With the exception of gastrointestinal stromal tumors (GIST), the prognostic of patients treated for locally advanced and metastatic STS remains poor. In all clinical guidelines, management of sarcoma patients should be discussed and carried out within a dedicated multidisciplinary team. Despite a management in expert centres, the median of overall survival is 18 months (12). The various palliative treatments may include surgery (for example, pulmonary metastasis), radiotherapy and especially chemotherapy (13). Nowadays, second lines after failure of doxorubicin and/or ifosfamide include others regimens like dacarbazine alone or in combination, eribulin for liposarcomas, gemcitabine alone or in combination with docetaxel for leiomyosarcomas and undifferentiated pleomorphic sarcomas, paclitaxel for angiosarcomas or pazopanib (14–18). In the end, few drugs have been approved for STS treatment, and the arrival of trabectedin in 2007 in second and/or beyond line treatment set a new course (19).
Our results, showing a median PFS of 3 months [CI95%: 2.3–4.7] and a median OS of 11.9 months [CI95%: 10.2–16.6], are fully compliant with the literature data. Demetri et al. published the results of a STS − 201 phase II study in 2009 that showed the superiority of trabectedin regimens every 3 weeks versus every week. They obtained a median PFS of 3.3 months and median OS of 13.9 months in 3 weeks 24-hour arm (10). Other phase II studies found similar results (20–22). Most recently, a prospective randomized phase III trial T-SAR comparing trabectedin versus best supportive care (BSC) in patients with pre-treated advanced STS, was presented to ESMO 2016 (23). Median PFS and median OS were 3.1 months [CI95%: 1.8–5.9] and 13.6 months [CI95%: 7.1–17.3] respectively in trabectedin arm. Finally, several RWE studies have already been published including a slightly larger population than in the pivotal studies: no limit neither for the number of lines of previous chemotherapy, nor for the histological subtypes (24–28). They found a median PFS between 3.6 and 5.9 months and a median OS between 11.9 and 21.3 months.
Based on the European Organization for Research and Treatment of Cancer (EORTC) experiment, Van Oosterom et al. and Van Glabbeke et al. proposed PFS rates at 3 and 6 months as end points for detected effective/ineffective treatments in phase II trials (19,20). A drug is considered active if it allows 3- and 6-month PFS rates > 39% and 14%, respectively, as second-line regimens. In our retrospective study, the 3-months PFS rate and the 6-months PFS rate were respectively 48.5% and 30.0%, comparable to the other publications.
To date, one major concern in STS treatment is that no predictive nor prognostic biomarker can be routinely used. In this retrospective analyse, no clinical or tumoral parameter was significantly associated with PFS and OS. Trabectedin administration resulted in a median PFS of 3.1 months [CI95%: 2.0–5.3] in L-sarcoma arm versus 2.8 months [CI95%: 2.3–5.0] in other subtypes arm, a median OS of 11.4 months [CI95%: 7.0–13.2] versus 13.8 months [CI95%: 9.3–18.8]. However, it is now assumed that the histological subtypes liposarcoma and leiomyosarcoma are particularly sensitive to trabectedin. This concept comes in particular from the results of pivotal phase II and III studies including exclusively patients treated for L-sarcomas (10,29). In the prospective multicenter phase III trial SAR – 3007, 518 patients were randomized between trabectedin to dacarbazine, after failure conventional chemotherapy. The median PFS was 4.2 months with a reduction in progression risk of 45% in favor of trabectedin. On the other hand, there is no difference on the median of OS. This sensitivity of L-sarcomas is well demonstrated in the T-SAR study (23). Inclusions were not limited to patients treated for liposarcomas or leiomyosarcomas. The median PFS of the trabectedin arm for all histological subtypes was 3.1 months [IC95%: 1.8–5.9], while it was 5.1 months [IC95%: 2-8.3] for L-sarcomas subtype (n = 32 patients) and 1.8 months [IC95%: 0.7–3.1] for others subtypes (n = 20 patients). There was no difference in OS between the different groups, trabectedin arm and BSC arm. Finally, all these results are really superposable, highlighting the ubiquitous action of trabectedin with more sensitive subtypes than others like L-sarcomas - especially myxoid liposarcomas (MLP) - and also solitary fibrous tumors (30–32). In our study, 11 patients were treated with trabectedin for MLP. The median PFS was remarkable with 13.3 months [CI95%: 2.3–18.7] versus 2.8 months [CI95%: 2.3–4.4] for all other patients with a significant improvement (p = 0.02). Equally interesting, this difference was also found for the median OS: 27.8 months [CI95%: 3.2–64.7] versus 11.1 months ([CI95%: 9.3–13.8], p = 0.02). Already in 2007, 51 patients with advanced pretreated MLP at five international institutions in a compassionate-use programme were analysed retrospectively (30). After a follow-up of 14.0 months, the objective response rate was 51% and the median of PFS was 14.0 months [IC95%: 13.1–21.0]. This specific subtype is characterized by the existence of the fusion gene DDIT3-FUS and the very high effectiveness of trabectedin against MLP seems to be correlated to its ability to counteract the activity of chimeric oncoprotein. By extrapolation, other histological subtypes with translocation-related sarcomas (synovial sarcoma, alveolar soft part sarcoma, endometrial stromal sarcoma and clear cell sarcoma) were evaluated in a retrospective analysis giving promising results (33).
Another highlight is the interest of ongoing treatment. Even if few patients were concerned, observations can be drawn. After the first 6 cycles, 2 opposing strategies can be chosen; either to stop or to continue trabectedin until intolerance or tumour progression. About a quarter of the patients received more than 6 cycles of trabectedin, which an acceptable toxicity allowing prolonged treatment. The evolution of the 2 groups differs: median PFS was doubled in favor of therapeutic maintenance. On the other hand, no difference on the median OS was observed. This finding is already documented by the T-DIS study, a randomised phase 2 trial (34). Median PFS was 7.2 months [CI95%: 4.0–12.7] in the continuation group versus 4.0 months [CI95%: 2.5 − 5.5] in the interruption group, with a significant improvement (HR: 1.97 [CI95%: 1.09–3.54], p = 0.02). Nonetheless, in our retrospective study, probable bias exists: no randomization, patients who continued chemotherapy were certainly patients where the benefit/tolerance balance was the most favourable, and few number of patients (29 patients in the maintenance group and 15 in the interruption group).
In terms of toxicities, most of them were expected and manageable, non - cumulative, mainly summarized by haematological and hepatological failures, and consistent with the previous T – SARC randomized trial (23). Thirty four percent of patients used granulocyte colony-stimulating factors which were 28% in the Demetri et al. study (10). Other toxicities were quite similarly to the results of the current publications (35–37).
This study had some limitations. The retrospective design made comparisons with previous randomized control trial difficult because the study population and follow-up may differ. For example, the enrolment and treatment criteria of our study were less restrictive with more pretreated patients and other histological subtypes than liposarcomas and leiomyosarcomas. In addition, dose reductions, treatment delays and tumor evaluations depended on the physician’s choice. The small number of patients and missing data limited the analysis, in particular the association between different histological subtypes and clinical outcomes. Finally, there was no provision for a quality-of-life assessment Despite these limitations, this study is representative of the general population because it included patients treated in 2 centers with different medical situations, and these results could be considered to represent assessment of trabectedin effectiveness in the real-world clinical conditions.