Dedifferentiated liposarcoma of the musculoskeletal system: expanded MR imaging spectrum from predominant fatty mass to non-fatty mass

Abstract

Background

The morphology of musculoskeletal dedifferentiated liposarcoma was previously reported as well-defined non-lipomatous masses juxtaposed with fatty tumors.

Purpose

To establish a new spectrum of magnetic resonance imaging (MRI) findings for dedifferentiated liposarcoma according to the quantity and distribution of the non-fatty component in the tumor.

Material and Methods

This study consisted of a retrospective review of MRI of 14 patients with a total of 18 histopathologically proven dedifferentiated liposarcoma lesions in the musculoskeletal area. Tumors were classified into four categories (I–IV) according to MR features: I = a well-defined fatty mass and juxtaposed well-defined non-fatty mass; II = a non-fatty component within a predominant fatty mass; III = a focal fatty component within a large non-fatty mass; and IV = a non-fatty mass. Tumor size and depth were also analyzed.

Results

The MR categorizations of the tumors were as follows: Category I, n = 3; Category II, n = 1; Category III, n = 7; and Category IV, n = 7. The mean tumor size was 11.9 cm (range = 5.0–22.2 cm). Most of the tumors showed a predominant non-fatty area and were deeply located at the intra- or intermuscular space.

Conclusion

Dedifferentiated liposarcoma lesions showed a diverse spectrum of MRI findings. Although the presence of a non-fatty region within or adjacent to a well-defined fatty mass can suggest dedifferentiated liposarcoma, non-fatty masses without a fatty component were frequent on MRI.

Keywords

Dedifferentiated liposarcoma MDM2 CDK4 immunohistochemical staining non-fatty tumor MR

Introduction

Liposarcoma is a frequent soft-tissue sarcoma, accounting for 16.9% of cases (1). The 2013 World Health Organization (WHO) classification subcategorized liposarcoma into well-differentiated liposarcoma/atypical lipomatous tumor (WDLPS/ALT), dedifferentiated liposarcoma (DDLPS), myxoid liposarcoma, and pleomorphic liposarcoma (2). Among these, WDLPS/ALT and DDLPS form the largest subgroups, and represent a morphologic and behavioral spectrum of the same disease entity (3,4). WDLPS/ALT usually manifests as a predominantly fat-containing mass, which can dedifferentiate to DDLPS, which behaves in a more aggressive manner with a propensity for local recurrence and the capacity to metastasize (5,6). As dedifferentiation always occurs within a WDLPS/ALT mass, DDLPS is suggested by a focal nodular non-lipomatous region (>1 cm) within or adjacent to a well-defined fatty mass on imaging. DDLPS shows less aggressiveness, less local recurrence or metastasis, and a better prognosis than other sarcomas such as undifferentiated pleomorphic sarcoma (5,7). However, histopathologically, DDLPS may show a markedly heterogeneous morphology; such diverse patterns of DDLPS can lead to difficulty in distinguishing it from other soft-tissue sarcomas, particularly in patients without an antecedent history of WDLPS/ALT, and in tumors with a predominant non-lipomatous component (5,8).

Recently, MDM2 and CDK4 immunohistochemical staining has been widely used for the histological diagnosis of WDLPS/ALT and DDLPS (9 –12). Tumors are diagnosed as DDLPS when they are positive for MDM2 and CDK4, even those tumors that do not show a WDLPS/ALT component (13,14). We have experienced several DDLPS cases that revealed a solid mass without a fatty component on magnetic resonance imaging (MRI); however, to the best of our knowledge, there are no radiologic reports on DDLPS in the musculoskeletal system without any demonstrable WDLPS component. Therefore, we conducted this study to further characterize the MRI spectrum of DDLPS lesions in the musculoskeletal system, from predominant lipomatous masses to completely non-lipomatous masses.

Material and Methods

Patient population

Our single-center study was approved by our institutional review board, which did not require informed consent because of the retrospective study design. By searching the electronic medical records of our institution between October 2009 and October 2017, 27 lesions were identified in 22 patients with pathologically proven DDLPS in the musculoskeletal area, including upper and lower extremities, chest wall, and abdominal wall. All tumors were histopathologically confirmed as DDLPS after wide excision (n = 26) or ultrasonography-guided core needle biopsy (n = 1) by one of three pathologists who majored in the musculoskeletal field at our institution. Of these, six lesions were excluded because of unavailable MRI studies, and a further three lesions were excluded because they were located at the intra-abdominal/retroperitoneal space with extension to the abdominal wall. Consequently, 14 patients with 18 lesions constituted the study population (9 men, 5 women; age range = 37–82 years; mean age = 62.9 ± 10.2 years). The most common clinical symptom at presentation was a palpable mass or swelling (n = 12). All tumors initially had dedifferentiated areas (de novo type) on histopathology.

All but one case underwent tumor excision within two months of MR examination. The patient who did not undergo surgical excision was diagnosed as DDLPS by ultrasonography-guided core needle biopsy, which was performed two days after MRI. The patient did not undergo resection because of an underlying lymphoma in the small intestine. Surgical treatment was followed by radiation therapy in 12 cases.

Twelve cases in 12 patients were primary lesions, but all the other six cases were metastatic lesions in two patients. There were four metastases in one patient, which developed sequentially over four years. The primary lesion was not studied with MRI, so it was excluded in this study. In another patient with two metastases, primary lesion was located in the retroperitoneal space and thus excluded from this study.

Initially, histopathological examination was performed in all cases, with additional immunohistochemical staining for MDM2 and CDK4 being added if needed. All primary lesions underwent immunohistochemical staining for MDM2 and six of these also underwent staining for CDK4. In the recurrent or metastatic cases, immunohistochemical staining for MDM2 and CDK4 was only performed in primary lesions, not in recurrent or metastatic masses. The results of the immunohistochemical staining were recorded by the pathologists. In the cases subjected to immunohistochemical staining, MDM2 and CDK4 were expressed in 100% of primary DDLPS lesions.

MRI protocols and analysis

The MR scanner and pulse sequences used in this study varied. All but two cases underwent MR studies in our institute, with the other two cases being referred to our institution after MRI. Most of the MR studies (15/18) were obtained using 3.0-T scanners at our institution: Ingenia and Achieva (Philips Healthcare, Best, Netherlands) scanners for eight and five cases, respectively, and a Skyra (Siemens Medical Solutions, Erlangen, Germany) scanner for two cases. One case (1/18) was acquired using a 1.5-T scanner at our institution (Avanto; Siemens Medical Solutions, Erlangen, Germany). One of the two MR studies from referring institutions was acquired using a 3.0-T scanner (Achieva; Philips Healthcare, Best, Netherlands) and one was acquired on a 1.5-T scanner (Avanto; Siemens Medical Solutions, Erlangen, Germany).

All MR images were reviewed by two musculoskeletal radiologists, with the findings being reported by means of consensus. As the MRI was not performed with the same protocol or equipment, we analyzed commonly obtained sequences. These were T1-weighted (T1W) and T2-weighted (T2W) images, fat-suppressed (STIR/fat saturation) T2W images, and post-contrast T1W images. To identify the MRI spectrum of DDLPS, we suggested modified categories (Fig. 1) of the morphologic categories for retroperitoneal DDLPS proposed by Hong et al. (15). These were as follows: Category I = a well-defined fatty mass and juxtaposed well-defined non-fatty mass; Category II = a non-fatty component within a predominantly fatty mass; Category III = a focal fatty component within a large non-fatty mass; and Category IV = a non-fatty mass. All cases were categorized into one of these four groups. Areas with signal characteristics similar to those of subcutaneous fat on T1W and T2W MR images and with dropped signal intensity on a fat-suppression sequence were considered as fatty components. In addition, assessed MR findings included the locations of the tumors by anatomical site and depth, and tumor size, shape, and margin. The diagnoses of MRI before surgery or biopsy were collected from the electronic medical records.

Fig. 1.

Schematic drawings of the modified morphologic categories of DDLPS in the musculoskeletal system: (a) Category I = a well-defined fatty mass and juxtaposed well-defined non-fatty mass; (b) Category II = a non-fatty component within a predominant fatty mass; (c) Category III = a focal fatty component within a large non-fatty mass; and (d) Category IV = a non-fatty mass.

Results

Table 1 reports the detailed topographic data. According to the morphological categorization, Category I (Fig. 2), which is defined as a well-defined non-lipomatous mass juxtaposed with fatty tumors, a traditionally known characteristic imaging feature of DDLPS, was identified in only three cases (16.7%). One case (5.6%) was identified as Category II (Fig. 3) and seven cases (38.9%) as Category III (Fig. 4); non-fatty masses were classified as Category IV in seven cases (38.9%; Fig. 5). Thus, Categories III and IV with sparse or no fatty components were the most frequent. All the preoperative MRI diagnoses of Category I and Category II lesions were liposarcoma and all of the category IV lesions, except metastasis and histopathologically proven DDLPS before MRI reading, were judged to be non-specific soft-tissue sarcoma. In the patient with four metastases (Patient 6), three lesions were III (Cases 6, 7, and 8) and the fourth was Category IV (Case 12). Two metastatic lesions of another patient (Patient 10) were all in Category IV (Cases 13 and 14).

Table 1.

A summary of the 18 DDLPS lesions in 14 patients.

Patient no.	Case no.	Age (years)/sex	Primary/metastasis	Location	Tumor size (cm)	Depth	Shape	Margin	Category	Preoperative diagnosis	MDM2	CDK4
1	1	59/F	Primary	Thigh	15.9	Deep	Lobulated	Well defined	I	Liposarcoma	+
2	2	66/F	Primary	Thigh	13.7	Deep	Lobulated	Well defined	I	DDLPS	+
3	3	51/M	Primary	Upper arm	15.4	Deep	Lobulated	Well defined	I	Myxoid liposarcoma	+
4	4	73/M	Primary	Upper arm	18.1	Deep	Lobulated	Well defined	II	WDLPS	+	+
5	5	62/M	Primary	Thigh	5.0	Deep	Round	Well defined	III	Soft-tissue sarcoma	+
6	6^*	60/M	Metastasis	Chest wall	8.9	Deep	Oval	Well defined	III	Metastatic DDLPS
6	7^*	61/M	Metastasis	Thigh	15.1	Deep	Lobulated	Well defined	III	Metastatic DDLPS
6	8^*	62/M	Metastasis	Thigh	14.8	Deep	Lobulated	Well defined	III	Metastatic DDLPS
7	9	53/M	Primary	Chest wall	7.8	Deep	Lobulated	Well defined	III	Soft-tissue sarcoma	+	+
8	10	78/M	Primary	Thigh	16.6	Deep	Round	Well defined	III	Soft-tissue sarcoma	+	+
9	11	82/M	Primary	Chest wall	22.2	Deep	Lobulated	Well defined	III	DDLPS^†	+	+
6	12^*	63/M	Metastasis	Upper arm	8.0	Deep	Round	Well defined	IV	Metastatic DDLPS
10	13^‡	70/F	Metastasis	Thigh	7.7	Deep	Round	Well defined	IV	Metastatic DDLPS
10	14^‡	70/F	Metastasis	Upper arm	7.4	Deep	Round	Well defined	IV	Metastatic DDLPS
11	15	37/F	Primary	Forearm	9.7	Superficial	Oval	Well defined	IV	DDLPS^†	+	+
12	16	58/M	Primary	Thigh	8.9	Deep	Round	Well defined	IV	DDLPS^†	+
13	17	63/M	Primary	Forearm	11	Deep	Lobulated	Well defined	IV	Soft-tissue sarcoma	+	+
14	18	62/F	Primary	Thigh	8.2	Deep	Round	Infiltrative	IV	Soft-tissue sarcoma	+

*The same patient presenting at different times.

^†Histopathologically proven DDLPS before MRI reading.

^‡Occurring in the same patient.

Fig. 2.

Category I. A 51-year-old man with DDLPS involving the left upper arm. Coronal T1W (a), T2W (b), and post-contrast fat-saturated T1W (c) images show a relatively well-defined pure fatty mass and juxtaposed solid mass. The non-fatty solid mass (arrows in c) shows heterogeneous enhancement with a non-enhancing myxoid component.

Fig. 3.

Category II. A 73-year-old man with DDLPS involving the left upper arm. Coronal T1W (a), T2W (b), and post-contrast fat-saturated T1W (c) images show a large mass with a predominant fatty component. An ill-defined inhomogeneously enhancing soft-tissue area (arrows in c) is seen within the mass.

Fig. 4.

Category III. A 62-year-old man with DDLPS in the right posterior thigh. Axial T1W (a), T2W (b), fat-saturated T2W (c), and post-contrast fat-saturated T1W (d) images show a non-fatty mass with peripheral and inhomogeneous internal enhancement. A small-sized nodular fatty component (arrow) is noted on T1W and T2W images. This portion is suppressed on fat-saturated images.

Fig. 5.

Category IV. A 69-year-old woman with DDLPS in the left forearm. Axial T1W (a), T2W (b), fat-saturated T2W (c), and post-contrast fat-saturated T1W (d) images show a well-margined inhomogeneously enhancing soft-tissue mass without identifiable intralesional fat.

The mean tumor size was 11.9 cm (range = 5.0–22.2 cm). The anatomical locations of the tumors included the thigh in nine cases, upper arm in four cases, chest wall in three cases, and forearm in two cases. The tumors were deep-seated (intra- or intermuscular space) in 17 cases (94.4%), but superficial in one case with the involvement of subcutaneous tissue. The tumors were lobulated (n = 9), round (n = 7), or oval (n = 2) in shape, and all but one (Patient 14) revealed a well-defined margin (n = 17). Table 1 reports the detailed topographic data.

Discussion

The DDLPS cases revealed diverse MRI findings in the musculoskeletal system. Although the morphology of musculoskeletal DDLPS lesions was previously reported as well-defined non-lipomatous masses juxtaposed with fatty tumors (6,8,15 –22), we found that completely non-fatty tumors existed. The present study found the traditionally described characteristic imaging features of DDLPS in only three cases (16.7%). Instead, non-lipomatous masses were common (38.9%) and these would not have been diagnosed as DDLPS if immunohistochemical staining had not been performed. Moreover, DDLPS with a focal fatty component within a large non-fatty mass was also quite common (38.9%). Therefore, according to the amount and distribution of the fat component on MR, we can expand the imaging spectrum of DDLPS from the previously reported characteristic fat-containing masses to completely solid non-fatty masses.

There are a few reports on the MRI features of DDLPS in the musculoskeletal system (6,8,16,17,19,20,22) and also in the retroperitoneum (6,8,15,18,21). In the musculoskeletal reports, the authors described only one kind of DDLPS morphology: that of a well-defined non-lipomatous mass juxtaposed with a fatty tumor. However, it was reported that retroperitoneal DDLPS showed diverse imaging morphology. Hong et al. (15) categorized retroperitoneal DDLPS into four morphological groups: Category I = a predominantly fatty mass; Category II = a predominantly non-fatty mass with a small fatty component; Category III = adjacent well-defined fatty and non-fatty masses; and Category IV = two adjacent predominantly non-fatty masses. They reported that, in cases of predominantly non-fatty masses with a small fatty component, DDLPS could be difficult to differentiate from myxoid liposarcoma, pleomorphic liposarcoma, or other non-specific soft-tissue sarcomas. Still, their retroperitoneal DDLPS cases revealed a fatty component in all the tumors, even if in a focal area. Our cases of musculoskeletal DDLPS showed several tumors without any fatty component on MRI. Therefore, to classify our musculoskeletal tumor cases into four categories, we modified Hong’s categories for retroperitoneal DDLPS. Thus, we found that non-fatty DDLPS tumors were common in the musculoskeletal system, accounting for seven (38.9%) of the 18 cases. To the best of our knowledge, this is the first radiological report to mention non-fatty musculoskeletal soft-tissue DDLPS tumors. Radiologists may have great difficulty in distinguishing non-fatty DDLPS from other soft-tissue sarcomas.

In our study, three primary cases of Category III did not have much fat, so we could not be sure of liposarcoma. They were diagnosed with soft-tissue sarcoma and included liposarcoma or DDLPS as another differential diagnosis. Two primary lesions of Category IV were also diagnosed as soft-tissue sarcoma, but the MRI finding were non-specific and we could not make a further differential diagnosis.

The diagnosis of DDLPS is important. Although DDLPS behaves as a high-grade sarcoma with a more aggressive local growth pattern and increased risk of local recurrence and tumor-related death compared with WDLPS/ALT, it tends to behave less aggressively than other pleomorphic sarcomas (5,7).

At present, the final diagnosis of DDLPS is made on histological evaluation (23). Histopathologically, DDLPS has been defined as the presence of a clear-cut WDLPS/ALT component clearly separated from the poorly differentiated component and/or occupying a large area beside the poorly differentiated component (13). Dedifferentiated areas usually consist of undifferentiated pleomorphic sarcoma or spindle cell sarcoma, with high to moderate cellularity and pleomorphism. A minority of DDLPS cases show low cellularity and consist of only histologically low-grade areas resembling fibromatosis or low-grade fibromyxoid sarcoma (5,16). DDLPS usually presents as a large tumor, with the proportions of the well-differentiated and dedifferentiated components being variable. Some may be entirely composed of a well-differentiated area, while others are composed of both dedifferentiated and well-differentiated components. Differentiation of the various histological types of liposarcoma is not always easy, even for an experienced pathologist, especially when only a small quantity of tissue is available (23). Therefore, extensive sampling of the mass is recommended to avoid missing any component. Sampling should be performed in both non-fatty and fatty tissues, because the diagnosis of WDLPS is regularly established in the latter (13). However, some may be entirely composed of a dedifferentiated component. In such cases, the differences between DDLPS and other soft-tissue sarcomas may be subtle and subjective. Fortunately, some scientific advances such as immunohistochemistry have allowed more accurate distinctions. The detection of MDM2 and CDK4 proteins is now becoming a good clue for identifying DDLPS (14). Cytogenetically, WDLPS and DDLPS characteristically harbor supernumerary ring and/or giant chromosomes containing amplicons of the 12q14-15 region, resulting in consistent amplification and overexpression of the MDM2 and CDK4 cell cycle oncogenes (24). The overexpression of MDM2 and CDK4 helps in distinguishing WDLPS from lipoma, as well as distinguishing DDLPS from other poorly differentiated malignancies (10,25). On immunohistochemistry, 97% and 92% of WDLPS/ALT/DDLPS tumors are positive for MDM2 and CDK4, respectively. The sensitivity and specificity of MDM2 immunostaining for WDLPS/ALT/DDLPS among other soft-tissue sarcomas are 97% and 92%, respectively, while the sensitivity and specificity of CDK4 immunostaining are 83% and 95%, respectively (10). Nowadays, diagnosis of DDLPS is made on histological evaluation with supporting immunohistochemistry.

MRI is the most suitable imaging modality to examine the relationship of the lesion with the surrounding structures, which is important in orienting the biopsy target and route, evaluating the operability of the tumor, and planning surgery. It is also the most reliable imaging method for follow-up of regional disease. Nowadays, MRI is used for the reference examination for the DDLPS diagnosis (23).

However, MRI is subject to some limitations concerning the precise diagnosis of DDLPS. The MRI features of DDLPS vary widely according to the spectra of pathologic appearances (17), as indicated in this study. Thus, histologic examination should be performed in soft-tissue tumors, because it is impossible to diagnosis DDLPS with MRI alone. However, sometimes a histologic examination may be insufficient because of inappropriate sampling or the scantiness of biopsy specimens. Therefore, integration between pathologists and radiologists is essential, combining the capabilities of two specialists for a diagnosis with greater accuracy.

This study has certain limitations. First, the retrospective nature of the study resulted in a lack of standardization in the MRI acquisition protocols. However, we considered the MRI qualities sufficient to detect and evaluate the tumors with fatty and non-fatty components. Second, not all lesions had undergone immunohistochemical staining for MDM2 and CDK4. Even though our hospital has been performing immunohistochemical staining for MDM2 and CDK4 since October 2009, one of the two immunohistochemical staining techniques had not been performed in some cases. This was because immunohistochemical staining for MDM2 or CDK4 was added if needed, after experienced musculoskeletal pathologists had performed a histopathological examination.

In conclusion, this study expands the spectrum of MR findings of DDLPS. The presence of intra-tumoral adipose tissue with a non-lipomatous component can suggest DDLPS; however, in some tumors, fat is not detectable at all on MRI. The challenge for radiologists dealing with DDLPS could be in recognizing the diverse imaging features of this tumor, which reflect its histologic heterogeneity.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Howlader

Noone

Krapcho

, et al. SEER Cancer Statistics Review, 1975-2014, National Cancer Institute. Available at: seer.cancer.gov/archive/csr/1975_2014/results_merged/sect_34_soft_tissue_sarcomas.pdf (accessed 2 April 2018).

Fletcher

CDM.

WHO Classification of Tumours of Soft Tissue and Bone. 4th ed. Lyon: IARC Press, 2013.

Nascimento

AG.

Dedifferentiated liposarcoma.

Semin Diagn Pathol 2001; 18:263–266.

Laurino

Furlanetto

Orvieto

, et al. Well-differentiated liposarcoma (atypical lipomatous tumors). Semin Diagn Pathol 2001; 18:258–262.

Thway

Jones

Noujaim

, et al. Dedifferentiated liposarcoma: updates on morphology, genetics, and therapeutic strategies. Adv Anat Pathol 2016; 23:30–40.

Gupta

Potti

Wuertzer

, et al. Spectrum of fat-containing soft-tissue masses at MR imaging: the common, the uncommon, the characteristic, and the sometimes confusing. Radiographics 2016; 36:753–766.

McCormick

Mentzel

Beham

, et al. Dedifferentiated liposarcoma. Clinicopathologic analysis of 32 cases suggesting a better prognostic subgroup among pleomorphic sarcomas. Am J Surg Pathol 1994; 18:1213–1223.

O’Regan

Jagannathan

Krajewski

, et al. Imaging of liposarcoma: classification, patterns of tumor recurrence, and response to treatment. Am J Roentgenol 2011; 197:W37–43.

Sirvent

Coindre

Maire

, et al. Detection of MDM2-CDK4 amplification by fluorescence in situ hybridization in 200 paraffin-embedded tumor samples: utility in diagnosing adipocytic lesions and comparison with immunohistochemistry and real-time PCR. Am J Surg Pathol 2007; 31:1476–1489.

10.

Binh

Sastre-Garau

Guillou

, 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; 29:1340–1347.

11.

Pilotti

Della Torre

Mezzelani

, et al. The expression of MDM2/CDK4 gene product in the differential diagnosis of well differentiated liposarcoma and large deep-seated lipoma. Br J Cancer 2000; 82:1271–1275.

12.

Boltze

Schneider-Stock

Jager

, et al. Distinction between lipoma and liposarcoma by MDM2 alterations: A case report of simultaneously occurring tumors and review of the literature. Pathol Res Pract 2001; 197:563–568.

13.

Coindre

Mariani

Chibon

, et al. Most malignant fibrous histiocytomas developed in the retroperitoneum are dedifferentiated liposarcomas: A review of 25 cases initially diagnosed as malignant fibrous histiocytoma. Modern Pathol 2003; 16:256–262.

14.

Coindre

Hostein

Maire

, et al. Inflammatory malignant fibrous histiocytomas and dedifferentiated liposarcomas: histological review, genomic profile, and MDM2 and CDK4 status favour a single entity. J Pathol 2004; 203:822–830.

15.

Hong

Kim

Woo

, et al. Dedifferentiated liposarcoma of retroperitoneum: spectrum of imaging findings in 15 patients. Clin imaging 2010; 34:203–210.

16.

Murphey

Arcara

Fanburg-Smith

From the archives of the AFIP: imaging of musculoskeletal liposarcoma with radiologic-pathologic correlation.

Radiographics 2005; 25:1371–1395.

17.

El Ouni

Jemni

Trabelsi

, et al. Liposarcoma of the extremities: MR imaging features and their correlation with pathologic data. Orthop Traumatol Surg Res 2010; 96:876–883.

18.

Song

Shen

Liang

, et al. Retroperitoneal liposarcoma: MR characteristics and pathological correlative analysis. Abdom Imaging 2007; 32:668–674.

19.

Drevelegas

Pilavaki

Chourmouzi

Lipomatous tumors of soft tissue: MR appearance with histological correlation.

Eur J Radiol 2004; 50:257–267.

20.

Peterson

Kransdorf

Bancroft

, et al. Malignant fatty tumors: classification, clinical course, imaging appearance and treatment. Skeletal Radiol 2003; 32:493–503.

21.

Tateishi

Hasegawa

Beppu

, et al. Primary dedifferentiated liposarcoma of the retroperitoneum. Prognostic significance of computed tomography and magnetic resonance imaging features. J Comput Assist Tomogr 2003; 27:799–804.

22.

Kransdorf

Meis

Jelinek

JS.

Dedifferentiated liposarcoma of the extremities: imaging findings in four patients.

Am J Roentgenol 1993; 161:127–130.

23.

Coran

Ortolan

Attar

, et al. Magnetic resonance imaging assessment of lipomatous soft-tissue tumors. In Vivo 2017; 31:387–395.

24.

Thway

Flora

Shah

, et al. Diagnostic utility of p16, CDK4, and MDM2 as an immunohistochemical panel in distinguishing well-differentiated and dedifferentiated liposarcomas from other adipocytic tumors. Am J Surg Pathol 2012; 36:462–469.

25.

Pires de Camargo

van de Rijn

Maestro

, et al. Other targetable sarcomas. Semin Oncol 2009; 36:358–371.