BACKGROUND: To report on long-term results of a phase 3 trial comparing three versus five cycles of adjuvant chemotherapy (CT) with full-dose epirubicin+ifosfamide in high-risk soft tissue sarcomas (STS).METHODS: Patients (pts) were randomized to receive three preoperative cycles of epirubicin 120 mg/m(2) and ifosfamide 9 g/m(2) (Arm A) or to receive the same three preoperative cycles plus two postoperative cycles (Arm B). Radiotherapy could be either delivered in the preoperative or in the postoperative setting. Non-inferiority of the primary end point, OS, was assessed by the confidence interval of the hazard ratio (HR; Arm A/Arm B) derived from Cox model.RESULTS: Between January 2002 and April 2007, 164 pts were assigned to arm A and 164 to arm B. At a median follow-up (FU) of 117 months (IQ range 103-135 months), 123 deaths were recorded: 58 in Arm A and 65 in Arm B. Ten-year OS was 61% for the entire group of patients: 64% in Arm A and 59% in Arm B. The intention-to-treat analysis confirmed that three cycles were not inferior to five cycles (one-sided 95% upper confidence limit was 1.24). A per protocol analysis was consistent with these results. Pts with leiomyosarcoma and undifferentiated pleomorphic sarcoma (UPS) had the lowest, and the highest response rates, respectively. Consistently, Leiomyosarcoma and UPS had the worse and the best prognosis, respectively.CONCLUSIONS: At a longer FU, the non-inferiority of three cycles of a full-dose conventional CT in comparison to five is confirmed. Response to therapy is also confirmed to be associated with better survival. This regimen is currently tested within an ongoing international trial against three cycles of a neoadjuvant histology-tailored CT (ClinicalTrials.gov Identifier: NCT01710176).