|
|
||||||||
Original Research |
1 Department of Radiology, Chelsea and Westminster Hospital, 369 Fulham Rd.,
London SW10 9NH, England.
2 Department of Surgery, Royal Marsden Hospital, London SW3 6JJ, England.
3 Department of Pathology, Royal Marsden Hospital, London SW3 6JJ,
England.
4 Department of Radiology, Royal Marsden Hospital, London SW3 6JJ,
England.
Received October 27, 2004;
accepted after revision January 4, 2005.
Address correspondence to J. C. Lee.
Abstract
|
|
|---|
MATERIALS AND METHODS. Sixty patients had imaging available for retrospective review of which 29 had preoperative MRI and final histopathologic diagnosis of aggressive fibromatosis.
RESULTS. The average age at diagnosis was 41.3 years with a female-to-male sex ratio of 1.2:1. Twenty lesions were extraabdominal; six, intraabdominal; and three, in the abdominal wall (classic desmoid). The average tumor size was 6.4 cm (range, 2.2-13.7 cm). Intraabdominal aggressive fibromatosis produced the largest tumors, averaging 9.5 cm. Most lesions were ovoid (52%) or infiltrative (34.5%) in outline with an irregular or lobulated contour (76%). The lesions crossed major fascial boundaries in 31% of cases overall and in 66% of patients referred for recurrent disease. On MRI, homogeneous isointensity or mild hyperintensity on T1-weighted images and heterogenous high signal on T2-weighted or STIR images were seen. All lesions enhanced after IV gadolinium, usually avidly. In contrast to previous reports, 38% of cases failed to show low signal on all pulse sequences and no abnormalities were seen in local bone structures. Histology showed sheets of bland spindle cells in dense collagen and did not vary with the MRI signal characteristics of the lesion. Patients referred for recurrent disease were most likely to have a recurrence after surgery. MRI and pathology findings did not predict recurrence.
CONCLUSION. Accurate diagnosis and staging of aggressive fibromatosis by MRI have important treatment and prognostic implications.
Keywords: desmoid tumor fibromatosis Gardner's syndrome mesenchymal tumor
|
|
|---|
Fibromatosis can be divided into two major groups with several subdivisions by clinical presentation, patient age, and natural history (Appendix 1). Occurring in almost any anatomic location, it can be divided into superficial and deep subtypes. The superficial fibromatoses are usually small (< 5 cm) slow-growing lesions that rarely involve deep structures. The deep group consists of rapidly growing lesions that often reach a large size and have a high tendency to recur after treatment, hence the term "aggressive fibromatosis." This group principally involves the musculature of the trunk and the extremities. The term "desmoid tumor" (desmos in Greek means band), first used by Mueller in 1838 to emphasize the bandlike or tendonlike consistency of these lesions, is synonymous with this type of fibromatosis [1]. However, for this study, we have defined the term "desmoid" to be synonymous with abdominal wall aggressive fibromatosis.
|
We reviewed the clinicoradiologic features of aggressive fibromatosis from the Royal Marsden Hospital Sarcoma Unit, which has one of the largest databases of soft-tissue tumors worldwide. We assessed multiple MRI parameters to characterize this condition and correlated each case with pathology and risk of postoperative recurrence. Our results are compared with previously published data on aggressive fibromatosis, and we discuss any relevant discrepancies.
|
|
|---|
Patient age at presentation and patient sex, tumor type (i.e., primary or recurrent), treatment, and outcome were obtained from the hospital electronic patient record system. The method of diagnosis was recordedthat is, percutaneous biopsy, incisional biopsy, or excisional biopsy. Tumor site, size, shape, contour, and margins; bone changes; crossing of fascial boundaries; and T1, T2, and STIR signal intensities and heterogeneity were assessed on MRI. Tumor size was recorded from the MR images as a mean of three orthogonal measurements. Two dedicated sarcoma radiologists reviewed each MRI with findings reached by consensus. The degree of low signal on all sequences, signal intensity from local bone structures, and patterns of behavior after administration of gadolinium were documented. The group was analyzed as a whole and divided into the subgroups previously described in Appendix 1.
Pathology was analyzed for cellularity, pleomorphism, clearance margins, vascularity, and presence of mitoses.
|
|
|---|
Of the 29 cases, 20 were extraabdominal; six, intraabdominal; and three, in the abdominal wall (desmoid). The average age was 41.3 years with a female-to-male ratio of 1.2:1. Table 1 summarizes the patient demographic data. It should be noted that the desmoid group is slightly younger on average and has a high female predominance (Figs. 1A and 1B).
|
|
|
The extraabdominal group included the limb girdle and extremity cases. Eleven tumors were located about the shoulder girdle (Fig. 2), five around the pelvic girdle or thigh, two on the trunk (Figs. 3A and 3B), and two in the popliteal fossa. The intraabdominal group included one case of Gardner's syndrome (colonic polyposis, osteomas, and abdominal fibromatosis) (Figs. 4A and 4B).
|
|
|
|
|
|
|
Moderate to marked low signal across all sequences was identified in nine (31%) of the 29 cases (Figs. 3A, 3B, 4A, and 4B). Another nine cases displayed only minimal low signal, and 11 cases (38%) showed no low signal.
After injection of IV gadolinium, all tumors showed enhancement that was either avid (15 cases) (Figs. 5A, 5B, and 6B) or moderate (three cases) (Fig. 1B) in intensity.
|
Histologic analysis showed sheets of bland spindle cells in dense collagen. Small slitlike vessels were seen, usually with mild perivascular edema and chronic inflammatory cells, especially mast cells. Nuclear atypia was not seen, and few or no mitoses were present (Fig. 7) in all cases. The histology was uniformly the same across the patient group with no differences in primary or recurrent tumors, site, patient sex or age, or MRI signal intensities or heterogeneity.
|
Outcome was influenced predominantly by whether the patient was referred for primary treatment or recurrence. Of the 23 cases referred with primary disease, 13 were treated surgically, nine conservatively, and one received adjuvant therapy only. Of the 13 surgically treated patients, three had postoperative recurrence. In the conservatively managed group, seven were stable, one had a decrease in size, and one had disappeared. The one patient treated with adjuvant therapy onlythat is, radiation therapy and chemotherapyhad stable disease 6 years after treatment. Of the six recurrent cases, four had surgery and two had adjuvant therapy. Five had progressive disease, and one was lost to follow-up.
|
|
|---|
|
MRI provides the optimum method of tumor evaluation, both before and after surgery because of its superior soft-tissue resolution compared with other imaging techniques. The relationship of the tumor to important local structures, such as nerves, vessels, deep viscera, and bone, can only be accurately assessed using MRI. Involvement of these structures is likely to influence initial patient management and may represent tumor recurrence if seen on follow-up scans.
The main clinical differential diagnosis of aggressive fibromatosis is malignant soft-tissue sarcoma, extranodal lymphoma, and benign myositis ossificans and arteriovenous malformation. The two benign diagnoses can be characterized on imaging by showing typical calcification on conventional radiography or CT in the former and signal flow voids and calcification on MRI in the latter. Extranodal lymphoma presenting as a muscular mass is rare.
Therefore, the main differential diagnosis of aggressive fibromatosis on imaging is malignant soft-tissue sarcoma. Can MRI accurately differentiate aggressive fibromatosis from soft-tissue sarcoma? As a general rule, most soft-tissue sarcomas grow as space-occupying intramuscular lesions, enlarging in a centripetal fashion that pushes rather than infiltrates local structures. They typically form a pseudocapsule as they enlarge by compressing normal tissue and usually respect fascial boundaries, remaining within anatomic compartments until late in their course. As they enlarge further, soft-tissue sarcomas may outstrip their blood supply resulting in central necrosis [5]. In contrast, our series reveals that aggressive fibromatosis typically shows an infiltrative growth pattern; crosses fascial boundaries; and does not show central necrosis, even in the largest lesions. Additional MRI findings that point toward the diagnosis of aggressive fibromatosis are bands of low signal intensity across all sequences and uniform moderate to avid enhancement after gadolinium administration. Low signal intensity on all pulse sequences is said to be characteristic of, but not specific for, aggressive fibromatosis [6]. Initial reports emphasized this finding and used it as a means of differentiating aggressive fibromatosis from malignant soft-tissue sarcomas [7]. However, low signal is not seen in all cases of aggressive fibromatosis, and prominent low signal intensity, particularly on T2-weighted images, may be seen in pigmented villous nodular synovitis, giant cell tumors of the tendon sheath, calcified masses, and malignancies such as fibrosarcoma or malignant fibrous histiocytoma [6]. These lesions complete the differential diagnosis.
|
It should be stressed that although MRI can provide clues to the correct diagnosis of aggressive fibromatosis, tissue confirmation is required for a final diagnosis in all cases. Aggressive fibromatosis generally presents as a palpable abnormality within the peripheral musculature, which is usually amenable to percutaneous biopsy. The majority of the patients in this series (22/29) underwent non-imaging-guided percutaneous core biopsy in the surgical outpatient clinic and avoided an invasive diagnostic procedure under anesthetic. For more complex masses involving the abdominal wall or peritoneum, imaging guidance may be required. In all cases, at least three 14-gauge cores should be obtained for accurate histologic assessment [8].
In our experience, the most common findings of aggressive fibromatosis on MRI are of an ovoid or infiltrative lesion that is lobulated or irregular in outline and crosses fascial boundaries in almost one in three cases. An infiltrative growth pattern is seen almost twice as often in the recurrent cases than in the primary cases, which may reflect the more aggressive nature of these lesions. Typically, aggressive fibromatosis lesions are homogeneously isointense on T1-weighted images, are heterogeneously high signal on T2-weighted and STIR images, and enhance avidly after IV gadolinium administration. Bands of low signal intensity within the lesion on all sequences were seen in 18 (62%) of the 29 cases. These findings are generally in concordance with previously published data [5, 6, 9-13].
Where this study differs from previously published work on aggressive fibromatosis is the absence of changes in local bone structures and in the degree of low signal within the lesion. Previous studies of aggressive fibromatosis suggest that bone changes occur in up to 37% of the cases [14], a finding not backed up by our series. It is recognized that bone changes are more common in recurrent aggressive fibromatosis [15], although this was not seen in the three extraabdominal recurrent cases in our study. No abnormalities of local bone structures were seen on MRI despite 20 of the cases being extraabdominal in location, with no lesion being farther than 4 cm from bone. The discordance in bone changes may be explained by the small numbers of cases involved in our study and previous studies, which vary from three [16] to 40 [9]. In addition, most of the previous descriptions are based on radiography findings rather than MRI findings [14].
Although bands of low signal intensity were seen on all sequences in 62% of the cases of this series, this finding was classified as marked in only three cases (10%). Furthermore, this finding was classified as minimal or absent in 20 (69%) of the 29 cases in our study. It has been hypothesized that the lack of low signal in some cases of aggressive fibromatosis may be due to the cellular component within these lesions. Sundaram et al. [16] attempted to explain this variability by comparing the T2-weighted MRI appearances with pathology in three cases of aggressive fibromatosis. Two of these lesions had decreased signal intensity on T2-weighted images, which was said to reflect hypocellularity and abundant collagen on histology. In the third case, high signal intensity on T2-weighted images was seen, and this correlated with marked cellularity [16]. In our study, as in larger subsequent studies, the MR-pathologic findings of aggressive fibromatosis are more variable [9, 10]. The histopathologic findings in our series were uniform across all cases, independent of signal characteristics on MRI. The heterogeneous T2 and STIR signal intensity pattern, which was the most common pattern in our study, may correspond to varying proportions of cellular tissue, myxoid tissue, and collagen in each lesion [9], although this could not be accurately quantified in our samples.
Compared with other forms of aggressive fibromatosis, abdominal wall desmoid tumors are known to have a lower postoperative recurrence rate and to occur in younger women, frequently around the time of pregnancy [17]. Only three cases were included in this study, making meaningful analysis of this subgroup difficult. The small number in this series presumably reflects common surgical practice in the United Kingdom, where these lesions are resected in secondary rather than tertiary centers.
The risk of recurrence was not related to patient age or sex or to tumor site or size. On MRI, tumor shape, contour, margins, T1 and T2 signal intensities, and contrast enhancement patterns did not predict recurrence but did help to characterize the lesion. As in other rare conditions, the relatively small numbers involved make statistical analysis difficult. The only major risk factor for postoperative recurrence was if the patient was referred for recurrent disease. In five of the six patients referred for recurrent disease in our study, disease recurred or progressed during the study interval.
Management of extraabdominal fibromatosis is a challenge because this disease does not respect the usual surgical rules relating to resection and recurrence. Surgical excision with clear margins is the goal. However, if the tumor is deemed inoperable because of involvement of significant neurovascular structures or poor patient performance status, then conservative management, adjuvant treatment with radiation therapy [18], chemotherapy, or hormonal antiestrogen treatment can be used [19]. It is important to note that after a period of rapid and alarming growth of either primary or recurrent disease, spontaneous growth arrest may take place, the so-called "plateau phase," and can be maintained indefinitely [20]. Results of treatment need to reflect this because growth arrest cannot be claimed as a therapeutic success. The most convincing supporting data come from the Mayo Clinic where stable disease was seen in 60 of 68 patients over an average follow-up period of 6.3 years [21].
In conclusion, aggressive fibromatosis is the likely diagnosis if, on MRI, the mass is ovoid or irregular, crosses fascial compartments, is isointense on T1- and heterogeneously hyperintense on T2-weighted images, and shows bands of low signal on all sequences. Local changes in bone are not typical. A core-cut biopsy, which may be imaging-guided, is sufficient for histopathologic characterization. MRI is the imaging method of choice for both preoperative planning and posttreatment monitoring. Aggressive fibromatosis is a challenging disease for both surgeons and radiologists alike. Awareness of this diagnosis in a young or middle-aged adult presenting with a mass, particularly around a limb girdle or in the abdominal wall of a woman, should enable optimum therapy to be instituted.
|
|
|---|
This article has been cited by other articles:
![]() |
E. Sutter, Y. Malik, M. Konar, S. Rufenacht, J. E.K. Howard, and M. M. Welle Fibromatosis in a young Bernese Mountain Dog: clinical, imaging, and histopathological findings J Vet Diagn Invest, November 1, 2009; 21(6): 895 - 900. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. D. Murphey, C. M. Ruble, S. M. Tyszko, A. M. Zbojniewicz, B. K. Potter, and M. Miettinen Continuing Medical Education@;DELIM@;From the Archives of the AFIP: Musculoskeletal Fibromatoses: Radiologic-Pathologic Correlation RadioGraphics, November 1, 2009; 29(7): 2143 - 2183. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. E. Laffan, B.-Y. Ngan, and O. M. Navarro Pediatric Soft-Tissue Tumors and Pseudotumors: MR Imaging Features with Pathologic Correlation: Part 2. Tumors of Fibroblastic/Myofibroblastic, So-called Fibrohistiocytic, Muscular, Lymphomatous, Neurogenic, Hair Matrix, and Uncertain Origin1 RadioGraphics, July 1, 2009; 29(4): e36 - e36. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. E. Laffan, B.-Y. Ngan, and O. M. Navarro Pediatric Soft-Tissue Tumors and Pseudotumors: MR Imaging Features with Pathologic Correlation: Part 2. Tumors of Fibroblastic/Myofibroblastic, So-called Fibrohistiocytic, Muscular, Lymphomatous, Neurogenic, Hair Matrix, and Uncertain Origin RadioGraphics, May 15, 2009; (2009) e36. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. S. McDonald, E. S. Yi, and D. E. Wenger Best Cases from the AFIP: Extraabdominal Desmoid-type Fibromatosis RadioGraphics, May 1, 2008; 28(3): 901 - 906. [Full Text] [PDF] |
||||
![]() |
H M O'DWYER, C F KEOGH, J X O'CONNELL, and P L MUNK A case report of synchronous osteoblastoma and fibromatosis Br. J. Radiol., March 1, 2008; 81(963): e68 - e71. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. B. McCarville, F. A. Hoffer, C. S. Adelman, J. D. Khoury, C. Li, and S. X. Skapek MRI and Biologic Behavior of Desmoid Tumors in Children Am. J. Roentgenol., September 1, 2007; 189(3): 633 - 640. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. A. Dinauer, C. J. Brixey, J. T. Moncur, J. C. Fanburg-Smith, and M. D. Murphey Pathologic and MR Imaging Features of Benign Fibrous Soft-Tissue Tumors in Adults RadioGraphics, January 1, 2007; 27(1): 173 - 187. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |