Approximately 5–10% of patients with DDLPS show heterogeneous differentiation (1). The most common differentiation is myogenic ( Figure 1 ); however, osteosarcomatous/chondrosarcomatous and angiosarcomatous elements, although very rare, have also been reported (1, 6). Myogenic differentiation in DDLPS has been investigated. Binh et al. (7) reported that in 27 cases of DDLPS, myogenic differentiation did not affect prognosis or metastasis compared to conventional DDLPS. In contrast, Gronchi et al. (6) reported that in 52 cases of DDLPS, myogenic differentiation in the retroperitoneum, in addition to the Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC) grade, affected overall survival (OS) and distant metastasis. The differences between these two reports are as follows: 1) case number and 2) tumor location (Binh’s cases originated in the internal trunk). Additionally, Kurzawa et al. proposed a combined myogenic differentiation score based on IHC staining of smooth muscle actin and desmin, evaluated by scoring intensity and focality. The combined myogenic differentiation score correlated with disease-free survival and OS, suggesting that further research could develop its clinical application (8). In osteogenic differentiation, 36 cases of DDLPS with osteogenic differentiation were reported (9); retroperitoneal location correlated with local recurrence-free survival and distant metastasis-free survival, but not with OS. In contrast, the FNCLCC grade correlated with OS and distant metastasis-free survival, suggesting a trend similar to that of myogenic differentiation (9). Further analyses with a large series may elucidate the nature of “osteogenic differentiation” in DDLPS.

MRI is a novel diagnostic modality for adipose tumors, regardless of whether they are benign or malignant. DDLPS often contains an ALT/WDLPS component, and the detection of the ALT/WDLPS component by MRI contributes largely to the differential diagnosis ( Figure 2 ) (2). Owing to the abundance of fat tissue in the retroperitoneum, the distinction between adipose tissue or lipoma and ALT/WDL, even by MRI alone, is often difficult; however, the combination of MRI and other factors, including diameter, the presence or absence of septa, and contrast effects, increases both the sensitivity and specificity for diagnosing ALT/WDLPS (10). In fact, in this case, fat tissue was present at the edge of the tumor or surrounding the tumor on computed tomography (CT). Considering that these fat tissues may be part of the tumor, the ALT/WDL component, MRI can be used to determine the extent of surgical resection. However, caution is required, as the degree of fat content on MRI varies in each case of DDLPS; it was reported that 24% of cases had a high fat content, while 44% of cases had no fat content at all (2). Based on these findings, it would be extremely difficult to diagnose DDLPS without a fat component using MRI alone, even if it is of retroperitoneal origin. On the other hand, undoubtedly, CT is beneficial for detecting ossification and calcification. In cases of ossification without a connection to the skeletal system, extraskeletal osteosarcoma (EO) should always be considered a differential diagnosis (11). Imaging features of EO, including various degrees of necrosis, hemorrhagic changes on MRI in almost all cases, and calcification on CT in approximately 60% of cases, have been reported (12). Thus, MRI in DDLPS is beneficial for differentiating EO.

The genetic features of DDLPS and ALT/WDLPS overlap; both share high-level amplification in the chromosomal 12q13-15 region, resulting in consistent amplification of MDM2 and CDK4 (3). Considering these features, IHC for MDM2 and CDK4 and FISH for MDM2 have great advantages for the diagnosis of both DDLPS and ALT/WDLPS, with a sensitivity of 100% for MDM2 and 83% for CDK4 in IHC ( Figure 3 ) (13). On the contrary, high MDM2 and low CDK4 were observed in some cases, suggesting that the expression level of MDM2 does not coincide with the level of CDK4 (14). Regarding MDM2, the specificity may be high in cases with differentiating myxoid liposarcoma and pleomorphic liposarcoma because these liposarcomas almost always show negativity (15), although they may be low in cases with differentiating myxofibrosarcoma and undifferentiated pleomorphic sarcoma because these sarcomas occasionally show positivity (13). Therefore, combining MDM2 and CDK4 with p16 contributes to an accurate diagnosis because of its higher specificity (13, 15). However, we must be very careful in its interpretation because p16 may be positive, even in benign lipomas (13). FISH is more beneficial than IHC in diagnosing DDLPS and ALT/WDL, with reported positivity rates of 93% and 95%, respectively (16). The correlation between MDM2 gene status and prognosis remains controversial because high MDM2 amplification correlates with a shorter OS, while MDM2 copy number (MDM2 to CEP12 ratio) does not correlate with metastasis and/or disease-related death (17, 18). Taken together, although further studies are needed, MDM2 may serve as a prognostic factor. Interestingly, a case of FISH positivity and IHC negativity for MDM2 has been reported, suggesting that when ALT/WDL or DDLPS is clinically suspected, not only IHC but also FISH should be performed (19). In cases with calcification, differentiation from EO is crucial not only in imaging but also in histopathology. Even with a pathological diagnosis, it is difficult to differentiate DDLPS from EO because the positivity rates of MDM2 and CDK4 in IHC and MDM2 in FISH for EO were reported to be 37%, 47%, and 38%, respectively, resulting in overlap between DDLPS and EO (20). On the other hand, the significance of MDM2 amplification in EO remains controversial because of its occurrence in low-grade EO and low-grade EO with high-grade transformation, as well as the lack of association in some low-grade EO cases. Larger studies are needed to clarify the role of 12q13-15 amplification, including MDM2, in EO (21). Although SATB2, an osteoblast marker, has been reported to be useful in EO (11), its usefulness and specificity in differentiating EO from DDLPS with osteosarcomatous elements remain unclear. In cases of DDLPS with no ALT/WDL component but with ossification, the diagnosis should be based on the overall histopathological findings with IHC and FISH, with attention to differentiation from EO.

The principal treatment is surgery, with R0 or R1 surgery preferred if possible; R0 and R1 surgeries are associated with lower recurrence and a better prognosis than R2 surgery (22). In the retroperitoneum, important organs, such as the kidneys and intestinal tract, often make it difficult to complete R0 surgery. Therefore, adjuvant therapies, including chemotherapy and radiation therapy, are often used. Preoperative radiotherapy contributed to better local control in grades 1 and 2 DDLPS but did not affect OS. Furthermore, preoperative radiotherapy did not affect local control or OS in patients with grade 3 DDLPS (23). Chemotherapy with doxorubicin, either as a single agent or in combination with ifosfamide, has been reported. Although the combination of doxorubicin and ifosfamide was expected to have clinical effects, there was no significant improvement in the OS; however, side effects, such as hematological toxicity, were frequently observed (4). Recently, clinical trials of agents targeting MDM2 and CDK4, which are characteristic of ALT/WDL and DDLPS, have been conducted (24). Vanni et al. identified CDK4 as a prognostic biomarker and reported the synergistic effect of CDK4 inhibitors with lenvatinib (14). Besides agents targeting MDM2 and CDK4, tyrosine kinase receptor inhibitors, peroxisome proliferator-activated receptor gamma agonists, and nelfinavir were considered candidates (24). If the efficacy of these novel agents is confirmed through further clinical trials, particularly when combined with radiation and chemotherapy, they could be applied to reduction surgery and inoperable cases.