Background: To further explore the activity of T as second/further line treatment in retroperitoneal leiomyosarcoma (LMS) and well differentiated/dedifferentiated liposarcoma (LPS). The primary endpoint of the study was the growth modulation rate (GMR) defined as the ratio between the time to progression under T (TTP) and during previous chemotherapy treatment (TTP-1). The secondary end-points were objective response rate as per RECIST and PFS. Methods: This was a multicenter, single-arm Phase 2 study, conducted in 20 Italian centers. Patients with locally relapsed or metastatic disease, already treated with one or more previous systemic treatments with anthracyclines and/or ifosfamide, were enrolled. T was administered at a dose of 1.3-1.5 mg/mq with a top dose of 2.6 mg per cycle. T was administered as a 24h continuous infusion until progressive disease, major toxicity, patient’s intolerance or medical decision. As per protocol, patients were considered responders if the GMR was > 1.33, non-responders if < 0.75 and neither if 0.76-1.32. Eighty evaluable patients were needed to detect an odds of trabectedin response ≥ 2.5, corresponding to 71.4% of patients with a GMR > 1.33 (80% power, one-sided alpha 2.5%). Results: From August 2014 to February 2019, 104 patients were registered and 91 were evaluable for the primary endpoint (32 pts with LMS and 59 with LPS). Overall, the median number of cycles received was 6.0 (q1-q3 3.0-12.0), the main reason for treatment discontinuation was disease progression in 72% of patients, followed by medical decision (8%). The median TTP was 6.0 months (6.2 and 6.0 for LMS and LPS), while the median TTP-1 was 7.5 months (8.1 and 6.4 for LMS and LPS). Thirty three patients (52% 95%CI: 36-58, p = 0.674, odds of response = 1.1) had a GMR > 1.33 (LMS: 46%, 95%CI 26-67,odds = 0.85; LPS 56%, 95%CI 40-72, odds = 1.3).Overall, response rate (CR+PR) was 16% (24% for LMS and 12% for LPS). Overall, in LPS we observed 15/47 patients with GMR < 0.5 and 15/47 with GMR > 2. Among LMS patients, 9/26 had a GMR < 0.5 and 10/26 > 2. Between LPS six patients had a GMR > 5. Previous treatment had been based on anthracyclines and/or ifosfamide in 85% of patients (91% in LPS population). Conclusions: While the primary end point of the study was not met, we noticed a subgroup of patients with a markedly discrepant TTP with T in comparison to previous therapy (GMR < 0.5 or > 2, the latter including some pts with a long TTP with T). Efforts are ongoing to assess the pathologic counterparts of such discrepancies. T seems to be selectively active in poorly understood subgroups, with a pattern of activity distinct from other available agents. Clinical trial information: 2012-005428-14.