The objective of this retrospective study was to assess the outcome of STS patients treated in the French ATU compassionate use program and to compare it with that of published clinical trials. Between 2003 and 2008, this program enabled the treatment of patients failing doxorubicin with trabectedin 1.5 mg/m2/21d. The inclusion criteria were the same than those of the EORTC trial, with the exception that all lines were allowed.
Not all patients could be retrospectively collected. Actually only centres that included more than 5 patients contributed to this analysis, and these included 181 patients in this retrospective study. This series represents therefore a selected subgroup of patients treated mostly in reference centres for sarcoma, and in experienced centres regarding trabectedin usage. Among the 11 centres participating to the study, 5 had participated to the phase II EORTC trial, reflecting the experience of the centres with this agent. This is therefore a selected subgroup of the ATU series, but this selection makes comparison with phase II data maybe more relevant. It would have been of interest to compare this series to that of patients treated in non-expert centres but this could not be obtained. In this group of 181 heavily pretreated STS sarcoma patients, either resistant or relapsing, trabectedin was received as a second line therapy for a majority of them and some patients received the treatment in 4th line. This is a more heavily pretreated patient population than that of the EORTC trials. Despite of this, the response rate (10%), stable disease rate (39%), PFS (median 3.6 months) and OS (median 16.1 months) were comparable to those observed with trabectedin in the phase II trials. According to the EORTC-STBSG (European Organisation for Research and Treatment of Cancer- Soft Tissue and Bone Sarcoma Group) criteria, because the 3-months progression free rate was largely superior to 40% and the 6-monthz PFR was superior to the threshold used to define an active treatment according to the EORTC STBSG
. It is however challenging to compare the present series with the EORTC database of the pre-trabectedin era
[2, 3] published since 1999 for several reasons : 1) the former series included mainly first line patients, while the present series gathers patients in all lines (from first line metastatic in patients pretreated in the adjuvant setting to fourth line patients. 2) histological classifications and inclusion criteria varied considerably between the 2 series; for instance, GIST were mixed amongst leiomyosarcoma in the former series. The exhaustive histological reviews of the former series were has not performed with the classifications of 2002 or 2013.
Possibly the best comparison can be obtained with the subsequent paper by Van Glabbeke et al., reporting separately second line + patients. In this case, the median progression free rate is 2.3 months, and a 1 year PFR rate of 7% in the whole series, and 12% for the series of patients treated with “active agents”
. The results observed with the present ATU series, median PFS of 3,6 months, and 12 months PFS close to 30% compare therefore favorably with these historical controls, despite all these limitations.
Detailed side-effects of trabectedin treatment were not collected in this retrospective study. Only toxicity leading to hospitalization was documented and remained limited, affecting only 17% of patients. As expected these patients had a smaller number of cycles delivered and, perhaps as a consequence, had a worse PFS. Overall survival was however not significantly different than that of on patients without toxicity-related hospitalization. Because most toxicities do not lead to rehospitalisation, no formal conclusion can be proposed on a possible lack of correlation with therapeutic efficacy, considering also the limited number of patients in this series.
As previously described, patients with myxoid liposarcomas had better response rates, PFS,OS, and was an independent prognostic factor for survival. The number of lines of chemotherapy administered before trabectedin also correlated significantly with longer PFS in the Cox model, but not for OS. Conversely, retroperitoneal location was were associated with a improved survival, possibly because of the low grade and loco-regional behavior; most are liposarcomas, a subset associated with a better outcome in large retrospective datasets in the present series as well
. Interestingly, OS and PFS of translocation-related sarcomas excluding myxoid liposarcoma was not different of that of other sarcoma types. Similar observations were made for response rate (not shown).
Thirty percent of the patients in the present study received more than 6 cycles of trabectedin, which underlines an acceptable toxicity profile allowing prolonged treatment. Long-term treatment is feasible with trabectedin, while this is not feasible with doxorubicin nor ifosfamide because of cumulative cardiac and renal toxicities. Prolonged trabectedin treatment thus allows testing the importance of maintenance treatment. Interestingly, among the 56 patients who were not progressing after 6 cycles, the 40 who continued treatment had a significantly better PFS and more surprisingly OS, with a more than doubling of the median OS. The retrospective nature of the study implies potentials biases in these observations, and therefore this cannot be considered as an evidence for the utility of prolonged treatment. However, these observations strengthen the rationale of the ongoing study randomizing treatment maintenance vs interruption after 6 cycles which is currently ongoing within the French Sarcoma Group (NCT01303094). It has previously been shown in a large randomized clinical trial (SUCCEED) that maintenance with an mTOR inhibitor enables to prolong PFS, but not OS (Demetri et al. submitted for publication). Maintenance therapy maybe a strategy worth further exploring in patients with advanced STS.