Making the Best of Available Options for Optimal Sarcoma Treatment

Until relatively recently, evolution in the treatment of advanced soft tissue sarcoma (STS) had been advancing somewhat glacially. Doxorubicin, which was introduced in the early 1970s, was and continues to be standard front-line therapy in palliative situations. Introduction of dacarbazine and ifosfamide in the early and late 1980s, respectively, was followed by a near 20-year gap before trabectedin was approved in 2007. Since then, pazopanib (2012) and eribulin (2015) have been approved and, in 2016, olaratumab (in combination with doxorubicin) received a positive opinion for marketing authorization from the European Medicine Agency’s Committee for Medicinal Products for Human Use. The future looks dramatically different, however, as multiple agents are in early phase development for treatment of advanced STS. Ongoing phase I/II trials are investigating immune-oncology therapies, vaccine therapy, tyrosine kinase inhibitors, targeted therapies, and new chemotherapeutic agents [1‒11].

As the armamentarium of effective treatments expands, sarcoma oncologists can expect to face additional pressure to treat the “right patient in the right place at the right time with the right diagnosis, in molecular subsets of histological subgroups of rare STS, with the right treatment sequence using the right individually-adapted schedule, in accordance with the patient’s wishes.” In the research setting, investigators are being challenged to individualise reproducible biomarkers of efficacy/resistance. As evidenced by proof-of-concept therapies in certain STS subtypes (Table 1), individualising targets can provide clues to the provenance of mesenchymal tumours. Investigating agents with target-driven activity in specific STS histotypes increases the probability of a positive outcome in phase II trials.

In the clinical trials framework in a disease as complex as STS, study outcomes can be influenced materially by aspects of study methodology.

Given the unique natural evolution of STS subtypes showing survival rates that vary by histotype (Fig. 1) [12], differences in the histological profiles of patient populations in phase II and subsequent confirmatory clinical trials can lead to inconclusive results. Indeed, histological variation is thought to explain in part the absence of efficacy confirmation in phase III studies of doxorubicin ± palifosfamide (an active metabolite of ifosfamide) [13], and doxorubicin ± evofosfamide (a hypoxia-activated prodrug of ifosfamide) [14], after positive results had been demonstrated in phase II studies.

Investigator decisions can also be critical to the success of clinical trials. Aldoxorubicin (an albumin-binding prodrug of doxorubicin) showed promising activity versus doxorubicin in a phase II study in previously untreated patients with advanced STS [15]. However, rather than replicate this methodology to confirm activity, the phase III trial is comparing aldoxorubicin to treatment of investi­gator’s choice (dacarbazine, pazopanib, ifosfamide and doxorubicin) in pretreated patients with advanced STS [16], ultimately affecting the conclusions that can be drawn.

Selecting adequate primary endpoints also has a huge impact on the success of clinical trials. For example, the applicability of “tumour shrinkage” as an endpoint in studies of advanced STS is questionable. In the EORTC 62012 study of doxorubicin ± ifosfamide as first-line treatment of advanced STS, analysis of overall survival (OS) according to Response Evaluation Criteria In Solid Tumors (RECIST) showed no difference between patients with a complete response (CR)/partial response (PR) and those with stable disease (SD): 18.7 vs. 19.6 months [17]. It has since been established that, in advanced STS, the clinical benefit rate (CR + PR + SD) is of greater value than attempting to identify complete or partial responders. The absence of RECIST progression defines prognosis in advanced STS.

Questions also exist about the suitability of median OS as a primary endpoint in studies of advanced STS. In the phase III study of eribulin and dacarbazine in patients with advanced leiomyosarcoma and liposarcoma, the 2-month benefit in median OS in favour of eribulin was statistically significant (13.5 vs. 11.5 months; p = 0.0169), in the absence of differences between treatment arms in median progression-free survival (PFS; 2.6 months with both drugs; p = 0.23) or in response and disease control rates (RR 4%/SD 56% with eribulin and RR 5%/SD 53% with dacarbazine) [18]. In contrast, in 2 other large phase III randomized trials (STS-201 and PALETTE), the 2-month OS benefit observed in each study was not statistically significant, despite statistical differences between treatment arms in PFS [19, 20].

Given that patients with advanced STS typically receive multiple consecutive lines of systemic treatment as well as loco-regional treatments (e.g., stereotaxic radiation therapy, radiofrequency ablation, palliative surgery), selecting OS as an endpoint in clinical trials of first-line treatment may be somewhat ambitious considering that the final result could be “diluted” by additional active agents (e.g., approved drugs, broad-label drugs, off-label drugs, experimental drugs) used in the second line and later. In this context, results of the large randomized phase III ANNOUNCE trial of doxorubicin plus olaratumab versus doxorubicin monotherapy in patients with advanced or metastatic STS are eagerly awaited (ClinicalTrials.gov identifier: NCT02451943) [21]. Immunotherapy agents are also likely to interfere with OS outcomes in future. Although these agents have only a minimal effect on tumour shrinkage, they appear to stimulate the immune system and change the natural evolution of the disease course [22].

Considerable scope also exists for improving the management of advanced STS in daily practice. An analysis of 2,225 patients with advanced STS included in the French CONTICANET database indicated that a non-negligible 28% (n = 625) had not received any first-line treatment [23]. The fact that most non-treated patients were elderly underscores the need to develop a therapeutic program for this population. Of 1,600 patients who had received first-line therapy, only 950 received second-line therapy, representing a 41% loss between the first and second lines. Although patient loss rates slowed between ensuing lines of therapy, they were still considerable (32% between second and third line, and 24% between third and fourth line).

Misclassification of sarcomas is another common problem in daily clinical practice. In the French Sarcoma Networks, for example, major discordance was found to be present in about 10% of cases [24]. This meant that each year, prior to implementing systematic review of pathological specimens, approximately 400 patients began treatment for a sarcoma that was not a sarcoma. Validating tumour histology prior to the start of treatment is essential for patient welfare and is also cost effective.

Importantly, decision making in advanced STS is highly complex and is dependent on diverse clinical presentations and tumour histologies [25]. Given the positive impact of clinical expertise on patient outcomes including improved survival [24], it is mandatory that systemic therapy be discussed in multidisciplinary tumour boards, especially in early lines of treatment in sarcoma patients [26].

The general treatment algorithm for STS is shown in Figure 2.

In patients with localised STS, first-line treatment is surgery ± radiotherapy ± chemotherapy. Recently, the ISG-STS 1001 study demonstrated the added value of administering 3 cycles of anthracycline + ifosfamide neoadjuvant chemotherapy in patients with localised high-grade STS of the extremities or trunk wall [27]. Accordingly, standard first-line systemic therapy in patients who relapse may need to change in future to account for previous exposure to anthracycline + ifosfamide.

In the advanced STS setting, single-agent doxorubicin remains the gold standard for chemotherapy, although multi-agent chemotherapy with adequate-dose anthracycline + ifosfamide may be an option in situations where a tumour response is considered potentially advantageous and the patient has a good performance status [26].

Patients with metastatic disease are candidates for palliation in cases where at least 2 different metastatic sites are involved (e.g., lung and liver, lung and bone) or for potentially curative therapy in cases of isolated (and resectable) lung metastases.

In the palliative setting of metastatic STS (80–90% of patients), first-line therapy is doxorubicin ± olaratumab (PDGFRα inhibitor). This new treatment paradigm is based on results of a phase II trial, which showed that median OS was significantly prolonged with the combination compared with doxorubicin alone (26.5 vs. 14.7 months; hazard ratio 0.46; 95% CI 0.30–0.71; p = 0.0003) [28]. Because tumour overexpression of PDGFRα had no effect on patient outcomes, mechanisms for the added value of this combination are not fully understood [26]. While results of the confirmatory phase III trial (ClinicalTrials.gov identifier: NCT02451943) [21] are awaited, olaratumab is currently available in some countries for use in combination with doxorubicin as first-line treatment of advanced STS.

In the potentially curative setting of metastatic STS (10–20% of patients), a more aggressive approach is required. First-line treatment is anthracycline-based multi-agent chemotherapy followed by surgery. Multi-agent chemotherapy with adequate-dose anthracyclines plus ifosfamide may be treatment of choice, particularly in subtypes sensitive to ifosfamide, when a tumour response is felt to be potentially advantageous. In advanced leiomyosarcomas, where the activity of ifosfamide is far less convincing based on available retrospective evidence, doxorubicin plus dacarbazine is the recommended option.

At the time of anthracycline failure, several treatment options exist and, increasingly, the choice of therapy is histology driven (Table 2) [26]. Steady improvement in the median OS of patients with metastatic STS [29] can be attributed to the introduction of new agents administered consecutively along multiple treatment lines and to optimising the sequential use of currently available agents. More extensive clinical experience has identified ways in which established agents can be used to advantage.

Increased experience with the use of a specific agent can have a positive impact on patient outcomes. This association became clear when examining outcomes of large European studies of trabectedin (Fig. 3) [30‒34]. The studies were conducted over an approximate 20-year timeframe and involved 1,361 patients with sarcoma. Over time, improvements were observed in median PFS, 6-month PFS rate, and median OS. Indeed, the median OS of 21.3 months achieved with single-agent trabectedin in a recent phase IV study represents a valuable step forward, and one that is difficult to attain when treating patients with advanced STS [34].

Several studies have provided background evidence for optimising use of trabectedin through improved drug management.

Post hoc analyses of pivotal clinical trials revealed that about one-third of patients treated with trabectedin in any line were able to benefit from long-term tumour control, remaining on treatment for at least 6 cycles [19, 35]. Importantly, this proportion increased to more than half of patients (53%) when trabectedin was administered early in treatment (i.e., in the second line) [35]. No other systemic therapy for advanced STS has been shown to provide such prolonged treatment [36].

The phase II T-DIS study demonstrated that, in patients who were progression-free after 6 cycles of trabectedin, continuing treatment until disease progression produced better outcomes than interrupting treatment and restarting at the time of progressive disease (median PFS 7.2 vs. 4.0 months; p = 0.02) [37]. The authors concluded that interrupting trabectedin treatment is not recommended in patients without progressive disease if tolerance is acceptable, even though trabectedin retains its activity at rechallenge [38].

Flexible administration is also a key element in the successful management of sarcoma with trabectedin. In a large retrospective analysis of advanced STS (n = 885), 47% of patients had their trabectedin dose reduced for adverse effects or other reasons [33]. By managing patients in this manner, the number of treatment-associated deaths was limited to 4 patients (0.5%).

The varying sensitivity of histological subtypes to chemotherapy and their differential prognoses are well recognised. Several studies of trabectedin involving pretreated patients with advanced STS of multiple subtypes indicated that median OS was longer in the leiomyosarcoma and liposarcoma subgroups than in other histologies [33, 39, 40], although clinical benefit with trabectedin has been demonstrated in numerous other histological subtypes (Table 3) [26, 33]. The best results to date with trabectedin in advanced sarcoma were reported in a recent phase IV study performed in a heterogeneous population of patients (n = 218) with multiple subtypes of advanced STS, most commonly leiomyosarcoma (42%), liposarcoma (23%) and synovial sarcoma (10%) [34]. Most patients (90%) had received a median of 1 previous line of chemotherapy. Patients received a median of 6 trabectedin cycles, with 124 patients (56.9%) receiving 6 or more cycles and up to a maximum of 44 cycles. In the response assessment, 58 patients (26.6%) had an objective response (CR + PR) and 143 patients (65.5%) achieved disease control (CR + PR + SD). The 3-month PFS rate was 70%, and 70% of patients were still alive 1 year after treatment; median OS was 21.3 months. As the study was non-interventional, these patient outcomes reflect real-world experience with trabectedin.

Outcomes with other approved drugs for advanced STS can also be influenced by factors such as treatment line, administration flexibility, and histological subtype. A post hoc analysis of the treatment benefit of pazopanib across key subgroups in the PALETTE registration study indicated that median PFS was prolonged when pazopanib was administered in the second line versus third and later lines (24.7 vs. 18.9 weeks) [41]. Likewise, median OS was prolonged when pazopanib was administered in the second line compared with more advanced lines (13.7 vs. 11.3 months). Although the PALETTE trial ultimately failed to show a statistically significant difference in median OS between pazopanib and placebo [20], it is thought that this may reflect a “rebound” effect accelerating progression after patients discontinue pazopanib therapy [42]. Adapting the pazopanib dose to manage toxicities was also beneficial in terms of improving outcomes: if ≥1 dose reduction, median PFS was prolonged versus no dose reduction (27.7 vs. 11.9 weeks); if ≥1 dose interruption, median PFS was prolonged versus no interruption (21.3 vs. 11.0 weeks) [41]. Interestingly, a retrospective review of phase II (EORTC 62043) and phase III (EORTC 62072) studies of pazopanib in advanced STS indicated that concomitant administration of gastric acid suppression therapy (e.g., proton pump inhibitors) for at least 80% of the time was associated with significantly decreased median PFS (2.8 vs. 4.6 months; p = 0.008) and median OS (8.0 vs. 12.6 months; p < 0.001) compared with non-use of gastric acid suppression therapy [43]. Pazopanib and the newer tyrosine kinase inhibitor, regorafenib, have both shown activity versus placebo in advanced STS, except in liposarcoma subgroups [44, 45]. Patients with metastatic adipocytic tumours were thus excluded from participation in the PALETTE study [20].

The identification by next-generation sequencing of a significant proportion of cases of actionable mutations among populations of patients with advanced STS suggests a move towards more personalized medicine in future [46, 47]. Next-generation sequencing has an important role in assisting diagnosis and selecting matched therapies [46]. Several proof-of-concept therapies already exist (Table 4). A new trial (MultiSarc) to be undertaken in France is the first study to implement exome and RNA sequencing to assist decision making in patients with advanced STS. Patients are to be randomised to either use or no use of next-generation sequencing which, in practice, is randomising patients between standard and adaptive approaches to treatment. Patients are to receive up to 6 cycles of a doxorubicin-based regimen. In the event of no disease progression and target identification in the next-generation sequencing arm, maintenance therapy will be proposed according to the specific genetic alteration.

The approach to treatment selection for upcoming generations of sarcoma patients is likely to be increasingly from the patient’s perspective (Fig. 4). Both now and in future, it is essential that the treatment plan be shared with the patient along with any additional information that he/she requires or requests and include accurate quality of life data that the patient can interpret and understand.

Writing assistance was provided by Content Ed Net (Madrid, Spain) with funding from PharmaMar, Madrid, Spain.

A.L.C. has received honoraria from Amgen, Bayer, Lilly, Novartis, Pfizer and PharmaMar.

The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.