborn WH. Malignant fibrous
histiocytoma of bone: report of two cases and re-
view of the literature. J Bone Joint Surg Am 1979;
61:939–942
19. Duong S, Sallis JG, Zee SY. Malignant fibrous histio-
cytoma arising within a bone infarct in a patient with

sickle cell trait. Int J Surg Pathol 2004; 12:67–73
20. Endo M, Yoshida T, Yamamoto H, et al. Low-
grade central osteosarcoma arising from bone in-
farct. Hum Pathol 2013; 44:1184–1189
21. Frierson HF Jr, Fechner RE, Stallings RG, Wang
GJ. Malignant fibrous histiocytoma in bone in-
Downloaded from www.ajronline.org by 180.244.112.35 on 11/30/18 from IP address 180.244.112.35. Copyright ARRS. For personal use only; all rights reserved
farct: association with sickle cell trait and alcohol
abuse. Cancer 1987; 59:496–500
22. Galli SJ, Weintraub HP, Proppe KH. Malignant
fibrous histiocytoma and pleomorphic sarcoma in
association with medullary bone infarcts. Cancer
1978; 41:607–619
23. Gaucher AA, Regent DM, Gillet PM, Pere PG,
Aymard BM, Clement V. Case report 656: malig-
nant fibrous histiocytoma in a previous bone in-
farct. Skeletal Radiol 1991; 20:137–140
24. Grootboom MJ, Charles RW. Sarcoma associated
with bone infarction: a report of 2 cases. South Afr
Med J 1989; 75:293–294
25. Heselson NG, Price SK, Mills EE, Conway SS,
Marks RK. Two malignant fibrous histiocytomas
in bone infarcts: case report. J Bone Joint Surg
Am 1983; 65:1166–1171
26. Heselson NG, Webber BL, Goldberg S, Mills EE.
Bone infarction complicated by sarcomatous
change: a case report. South Afr Med J 1976;
50:1942–1944
27. Johnson LC, Vetter H, Putschar WG. Sarcomas
arising in bone cysts. Virchows Arch Pathol Anat
Physiol Klin Med 1962; 335:428–451
28. Kenan S, Abdelwahab IF, Hermann G, Klein MJ.
Malignant fibrous histiocytoma associated with a
bone infarct in a patient with hereditary bone dys-
plasia. Skeletal Radiol 1998; 27:463–467
29. Kitano M, Iwasaki H, Yoh SS, Kuroda K, Hayashi
K. Malignant fibrous histiocytoma at site of bone
infarction in association with DCS. Undersea
AJR:205, October 2015 W441
Infarct-Associated Bone Sarcomas
Biomed Res 1984; 11:305–314
30. Martinez-Tello FJ, Navas-Palacios JJ, Calvo-
Asensio M, Loizaga-Iriondo JM. Malignant fi-
brous histiocytoma of bone: a clinico-pathologi-
cal and electron microscopical study. Pathol Res
Pract 1981; 173:141–158
31. Matsuno T, Kaneda K, Takeda N. Development of
angiosarcoma at the site of a bone infarct. Clin
Orthop Relat Res 1996; 259–263
32. Mirra JM, Bullough PG, Marcove RC, Jacobs B,
Huvos AG. Malignant fibrous histiocytoma and
osteosarcoma in association with bone infarcts:
report of four cases, two in caisson workers.
J  Bone Joint Surg Am 1974; 56:932–940
33. Mirra JM, Gold RH, Marafiote R. Malignant (fi-
brous) histiocytoma arising in association with a
bone infarct in sickle-cell disease: coincidence or
cause-and-effect? Cancer 1977; 39:186–194
34. Péré P, Adolphe J, Delgoffe C, Froment N, Gaucher
A. Epiphyseal osteonecrosis, multiple bone infarc-
tions and malignant fibrous histiocytoma [in French].
Rev Rhum Mal Osteoartic 1984; 51:427–430
35. Petra M, Gibbons CL, Athanasou NA. Leiomyo-
sarcoma of bone arising in association with a bone
infarct. Sarcoma 2002; 6:47–50
36. Pins MR, Mankin HJ, Xavier RJ, Rosenthal DI,
Dickersin GR, Rosenberg AE. Malignant epithelioid
hemangioendothelioma of the tibia associated with a
bone infarct in a patient who had Gaucher disease: a
case report. J Bone Joint Surg Am 1995; 77:777–781
37. Resnik CS, Aisner SC, Young JW, Levine A. Case
report 767: osteosarcoma arising in bone infarc-
tion. Skeletal Radiol 1993; 22:58–61
38. Robbin MR, Pavlidakey PG, Getty P, ­Yarmohammadi
H. Chronic knee pain in an 80-year-old woman. Clin
Orthop Relat Res 2013; 471:3062–3068
39. Spainer SS. Malignant fibrous histiocytoma of
bone. Orthop Clin North Am 1977; 8:947–961
40. Torres FX, Kyriakos M. Bone infarct-associated
osteosarcoma. Cancer 1992; 70:2418–2430
41. Troop JK, Mallory TH, Fisher DA, Vaughn BK.
Malignant fibrous histiocytoma after total hip ar-
throplasty: a case report. Clin Orthop Relat Res
1990; 297–300
42. Yamamoto Y, Takakuwa Y, Kuroda M, et al. A
p53 gene mutation in malignant fibrous histiocy-
toma associated with bone infarction. Tohoku J
Exp Med 2011; 225:215–220
43. Michael RH, Dorfman HD. Malignant fibrous
histiocytoma associated with bone infarcts. Clin
Orthop Relat Res 1976; 118:180–183
44. Davies AM, Evans N, Grimer RJ. The jockey with
a mass in the thigh. Br J Radiol 1989; 62:183–184
45. Fletcher CD, Unni KK, Mertens F. World Health
Organization classification of tumours: ­pathology
and genetics of tumours of soft tissue and bone.
Lyon, France: IARC Press, 2002
46. Wintrobe MM, Greer JP. Wintrobe’s clinical
­hematology, 12th ed. Philadelphia, PA: Wolters
­Kluwer Health/Lippincott Williams & Wilkins, 2009
47. Murphey MD, Nomikos GC, Flemming DJ, Gannon
FH, Temple HT, Kransdorf MJ. From the archives of
AFIP: imaging of giant cell tumor and giant cell re-
parative granuloma of bone—radiologic-pathologic
correlation. RadioGraphics 2001; 21:1283–1309
48. Mertens F, Romeo S, Bovée JV, et al. Reclassifica-
tion and subtyping of so-called malignant fibrous
histiocytoma of bone: comparison with cytoge-
netic features. Clin Sarcoma Res 2011; 1:10
49. Romeo S, Bovée JV, Kroon HM, et al. Malignant
fibrous histiocytoma and fibrosarcoma of bone: a
re-assessment in the light of currently employed
morphological, immunohistochemical and molecu-
lar approaches. Virchows Arch 2012; 461:561–570