Skeletal Radiol (2013) 42:635–647 DOI 10.1007/s00256-012-1517-z

SCIENTIFIC ARTICLE

MRI characteristics of lipoma and atypical lipomatous tumor/well-differentiated liposarcoma: retrospective comparison with histology and MDM2 gene amplification Mélanie Brisson & Takeshi Kashima & David Delaney & Roberto Tirabosco & Andrew Clarke & Suzie Cro & Adrienne M. Flanagan & Paul O’Donnell

Received: 18 March 2012 / Revised: 28 July 2012 / Accepted: 27 August 2012 / Published online: 18 September 2012 # ISS 2012

Abstract Purpose To review the reliability of MR imaging features for the purpose of distinguishing lipoma and atypical lipomatous tumor/well-differentiated liposarcoma (ALT/WDL). Materials and methods A retrospective review of 87 patients with histologically proven lipomatous tumors was performed. All underwent MR imaging, assessing lipomatous content, septation, and nodules. The associations between these features and tumor diagnosis based on morphology and the presence or absence of MDM2 amplification were explored. The age of the patient and the size and location of the lesion were also recorded for statistical analysis. Results Of the 87 patients, 54 were classified as lipomas and 33 as ALT/WDL. MR identified ALT/WDL with a sensitivity of 90.9 % (CI 74.5–97.6) and a specificity of 37.0 % (CI M. Brisson : A. Clarke : P. O’Donnell (*) Department of Radiology, Royal National Orthopaedic Hospital, Stanmore, Middlesex HA7 4LP, UK e-mail: [email protected] T. Kashima : D. Delaney : R. Tirabosco : A. M. Flanagan Department of Histopathology, Royal National Orthopaedic Hospital NHS Trust, Stanmore, Middlesex HA7 4LP, UK A. M. Flanagan : P. O’Donnell Institute of Orthopaedics and Musculoskeletal Science, UCL, Stanmore, UK S. Cro Clinical Trials Unit, Medical Research Council, London, UK A. M. Flanagan UCL Cancer Institute, Huntley Street, London WC1E 6BT, UK

24.6–51.3). The positive and negative predictive values were 46.9 % (CI 34.5–59.7) and 86.9 % (CI 65.3–96.6), respectively. The mean age of patients with ALT/WDL was greater (60 years [range 40–83 years]) than those with lipoma (52 years [range 10–79 years]) (p00.025). The mean size of ALT/WDL (18.7 cm [range 5–36 cm]) was significantly greater than lipoma (13.9 cm [range 3–32 cm]) (p0 0.003). Features that increased the likelihood of ALT/WDL included: patient age over 60 years, maximal lesion dimension over 10 cm, location in lower limb, and presence of non-fatty areas, by a factor of 2.61–6.25 times. Conclusions ALT/WDL and lipoma have overlapping MR imaging characteristics. The most reliable imaging discriminators of ALT/WDL were size of lesion and lipomatous content, but due to the overlap in the MRI appearances of lipoma and ALT/WDL, discrimination should be based on molecular pathology rather than imaging. Keywords Atypical lipomatous tumor . Well-differentiated liposarcoma . Lipoma . Magnetic resonance imaging . MDM2

Introduction Liposarcoma is the most frequent soft tissue sarcoma, accounting for about 20 % of all mesenchymal tumors [1]. The World Health Organization (WHO) Committee for Classification of Soft tissue Tumors in 2002 subdivided soft-tissue liposarcomas into five categories: well-differentiated, dedifferentiated, myxoid, pleomorphic, and mixed liposarcomas [2]. Atypical lipomatous tumor/well-differentiated liposarcoma (ALT/WDL) represent 40–45 % of liposarcomas and is the most frequent subcategory of liposarcoma [2, 3]. They occur most frequently

636

Skeletal Radiol (2013) 42:635–647

during the sixth decade of life, with no sex predilection [1, 4], and arise most commonly in the deep soft tissues of the limbs, especially the thigh, followed by the retroperitoneum, the paratesticular area, the mediastinum, and the subcutaneous tissue. They are locally aggressive, with no potential for metastasis unless they contain a dedifferentiated component [2]. The term ALT is preferred to WDL when referring to lesions arising in the limbs and trunk, as a wide excision margin is usually curative [5]. In contrast, when tumors are located in the retroperitoneum and mediastinum, it is commonly impossible to obtain a wide excision margin, and local recurrence, which is frequent, may lead to death. In these sites, the term WDL is often used, but the application of this terminology can be confusing. In this study, the terms ALT/ WDL are considered as synonymous as both have identical morphology, karyotype, and biological behavior [2]. It is generally possible to distinguish lipoma and ALT/ WDL by histology alone, as the latter consists of a mature adipocytic neoplasm showing unequivocally atypical hyperchromatic nuclei. [2]. Unlike lipomas, these tumors are often traversed by fibrous septa, where the atypical cells tend to be more numerous and difficult to identify. However, the diagnosis is sometimes difficult because the atypical stromal cells can be infrequent and scattered throughout the lesion. In some cases, the identification of unequivocal atypical cells requires extensive sampling of the tumor [3]. In those particular cases, the difference between lipoma and ALT/WDL may be subtle and subjective [3]. A number of ancillary tools have been described to distinguish these tumors: immunohistochemistry, reverse transcription polymerase chain reaction (RT-PCR) [6, 7], fluorescence in situ hybridization (FISH) [8], and antibodies to MDM2 and CDK4 [9]. FISH has been shown to

Fig. 1 Transmitted and fluorescent light microscopy of mature lipomatous tumors. Hematoxylin and eosin (H&E)stained slides showing representative tissue sections of mature lipomatous tumor: adulttype fat without atypical cells (a) and with a single cell with an enlarged hyperchromatic nucleus (black arrow) (b). The corresponding images showing representative fluorescent in situ hybridization of a nucleus from the lipoma showing two red signals and two green signals (c), and the ALT/WDL with two red signals and >30 green signals (d). The red signals represent the centromere and the green signals represent MDM2. Bar050 µm

provide a near-perfect correlation between the presence of supernumerary ring or giant marker chromosomes in the 12q13-15 region and the morphological features of ALT/ WDL [10, 11]. These abnormalities result in amplification of several genes (CDK4, SAS, HMGIC, and others) with MDM2, the most frequently amplified gene in ALT/WDL [3, 10, 12]. A recent publication shows that in a large series of mature lipomatous tumors, MDM2 is a highly sensitive marker for ALT/WDL and that pathologists had a tendency to “overcall” lipomas. The findings were correlated with clinical outcome providing additional value to the study [13]. Magnetic resonance imaging has been described as a useful tool for diagnosing lipomatous tumors [4, 14–17]. It is reported that a confident diagnosis of lipoma can be made on MR imaging when the entire lesion is isointense to subcutaneous fat on all sequences, there is complete suppression of fatty signal on fat-saturation techniques, and the mass contains only a few thin septa [4, 14]. An intramuscular lipoma may present with irregular margins, interdigitating with skeletal muscle, with traversing muscle fibers typically causing a striated appearance, which is not seen with other lipomas [4, 18]. Several imaging features are currently thought to suggest a diagnosis of ALT/WDL and are widely used by radiologists: these include the size of the lesion (more than 10 cm), thick irregular/nodular septa, nonadipose areas or nodules, and fat content less than 75 % [16]. Other imaging characteristics have also been described as useful in attempting to distinguish these two lesions, such as prominent foci of high signal on fluid sensitive sequences [19, 20] and septal enhancement [21, 22]. Just as there are histological similarities between lipomas and ALT/WDL, there is also imaging overlap, with some

A

B

C

D

Skeletal Radiol (2013) 42:635–647

637

Table 1 MDM2 amplification status by FISH and histological diagnosis for each case (n087) MDM2 amplification status

Histological diagnosis

Number of cases

MDM2 + MDM2 −

ALT/WDL Lipoma Total

33 (38 %) 54 (62 %) 87

Data is presented asn(%) Fig. 2 Lipomatous content. Homogeneously lipomatous (completely fatty) tumor. a Coronal T1-weighted (T1W) and b T2-weighted MR images with fat-saturation (T2W-FS) showing a superficial lipomatous mass in the right thigh of a 37-year-old male that is completely isointense to subcutaneous fat in both sequences (black arrows). Diagnosis: lipoma. Amorphous, non-fatty areas c Coronal T1W and d Short-tau inversion recovery (STIR) MR images of a deep heterogeneous lipomatous mass in the anterior compartment of the left thigh of a 73-year-old male with nonfatty amorphous areas within the mass (*). Diagnosis: lipoma. Faint hyperintensity (“altered fat”) e Coronal T1W and f STIR MR images of the left thigh of a 27-year-old female showing faint hyperintensity in a lipomatous mass in the quadriceps compartment (“altered fat” signal, white arrow). Diagnosis: lipoma

lipomas sharing the same imaging characteristics traditionally described in ALT/WDL [16]. Considering the recent advances in molecular genetics, which have been exploited as ancillary tools for diagnosing specific tumors, a more robust diagnosis is now available to distinguish lipoma from ALT/WDL [13]. The purpose of this study is to review the reliability of MR imaging characteristics of lipoma and ALT/WDL, diagnosed on the basis of histopathological features and MDM2 amplification.

A

B

C

D

*

*

E

F

638

Skeletal Radiol (2013) 42:635–647

Materials and methods This was a retrospective MR imaging review of 102 cases of primary soft tissue lipomatous tumors of the extremity and trunk with a histologically confirmed diagnosis of lipoma or ALT/WDL. Cases were retrieved from the histopathology database of the Royal National Orthopaedic Hospital between 2003 and 2008. Non-imaging criteria (patient age and gender) were recorded for further statistical analysis but were not used by the radiologist to distinguish lipoma from ALT/WDL. Recurrent lipomatous tumors and those with inadequate MR imaging were excluded. This study was deemed an audit of current practice and thus patient consent and institutional review was not required. Histology with MDM2 gene amplification analysis Representative tissue sections from tissue blocks were examined by fluorescence in situ hybridization (FISH) using the ZytoLight SPEC MDM2/CEN12 Dual Colour Probe kit (ZytoVision GmbH, Bremerhaven, Germany), according to the protocol provided by the manufacturer. This probe cocktail highlighted the chromosomal region of the human MDM2 gene as a green signal. The centromere of chromosome 12 (CEN12) was detected as a strong and intense red signal (Fig. 1). Fig. 3 Septations. Thin (“pencil-line”) septa a Axial T1W and b STIR MR images showing a lipomatous mass in the midline posterior trunk of a 43-year-old female that remains isointense to subcutaneous fat in both sequences and that contains thin, delicate septa (black arrows). Diagnosis: lipoma. Thick septa c Coronal T1W and d STIR MR images of a deep heterogeneous lipomatous mass of the posterior compartment of the left thigh of a 55-year-old female that is predominantly isointense to subcutaneous fat in both pulse sequences. Thick septations are also evident (white arrows). Diagnosis: lipoma

Evaluation of the sampled tissue sections was carried out using fluorescence microscopy on an Olympus BX-50 microscope or on a Mirax viewer (3DHISTECH Kft., Budapest, Hungary) following the digitization of slides using a Zeiss Mirax scanner. Counting a minimum of 50 non-overlapping nuclei per case, the number of each signal for MDM2 and CEN12 were recorded. A ratio greater than 2.0 was considered to represent MDM2 amplification consistent with a diagnosis of ALT/ WDL. A ratio of 2 or less was considered to be non-amplified. Four pathologists, experienced in the examination of soft tissue tumors, reviewed the histological sections taken at the time of resection and provided a consensus diagnosis based on the World Health Organization criteria [2], in the absence of knowledge of the relevant clinical information, including tumor location and size. The microscopic diagnosis was compared with the fluorescence in situ hybridization (FISH) results and where the histopathology diagnosis did not concur with the FISH, the slides were reviewed. The pathologists diagnosed one lipoma that showed extensive MDM2 amplification and this was reclassified as an ALT/ WDL and included in the study. There were six tumors diagnosed as ALT/WDL, which revealed normal MDM2 and CDK4 copy number and these were excluded from the study because for a variety of reasons it was not clear how best to classify them. Hence, all tumors included in this

A

B

C

D

Skeletal Radiol (2013) 42:635–647

639

imaging study that revealed MDM2 amplification (according to the above definition) were classified as ALT/WDL, and those without atypia (by consensus) and without this genetic abnormality were classified as lipoma. From the initial group of 102 cases where imaging was available for review (Table 1), there were 33 cases where MDM2 was amplified, 32 of which showed unequivocal cellular atypia and one (mentioned above) with minimal atypia but extensive MDM2 amplification, consistent with a diagnosis of ALT/WDL. There were 54 cases where MDM2 was not amplified and histological analysis was consistent with a diagnosis of lipoma. Finally, 15 cases were excluded (six unclassified tumors with normal MDM2 copy number (as above) and nine cases of dedifferentiated liposarcoma [DLS]). A total of 87 cases were included in this study: 50 female and 37 male with a mean age of 55 years (range 10–83 years). Fig. 4 Nodules. a Axial T1W and b STIR MR images showing a deep lipomatous mass in the left gluteal region of a 68-year-old female with a non-specific nodule (defined area of non-lipomatous tissue, hypointense on T1W and hyperintense on STIR (open white arrows). Diagnosis: ALT/WDL. c Axial T1W and d STIR MR images of a deep heterogeneous lipomatous mass of the posterior compartment of the right thigh of a 56-year-old male presenting with a fatty nodule (open white arrows). Diagnosis ALT/WDL. e Lateral plain film in the same patient confirms a rim of ossification (black arrow). f Coronal T1W, g T2WFS MR and h axial CT images of the anterior compartment of the right thigh in a 64-year-old male patient showing a densely ossified nodule in a heterogeneous, predominantly fatty mass (black arrows). Diagnosis: ALT/WDL

MR image analysis One radiologist, experienced in bone and soft tissue tumor MRI, without knowledge of the patient history or final pathological diagnosis, reviewed MR images of the 87 cases. A second radiologist, also blinded to patient history and final diagnosis, independently reviewed 90 % of the cases to establish interobserver reliability for the diagnosis of lipoma or ALT/ WDL using these MRI criteria. The first radiologist assessed the images for a second time so that intra-observer variability could be additionally assessed. MR imaging was performed on a variety of high-field systems, as many patients had been scanned before referral. The MR imaging protocols included a variety of sequences in sagittal, coronal, and axial planes using T1-weighted spin echo (T1 W SE), T2-weighted fast spin echo with fat suppression (T2 W FSE FS), and short tau inversion recovery (STIR) sequences. Contrast enhancement

A

B

C

D

F

G

E

H

640

Skeletal Radiol (2013) 42:635–647

Fig. 5 Intramuscular lipoma. a Coronal and b axial T1W images of a 62-year-old female showing a predominantly fatty lesion in the left deltoid with muscle fibers running through the lesion (open black arrows). Diagnosis: lipoma

A

was not used in the majority of patients in this study as patients were referred from a number of hospitals, using different imaging protocols and was not evaluated in this study. Established MR imaging features were assessed: lipomatous content (Fig. 2) ([a and b] completely fatty; [c and d] amorphous, non-fatty areas within the lipomatous tumor and even faint areas of hyperintensity (“altered fat” signal [e and f]) on fluidsensitive sequences; solid, non-fatty masses); septa (Fig. 3) (absence, thin (“pencil-line”) [a and b] or thick/nodular [c and d]) and nodules (Fig. 4) (absence or presence, size and signal). The location (upper limb, lower limb, trunk), depth (subcutaneous or deep to fascia) and maximum dimension of tumor were recorded but, as with non-imaging criteria, not used to distinguish lipoma from ALT/WDL. Table 2 Summary of results for each imaging and non-imaging variable in group with ALT/ WDL and lipoma

Variable Gender Age Lesion size Location Septation Nodules Nodules Fat content

Omitted as nodules ≥1 cm predicts ALT/WDL perfectly

a

Signal intensity

Male Female >60 years ≤60 years >10 cm ≤10 cm Lower limb Upper limb/trunk Thick Absence/thin Presence Absence ≥1 cm Absence/60 years), size of lesion (>10 cm), location, and fat content were statistically

Maximum dimension of tumor (cm)

Every test performed was two-sided and p values less than 0.05 were considered to indicate a statistically significant difference. The MR imaging diagnoses suggested by the radiologist were compared to the pathological diagnoses based on histopathology and MDM2 amplification and the following analyses were performed: sensitivity (the proportion of histology/MDM2-amplification positive cases diagnosed as ALT/WDL, compared with MR diagnosis), specificity (the proportion of histology/MDM2-amplification negative cases diagnosed as lipoma, compared with MR diagnosis), positive predictive value (proportion of ALT/WDL that were correctly diagnosed using MR criteria) and negative predictive value (proportion of lipomas that were correctly diagnosed using MR criteria). Comparisons of patient age and size of lesion between both groups were performed using unpaired Wilcoxon ranksum tests for non-normally distributed continuous variables. The univariable association between each of the binary imaging and non-imaging variables under study and ALT/ WDL was analyzed by logistic regression. Odds ratios (OR) were calculated along with 95 % confidence intervals to quantify the relative odds of having an ALT/WDL for an individual of the marked category, in comparison to the reference category. An OR greater than 1 indicates increased odds of ALT/WDL while an OR less than 1 indicates decreased odds of ALT/WDL. A multiple binary logistic regression model was built using a stepwise selection method to confirm the multivariable contribution of the observed imaging and non-imaging variables and determine which study variables best predicted a diagnosis of ALT/WDL. Reproducibility between each of the two observers was calculated using Cohen’s kappa statistics with correction for expected chance agreement. Data analysis was performed using Prism version 4 for Macintosh (GraphPad Software, San Diego, CA), except for kappa and odds ratios statistics that were calculated using Stata/IC v12.0, StataCorp, College Station, TX, USA.

641

* 40 30 20 10 0 lipoma (n=54)

ALT/WDL (n=33)

Fig. 7 Comparison of the size of lipomas (n054) and ALT/WDL(n0 33). The continuous lines depict the average dimensions in both groups. The average tumor dimensions were 13.9 cm (range 3– 32 cm) for lipomas and 18.7 cm (range 5–36 cm) for ALT/WDL. The break lines represent the number of tumors measuring 10 cm and less (19 lipomas and three ALT/WDL). * indicates p00.003 with unpaired Wilcoxon rank-sum test

642

Skeletal Radiol (2013) 42:635–647

Table 3 Body distribution for each case of ALT/WDL (n033) and lipoma (n054) Distribution

ALT/WDL

Lipoma

Lower limb Upper limb/trunk

28 (85 %) 5 (15 %)

30 (56 %) 24 (44 %)

Data is presented as n(%)

significant with p values of 0.035, 0.011, 0.007, and 0.002, respectively (Figs. 6 and 7). The distribution of tumors is presented in Table 3. ALT/ WDL were frequently located in the lower limb (n028 (85 %)) while lipomas were slightly more evenly distributed between the lower limb (n030 (56 %)) and the upper limb/ trunk (n024 (44 %)). In this study, all ALT/WDL were located in deep locations with only seven of the 54 lipomas sited subcutaneously, five located in the trunk/upper limb, and two in the lower limb. A purely fatty stroma on MRI (100 % lipomatous content), without altered fat signal or non-fatty masses, and containing either no or only thin septa was observed in both lipomas (Fig. 8) and ALT/WDL. Specifically, this pattern was seen in four cases (12 %) of ALT/WDL (Fig. 9a, b), and of these, one case showed muscle fibers running through the lesion and was diagnosed as an intramuscular lipoma based on MR criteria (Fig. 9c, d). In contrast, 24 lipomas (44 %) showed these appearances. Areas of increased signal intensity relative to subcutaneous fat on fluid-sensitive sequences, and/or non-fatty foci were recorded in the majority of ALT/WDL (n029, 88 %) but were also seen in 29 (54 %) of lipomas. Thereafter, we subdivided the lesions containing amorphous areas into two categories: amorphous areas occupying less or more than

A

Fig. 8 Lipoma with MR imaging features of lipoma. a Coronal T1W and b STIR MR images showing a deep lipomatous mass in the adductor compartment of left thigh of a 38-year-old male. The mass is homogenous, completely isointense to subcutaneous fat on T1W images, and is completely suppressed on the STIR sequence, consistent

50 % of the mass. An amorphous signal occupying more than 50 % of the tumor was more frequently seen in the group with ALT/WDL (ten cases, 30 % of the total number of ALT) than in the group with lipoma (six cases, 11 %). There were two cases (one lipoma and one ALT/WDL) that showed large solid masses within the fatty tumor, suggesting a dedifferentiated component (Fig. 10). Histologically, these non-fatty components were described as lipomatous tumors with a major inflammatory component and, in one case, there was also focal metaplastic bone. Subtle, amorphous, and faint hyperintensity within the tumor (“altered fat” signal) on T2 W FSE FS or STIR sequences was also assessed as a potential sign of ALT/WDL (Fig. 2e and f). There were 14 cases where a radiological diagnosis of ALT/ WDL was based solely on the presence of a subtle amorphous area of faint increased signal. Nine out of these 14 cases were diagnosed as lipomas based on histology and MDM2 amplification status. Absence, or presence of only thin “pencil-line” septation was identified in 39 (72 %) lipomas and 18 (55 %) ALT/WDL, while thick/nodular septa were present in 15 (28 %) lipomas and 15 (45 %) ALT/WDL. The presence of nodules was also described in both groups, but nodules measuring more than 1 cm were only seen in nine (27 %) cases of ALT/WDL and were never observed in lipomas. The signal of individual nodules was unhelpful as we found mineralized (corticated margin with fatty marrow centrally or diffusely calcified [Fig. 4c–h]) and non-specific soft tissue nodules (hypointense on T1-weighted and hyperintense on STIR images relative to fat [Fig. 4a/b]) in both ALT/WDL and lipomas. In total, 30 out of 54 lipomas (MDM2 amplification negative) (56 %) presented with non-fatty areas, nodules, or thick septa. All of these cases revealed unequivocal absence of cellular atypia although they contained areas of

B

with a radiological diagnosis of lipoma. Some muscle fibers are visible at the edge of the mass (white arrows). No cellular atypia were reported and MDM2 was not amplified, which is consistent with a pathological diagnosis of lipoma

Skeletal Radiol (2013) 42:635–647 Fig. 9 ALT/WDL with MR imaging features of lipoma. a Axial T1W and b STIR MR images showing a deep lipomatous mass in the adductor compartment of the right thigh of a 44-year-old male. The mass is homogeneous, completely isointense to subcutaneous fat on T1W images, and is completely suppressed on the STIR sequence, which is consistent with a radiological diagnosis of lipoma (white arrows). c Sagittal T1W and d STIR images show a lesion in the anterior compartment of the arm of a 51-yearold female with signal identical to subcutaneous fat, containing muscle fibers, consistent with a radiological diagnosis of intramuscular lipoma (black arrows). FISH performed on these two cases showed numerous cells with unequivocal MDM2 amplification consistent with pathological diagnosis of ALT/WDL

643

A

B

C

D

fat necrosis or fibrosis (Fig. 11), resulting in similar MRI appearances to ALT/WDL (Fig. 12). The odds ratios used to assess the risk of having an ALT/ WDL versus lipoma are detailed in Table 2. The nonimaging and imaging features favoring a diagnosis of ALT/WDL in univariable analysis were: age over 60 years Fig. 10 ALT/WDL with an MR imaging appearance of dedifferentiated liposarcoma (DLS). a Sagittal T1W and b STIR MR images showing a deep lipomatous mass in the anterior compartment of the thigh of a 75-year-old female. The mass appears heterogeneous, with a large area of more solid, non-lipomatous tumor inferiorly (black arrows), suggesting a diagnosis of dedifferentiated liposarcoma. MDM2 showed amplification and histology an atypical lipomatous tumor with a major inflammatory component, consistent with the diagnosis of ALT/WDL

A

(OR02.61, 95 % CI 1.07–6.39, p00.035), lesion size of more than 10 cm (OR05.43, 95 % CI 1.46–20.15, p0 0.011), mass located in the lower limb (OR04.48, 95 % CI 1.50–13.36, p00.007) and presence of a non-fatty area (solid or amorphous) on MR imaging (OR06.25, 95 % CI 1.93–20.22, p00.002).

B

644 Fig. 11 Lipoma with MR imaging features of ALT/WDL. a Coronal T1W and b STIR MR images showing a deep lipomatous mass in the right deltoid of a 53-year-old male, which contains an area of heterogeneous signal suggestive of ALT/WDL based on radiological criteria (white arrows). On histology review, the lesion showed areas of fibrosis and necrosis without MDM2 amplification consistent with a diagnosis of lipoma

Skeletal Radiol (2013) 42:635–647

A

Using the imaging and non-imaging criteria, a final multiple logistic regression model was built and is shown in Table 4. The size of the lesion (cm) and lipomatous content (solid/ amorphous relative to completely fatty) were the strongest predictors of ALT/WDL in the multivariable analysis and, after adjusting for these two variables, none of the others retained significance. The adjusted OR for size (cm) was 1.09 (95 % CI 1.02–1.18, p00.016), and the adjusted OR for lipomatous content 5.15 (95 % CI 1.54–17.19), p00.008). The radiological inter-observer reliability was substantial and the intra-observer reliability excellent with Kappa values of 0.71 and 0.89, respectively. Analysis of large, deep lesions in older patients A total of 55 cases (32 female, 23 male; mean age 61 years, range 41–83 years) were included in the subgroup analysis of older patients (over 40 years old), with large (greater than 10 cm), deep (subfascial) tumors. Of the 55 patients, 26 were classified as lipomas and 29 as ALT/WDL using a combination of histology and FISH analysis. Using the strict MR imaging criteria described in the Materials and methods section, 46 Fig. 12 ALT/WDL with MR imaging features of ALT/WDL. a Coronal T1W and b STIR MR images showing a deep lipomatous mass in the adductor compartment of the right thigh of a 43-year-old female. The mass is heterogeneous and shows extensive, amorphous, non-fatty signal consistent with an MR diagnosis of ALT/WDL (open white arrows)

A

B

ALT/WDL and nine lipomas were diagnosed by the radiologist. MR identified ALT/WDL with a sensitivity of 96.6 % (CI 80.4–99.8) and a specificity of 30.8 % (CI 15.1–51.9). The positive and negative predictive values were 60.9 % (CI 45.4– 74.5) and 88.9 % (CI 50.7–99.4), respectively. The mean age of patients with ALT/WDL was greater (mean062 years, median065 years [range 41–83 years]) than those with lipoma (mean060 years, median063 years [range 42–79 years]). However, this was not a significant difference (p00.473). The mean size of ALT/WDL (mean0 20.3 cm, median019 cm [range 12–36 cm]) was not significantly greater than lipoma (mean018.1 cm, median018 cm [range 11–32 cm]) (p00.221). Table 5 summarizes the associations between ALT/WDL (as assessed using the combination of histology and FISH analysis) and the important imaging and non-imaging variables of interest in this study subgroup. The difference between the lipoma and ALT/WDL groups in fat content was statistically significant with a p value of 0.006. The presence of non-fatty (solid/amorphous) areas increased the likelihood of ALT/WDL by a factor of 9.90 times (95 % CI 1.93–50.71). This was the strongest predictor of ALT/WDL

B

Skeletal Radiol (2013) 42:635–647

645

Table 4 Multiple logistic regression model to predict ALT/WDLa Odds ratio Size of lesion (cm) Fat content (solid/amorphous relative to completely fatty)

1.09 5.15

95 % CI

p value

(1.02–1.18) (1.54–17.19)

0.016 0.008

a

Predictors of ALT/WDL selected from Table 2 by using stepwise logistic regression

and, after adjusting for lipomatous content using a multiple logistic regression model, none of the others variables were significant predictors.

Discussion Just as distinction between lipoma and ALT/WDL on MR imaging often represents a challenge, the same difficulty is not infrequently encountered by histopathologists. Herein we have described the MR imaging features of lipomas and ALT/WDL based on a pathological diagnosis using a combination of histology, and genetic testing (MDM2-amplification status), which currently represents the most objective method for reaching a diagnosis [13]. The MRI appearances of ALT/WDL in this study are similar to those previously described [16] and most frequently this tumor presents as a heterogeneous fatty mass similar to lipoma, containing areas of non-adipose tissue or nodules and areas of non-specific increased signal intensity on STIR or T2 W FSE FS images. However, the presence of thick and irregular/nodular septa traditionally described in ALT/WDL was not statistically significant in this study, although more lesions with thick septa were seen in this group [16, 22]. This finding could be due to a referral bias, Table 5 Subgroup analysis for patients >40 years, tumor size >10 cm, deep (subfascial) location. Summary of results for each imaging and non-imaging variable in group with ALT/ WDL and lipoma

Variable Gender Age Location Septation Nodules Nodules Fat content

Omitted as nodules ≥1 cm predicted ALT/WDL perfectly

a

as lipomas presenting with atypical features are more likely to be referred for consultation. Hosono et al. [22] suggested that irregular, thick septa in ALT/WDL showed marked enhancement following intravenous injection of gadopentetate dimeglumine as opposed to faint enhancement of thin, regular septa in lipomas. Septal enhancement was not investigated in our study as too few patients were imaged following intravenous injection of contrast, but may be a method of further discriminating tumors with thick septa, as septal morphology was unhelpful. It is of interest that four of 33 (12 %) ALT/WDL showed completely fatty masses with none or only pencil-line septa on MR imaging. Although these cases were classified as ALT/ WDL by the histopathologists on the basis of cellular atypia, which was supported by the presence of numerous cells being MDM2 amplification-positive, these tumors revealed virtually no fibrosis microscopically. The MRI appearances of one of these cases presented with the typical appearances of an intramuscular lipoma. Recently, similar findings were reported in two cases of genetically proven intramuscular ALT/WDL [23]. These findings imply that the presence of intralesional muscle fibres is not a reliable sign of a benign lesion as traditionally described [18]. In contrast, nodules measuring more than 1 cm in diameter seem to be an important discriminator and favors ALT/WDL over a lipoma, although the signal of the nodules was not found to be helpful. Traditionally, lipomas have been described as showing homogeneous signal intensity identical to subcutaneous fat in all pulse sequences, with complete loss of signal after fat suppression. Forty-four percent (n024) of tumors classified as lipomas in this study presented as a completely fatty mass on imaging without any, or only thin (“pencil-line”) septa. However, non-fatty areas, nodules, or thick septa were identified in other lipomas (56 %, n030) and correlation with the

Signal Intensity

Male Female >60 years ≤60 years Lower limb Upper limb/Trunk Thick Absence/Thin Presence Absence ≥1 cm Absence/editors. World Health Organization Classification of Tumor. Pathology and Genetics of Tumors of Soft Tissue and Bone. Lyon: IARC Press; 2002. 3. Coindre JM, Pedeutour F, Aurias A. Well-differentiated and dedifferentiated liposarcomas. Virchows Arch. 2009 Aug 18. 4. Murphey MD, Arcara LK, Fanburg-Smith J. From the archives of the AFIP: imaging of musculoskeletal liposarcoma with radiologicpathologic correlation. Radiographics. 2005;25(5):1371–95. 5. Evans HL. Atypical lipomatous tumor, its variants, and its combined forms: a study of 61 cases, with a minimum follow-up of 10 years. Am J Surg Pathol. 2007;31(1):1–14. 6. Shimada S, Ishizawa T, Ishizawa K, Matsumura T, Hasegawa T, Hirose T. The value of MDM2 and CDK4 amplification levels using real-time polymerase chain reaction for the differential diagnosis of liposarcomas and their histologic mimickers. Hum Pathol. 2006;37(9):1123–9. 7. Hostein I, Pelmus M, Aurias A, Pedeutour F, Mathoulin-Pelissier S, Coindre JM. Evaluation of MDM2 and CDK4 amplification by real-time PCR on paraffin wax-embedded material: a potential tool for the diagnosis of atypical lipomatous tumours/well-differentiated liposarcomas. J Pathol. 2004;202(1):95–102. 8. Sirvent N, Coindre JM, Maire G, Hostein I, Keslair F, Guillou L, et al. Detection of MDM2-CDK4 amplification by fluorescence in situ hybridization in 200 paraffin-embedded tumor samples: utility in

Skeletal Radiol (2013) 42:635–647

9.

10.

11.

12.

13.

14.

15.

diagnosing adipocytic lesions and comparison with immunohistochemistry and real-time PCR. Am J Surg Pathol. 2007;31(10):1476– 89. Binh MB, Sastre-Garau X, Guillou L, de Pinieux G, Terrier P, Lagace R, et al. MDM2 and CDK4 immunostainings are useful adjuncts in diagnosing well-differentiated and dedifferentiated liposarcoma subtypes: a comparative analysis of 559 soft tissue neoplasms with genetic data. Am J Surg Pathol. 2005;29(10):1340–7. Kashima T, Halai D, Ye H, Hing SN, Delaney D, Pollock R, et al. Sensitivity of MDM2 amplification and unexpected multiple faint alphoid 12 (alpha 12 satellite sequences) signals in atypical lipomatous tumor. Mod Pathol. 2012. doi:10.1038/modpathol.2012.90 [Epub ahead of print]. Weaver J, Downs-Kelly E, Goldblum JR, Turner S, Kulkarni S, Tubbs RR, et al. Fluorescence in situ hybridization for MDM2 gene amplification as a diagnostic tool in lipomatous neoplasms. Mod Pathol. 2008;21(8):943–9. Dei Tos AP, Doglioni C, Piccinin S, Sciot R, Furlanetto A, Boiocchi M, et al. Coordinated expression and amplification of the MDM2, CDK4, and HMGI-C genes in atypical lipomatous tumours. J Pathol. 2000;190(5):531–6. Zhang H, Erickson-Johnson M, Wang X, Oliveira JL, Nascimento AG, Sim FH, et al. Molecular testing for lipomatous tumors: critical analysis and test recommendations based on the analysis of 405 extremity-based tumors. Am J Surg Pathol. 2010;34(9):1304–11. Nishida J, Morita T, Ogose A, Okada K, Kakizaki H, Tajino T, et al. Imaging characteristics of deep-seated lipomatous tumors: intramuscular lipoma, intermuscular lipoma, and lipoma-like liposarcoma. J Orthop Sci. 2007;12(6):533–41. Murphey MD, Carroll JF, Flemming DJ, Pope TL, Gannon FH, Kransdorf MJ. From the archives of the AFIP: benign musculoskeletal lipomatous lesions. Radiographics. 2004;24(5):1433–66.

647 16. 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(1):99– 104. 17. Gaskin CM, Helms CA. Lipomas, lipoma variants, and welldifferentiated liposarcomas (atypical lipomas): results of MRI evaluations of 126 consecutive fatty masses. AJR Am J Roentgenol. 2004;182(3):733–9. 18. Matsumoto K, Hukuda S, Ishizawa M, Chano T, Okabe H. MRI findings in intramuscular lipomas. Skeletal Radiol. 1999;28 (3):145–52. 19. Galant J, Marti-Bonmati L, Saez F, Soler R, Alcala-Santaella R, Navarro M. The value of fat-suppressed T2 or STIR sequences in distinguishing lipoma from well-differentiated liposarcoma. Eur Radiol. 2003;13(2):337–43. 20. Doyle AJ, Pang AK, Miller MV, French JG. Magnetic resonance imaging of lipoma and atypical lipomatous tumour/well-differentiated liposarcoma: observer performance using T1-weighted and fluid-sensitive MRI. J Med Imaging Radiat Oncol. 2008;52 (1):44–8. 21. Ohguri T, Aoki T, Hisaoka M, Watanabe H, Nakamura K, Hashimoto H, et al. Differential diagnosis of benign peripheral lipoma from well-differentiated liposarcoma on MR imaging: is comparison of margins and internal characteristics useful? AJR Am J Roentgenol. 2003;180(6):1689–94. 22. Hosono M, Kobayashi H, Fujimoto R, Kotoura Y, Tsuboyama T, Matsusue Y, et al. Septum-like structures in lipoma and liposarcoma: MR imaging and pathologic correlation. Skeletal Radiol. 1997;26(3):150–4. 23. Donato M, Vanel D, Alberghini M, Mercuri M. Muscle fibers inside a fat tumor: a non-specific imaging finding of benignancy. Eur J Radiol. 2009;72(1):27–9.