Established and Experimental Systemic Treatment Options for Advanced Liposarcoma

Soft-tissue sarcomas (STSs) are rare tumors accounting for about 1% of adult cancers, and they comprise of more than 80 different histological subtypes [1‒4]. While various STSs have the commonality of developing from pluripotent mesenchymal stem cells (shown in Fig. 1), their clinical features, genetic profiles, and histology are very heterogenous [5‒7].

Adipocytic tumors represent more common STS tumors, and they may have a benign, intermediate, or malignant behavior [8‒11]. The most common benign adipocytic tumors are lipomas, and other less common types include lipomatosis spindle cell/pleomorphic lipoma, hibernoma, lipoblastoma, myolipoma, and chondroid lipoma [8, 9]. Atypical lipomatous tumors fall into the intermediate group (i.e., locally aggressive) [11]. Liposarcomas are defined as malignant adipocytic tumors and are among the most common types of STS, accounting for about 15–20% of all STSs, with an age-adjusted incidence of 0.4–1.1/100,000 persons/year [1, 12‒16]. Liposarcomas are divided into four major subtypes: (1) atypical lipomatous tumor/well-differentiated liposarcoma (ALT/WDLPS); (2) dedifferentiated liposarcoma (DDLPS); (3) myxoid/round-cell liposarcoma (MLPS); and (4) pleomorphic liposarcoma (PLPS) [12‒15]. Of these four liposarcoma subtypes, ALT/WDLPS are the most common (shown in Fig. 2) and are the least aggressive of the malignant forms of liposarcomas with little to no metastatic potential, while DDLPS, MLPS, and PLPS are more aggressive liposarcomas [9, 14, 15]. Liposarcomas are typically diagnosed around the age of 50 years, and their incidence varies as per the age-group (Table 1) [15, 17]. The etiology of liposarcoma is not clearly known, although risk factors may include specific gene mutations, exposure to radiation (including radiotherapy to treat other tumors), or toxic chemicals [17, 18]. The four different subtypes of liposarcomas have varying underlying molecular pathology, clinical behavior, and treatment sensitivity, and thus, it is important to identify the type of liposarcoma to aid the treatment approach [8, 15].

ALT/WDLPS is a locally aggressive neoplasm. Morphologically, this tumor shows focal nuclear atypia in adipocytes and stromal cells (Table 1 and shown in Fig. 3) [8, 19]. Cytogenetically, the tumor is characterized by the presence of supernumerary ring and giant marker chromosomes (shown in Fig. 4a) [19]. Amplification of mouse double minute 2 homolog (MDM2) (shown in Fig. 5a) and/or cyclin-dependent kinase 4 (CDK4) is almost always present [8, 19]. There are four main subtypes of ALTs/WDLPSs: lipoma-like, sclerosing, inflammatory, spindle cell subtypes [8]. ALT/WDLPS presents as a slow growing mass, but it can attain a large size and most commonly occurs in the extremities, trunk, and retroperitoneum [8, 15, 19]. ALTs/WDLPSs are non-invasive and do not metastasize unless they undergo dedifferentiation (see DDLPS) [8, 19]. They often reoccur locally after local treatment, with recurrence being more common when the tumor arises in the retroperitoneum, paratesticular, or mediastinum region [14, 15]. The prognosis is dependent on the anatomical location , with mortality being 0% in extremities (where complete surgical resection is often feasible) but >80% in tumors occurring in deep anatomical sites such as the retroperitoneum (which have a high risk of recurrence, intraperitoneal spread, and dedifferentiation) [8, 19]. Overall, for patients with ALT/WDLPS, 6–11 years is the median interval between diagnosis and death [8, 19].

DDLPSs are ALTs/WDLPSs with features of progression to nonlipogenic sarcomas of variable grade (Table 1 and shown in Fig. 3) [8, 20]. Amplification of MDM2 and/or CDK4 is present in most cases, although genomic alterations in DDLPS are often more complex than in WDLPS [8, 20, 21]. In DDLPs, a well-differentiated component might not be seen, and the level of dedifferentiation may be variable [8, 20]. DDLPS is often high-grade and can have a high rate of local recurrence (at least 40%) and distant metastases (which can occur in 15–30% of cases) [14, 15, 20]. Anatomical location of the DDLPS is the most important prognostic factor, with patients with retroperitoneal tumors having a poor prognosis [20]. The risk of death is 6-fold greater with DDLPS than with ALT/WDLPS, and patients with distant metastases at 5-year follow-up have a mortality rate of 28–30% [14, 15, 20].

MLPSs consist of uniform, small, round to oval cells with a variable number of immature lipoblasts set in a prominent myxoid stroma (Table 1 and shown in Fig. 3) [15, 22]. MLPS is characterized genetically by translocations resulting in FUS-DDIT3 or EWSR1-DDIT3 fusions (shown in Fig. 4b, 5b) [8, 15, 22]. MLPSs are usually found in the extremities in deep soft tissue, with most cases originating in the thigh, and are rarely found in the retroperitoneum [8, 15, 22]. Local recurrences (about 12–25% of cases) and metastases (about 30–60% of cases) are common [8, 15, 22]. Hypercellular variants (round-cell liposarcomas) are more high-grade and are associated with a higher rate of metastasis and risk of death [22]. In many patients with MLPS, recurrences become increasingly hypercellular and aggressive over time, which is reflected by the myxoid versus round-cell terminology used for this entity. In patients with pure MLPS, the 5-year survival rate is 90%, but where the high-grade round-cell component is present in ≥25% of the tumor, the 5-year survival rate decreases to about 25% [8].

PLPSs are the rarest type of liposarcoma and are high-grade aggressive liposarcomas with a variable number of large, atypical lipoblasts (Table 1 and shown in Fig. 3) [8, 14, 21, 23]. Areas of ALT/WDLPS or other lines of differentiation are not present in PLPS [23]. The molecular profile of PLPS is complex, resembling other pleomorphic sarcomas, with unidentifiable marker chromosomes, high intracellular heterogeneity, and polyploidy (shown in Fig. 4c) [23]. PLPSs have an absence of MDM2 amplification, and typically the staining for CDK4 is also negative. Mutations involving TP53 and NF1 have been described in PLPS as in other sarcomas [8, 15, 23]. Most PLPSs occur in the extremities, mostly in the lower limbs than the upper limbs; other areas that are less frequently affected include the trunk wall, mediastinum, and retroperitoneum [8, 23]. These adipocytic tumors usually develop in deep soft tissue, only about 25% arise in subcutaneous fat, and they are rarely purely dermal [8, 23]. PLPS is an aggressive liposarcoma with a high local recurrence and metastatic rate (of around 30–50% each) and a 60% 5-year survival rate [8, 15, 23].

For patients with localized liposarcomas, surgical resection is the mainstay of treatment [1, 2, 12, 25]. The primary goal of surgery is complete tumor resection with wide margins (R0), which can potentially be curative [1, 21, 25, 26]. In patients with WDLPS or DDLPS in the retroperitoneum, cure rates may be increased with multivisceral resections [19, 20]. For patients with MLPS, the current cornerstone of treatment is surgical wide excision [14, 22]. PLPS are aggressive sarcomas, and early diagnosis is essential for enhancing surgical outcomes [23].

However, local recurrences can occur and are frequent in some liposarcoma subtypes, and radiotherapy may improve local disease control [2, 21, 25, 26]. Radiotherapy is often used in conjunction with surgery, and preoperative, intraoperative, or postoperative radiotherapy can be considered [1, 2, 4, 25]. Preoperative radiotherapy may help nonresectable tumors become resectable [2]. Radiotherapy, in conjunction with surgery, may be necessary for high-grade, deep-seated lesions [1, 17]. Radiosensitivity does vary by liposarcoma subtype, with MLPS being highly radiosensitive, while the other liposarcoma subtypes are moderately radiosensitive, similar to most other STSs (Table 1) [21, 22].

Unresectable, advanced, or metastatic liposarcoma is generally considered incurable, and currently available treatment options achieve unsatisfactory outcome, with few exceptions. Chemotherapy is often used in patients with liposarcoma, the effectiveness of which can vary by the liposarcoma subtype [14, 15, 21]. MLPS, especially the hypercellular round-cell variant, is highly chemo-sensitive; PLPS is considered relatively chemo-sensitive; DDLPS can respond sporadically to chemotherapy, while ALT/WDLPS is generally perceived as chemo-insensitive (Table 1) [14, 15, 21]. Even in the more chemo-sensitive subtypes, the median survival of patients with advanced disease is still low and is around 2 years, which resembles survival outcomes of other STSs [14].

First-line systemic therapy usually involves doxorubicin or ifosfamide or a combination of both [1, 13, 21, 27]. As a palliative first-line treatment for STS in general, the overall response rate was 11.8% with doxorubicin and 5.5–8.4% with ifosfamide; the median progression-free survival (mPFS) was 2.5 months with doxorubicin and 2.2–3.0 months with ifosfamide; and the median overall survival (mOS) was 12.0 months with doxorubicin and 10.9 months with ifosfamide [28]. A phase 2 study with two regimens of ifosfamide, as a single agent for the treatment of 98 patients with STS, showed a response rate of 10–25% as a first-line treatment [29]. However, these studies include “all-comer” STS data and did not provide specific data for liposarcoma, which is a limitation of many historical trials in STS.

Table 2 shows first-line chemotherapy results for the treatment of liposarcomas. In a retrospective study of 88 liposarcoma patients treated at the Royal Marsden Hospital, patients with MLPS had a significantly higher response rate to first-line chemotherapy than in patients with WDLPS or DDLPS (48 vs. 11%, p = 0.005), underscoring the different sensitivity of liposarcoma subtypes to systemic therapy [30]. In a phase 3, randomized, multicenter study, doxorubicin plus ifosfamide (n = 227 patients with STS, including 31 liposarcomas) achieved a significantly higher overall response and mPFS than single-agent doxorubicin (n = 228 patients with STS, including 26 liposarcomas) for first-line treatment of STS; however, there was no statistically significant difference in median overall survival (mOS), and the doxorubicin-ifosfamide combination was associated with a higher grade 3–4 toxicity than single-agent doxorubicin [31]. In a retrospective analysis at the University Hospitals Leuven, a 17% objective response (OR) was observed with first-line chemotherapy in 65 patients with liposarcoma (DDLPS, MLPS, PLPS), with a mOS of 13 months. Overall survival varied by the liposarcoma subtype, with patients with MPLS having significantly longer survival than those with DDLPS or PLPS [13]. In this study, doxorubicin was the most commonly used first-line therapy (55%) followed by a combination of ifosfamide and anthracyclines (23%), and overall survival was not influenced by treatment regimen [13]. Similarly, Italiano et al. [32] observed in a retrospective analysis (N = 208 patients with LPS) that first-line combination chemotherapy achieved a significantly higher OR rate (ORR) than single-agent chemotherapy in patients with LPS (18 vs. 7.5%, p = 0.04), but this did not impact on OS or PFS; in this analysis, the ORR was similar between patients with WDLPS and DDLPS (13 vs. 12%), but the mPFS was significantly different between patients with WDLPS and DDLPS (8.7 months vs. 4 months, p = 0.05). In the retrospective analysis of patients with PLPS (N = 39), Italiano and colleagues [32, 33] reported an ORR of 37% in PLPS, which was higher than the previous series in WDLPS and DDLPS (of 13% and 12%, respectively), while the mPFS and mOS from both series were similar. Compared with other sarcomas, MPLS is relative sensitive to anthracyclines and likely also to other cytotoxic agents [22]. Livingston et al. [34] reported an ORR of 21% for first-line chemotherapy in 82 evaluable patients with retroperitoneal WDLPS/DDLPS, with a mPFS of 4 months and mOS of 29 months. In 109 patients with WDLPS/DDLPS of intra-abdominal origin, Stacchiotti and colleagues also found that cytotoxic chemotherapy had limited activity in these patients (ORR 9%, mPFS 4 months, and mOS 19 months), although the doxorubicin-ifosfamide combination did have a higher ORR of 22% [35]. In summary, outside of clinical trials, for the first-line treatment of advanced, inoperable metastatic liposarcoma, the common chemotherapy treatments include doxorubicin +/− local treatment or ifosfamide +/− local treatment or a doxorubicin-ifosfamide combination +/− local treatment. Response rates and survival seem to depend on liposarcoma subtype, with MLPS and PLPS tending to be more chemo-sensitive than other entities. Overall survival also differs between liposarcoma variants, possibly reflecting differences in the natural evolution of the diseases rather than an effect of the actual treatment.

Chemotherapy. After progression during or after first-line treatment, typically anthracycline-based, potential second-line treatments include non-anthracycline chemotherapy regimens or targeted therapies [1]. Non-anthracycline chemotherapy regimens that have been evaluated in STS and liposarcomas include ifosfamide, dacarbazine, gemcitabine-docetaxel, and gemcitabine-dacarbazine (Table 2). Regarding ifosfamide, a phase 2 study with two regimens of ifosfamide, as a single agent in 76 patients with STS, showed a response rate of 6–8% as a second-line treatment [29]. In a very recent randomized European Organization for Research and Treatment of Cancer (EORTC) study, we obtained a similar ORR of 5% in the ifosfamide control arm in 40 STS patients (11 of whom had DDLPS), the mPFS was 4.4 months, and the mOS was 24.1 months [36]. A study by Maki et al. [37] demonstrated that the ORR, mPFS, and mOS were higher with a gemcitabine-docetaxel combination than single-agent gemcitabine (16 vs. 8%, 6.2 vs. 3.0 months, and 17.9 vs. 11.5 months, respectively) in 122 STS patients (20 of whom had liposarcomas). Similarly, the report by García-del-Muro et al. [38] showed that a gemcitabine-dacarbazine combination had a higher ORR, mPFS, and mOS than single-agent dacarbazine (12 vs. 4%, 4.2 vs. 2 months, and 16.8 vs. 8.2 months, respectively) in 109 STS patients (19 of whom had liposarcomas). In Italiano and colleagues’ [33] assessment in patients with PLPS, 19 of these patients went on to receive ≥ second-line chemotherapy; 3 patients receiving second-line gemcitabine-docetaxel combination achieved a partial response, and one patient receiving third-line trabectedin had an OR. In the Livingston et al. [34] analysis, 18% of 39 evaluable patients with retroperitoneal WDLPS/DDLPS who received second-line chemotherapy (which included gemcitabine-docetaxel, doxorubicin-dacarbazine, or bevacizumab-temozolamide) achieved a partial response.

Therapies such as pazopanib, trabectedin, and eribulin are approved for STS and have different activity in liposarcoma subtypes (Table 3) [12‒15, 21, 26, 39, 40]. As second-line treatments, trabectedin and eribulin are specifically approved for unresectable/metastatic liposarcoma [12].

Trabectedin. Trabectedin is a marine-derived cytotoxic treatment that binds to the minor groove of DNA which impacts on the function of DNA and induces p53-independent apoptosis [41]. A phase 2 randomized study assessed two different dose schedules of trabectedin in patients with advanced or metastatic liposarcomas and leiomyosarcomas (N = 270, of whom 28% had liposarcomas) after failure of prior anthracyclines and ifosfamide, while the trabectedin 24 h intravenous infusion once every 3 weeks was more effective than the trabectedin 3 h intravenous infusion every week for 3 weeks of a 4-week cycle, both regimens demonstrated activity relative to historical comparisons at that time (2009), and trabectedin showed antitumor activity in patients with liposarcomas after failure of conventional treatments (Table 3) [42]. Interestingly, this trabectedin versus trabectedin comparison was the basis for European approval of trabectedin for lipo- and leiomyosarcoma.

In a later phase 3, randomized multicenter study of trabectedin versus dacarbazine in patients with metastatic liposarcomas or leiomyosarcomas after failure of conventional chemotherapy, the overall ORR was 10% versus 7%, the mPFS was 4.2 months versus 1.5 months, and the mPFS was 12.4 months versus 12.9 months, respectively (Table 3) [41]. In this study, 140 of the 518 patients had liposarcomas, of whom 93 patients received trabectedin (13% DDLPS, 11% MLPS, and 3% PLPS), and 47 patients (15% DDLPS, 11% MLPS, and 2% PLPS) received dacarbazine. The mPFS in the patients with liposarcoma who received trabectedin was longer compared with those receiving dacarbazine (3.0 vs. 1.5 months); similarly, the mPFS was longer for trabectedin than dacarbazine in the liposarcoma subgroups (DDLPS: 2.2 vs. 1.9 months; MLPS: 5.6 vs. 1.5 months; and PLPS: 1.5 vs. 1.4 months) [41]. Overall, the most frequent (all grade) adverse events (AEs) for trabectedin were nausea (73%), fatigue (67%), and neutropenia (49%), while for dacarbazine, they were fatigue (51%), nausea (49%), and thrombocytopenia (36%) [41]. Based on this more definitive phase 3 study, trabectedin finally also received US Food and Drug Administration (FDA) approval for the treatment of patients with unresectable or metastatic liposarcoma who received a prior anthracycline-containing regimen [43]. In the final OS and subgroup analysis of this phase 3 study, in patients with liposarcoma, trabectedin achieved a statistically significant improvement in mPFS compared with dacarbazine (3.0 vs. 1.5 months, p = 0.009); however, the mOS was comparable (Table 3) [44]. The original study design (April 2011) of this trial had an evaluation of OS as the primary endpoint with a time frame of approximately 3 years, and the primary endpoint in December 2015 was OS with a time frame of approximately 3 years 8 months (i.e., from study start date to final analysis data cutoff).

In a retrospective analysis of 49 patients with advanced WDLPS/DDLPS, trabectedin was more active against WDLPS/low-grade DDLPS than high-grade DDLPS (mPFS: 13.7 vs. 3.2 months, p = 0.005; mOS: 16.1 vs. 10.2 months, p = 0.044 [Table 3]), which may provide a stratification tool for personalized medicine if confirmed in a large study [45]. The activity of trabectedin in adipocytic tumors seems to be dependent on the histological subtype of liposarcoma, with the most spectacular volumetric responses observed in MLPS and its round-cell variant. In one study with 32 patients with advanced pretreated MLPS, the ORR was 50% (including two complete remissions), the disease control rate was 90%, the mPFS was 17 months, the 6-month PFS rate was 90%, and the mOS was not reached in patients treated with trabectedin [46, 47]. In a study with 52 patients with advanced pretreated MLPS, the ORR was 51% (including two complete remissions), the mPFS was 14 months, and the 6-month PFS rate was 88% in patients treated with trabectedin [47, 48].

As trabectedin is not neuro- or cardiotoxic, and the hepatic toxicity and neutropenia that may occur are not cumulative, trabectedin as a prolonged treatment may be appropriate and should also be evaluated in combination with other treatments [27]. Trabectedin, even though administered as a 24-h infusion, can also be administered using disposable elastomeric pumps, facilitating outpatient treatment [49].

Eribulin. Eribulin is another compound related to a cytotoxic agent with marine origin, a nontaxane microtubule dynamics inhibitor with a distinct mode of action. Eribulin binds to specific sites on the growing positive ends of microtubules, thereby inhibiting their growth [3, 40]. First hints of activity in sarcoma were observed in a phase 2 multi-sarcoma trial performed by EORTC, and lipo- and leiomyosarcomas were prioritized for further clinical testing in a more definitive trial [50]. In the following phase 3, multicenter, open-label study, previously treated patients with advanced liposarcoma or leiomyosarcoma were randomized to receive either eribulin (N = 228, of whom 33% had liposarcoma [14% DDLPS, 13% MLPS, and 6% PPLS]) or dacarbazine (N = 224, of whom 35% had liposarcoma [17% DDLPS, 12% MLPS, and 7% PPLS]) [3]. In this study, the mOS was significantly improved with eribulin compared with dacarbazine in the entire study population (13.5 vs. 11.5, respectively [p = 0.0169]), and mOS was longer with eribulin versus dacarbazine in the liposarcoma population (15.6 months vs. 8.4 months, respectively), with a similar ORR and mPFS in both treatment groups (Table 3) [3]. Treatment-related AEs (all grades) occurred in 93% of patients receiving eribulin and 91% of patients receiving dacarbazine. The most common AEs with eribulin were fatigue (44%), neutropenia (44%), and nausea (40%) and with dacarbazine were nausea (47%), anemia (31%), and thrombocytopenia (28%) [3].

In a subgroup analysis of the above mentioned phase 3 study in the liposarcoma patients, mOS was significantly longer with eribulin than dacarbazine (15.6 vs. 8.4 months, respectively [p < 0.001]), and the mOS was statistically significantly longer with eribulin versus dacarbazine in patients with DDLPS and PLPS and numerically longer in patients with MLPS (Table 3) [40]. In the liposarcoma subgroup, the mPFS was significantly improved in patients receiving eribulin compared with dacarbazine (2.9 vs. 1.7 months, respectively [p = 0.0015]), and the mPFS was longer with eribulin in patients with MLPS and PLPS (Table 3) [40]. The most impressive survival-prolonging effects were observed in PLPS, with a median OS of 22.2 months with eribulin as compared to 6.7 months with dacarbazine; similarly, the mOS was longer with eribulin compared with dacarbazine in DDLPS and MLPS (18.0 vs. 8.1 months, and 13.5 vs. 9.6 months, respectively) [40]. As dacarbazine is perceived by many sarcoma experts as an inactive agent in liposarcoma, this comparison provides some evidence to what extent a drug like eribulin can provide true clinical benefit in pretreated patients in liposarcoma.

In this liposarcoma subgroup, 96% and 85% of patients in the eribulin and dacarbazine arms, respectively, had treatment-related AEs [40]. The most common AEs in this liposarcoma subgroup in the eribulin arm were alopecia (40%), fatigue (40%), and neutropenia (39%) and in the dacarbazine arm were nausea (44%), anemia (35%), and fatigue (32%) [40].

As a single agent, eribulin is the first monotherapy to demonstrate in a phase 3 study an improvement in OS and PFS in patients with advanced LPS. Based on the results of this phase 3 study, the US FDA and EMA approved eribulin for the treatment of advanced liposarcoma after failure of prior anthracycline-containing regimens. Further explorations of eribulin in the treatment of liposarcomas, either as a single agent or as in combination regimens, are warranted [40, 51].

Pazopanib. Pazopanib is a multitargeted tyrosine kinase inhibitor [52]. Based on findings from the original phase 2 study performed by EORTC [53], liposarcoma was deprioritized from further clinical testing of the angiogenesis inhibitor in sarcoma. The PALETTE registration trial excluded liposarcoma patients from study entry [52]. In this study of pazopanib versus placebo in patients with STS, the best overall response was 6% versus 0%, mPFS was 4.6 months versus 1.6 months (p < 0.0001), and the mOS was 12.5 months versus 10.7 months (p = 0.25) [52]. In the pazopanib versus placebo groups, the most common AEs were fatigue (65 vs. 49%), diarrhea (58 vs. 16%), and nausea (54 vs. 28%) [52]. Interestingly, a post hoc analysis of the original phase 2 study with the oral compound revealed that some sarcomas were misclassified in the original trial. Based on the correct histological classification of patients entered in this study, liposarcoma should not have been excluded from further development of the compound.

In a later open-label phase 2 liposarcoma-specific study with pazopanib in 52 patients (71% WDLPS/DDLPS, and 29% MLPS), there were no ORs; the mPFS and mOS for WDLPS/DDLPS were 3.5 months and 16.4 months, respectively; and for MLPS, they were 1.9 months and 22.3 months, respectively (Table 3) [54]. In this study, due to lack of efficacy, the MLPS cohort was closed after 15 patients, while the WDLPS/DDLPS cohort continued, and the authors concluded that pazopanib was active in WDLPS/DDLPS [54].

In another open-label, single-arm phase 2 study with pazopanib in 42 liposarcoma patients (66% DDLPS, 29% MLPS, and 5% PLPS), 17% of patients were entered without prior systemic therapy, 24% had one prior therapy, and 59% had at least two prior therapies. The OR in this study was 2.4%, the mPFS was 4.4 months, and the mOS was 16.2 months [55] (Table 3). The most common AEs (all grades) were nausea (39%), hypertension (37%), diarrhea (34%), and fatigue (29%) [55]. The data again supported the concept that pazopanib has some activity in liposarcoma [55]. Some authors conclude that pazopanib shows promise in the treatment of intermediate and high-grade liposarcomas [56].

New treatments are needed to effectively treat liposarcomas, and there are a number of clinical trials that are ongoing investigating different targets and treatment options for patients with liposarcoma. Most historical trials have pooled adipocytic tumors together with other STS, which makes the interpretation of results very difficult, if not to say impossible. More recently a number of trials have focused on liposarcoma in general or on specific liposarcoma subtypes in particular. Liposarcoma is also one of the few STS where subtype-specific randomized phase 3 trials have been performed and where drugs have achieved regulatory approval specifically for liposarcoma, as illustrated by the two marine compounds described above. Many liposarcoma-specific trials have focused on more aggressive variants of adipocytic tumors and excluded WDLPS. Most trials have been performed in DDLPS, a few have focused exclusively on MLPS, and there have been no subtype-specific approaches for PLPS.

In both WDLPS and DDLPS, MDM2 is involved in tumorigenesis and is thus considered a very promising treatment target for novel therapies [13‒15, 18]. MDM2 gene amplification and MDM2 overexpression is a critical component in tumorigenesis in WDLPS and DDLPS, whereby amplification of MDM2 inhibits the tumor suppressor, p53 [14, 39, 43, 57, 58]. Through multiple mechanisms, MDM2 controls the cellular levels of p53, and overexpression of MDM2 and its inhibition of p53 result in decreased apoptosis and increased cell survival and proliferation [39, 57, 59]. MDM2 inhibitors disrupt the MDM2-p53 interaction, thereby restoring p53 activity, allowing p53 to induce apoptosis, cell cycle arrest, and other functional downstream signaling effects [15, 39, 57]. There are several MDM2 inhibitors in clinical development.

BI 907828 is an oral MDM2-p53 antagonist that has shown antitumor activity in vivo in DDLPS patient-derived xenografts harboring TP53 wild-type and MDM2 amplification [60, 61]. In an ongoing, dose escalation phase 1 study (NCT03449381) of BI 907828 in patients with advanced solid tumors, in 54 patients who were enrolled by July 2021 (8 of whom had WDLPS and 11 had DDLPS), BI 907828 had a manageable safety profile with the DLTs and CTCAE grade 3–4 AEs, thus far being mostly neutropenia and thrombocytopenia [61]. In this study, there was a 38% PR (3/8) in patients with WDLPS, 100% SD (11/11) in patients with DDLPS, the estimated median PFS was 10.8 months (range, 1.3–21.0 months), and the phase 1b dose expansion part of the study is ongoing at the recommended dose of 45 mg every 3 weeks [61]. As shown in Figure 6, a pronounced response was observed in a patient with DDLPS in this phase 1 study after only two oral administrations of BI 907828 at a dose of 45 mg. Additionally, BI 907828 is being evaluated in combination with BI 754091 (ezabenlimab, an anti-programmed cell death protein 1 [PD-1] antibody) and BI 754111 (an anti-LAG-3 antibody) in a phase 1a/1b study (NCT03964233) in patients with advanced solid tumors, with the primary endpoint in the phase 1a part being the maximum tolerated dose (MTD) based on DLTs during the first treatment cycle and in the phase 1b part being OR [62]. This study is currently ongoing, began in June 2019, and the estimated completion date is June 2025; and the phase 1b (dose expansion) part of the study will include 4 cohorts of patients, of which cohort 3 will include liposarcoma patients [62]. BRIGHTLINE-1 (NCT05218499) is an ongoing phase 2/3 randomized, open-label, multicenter trial comparing BI 907828 with doxorubicin as a first-line treatment for patients with advanced DDLPS, , with 300 patients planned. In this study, eligible patients include adult patients with advanced or metastatic, unresectable, progressive, or recurrent DDLPS with MDM2 amplification and the primary endpoint is PFS.

Milademetan (DS-3032, RAIN-32) is an orally bioavailable MDM2 inhibitor that demonstrated in a first-in-human phase 1 study (NCT01877382) involving 107 patients with solid tumors or lymphomas (50% of whom had WDLPS or DDLPS) ORs and durable stable disease (SD) in 53 patients with WDLPS/DDLPS, with 3/53 patients (6%) achieving a PR, resulting in a disease control rate of 59% [57, 63, 64]. In this phase 1 study, the most common AEs were thrombocytopenia, anemia, neutropenia, nausea, fatigue, diarrhea, and vomiting [63]. The intermittent schedule of milademetan taken once a day (qd) 3/14 days twice in a cycle (Schedule D) had the lowest rate of treatment-related AEs, and no dose-limiting toxicities (DLTs) were observed at the MTD of 260 mg [63]. The activity observed in the liposarcoma patients of this phase 1 trial and the acceptable safety profile warranted further study of milademetan in liposarcoma [65]. Currently, a phase 3, randomized, multicenter, open-label, registration trial (MANTRA, NCT04979442) is ongoing comparing the efficacy and safety of milademetan to trabectedin in patients with unresectable or metastatic DDLPS that has progressed on ≥1 prior systemic therapy, including at least one anthracycline-based therapy. In the MANTRA study, approximately 160 patients are randomized in a 1:1 ratio to either treatment, the primary objective is PFS, the study started in July 2021, and the estimated completion date is July 2025.

Nutlins, a class of imidazole compounds, have activity against MDM2 [14, 18]. Nutlin-3 is a MDM2 antagonist and disrupts the MDM2-p53 interaction, thereby inducing the p53 pathway and has shown potent in vitroeffects against liposarcoma cells [14, 58]. Radiotherapy is also known to induce p53, and after the co-treatment with nutlin-3 and radiotherapy in liposarcoma cell lines, nutlin-3 increased the sensitivity of liposarcoma cells to radiotherapy and augmented the activation of p53 [58]. RG7112, a member of the nutlin family that inhibits MDM2, was the first to be tested in a clinical trial and a proof-of-mechanism study in 20 chemotherapy-naive primary or relapsed WDLPS or DDLPS with MDM2 amplification, and this study confirmed that RG7112 inhibited MDM2 and activated the p53 pathway, and one patient had a PR, and 14 patients had SD [66].

Another MDM2 inhibitor is SAR405838, an oral spiro-oxindole derivative, which was shown to induce significant apoptosis in DDLPS cell lines [14, 15, 67]. A first-in-human phase 1 trial (NCT01636479) with SAR405838 in 74 patients with advanced solid tumors showed that the safety profile was acceptable (the main DLT was thrombocytopenia); and in the DDLPS cohort of patients, 56% of patients had SD (there were no PRs), and at 3 months, the progression-free rate was 32% [14, 43, 68].

Other MDM2 inhibitors in clinical development include MK-8242 or AMG 232, both of which are orally administered [69, 70]. MK-8242 showed activity in a phase 1 study in 47 patients with advanced solid tumors (27 of whom had liposarcomas [9 WDLPS, 17 DDLPS, and 1 unknown]): 11% (3/27) PR, 5 patients with liposarcoma had prolonged SD, and the mPFS was 7.8 months in patients with liposarcoma [69]. In the dose expansion part of the phase 1 study with AMG 232, 100% (10/10) of patients with WDLPS and 70% (7/10) of patients with DDLPS had SD for a median of 3.9 months and 2.0 months, respectively [70].

CDK4 amplification is common in WDLPS and DDLPS, and CDK4/6 inhibitors, such as palbociclib, abemaciclib, and ribociclib have shown activity in liposarcoma [43, 71]. A phase 2, nonrandomized, open-label trial (NCT01209598) with palbociclib in 60 patients with advanced WDPLS (n = 13) or DDLPS (n = 47) showed that palbociclib had a favorable PFS (mPFS was 17.9 weeks); one patient had a CR, and the most common AEs were neutropenia, anemia, and thrombocytopenia (however, there were no episodes of neutropenic fever) [72]. Currently, palbociclib is being evaluated in combination with the anti-PD-1 monoclonal antibody INCMGA00012 (retifanlimab) in a phase 2 open-label study in patients with advanced WDLPS or DDLPS (NCT04438824). This combination study started in June 2020, the estimated completion date is June 2023, and the primary endpoint is to confirm the recommended phase 2 dose and the best ORR.

An open-label phase 2 study (NCT02846987) evaluating abemaciclib in 30 patients (29 evaluable) with DDLPS showed abemaciclib had a favorable PFS (mPFS was 30.4 weeks), with one PR, and grade 3–4 AEs included anemia, neutropenia, thrombocytopenia, and diarrhea [73]. Currently, a randomized, double-blind, phase 3 study is ongoing comparing abemaciclib to placebo (randomized 1:1) in patients with advanced, recurrent, and/or metastatic DDLPS (NCT04967521, SARC041). This phase 3 study started in November 2021, and the estimated completion date is November 2024, and the primary endpoint is PFS.

A phase 1 dose escalation study with ribociclib (LEE011, NCT01237236) enrolled 132 patients, of whom 39 patients had liposarcomas [74]. In this study, 6 patients with liposarcoma had prolonged SD [74]. A two-center, two-arm, phase 2 study is ongoing evaluating the combination of ribociclib and everolimus in patients with advanced DDLPS or leiomyosarcomas (NCT03114527); this study started in 2017, and the estimated completion date is December 2023, and the primary endpoint is progression-free rate. A phase 1b study (NCT03009201) assessing the safety and efficacy of ribociclib in combination with doxorubicin in patients with advanced or metastatic STS (which includes patients with liposarcomas) that cannot be removed by surgery commenced in 2017, and the estimated completion date is June 2022, with the primary endpoint being the incidence of DLTs.

In WDLPS or DDLPS, MDM2 and CDK4 are commonly co-amplified, and in vivo evaluations of an MDM2 inhibitor (RG7388) combined with a CDK4 inhibitor (palbociclib) have shown significantly greater apoptosis compared to single agents [75]. A phase 1 study (NCT02343172) has been conducted assessing the safety and efficacy of a combination of a MDM2 inhibitor (HDM201) and the CDK4 inhibitor ribociclib in patients with liposarcomas, and the study results are awaited.

Immunotargets include PD-1 and its ligand (PD-L1) and cytotoxic T-lymphocyte-associated antigen-4, and clinical trials are ongoing in patients with liposarcomas to investigate the potential of mAbs against these immunotargets [7, 43]. Furthermore, PD-L1 is highly expressed in some DDLPS cell lines [76‒78].

In a phase 2, single-arm, multicenter study (SARC028, NCT02301039) with the anti-PD-1 antibody pembrolizumab in 86 patients with advanced sarcomas, there was a 20% PR (2/10) and 40% (4/10) SD in evaluable liposarcoma patients with a mPFS of 25 weeks in these patients [79]. Due to this promising activity, an expansion cohort of this SARC028 study included a liposarcoma cohort [79]; the liposarcoma (DDLPS/PLPS) expansion cohort showed an ORR of 10% (4/39) of evaluable patients, the mPFS was 2 months, and the mOS was 13 months. There were no unexpected toxicities [80]. A randomized phase 2 trial (SU2C-SARC032, NCT03092323) is ongoing evaluating neoadjuvant radiotherapy followed by surgical resection versus neoadjuvant pembrolizumab plus radiotherapy followed by surgical resection and adjuvant pembrolizumab in patients with high-risk STS (which will also allow for the inclusion of patients with DDLPS or PLPS); the study started in July 2017, the estimated completion date is July 2025, and the primary endpoint is disease-free survival [81]. A phase 2 study (NCT03899805) with pembrolizumab and eribulin in patients with STS is currently ongoing (study start date: May 2019, estimated study completion date: August 2024, and primary endpoint: rate of PFS at 12 weeks), and results from the liposarcoma cohort are awaited. A phase 1–2 study (NCT03611868) is ongoing with pembrolizumab and the MDM2 inhibitor APG-115 in patients with metastatic melanomas or advanced solid tumors which may include liposarcoma (study start date: August 2018; estimated study completion date: August 2023; and primary endpoints: MTD, recommended phase 2 dose, and ORR).

Nivolumab (anti-PD-1 monoclonal antibody) ± ipilimumab (targeting cytotoxic T-lymphocyte-associated antigen-4) has been evaluated in a multicenter, randomized, phase 2 study (Alliance A091401, NCT02500797) in 85 patients with metastatic sarcoma [82]. In the first 76 evaluable patients on this study, the ORR was 5% for nivolumab monotherapy and 16% for the combination therapy; however, there were only 5 patients in total who had liposarcoma (WDLPS/DDLPS) in this study, and as such, results were not reported for this subgroup, and an expansion cohort was planned to better characterize activity in specific patient subgroups including liposarcoma [82]. In the expansion part of the study in 24 DDLPS patients (nivolumab monotherapy [n = 12] vs. nivolumab-ipilimumab combination [n = 12]), the 6-month RR was 7% versus 14%, the mPFS was 4.6 months versus 5.5 months, and the mOS was 8.1 months versus 13.1 months, respectively [83]. In a phase 2 study (NCT03307616), nivolumab ± ipilimumab is being evaluated as a neoadjuvant therapy in patients with recurrent or resectable undifferentiated pleomorphic sarcoma (UPS) or DDLPS before surgery [25, 84]. In this study, preliminary results in 14 patients with DDLPS showed a pathological response rate of 23% and a median change in tumor size of +9%, and further results are awaited [84]. An ongoing phase 1 study (NCT04420975), which started in October 2020, is evaluating nivolumab and BO 112 (a noncoding double-stranded synthetic RNA, which may induce the body’s immune system) in combination before surgery in patients with resectable STS (which may include resectable DDLPS); the estimated study completion date is January 2025, and the primary endpoint is frequency and severity of AEs and DLTs.

In a pooled analysis of the phase 2 trials targeting PD-1/PD-L1 in DDLPS, the ORR was 7%, and the nonprogression rate was 55%, suggesting alternative strategies may be needed to increase the immune response in this entity [85]. The cancer-testis antigen New York Esophageal Squamous Cell Carcinoma-1 (NY-ESO-1) is expressed in 80–90% of MLPS, and a chimeric antigen receptor T cell therapy open-label, single-arm, phase 1–2 pilot study (NCT02992743) is evaluating the efficacy and safety of letetresgene autoleucel (GSK3377794, autologous NY-ESO-1^c259 T cells) in MPLS. Initial data suggest that treatment with NY-ESO-1^c259 T cells is feasible, and full results are awaited (estimated study completion date: August 2022) [7, 86]. As already described previously, the MDM2 antagonist BI 907828 is being evaluated in combination with BI 754091 (ezabenlimab, an anti-PD-1 antibody) and BI 754111 (an anti-LAG-3 antibody) in an ongoing phase 1a/1b study (NCT03964233). A single-arm phase 2 study of the combination of sintilimab (anti-PD-1 antibody) and doxorubicin and ifosfamide as a first-line treatment in patients with advanced STS is being evaluated (NCT04356872); in the first stage of the study, the ORR was 63% in 16 evaluable patients with STS (including 100% [2/2] MLPS and 17% [1/6] DDLPS), and the second stage of the study is ongoing [87].

Exportin (XPO1) is an export receptor that is responsible for the nuclear to cytoplasm transportation of more than 200 proteins and multiple RNA molecules [15, 88, 89]. XPO1 is overexpressed in liposarcoma; in liposarcoma cells, knockdown of XPO1 induces apoptosis and inhibits tumor growth [15, 89]. Selinexor (KPT-330) is an orally bioavailable XPO1 inhibitor which has been evaluated in a phase 1b study (NCT01896505) in 54 patients with sarcoma (which included 16 patients with WDLPS/DDLPS and 3 patients with MLPS) [90]. In this study, in 52 evaluable sarcoma patients, no ORR was observed; however, 75% (30/52) of patients had SD, and 33% (17/52) had prolonged SD of ≥4 months in particular in patients with DDLPS (40% [6/15] DDLPS patients had a reduction from baseline in target lesion size, and 47% [7/15] of DDLPS patients had a SD for ≥4 months) [90]. A multicenter, randomized, double-blind phase 2–3 study (SEAL, NCT02606461) evaluating selinexor versus placebo in patients with advanced unresectable DDLPS was conducted, and the phase 2 results of this study in 56 evaluable patients showed that common AEs on both arms of the study were nausea, anorexia, and fatigue; and the PFS (selinexor vs. placebo) was 5.6 months versus 1.8 months, respectively, by RECIST v1.1 [91]. The SEAL trial is one of the largest global phase 3 trials in patients with relapsed DDLPS, and this study showed a significant improvement in PFS with selinexor versus placebo (HR: 0.70; median 2.83 vs. 2.07, p = 0.0228) [92]. The clinical relevance of the magnitude of PFS prolongation is challenged by experts in the field. In the 277 patients enrolled in the phase 3 part of the SEAL trial, 255 patients completed quality of life assessments, and patients receiving selinexor reported lower rates of pain and slower worsening of pain versus placebo [92].

The multi-kinase inhibitor regorafenib has been evaluated in a randomized, double-blind, placebo-controlled phase 2 study (REGOSARC, NCT01900743) in 4 cohorts of patients with metastatic STS [93]. In the liposarcoma cohort, the PFS was 1.1 month with regorafenib versus 1.7 months with placebo (p = 0.70) [93]. Another randomized, double-blind, placebo-controlled phase 2 study (SARC024, NCT02048371) with regorafenib in patients with selected sarcoma subtypes, in 48 patients (34 DDLPS, 12 MLPS, and 2 PLPS) with advanced/metastatic treatment-refractory liposarcoma showed for regorafenib versus placebo, the mPFS was 1.87 months versus 2.07 months (p = 0.62), PR was 0% versus 5% (1/22), and mOS was 6.46 months versus 4.89 months (p = 0.28), respectively, and the authors concluded that the current published data do not support the routine use of regorafenib in patients with liposarcoma [94].

Peroxisome proliferator-activated receptor gamma (PPARγ) is a nuclear receptor that plays a key role in regulating adipocyte differentiation [15, 18, 95]. PPARγ agonists have been evaluated for the treatment of patients with liposarcoma; however, some of these studies have had disappointing results, although more recent results with the combination of pioglitazone and trabectedin in MLPS patient-derived xenografts suggest that the combination is able to reactivate adipocytic differentiation, thereby overcoming resistance to trabectedin with curative effects (complete pathological response) observed in some mice [15, 96]. A pilot phase 2 study (TRABEPIO, NCT04794127) is planned with the pioglitazone and trabectedin combination in patients with MLPS with SD after trabectedin monotherapy (estimated study start date: April 2021, estimated study completion date: April 2023, and primary endpoint: OR).

An open-label phase 2 study (NCT00400569) with the multitargeted tyrosine kinase inhibitor sunitinib in patients with relapsed or refractory STS showed that in the liposarcoma subgroup (n = 18), the mPFS was 3.9 months, and the mOS was 18.6 months. There were no PRs in 17 evaluable patients, but SD was observed in 82% (14/17) of patients [97].

Other agents in development include the Janus kinase 1 inhibitor itacitinib, the lysine-specific demethylase 1 inhibitor seclidemstat, and the deoxycholic acid ATX-101 [98‒100]. A phase 1 study (NCT03670069) started in September 2019 that is evaluating itacitinib in patients with refractory metastatic/advanced sarcomas which may include liposarcomas (MLPS, PLPS), and the estimated study completion date is June 2024, with a primary endpoint being the difference in the percentage of immune inhibitory cells from pretreatment to first posttreatment biopsy. The phase 1 study (NCT03600649) with seclidemstat in patients with relapsed or refractory Ewing or Ewing-related sarcomas permitted the inclusion of MLPS, and the first 27 patients with Ewing sarcoma enrolled in this study showed preliminary activity, supporting the planned phase 2 dose expansion with single-agent seclidemstat [99]. ATX-101 is being assessed in a phase 2 study (NCT05116683) in patients with advanced DDLPS or leiomyosarcoma; the study started in November 2021, and the estimated study completion date is February 2026, with the primary endpoint being the progression-free rate.

For localized disease, surgery remains the gold standard of treatment in adipocytic tumors. Radiotherapy also plays an important role in the management of liposarcoma. Outside of clinical trials, first-line treatment options for patients with liposarcoma include doxorubicin ± local treatments, ifosfamide ± local treatments, or doxorubicin + ifosfamide ± local treatments, while current second-line (and beyond) treatment options for liposarcoma include ifosfamide, gemcitabine-based combinations, trabectedin, eribulin, and possibly pazopanib. Emerging experimental treatment options for liposarcoma include MDM2 inhibitors, CDK4/6 inhibitors, immune checkpoint modulators, XPO1 inhibitors, PPARγ agonists, or combinations thereof, and the results of ongoing trials will further define the role these treatments will have in the management of the different liposarcoma subtypes.

Patrick Schöffski funded editorial/medical writing support, and J. O’Regan (Bingham Mayne and Smith) provided editing/medical writing support.

Patrick Schöffski is/was an active investigator on doxorubicin, ifosfamide, trabectedin, eribulin, pazopanib, BI 907828, ezabenlimab and milademetan trials involving liposarcoma patients. His institution received institutional compensation for entry of patients in such clinical trials. Patrick Schöffski received institutional honoraria for performing preclinical work with trabectedin, eribulin, pazopanib, and BI 907828 in his Laboratory of Experimental Oncology, which included the testing of BI 907828 in patient-derived mouse xenografts. He is the lead investigator on Boehringer Ingelheim’s NCT05218499 trial and receives consulting honoraria for this and other BI 907828-related activities. Consulting and advisory activities outside of the scope of this manuscript are performed for Blueprint Medicines, Deciphera, Ellipses Pharma, Transgene, Exelixis, PharmaMar, Boehringer Ingelheim, SQZ Biotechnology, Adcendo, and Merck Healthcare. Research funding was obtained through KU Leuven from CoBioRes, Eisai, G1 Therapeutics, Pharmamar, Genmab, Merck, Sartar Therapeutics, and Ona Therapeutics.

The preparation of this manuscript received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Patrick Schöffski proposed and developed the manuscript. Patrick Schöffski approved the final version of this manuscript and was responsible for the decision to submit the manuscript.