Sclerosing Variant of Well- Differentiated Liposarcoma: Relative Prevalence and Spectrum of CT and MRI Features

Musculoskeletal Imaging Original Research Bestic et al. CT and MRI of Liposarcoma Musculoskeletal Imaging Original Research Joseph M. Bestic 1 Mark J. Kransdorf 1,2 Lawrence M. White 3 Mellena D. Bridges 1 Mark D. Murphey 4 Jeffrey J. Peterson 1 Hillary W. Garner 1 Bestic JM, Kransdorf MJ, White LM, et al. Keywords: CT, liposarcoma, MRI, musculoskeletal imaging, musculoskeletal system DOI:10.2214/AJR.12.9462 Received June 22, 2012; accepted after revision October 16, 2012. Preliminary data presented at the 2010 annual meeting of the Society of Skeletal Radiology, Las Vegas, NV. 1 Department of Radiology, Mayo Clinic, 4500 San Pablo Rd, Jacksonville, FL 32224-3899. Address correspondence to J. M. Bestic (bestic.joseph@mayo.edu). 2 Present address: Department of Radiology, Mayo Clinic, Scottsdale, AZ. 3 Department of Radiology, Mount Sinai Hospital University of Toronto, Toronto, ON, Canada. 4 American Institute for Radiologic Pathology, Silver Spring, MD. AJR 2013; 201:154 161 0361 803X/13/2011 154 American Roentgen Ray Society Sclerosing Variant of Well- Differentiated Liposarcoma: Relative Prevalence and Spectrum of CT and MRI Features OBJECTIVE. The purpose of this study was to determine the relative prevalence of the sclerosing variant of well-differentiated liposarcoma at one institution and to elucidate the CT and MRI characteristics of this subtype of well-differentiated liposarcoma. MATERIALS AND METHODS. A retrospective computerized search was conducted to calculate the relative prevalence of the sclerosing variant of well-differentiated liposarcoma among all well-differentiated liposarcoma subtypes at one institution. The MRI and CT features of a total of 19 cases of pathologically proven sclerosing variant of welldifferentiated liposarcoma were evaluated (seven identified from the study institution database and 12 cases contributed by other institutions). RESULTS. The cases of a total of 36 patients with well-differentiated liposarcoma were identified in the pathology database; six (17%) cases had evidence of dedifferentiation. Seven (19%) cases of sclerosing variant of well-differentiated liposarcoma were identified. Of these, three (43%) had evidence of dedifferentiation. On images, the sclerosing variant of well-differentiated liposarcoma typically presented as a large (average, 16.6 cm) well-circumscribed heterogeneous mass most commonly situated in the retroperitoneum (58%). Sixteen of the 19 tumors evaluated (84%) had predominantly well-circumscribed margins. Tumor composition ranged from predominantly fatty to entirely devoid of macroscopic fat; only three (16%) were composed of more than 75% fat. Variable amounts of nonlipomatous elements were identified in all cases. Enhancement of these elements was evident at CT or MRI in all 14 cases in which enhancement could be reliably assessed. CONCLUSION. The sclerosing variant of well-differentiated liposarcoma should be included in the differential diagnosis of any well-circumscribed lipomatous mass containing variable amounts of nonlipomatous elements, particularly when located in the retroperitoneum. Unlike other subtypes of well-differentiated liposarcoma, the sclerosing variant is less likely to be composed predominantly of fat and may be associated with an increased propensity for dedifferentiation. L iposarcoma is one of the most common soft-tissue sarcomas in the adult population [1 5]. The World Health Organization separates liposarcomas into four distinct subtypes based on morphologic characteristics and biologic behavior. These include atypical lipomatous tumor or well-differentiated liposarcoma, myoxid liposarcoma, pleomorphic liposarcoma, and dedifferentiated liposarcoma [2]. The atypical lipomatous tumor or well-differentiated liposarcoma is the most common, accounting for approximately 40 54% of such tumors [1, 2, 5]. Of these, the term well-differentiated liposarcoma is reserved for lesions located where a wide surgical margin cannot be obtained (such as the retroperitoneum and mediastinum), and the term atypical lipomatous tumor is applied to lesions in the somatic soft tissues (extremities and trunk), where wide excision should be curative [2]. Nonetheless, these two lesions are identical morphologically and karyotypically, and we use the term well-differentiated liposarcoma to encompass both manifestations. Well-differentiated liposarcoma has a peak incidence in late adult life and is generally considered a locally aggressive nonmetastasizing neoplasm that is prone to local recurrence. Microscopically, well-differentiated liposarcoma is characterized by proliferation of mature adipocytes that vary in cell size and nuclear atypia. A variable number of vacuolated lipoblasts may be evident [2]. Longstanding lesions may ultimately undergo dedifferentiation with development of a juxtaposed high-grade nonlipogenic sarcoma, most often a fibrosarcoma or high-grade undifferentiated pleomorphic 154 AJR:201, July 2013

CT and MRI of Liposarcoma sarcoma (previously called malignant fibrous histiocytoma) [1, 2, 6]. The incidence of dedifferentiation is reported to be 10 15% for retroperitoneal tumors, and dedifferentiation is considered to be a time-dependent phenomenon [1, 2, 6 9]. Although radiologists are generally familiar with the locally aggressive behavior of well-differentiated liposarcoma and its potential for dedifferentiation, they are usually not aware of its histologic diversity and the influence of this diversity on the imaging appearance. The World Health Organization subdivides well-differentiated liposarcoma into four morphologic subtypes: lipomalike (adipocytic), sclerosing, inflammatory, and spindle cell [2]. Lipomalike well-differentiated liposarcoma, overwhelmingly the most common subtype, is generally accepted as the typical appearance of well-differentiated liposarcoma, and its imaging appearance has been well reported. The sclerosing variant is second in frequency only to the lipomalike form, and it is not rare for these two subtypes to coexist [1, 2, 10, 11]. The sclerosing subtype differs from the lipomalike subtype by the presence of areas of dense stromal fibrosis, which may constitute most of the lesion and mask its lipomatous nature [1, 2, 7, 11]. Microscopically, the sclerosing variant of well-differentiated liposarcoma is characterized by the presence of multivacuolated lipoblasts, atypical fibroblasts, primitive mesenchymal cells, and abundant strands of collagen [2, 12]. The inflammatory and spindle cell subtypes of welldifferentiated liposarcoma are quite rare and were not specifically addressed in this study. There is scant literature pertaining to the imaging features of the sclerosing variant of welldifferentiated liposarcoma. The purposes of our study were to determine the relative prevalence of the sclerosing variant of well-differentiated liposarcoma and to define the CT and MRI characteristics of this liposarcoma subtype. A further purpose was to review the demographic and pathologic data for a large group of patients with this variant to better define the spectrum of patient presentations, anatomic distribution, and prevalence of dedifferentiation. Materials and Methods Institutional review board approval was obtained before the study. In keeping with the retrospective nature of the study, informed consent was not required, and management of patient information was compliant with HIPAA guidelines. The prevalence study group was identified from a retrospective computerized search of our institutional pathology database from July 1, 1999, through June 30, 2009, with the keywords welldifferentiated liposarcoma. From this population, a secondary search with the keyword sclerosing was performed. Inclusion criteria for the study were histologic proof of diagnosis based on material obtained from biopsy or resection, availability of adequate CT or MR images of the pathologically proven tumor, and documentation of tumor location. Each lesion was assessed for location, status (primary or recurrent), subtype (lipomalike or sclerosing), and the presence or absence of dedifferentiation. Patient age and sex were also recorded. The well-differentiated liposarcoma prevalence group consisted of 36 patients, including seven patients with the sclerosing subtype. The aim of the second portion of our study was to evaluate the CT and MRI features of the sclerosing variant of well-differentiated liposarcoma. Cases included in the imaging study group were drawn from the prevalence study group (n = 7); two supplementary cases from our institutional teaching file also were included. Additional pathologically proven cases of the sclerosing variant of well-differentiated liposarcoma were obtained from the Department of Radiologic Pathology, American Institute for Radiologic Pathology (n = 6) and the University of Toronto, Ontario, Canada (n = 4). This expanded imaging study group therefore included a total of 19 cases, consisting of retroperitoneal, extremity, and peritoneal tumors. A total of 25 imaging studies were available for review, including 15 CT examinations (11 with contrast administration) and 10 MRI examinations (nine with contrast administration). The MRI equipment and imaging parameters varied both within and between institutions. With the exception of a single patient imaged at 3 T, all MRI studies were performed on 1.5-T systems. All studies were performed with T1-weighted or volume-interpolated breath-hold (VIBE) and fluidsensitive sequences. The latter included spin-echo T2-weighted (n = 4, two with fat suppression) and turbo or fast spin-echo T2-weighted (n = 6, five with fat suppression), HASTE (n = 3), and STIR (n = 4) pulse sequences. Seven of 10 MRI examinations included fluid-sensitive sequences without fat suppression. Contrast-enhanced MRI included T1-weighted (n = 6) and VIBE (n = 3) sequences, all performed with fat suppression. One patient underwent MR angiography with a dynamic 3D technique: time-resolved angiography with stochastic trajectories (TWIST). In addition to parameters evaluated in the prevalence study group, the following additional information was assessed: lesion size (maximal dimension), tumor margins, estimated fat content, CT attenuation, MR signal-intensity characteristics of the nonadipose elements, enhancement behavior, and presence or absence of calcification or ossification. Tumor margins were characterized as generally well circumscribed, ill-defined, or grossly infiltrative. The percentage of fat within each tumor was subjectively graded as less than 25%, 25 50%, 50 75%, and greater than 75%. Lesion MR signal intensity was compared with internal standards of fat, muscle, and fluid. Enhancement of the nonadipose component was subjectively graded as mild, moderate, or intense. Results A total of 43 lesions of well-differentiated liposarcoma in 36 patients were identified from the retrospective computerized search of our pathology database. Of these cases, 36 were primary tumors, and seven were recurrent lesions. Twenty-nine (81%) patients had lipomalike liposarcomas, and seven (19%) had the sclerosing variant of well-differentiated liposarcoma. No inflammatory or spindle cell variants were identified. Six (17%) welldifferentiated liposarcomas had evidence of dedifferentiation, and three (43%) of the sclerosing variants had evidence of dedifferentiation. The sclerosing variant of well-differentiated liposarcoma subgroup comprised four men and three women (median age, 64 years; range, 60 76 years). The size range of the tumors was 1.5 22 cm (mean, 12.5 cm). Five (72%) of the sclerosing variant tumors were situated in the retroperitoneum, one (14%) in the extremities (upper arm), and one (14%) in the peritoneal cavity. The imaging study group consisted of 19 patients (11 men, eight women; median age, 62 years; range, 42 76 years). Sixteen patients had primary tumors, and three had recurrent disease. The size range of the tumors was 1.9 40 cm (average, 16.6 cm). The tumors were located in the retroperitoneum (n = 11; 58%), extremities (n = 7; 37%), and peritoneum (n = 1; 5%). Of tumors located in the extremities, four were situated in the thigh, two in the upper arm, and one in the buttock. Sixteen (84%) of the tumors had predominantly well-circumscribed margins, and three (16%) were grossly infiltrative. Seven (37%) tumors had less than 25% fat, three (16%) had 25 50% fat, six (31%) had 50 75% fat, and three (16%) had more than 75% fat. The variable amount of nonlipomatous tissue most commonly appeared as strands, septations, or nodules of soft-tissue attenuation (15 50 HU) with intermediate T1 signal intensity and variable hyperintensity (relative to muscle and fat) on images obtained with fluid-sensitive sequences (Fig. 1). Of the 10 tumors evaluated with MRI, nonlipomatous AJR:201, July 2013 155

Bestic et al. A C Fig. 1 62-year-old woman with recurrent retroperitoneal sclerosing variant of well-differentiated liposarcoma mimicking more aggressive liposarcoma. A and B, Axial enhanced CT scan (A) and corresponding axial T1-weighted in phase spin-echo MR image (TR/TE, 99/4.8) (B) show large retroperitoneal mass (asterisks) and relatively small fatty component (long arrow). Short arrow indicates clips from previous surgery. C, Coronal ultrafast HASTE MR image (TR/TE, 1160/84) shows mass (asterisk) has signal intensity greater than that of surrounding fat. D, Axial fat-suppressed IV contrast-enhanced volume-interpolated breath-hold MR image (TR/TE, 3.9/1.7) shows heterogeneous enhancement of nonadipose component (asterisks). Fatty component (arrow) is less conspicuous owing to presence of associated enhancing fibrous tissue. elements in eight (80%) were predominantly equal to or higher in signal intensity than skeletal muscle on fluid-sensitive images and hypointense relative to fluid in all cases. Seven of 10 MRI examinations included fluid-sensitive sequences without fat suppression techniques. Nonlipomatous elements had signal intensity equal to or greater than fat in five of these examinations. Two tumors contained nonlipomatous elements nearly devoid of signal on both T1-weighted and fluid-sensitive images. Enhancement of the nonlipomatous elements was clearly evident on CT or MR images in 14 of 20 (70%) contrast-enhanced examinations. In the other six examinations, enhancement could not be reliably assessed. When present, enhancement was typically heterogeneous and confined to the nonlipomatous elements of the tumor. The degree of enhancement ranged from moderate to intense and was noted to be progressive when delayed images were available for review. Most of the tumors displaced adjacent vasculature, although dynamic MR angiography showed that one tumor appeared to have dedicated feeding vessels (Fig. 2). Two tumors had nonenhancing bands and confluent areas of very low signal intensity believed to represent densely fibrous elements on both T1-weighted and fluid-sensitive images. A few scattered and clustered small coarse calcifications were evident in three (16%) tumors. Calcification was not a predominant feature in any of the tumors examined. Demographic data, tumor characteristics, and key imaging findings on all 19 patients with the sclerosing variant of well-differentiated liposarcoma are summarized in Table 1. Five of 19 (26%) cases reviewed in the imaging subgroup had pathologic evidence of B D 156 AJR:201, July 2013

CT and MRI of Liposarcoma A D Fig. 2 75-year-old man with biopsy-proven sclerosing variant of well-differentiated liposarcoma in vastus medialis musculature of mid thigh. A, Axial unenhanced CT image shows well-circumscribed intramuscular mass predominantly composed of fatty tissue. Scattered strands of nonlipomatous tissue with attenuation values similar to skeletal muscle are distributed throughout mass (arrow). B, Axial T1-weighted MR image (TR/TE, 800/9.9) shows predominantly fatty mass (arrow) with scattered areas of nonlipomatous tissue isointense to skeletal muscle. C, Axial T2-weighted MR image (TR/TE, 3000/103) shows scattered areas of nonlipomatous tissue (arrow) hyperintense relative to skeletal muscle but hypointense relative to subcutaneous fat. D, Axial contrast-enhanced T1-weighted fat-suppressed MR image (TR/TE, 702/9.9) shows intense enhancement of nonlipomatous tissues (arrow). E, Coronal maximum-intensity-projection MR angiogram (TR/TE, 3.59/1.17) through mid thigh shows multiple arterial vessels directly supplying mass with globular enhancement of nonlipomatous tissues. B E dedifferentiation. Enhancement of the nonlipomatous elements of tumors with dedifferentiation ranged from moderate to intense, most commonly heterogeneous and progressive in nature. Four of these lesions had large confluent areas of soft-tissue attenuation or low T1 and heterogeneously increased signal intensity in fluid-sensitive sequences (relative to muscle and fat). Three of five tumors had an estimated fat content less than 25%. One of the other two tumors contained 25 50% fat and the other more than 75% fat. Targeted percutaneous needle biopsy of the nonlipomatous component of the latter tumor revealed sclerosing histologic features (Fig. 3). At complete resection, the bulk of the remaining tumor was interpreted as lipomalike well-differentiated liposarcoma. Pathologic evidence of a less than 5 mm focus of dedifferentiation was found in the same tumor. Such findings underscore the well-documented propensity of liposarcomas to have more than one histologic pattern in the same lesion. Notable among the 19 cases of sclerosing variant of well-differentiated liposarcoma examined was a particularly interesting case in which a large retroperitoneal sclerosing liposarcoma with histologically proven low-grade dedifferentiation was found in association with a spatially separate mesenteric mass, representing biopsy-proven well-differentiated liposarcoma that lacked any sclerosing histologic features (Fig. 4). The well-circumscribed mesenteric mass had soft-tissue attenuation similar to that of the dominant retroperitoneal tumor (average, 40 HU). Interestingly, both tumors were hypermetabolic at PET (maximum standardized uptake values, 7.1 and 5.5 within the retroperitoneal and mesenteric tumors). We postulate the smaller mesenteric well-differentiated liposarcoma represents a metastatic deposit associated with dedifferentiation of the dominant retroperitoneal tumor rather than a satellite tumor of multifocal liposarcoma. Discussion The inherent structural and behavioral diversity of liposarcoma is well recognized, and knowledge of this complex subset of soft-tissue tumors continues to expand. The purposes of our study were to further elucidate the relative prevalence of the sclerosing variant of well-differentiated liposarcoma and to de- C AJR:201, July 2013 157

Bestic et al. TABLE 1: Summary of Demographic Data, Tumor Characteristics, and Key Imaging Findings in 19 Patients With the Sclerosing Variant of Well-Differentiated Liposarcoma Signal Intensity of Nonlipomatous Elements Relative to Fat on Fluid-Sensitive Images Fluid-Sensitive Sequence Enhancement of Nonlipomatous Elements Fat Content (%) Margins Calcification Tumor Status Dedifferentiation Location Size (cm) Patient No. Age Sex 1 62 F Primary No Retroperitoneal 17.0 50 75 Well circumscribed No Moderate HASTE High 2 64 F Primary No Upper extremity 1.9 < 25 Well circumscribed N/A Moderate STIR T2 SE High 3 75 M Primary No Retroperitoneal 8.0 50 75 Well circumscribed No N/A N/A N/A 4 67 M Primary Yes Retroperitoneal 22.0 < 25 Well circumscribed Yes Moderate HASTE T2 FS TSE High 5 60 M Recurrent Yes Retroperitoneal 22.0 25 50 Well circumscribed Yes Intense N/A N/A 6 76 M Recurrent Yes Retroperitoneal 9.4 < 25 Well circumscribed No Moderate HASTE T2 FS SE High 7 63 F Recurrent No Peritoneal 7.0 < 25 Well circumscribed No Moderate N/A N/A 8 67 M Primary Yes Retroperitoneal 15.5 < 25 Infiltrative No Moderate STIR T2 FS SE T2 TSE Similar 9 75 M Primary No Lower extremity 7.0 50 75 Infiltrative No Moderate STIR T2 SE Low 10 70 M Primary No Retroperitoneal 37.0 > 75 Well circumscribed No N/A N/A N/A 11 50 M Primary No Retroperitoneal 25.0 25 50 Well circumscribed No N/A N/A N/A 12 49 F Primary N/A Retroperitoneal 38.0 > 75 Well circumscribed No Moderate N/A N/A 13 53 M Primary N/A Lower extremity 40.0 50 75 Infiltrative Yes N/A N/A N/A 14 62 F Primary N/A Retroperitoneal 15.0 < 25 Well circumscribed No Moderate N/A N/A 15 57 F Primary N/A Retroperitoneal 18.0 50 75 Well circumscribed No N/A N/A N/A 16 53 F Primary No Lower extremity 14.0 50 75 Well circumscribed N/A N/A T2 FS FSE High 17 42 M Primary No Upper extremity 8.2 25 50 Well circumscribed N/A Moderate T2 FS TSE High 18 44 F Primary No Lower extremity 12.5 < 25 Well circumscribed N/A Moderate T2 FSE T2 FS FSE Low 19 60 M Primary Yes Lower extremity 19.2 > 75 Well circumscribed N/A Moderate STIR High Note T2 = T2-weighted, SE = spin-echo, FS = fat-suppressed, TSE = turbo spin-echo, FSE = fast spin-echo, N/A = not applicable or unable to assess. scribe its imaging features. In the review of our institutional database, we found that 19% of all well-differentiated liposarcomas reviewed were of the sclerosing variant. As previously noted [1, 2, 10, 11], it is not rare for the lipomalike and sclerosing subtypes to coexist. Despite the well-defined histologic criteria for each subtype, the dividing line between a lipomalike liposarcoma with sclerosing elements and a sclerosing well-differentiated liposarcoma is not clearly drawn. This is likely the cause of the variation from the previously reported prevalence of 40% [6]. Accurately predicting the histologic subtype of liposarcoma is important in establishing its likely biologic behavior and natural history. Characteristic imaging features have been described for many of the liposarcoma subtypes. The goal is stratification of tumors into those that are high grade, with metastatic potential, and those that are low grade, with the threat of only local recurrence [10, 13 18]. In addition to other forms of liposarcoma, including well-differentiated, dedifferentiated, myxoid, and pleomorphic varieties, the differential imaging diagnosis of the sclerosing variant of well-differentiated liposarcoma is quite broad, including both neoplastic and nonneoplastic entities. Other fat-containing lesions that may have similar imaging findings include fat necrosis, myolipoma, angiomyolipoma, fibrolipoma, spindle cell lipoma, and hibernoma. Furthermore, nonlipomatous tumors can occasionally engulf fat and simulate a lipomatous lesion. The classic lipomalike well-differentiated liposarcoma has typically been characterized as a predominantly fatty mass (> 75% fat) with superimposed thickened, nodular, irregular septations or nodular or globular nonadipose areas [13 17]. In contrast, only 3 of 19 (16%) of the sclerosing variant of well-differentiated liposarcoma were composed of greater than 75% fat. Lack of detectable fat within such a tumor could lead to exclusion of well-differentiated liposarcoma from the list of differential considerations. Kim et al. [10] evaluated the CT and MRI characteristics of abdominal liposarcoma and described the imaging findings in five cases that included sclerosing components. As we did, those authors found that the sclerosing subtype could present with a paucity of fat, mimicking highergrade liposarcomas such as pleomorphic liposarcoma. In the same study, the authors observed that the sclerosing components were low in signal intensity on T2-weighted images and cited this feature as a means of differenti- 158 AJR:201, July 2013

CT and MRI of Liposarcoma ating the sclerosing variant of well-differentiated liposarcoma from pleomorphic liposarcoma, in which the nonlipomatous elements have high T2 signal intensity. In contrast, we found that the nonlipomatous elements of the sclerosing variant of liposarcoma most commonly had variably increased signal intensity relative to skeletal muscle and fat in fluid-sensitive sequences, and we were unable to differentiate these lesions solely on the basis of signal-intensity characteristics. However, the variable hyperintensity of nonlipomatous elements in fluid-sensitive sequences may facilitate differentiation of the sclerosing variant of well-differentiated liposarcoma from the homogeneous high signal intensity typical of the myxoid subtype of liposarcoma. Of greater clinical significance is the observation that 43% of the cases of sclerosing variant of well-differentiated liposarcoma examined at our institution had histologic proof of B A dedifferentiation. This constitutes more than a 2.5-fold increase over the 17% prevalence of dedifferentiation among all well-differentiated liposarcomas at our institution. The generally accepted incidence of dedifferentiation of all well-differentiated liposarcomas is 10 15% for retroperitoneal tumors and less than 5% for extremity tumors [1]. The propensity for well-differentiated liposarcoma to dedifferentiate is well established and now known to be time dependent, the risk increasing in longstanding lesions [9]. Although Kooby et al. [8] suggested that sclerosing histologic features in a well-differentiated liposarcoma are associated with increased risk of local recurrence, a significantly increased likelihood of dedifferentiation has not previously been reported. Whether this apparent increased propensity for dedifferentiation will have implications in management and follow-up of these tumors necessitates further investigation. Fig. 3 60-year-old man with predominantly fatty mass surrounding gracilis musculature along medial aspect of upper thigh. A, Axial unenhanced CT image shows focal area of nonlipomatous tissue in anterior aspect of predominantly fatty mass (arrow). B, Axial T1-weighted MR image (TR/TE, 550.0/8.0) shows predominantly fatty mass with focal area of nonlipomatous tissue hyperintense relative to skeletal muscle (arrow). C, Axial fat-suppressed T2-weighted MR image (TR/TE, 5600.0/69.6) clearly shows hyperintense focus of nonlipomatous tissue (arrow). D, Axial contrast-enhanced fat-suppressed T1-weighted image shows moderate enhancement of nonlipomatous tissues (arrow). Percutaneous needle biopsy of this focal area of nonlipomatous tissue revealed sclerosing histologic features. At complete resection, bulk of remaining tumor was interpreted as lipomalike well-differentiated liposarcoma. Pathologic evidence of less than 5-mm focus of dedifferentiation was found in same tumor. C We acknowledge several inherent limitations of our study, including its retrospective nature and relatively small number of patients. The potential for referral bias existed because all cases were drawn from large referral populations. Lack of standardized imaging parameters and techniques, both within and between institutions, necessitated subjective comparison of imaging features among cases, particularly in regard to signal intensity on MR images. The propensity of liposarcomas to have heterogeneous histologic features, as previously described, makes accurate pathologic diagnosis a challenge [1, 2, 11]. Moreover, in the absence of strict diagnostic criteria, pathologic interpretation may be influenced by the experience and personal preferences of individual pathologists. Our study relied on available pathologic interpretations as the reference standard, subjecting our results to the aforementioned variables. D AJR:201, July 2013 159

Bestic et al. A C Fig. 4 67-year-old man with large infiltrative right retroperitoneal mass and smaller well-circumscribed mesenteric mass. A, Axial contrast-enhanced CT image through mid abdomen shows large retroperitoneal mass (asterisk) with infiltrating margins and attenuation similar to that of muscle (average, 40 HU). Additional well-circumscribed mass (arrow) of similar attenuation is situated in mesentery. B, Axial PET/CT fusion image shows increased metabolic activity in both retroperitoneal (asterisk) and mesenteric (arrow) masses (maximum standardized uptake value, 7.1 and 5.5). Biopsy of large retroperitoneal mass revealed sclerosing histologic features with evidence of low-grade dedifferentiation. Smaller mesenteric mass was consistent with well-differentiated liposarcoma without evidence of sclerosing histologic features. We postulate smaller mesenteric mass represents metastatic deposit associated with dedifferentiation of dominant retroperitoneal tumor rather than satellite tumor accompanying multifocal liposarcoma. C, Coronal T1-weighted MR image (TR/TE, 562/14) through abdomen shows large infiltrating retroperitoneal mass superiorly displacing right kidney. Mass is predominantly isointense relative to skeletal muscle with strands of low signal intensity (arrow) corresponding to dense fibrous elements. Small focus of internal fat is evident within inferior aspect of mass (arrowhead). D, Photomicrograph (H and E, 10) of core needle biopsy specimen from large retroperitoneal mass shows features consistent with low-grade sarcoma with sclerosis (fibrosis). Fluorescence in situ hybridization (FISH) assay showed amplification of CPM gene locus on 12q13-15, which supports diagnosis of low-grade dedifferentiated liposarcoma. In summary, the sclerosing variant of welldifferentiated liposarcoma is less likely to have a predominantly fatty composition and may simulate a more aggressive tumor, such as pleomorphic or dedifferentiated liposarcoma. Furthermore, the sclerosing variant of well-differentiated liposarcoma may be associated with an increased tendency toward dedifferentiation, inherent implications in management and follow-up of these tumors being yet to be determined. As a common variant of well-differentiated liposarcoma, sclerosing well-differentiated liposarcoma should be considered in the differential diagnosis of any well-circumscribed lipomatous mass containing variable amounts of nonlipoma- tous elements, particularly when located in the retroperitoneum. Acknowledgment We thank Cherise Cortese, Department of Pathology, Mayo Clinic Florida, for assistance in pathologic imaging and interpretive guidance. B D 160 AJR:201, July 2013

CT and MRI of Liposarcoma References 1. Weiss SW, Goldblum JR. Liposarcoma. In: Weiss SW, Goldblum JR, eds. Enzinger and Weiss s soft tissue tumors, 5th ed. Philadelphia, PA: Mosby Elsevier, 2008:477 498 2. Dei Tos AP, Pedeutour F. Atypical lipomatous tumour/well differentiated liposarcoma. In: Fletcher CD, Unni KK, Mertens F, eds. World Health Organization classification of tumours. pathology and genetics of tumours of soft tissue and bone. Lyon, France: IARC Press, 2002:35 37 3. Enzinger FM, Winslow DJ. Liposarcoma: a study of 103 cases. Virchows Arch Pathol Anat Physiol Klin Med 1962; 335:367 388 4. Dei Tos AP. Liposarcoma: new entities and evolving concepts. Ann Diagn Pathol 2000; 4:252 266 5. Kransdorf MJ. Malignant soft-tissue tumors in a large referral population: distribution of diagnoses by age, sex, and location. AJR 1995; 164:129 134 6. Lucas DR, Nascimento AG, Sanjay BK, Rock MG. Well-differentiated liposarcoma: the Mayo Clinic experience with 58 cases. Am J Clin Pathol 1994; 102:677 683 7. Laurino L, Furlanetto A, Orvieto E, Dei Tos AP. Well-differentiated liposarcoma (atypical lipomatous tumors). Semin Diagn Pathol 2001; 18:258 262 8. Kooby DA, Antonescu CR, Brennan MF, Singer S. Atypical lipomatous tumor/well-differentiated liposarcoma of the extremity and trunk wall: importance of histological subtype with treatment recommendations. Ann Surg Oncol 2004; 11:78 84 9. Weiss SW, Rao VK. Well-differentiated liposarcoma (atypical lipoma) of deep soft tissue of the extremities, retroperitoneum, and miscellaneous sites: a follow-up study of 92 cases with analysis of the incidence of dedifferentiation. Am J Surg Pathol 1992; 16:1051 1058 10. Kim T, Murakami T, Oi H, et al. CT and MR imaging of abdominal liposarcoma. AJR 1996; 166:829 833 11. Suster S, Morrison C. Sclerosing poorly differentiated liposarcoma: clinicopathological, immunohistochemical and molecular analysis of a distinct morphological subtype of lipomatous tumour of soft tissue. Histopathology 2008; 52:283 293 12. Shattuck MC, Victor TA. Cytologic features of well-differentiated sclerosing liposarcoma in aspirated samples. Acta Cytol 1988; 32:896 901 13. Murphey MD, Arcara LK, Fanburg-Smith J. From the archives of the AFIP: imaging of musculoskeletal liposarcoma with radiologic-pathologic correlation. RadioGraphics 2005; 25:1371 1395 14. Kransdorf MJ, Bancroft LW, Peterson JJ, Murphey MD, Foster WC, Temple HT. Imaging of fatty tumors: distinction of lipoma and well-differentiated liposarcoma. Radiology 2002; 224:99 104 15. Song T, Shen J, Liang BL, Mai WW, Li Y, Guo HC. Retroperitoneal liposarcoma: MR characteristics and pathological correlative analysis. Abdom Imaging 2007; 32:668 674 16. Arkun R, Memis A, Akalin T, Ustun EE, Sabah D, Kandiloglu G. Liposarcoma of soft tissue: MRI findings with pathologic correlation. Skeletal Radiol 1997; 26:167 172 17. Jelinek JS, Kransdorf MJ, Shmookler BM, Aboulafia AJ, Malawer MM. Liposarcoma of the extremities: MR and CT findings in the histologic subtypes. Radiology 1993; 186:455 459 18. Kransdorf MJ, Meis JM, Jelinek JS. Dedifferentiated liposarcoma of the extremities: imaging findings in four patients. AJR 1993; 161:127 130 FOR YOUR INFORMATION The AJR has made getting the articles you really want really easy with an online tool, Really Simple Syndication, available at www.ajronline.org. It s simple. Click the RSS button located in the menu on the right side of the page. You ll be on your way to syndicating your AJR content in no time. AJR:201, July 2013 161