AJR 2005; 184:1762-1767
© American Roentgen Ray Society
Imaging Findings in Desmoplastic Fibroma of Bone: Distinctive T2 Characteristics
Matthew A. Frick1,
Murali Sundaram1,2,
Krishnan K. Unni3,
Carrie Y. Inwards3,
Nicola Fabbri4,
Federico Trentani4,
Patrizia Baccini5 and
Franco Bertoni5
1 Department of Radiology, Mayo Clinic, West-2, 200 First Street, SW, Rochester,
MN 55905.
3 Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester,
Minnesota.
4 Servizio di Ortopedico, Istituto Ortopedico Rizzoli, Bologna, Italy.
5 Servizio di Anatomia e Istologia Patologica, Istituto Ortopedico Rizzoli,
Bologna, Italy.
Received July 21, 2004;
accepted after revision September 9, 2004.
Address correspondence to M. A. Frick.
2 Present Address: Department of Radiology, Cleveland Clinic Foundation, 9500
Euclid Ave., Cleveland, OH 44915.
Abstract
OBJECTIVE. The purpose of this study was to evaluate the imaging
features of desmoplastic fibroma of the bone, with an emphasis on MRI signal
characteristics.
CONCLUSION. Significant T2 shortening of a nonsclerotic fibroosseous
lesion should place desmoplastic fibroma high among the diagnostic
considerations.
Introduction
In 1958, Jaffe [1]
first described desmoplastic fibroma of the bone as a distinct entity when he
documented five cases of a previously unclassified osseous fibrous tumor that
histologically was similar to abdominal desmoid tumor. Desmoplastic fibroma of
the bone is now considered the intraosseous counterpart of the common
soft-tissue desmoid or fibromatoses
[2]. This rare bone tumor is
most frequently described in single case reports. The largest series in the
radiology literature [3]
consists of 18 cases and primarily discusses the radiographic features of the
tumor. To our knowledge, no series has analyzed the MRI characteristics of
desmoplastic fibroma of the bone. On the basis of material obtained from the
pathology files at two institutions, we describe and discuss the imaging
features of desmoplastic fibroma of the bone, with an emphasis on radiologic
and MRI characteristics. We compared our MRI findings with descriptions in a
previous case report [4],
paying special attention to the frequency and extent of the T2 shortening
within the lesion. A short T2 is rarely encountered in lytic bone lesions and
soft-tissue masses. One common soft-tissue lesion that may have a short T2 is
fibromatosis [5,
6]. The purpose of this study
was to determine whether a short T2 applies to its intraosseous counterpart
with any frequency and, if so, to semiquantitatively describe its extent
within a lesion and discuss its value as a diagnostic sign in the realm of
fibroosseous lesions of the bone.
Materials and Methods
We retrospectively reviewed the pathology files at two institutions with an
orthopedic oncology referral practice for cases of tumors classified as
desmoplastic fibroma of the bone. We found 96 cases. The histology in all
cases at both locations was reviewed by four pathologists, two from each
institution. One case was excluded because the pathologists thought it
represented a nonossifying fibroma. This left 95 patients in the study. The
medical records or consultation letters were reviewed to determine the
patients' sex and age at the time of lesion discovery. Of the 95 patients, 45
had some form of imaging study (i.e., radiology, CT, MRI, or a combination)
available for review. Radiographs of 38 patients were evaluated to determine
the lesion's location, site within the bone, matrix, cortical integrity,
margins, and size. CT images were available for 17 patients, 14 of whom had
corresponding radiographs and six, corollary MR images. The CT images were
evaluated for lesion matrix, cortical integrity, and the presence of
associated soft-tissue mass. A soft-tissue mass was considered present when
extraosseous extent of tumor with breeched cortex and periosteum were noted.
MRI was performed on multiple magnets on two continents; conventional
T1-weighted images were available for all 14 patients. The files of nine
patients also included conventional T2-weighted images; in two cases the exact
TR and TE were unknown, but images were determined to be T2. The files of four
patients included proton-density (PD) images and the file of one patient,
T1-weighted images in isolation. Of patients with MR images, 10 had
corresponding radiographs and six, corresponding CT images. MR images were
reviewed to determine cortical integrity, soft-tissue mass and concordance
with radiographic dimensions of the intraosseous component of the lesion, and
the signal relationship to muscle on T1- and T2-weighted sequences. In lesions
with areas of short T2, the extent of the area as a percentage of lesion size
was estimated as less than 25%, 25-50%, 50-75%, or greater than 75%. Exact MR
scanning parameters were available for 11 patients and were partially known in
another case. T1-weighted images ranged from 430-683 msec (TR) and 9-25 msec
(TE). T2-weighted sequences ranged from 1,800-5,300 msec and 90-117 msec,
respectively. Proton density-weighted images were available for four patients;
the TR ranging from 2,000-2,240 msec and the TE, from 30-39 msec. Spoiled
gradient recalled echoes were also available for one patient, fat-saturated T2
sequences for three patients, and PD fat-saturated sequences for one patient.
Gadolinium-enhanced images were available for two patients. The images were
also reviewed for the presence of associated pathologic fractures. The imaging
findings were documented by two radiologists working in consensus who were
aware of the diagnosis at time of review. Institutional review board approval
of the study was obtained.
View larger version (142K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1. —Anteroposterior radiograph of pelvis of 30-year-old man shows
well-marginated, nonsclerotic, osteolytic lesion (white box) of left
superior pelvis that extends into left superior pubic ramus abutting pubic
symphysis. A few thin, internal bony ridges are present.
|
|
View larger version (58K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —Lateral radiograph of right forearm of 75-year-old man shows
long, well-marginated, unmineralized, ulnar diaphyseal lesion with associated
endosteal scalloping. Lesion has no significant trabeculations.
|
|
Results
The sex of 93 patients was known; there was a slight predominance of men
(56 versus and 37 women). At time of diagnosis, the age of the patients, which
was not known in two cases, was between less than 1 year and 86 years (mean,
26.3 years).
Based only on available imaging, there were 27 lesions within the tubular
bones; four were diaphyseal, nine metadiaphyseal, nine metaphyseal, and five
epiphyseal (juxtaarticular). The lesions were located centrally in 61% of
patients and eccentrically in 39%. The host bone in the remaining 18 cases was
the calcaneus (n = 4), pelvis (n = 6), mandible (n
= 4), vertebrae (n = 3), and skull base (n = 1).
Radiographs showed the matrix was osteolytic in 9 lesions (24%) (Figs.
1 and
2), osteolytic with coarsened
ridge-like trabeculae in 24 lesions (63%)
(Fig. 3), and mixed lytic and
mildly sclerotic in five lesions (13%) (Figs.
4A and
4B). Cortical breaching was
present in 20 (53%) of 38 lesions (Fig.
3). Radiology showed the margins were well defined in 13 lesions
(34%) (Figs. 1 and
2), partially well defined in
23 lesions (61%), and not well seen in two lesions (5%). The margins were
sclerotic in 46% of lesions. The average size of the lesions was 8.1 x
3.9 cm in greatest longitudinal and transverse dimensions, respectively
(range, 3-16 cm and 1.5-9 cm, respectively). Five patients (13%) had evidence
of pathologic fracture.
View larger version (100K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3. — Anteroposterior (A) and lateral (B)
radiographs of distal right femur in 30-year-old man show large eccentric
lesion in femur that, unlike in Figs.
1 and
2, reveal markedly coarse
trabeculations within lesion and cortical breaching posterolaterally.
|
|
View larger version (122K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4. —Anteroposterior (A) and lateral (B) radiographs
of upper lumbar spine in 24-year-old man show well-defined osteolytic lesion
of L2 vertebra with extension into pedicle. A few thin, internal trabeculae
also are present.
|
|
CT revealed osteolytic (65%) (Fig.
4C) or mixed osteolytic and mildly sclerotic (35%) matrix
patterns. Cortical destruction was seen on CT in 15 (88%) of 17 patients
(Fig. 4C) and a soft-tissue
mass extending beyond the original destroyed cortex, in seven patients
(41%).
In all cases, T1-weighted sequences showed that the signal intensities
within the lesions were isointense or hypointense to adjacent normal muscle
(Figs. 5 and
6). Lesions were at least in
part isointense or hypointense to muscle in eight (89%) of nine cases for
which conventional T2-weighted spin-echo or fat-saturated sequences were
available (Figs. 7,
8,
9,
10,
11). The one case in which the
lesion did not contain areas of low T2 signal was confounded by the presence
of an associated pathologic fracture, and much of the increased signal
intensity may have been caused by edema. In the eight lesions with short T2,
the low signal encompassed more than 75% of the tumor in five cases and
between 50% and 75% in three cases. In no case with a low T2 signal in the
lesion did the amount of low signal encompass less than 50% of the tumor. In
one patient, PD sequences showed high signal. Conventional and other
T2-weighted sequences were not available for this patient.
View larger version (79K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7. —Coronal PD-weighted image (TR/TE, 4,560/39) image of
30-year-old man also shown in Figs.
1 and
5. More than 75% of lesion has
T2 shortening, making it virtually indistinguishable from surrounding muscle.
Peripheral signal intensity is high.
|
|
View larger version (134K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 9. —Sagittal T2-weighted image (TR/TE, 2,000/80) of lumbar spine
of 24-year-old man also shown in Figs.
4A,
4B, and
4C shows T2 shortening of
almost entire lesion. Entire involved vertebral body is slightly lower in
signal intensity than surrounding normal vertebral bodies.
|
|
View larger version (82K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 10. —42-year-old man with vertebral lesion that had been
percutaneously biopsied. Histologic differential diagnosis was fibrous
dysplasia versus desmoplastic fibroma. Axial T2-weighted image (TR/TE,
4,000/88) shows more than 90% of lesion is hypointense to muscle, favoring
diagnosis of desmoplastic fibroma.
|
|
On the fat-suppressed sequences, the lesions remained hypointense to muscle
(Figs. 7 and
11). In two patients for whom
gadolinium-enhanced images were available, the lesions had prominent but
somewhat heterogeneous enhancement.
There was no significant discordance between radiographic and MR images in
the intraosseous extent of the tumor.
Discussion
Desmoplastic fibroma of the bone, the intraosseous counterpart of
soft-tissue fibromatoses, also is known as desmoid tumor of bone. Desmoplastic
fibromas account for up to 0.1% of primary bone tumors
[2]. Recognition of
desmoplastic fibroma is important because on radiology and histology, the
lesion may be mistaken for an indolent, benign fibrous lesion or more
aggressive spindle-cell sarcomas
[7]. Because of its relative
rarity, desmoplastic fibroma of the bone has been described in only a few
small series in the orthopedic, pathology, and radiology literature.
Cumulatively, these series are far out-numbered by single case reports.
In our series comprising a selected population from two referral centers,
the tumor was seen most frequently in patients in the third decade of life. In
other series, they have been found in the third to fourth decade
[3,
7,
8]. Reports of the sex
distribution of patients with this tumor are conflicting. Crim et al.
[8] reviewed 107 cases in the
literature, in addition to their seven cases, and did not find a
predisposition based on sex predilection. Results in reviews of fewer patients
suggest a slight predominance of men
[5]. In our series, the sex was
known in 93 patients and the male-female ratio was 1.5:1.
The only two series examining the radiologic features of desmoplastic
fibroma of bone [3,
8] emphasize that the lesion is
osteolytic and does not contain significant mineralized matrix (Figs.
1 and
2) with a favored metaphyseal
origin. In a review of 79 published radiographs and their seven cases, Crim et
al. [8] found only six cases in
the diaphysis. The authors also emphasize the presence of ridges of intact
bone near the periphery that are the result of uneven bone destruction
(Fig. 3). These observations
were largely borne out by our review, in which metaphysis and diametaphysis
were equally involved and exclusive diaphyseal involvement of a tubular bone
(4 of 27, 15%) was the rarest site (Fig.
2). In the series by Taconis et al.
[3], the most common location
in a long bone was the diametaphysis; only one lesion included in their series
was exclusively diaphyseal in location.
At the time of discovery, the average size of the lesions in our study was
approximately 8 x 4 cm, with cortical breaching a common finding (53%).
Cross-sectional imaging revealed a soft-tissue mass in 41% of cases on CT and
in 57% of cases on MRI. The morphologic appearance on the images was that of a
relatively slow-growing lesion with focal aggressive features. The spectrum of
radiologic features of desmoplastic fibroma of the bone overlap the
radiographic patterns described for low-grade central osteosarcoma
[9], and in some cases without
cortical breaching may resemble fibrous dysplasia. The histologic distinction
between these entities and fibrosarcoma, especially on a percutaneous bone
biopsy, can be difficult, which has obvious management ramifications.
Case reports of desmoplastic fibroma of the bone mention the presence of
low to intermediate signal intensity on T2-weighted MRI sequences and vary in
the degree of emphasis on this observation. In 2000, Vanhoenacker et al.
[4] described a large
desmoplastic fibroma of the femur and also tabulated the MRI features in 17
additional cases. In six of the cases, the T2-weighted sequences were not
illustrated; in seven cases, the authors refer to the presence of a low or
intermediate signal on T2-weighted sequences. Subsequently, two reports
[10,
11] have described this tumor
in vertebra, highlighting the rare location. In both cases, T2-weighted
sequences showed a large lesion that was isointense and minimally hyperintense
to muscle. The three patients with vertebral lesions in our series had marked
T2 shortening (Figs.7,
8,
9,
10). In 11 of 14 patients for
whom MR images with T2-weighted sequences were available for review, the most
common signal of the lesion was low, occupying more than 75% of the lesion in
six patients and between 50% and 75% in three patients. Irrespective of
location (i.e., long tubular bone, small tubular bone, pelvis, or vertebra), a
short T2 was the predominant signal pattern (Figs.
5,
6,
7,
8,
9,
10).
Foci of short T2 in osteolytic lesions of the bone have been described in
giant-cell tumors, fibrous dysplasia, lymphoma, and leiomyosarcoma
[12-15].
T2 shortening in giant cell tumors was attributed to hemosiderin
[13]. Foci of collagen were
believed to contribute to short T2 in fibrous dysplasia
[12]. The cause in lymphoma
was unresolved because there was no histologic correlation with fibrosis
[14]. In leiomyosarcoma, it
was speculated that the muscle component of the tumor was responsible for the
short T2 [15]. A review of the
osseous component of the lesion in illustrations in these papers suggests that
a predominance of T2 signal is most apparent in giant-cell tumors
[9] and leiomyosarcomas.
Histologically, neither type of tumor enters into the differential diagnosis
of osseous desmoid and thus would not be mistaken for that entity. The short
T2 in patients with leiomyosarcoma of the bone was noted only on
non-fat-suppressed, T2-weighted, spin-echo sequences; on the fat-suppressed
sequences, all tumors were hyperintense. In our series, four patients had
T2-weighted, fat-saturated sequences; two had more than 50% and two more than
75% of T2 shortening within the tumor.
A predominantly osteolytic lesion with a large area of T2 shortening has a
limited differential diagnosis. It appears this feature in osseous desmoid
distinguishes it from lesions that it could be mistaken for tissue from a
percutaneous biopsy. In this series, a diagnosis of desmoplastic fibroma could
be made in one patient (Fig.
10) seen in consultation by one of us, even though it was unclear
whether there was a histologic difference between the two lesions. Surgery was
appropriately performed, and desmoplastic fibroma was histologically
confirmed.
The tumor in one patient in our series had a long T2; however, it was
associated with a pathologic fracture. Another patient had a long T2 on PD
images, and the lesion would probably have been hyperintense on a T2-weighted
sequence. It is likely the lesion was more cellular than the others in our
series, although the pathologist could not confirm this speculation. This
patient had been seen by pathologists at other institutions whose opinions
were divided between desmoplastic fibroma and grade 1 fibrosarcoma (Unni K,
personal communication, March 1, 2002). In either case, the management for the
two types of tumors is similar.
Although the preponderance of lesions with T2 weighting in our series (90%)
and in previous case reports had significant T2 shortening, we anticipate the
finding will be absent in some patients, as was the case with the evolutionary
MRI experience of soft-tissue desmoids. This study, however, confirms that the
MRI signal characteristics of desmoplastic fibroma of the bone is shared by a
subset of its soft-tissue counterpart, analogous to the histologic
similarities of this tumor in bone and soft tissue. The radiologic features of
a predominantly osteolytic lesion with prominent T2 shortening on MR images
make a diagnosis of intraosseous desmoid plausible.
Acknowledgments
We would like to pay special thanks to Scott Stacy of the University of
Chicago for his assistance in providing two cases for this review.
References
- Jaffe HL. Tumors and tumorous conditions of the bones
and joints. Philadelphia, PA: Lea and Febiger, 1958:298
-303
- Fornasico V, Pritzker KPH, Bridge JA. Desmoplastic fibroma of bone.
In: Fletcher CDM, Unni KK, Merten SF, eds. The World Health
Organization classification of tumours. Pathology and genetics of tumours of
soft tissue and bone. Lyon: IARC Press, 2002:288
- Taconis WK, Schütte HE, van der Heul RO. Desmoplastic fibroma
of bone: a report of 18 cases. Skeletal Radiol1994; 23:283
-288[Medline]
- Vanhoenacker FM, Hauben E, De Beuckeleer LH, Willemen D, Van Marck
E, De Schepper AM. Desmoplastic fibroma of bone: MRI features.
Skeletal Radiol2000; 29:171
-175[Medline]
- Sundaram M, McGuire MH, Schajowicz F. Soft-tissue masses:
histologic basis for decreased signal (short T2) on T2-weighted MR images.
AJR 1987;148:1247
-1250[Abstract/Free Full Text]
- Vandevenne JE, De Schepper AM, De Beuckeleer L, et al. New concepts
in understanding evolution of desmoid tumors: MR imaging of 30 lesions.
Eur Radiol 1997;7:1013
-1019[Medline]
- Inwards CY, Unni KK, Beabout JW, Sim FH. Desmoplastic fibroma of
bone: radiographic analysis. Cancer1991; 68:1978
-1983[Medline]
- Crim JR, Gold RH, Mirra JM, Eckardt JJ, Bassett LW. Desmoplastic
fibroma of bone: radiographic analysis. Radiology1989; 172:827
-832[Abstract/Free Full Text]
- Andresen KJ, Sundaram M, Unni KK, Sim FH. Imaging features of
low-grade central osteosarcoma of the long bones and pelvis.
Skeletal Radiol2004; 33:373
-379[Medline]
- Kong KY, Kang HS, Jung HW, Kim JJ, Lee CK. MR findings of
desmoplastic fibroma of the spine. A case report. Acta
Radiol 2000;41:89
-91[Medline]
- Juergens KU, Bullmann V, Link TM, Brinkschmidt C, Heindel W.
Desmoplastic fibroma in the thoracic spine: an unusual localization of a rare
primary bone tumor. Eur Radiol2001; 11:273
-275[Medline]
- Norris MA, Kaplan PA, Pathria M, Greenway G. Fibrous dysplasia:
magnetic resonance imaging appearance at 1.5 Tesla. Clin
Imaging 1990;14:211
-215[Medline]
- Aoki J, Tanikawa H, Ishii K, et al. MR findings indicative of
hemosiderin in giant-cell tumor of bone: frequency, cause, and diagnostic
significance. AJR1996; 166:145
-148[Abstract/Free Full Text]
- White LM, Schweitzer ME, Khalili K, Howarth DJ, Wunder JS, Bell RS.
MR imaging of primary lymphoma of bone: variability of T2-weighted signal
intensity. AJR1998; 170:1243
-1247[Abstract/Free Full Text]
- Sundaram M, Akduman I, White LM, McDonald DJ, Kandel R, Janney C.
Primary leiomyosarcoma of bone. AJR1999; 172:771
-776[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
