Background: Vincristine, irinotecan and temozolomide (VIT) has provided an effective regimen in relapsed or refractory Ewing sarcoma patients. Optimal dose schedule of VIT is still undefined. Methods: We did this phase 2 randomized controlled trial and included patients with relapsed or refractory Ewing sarcoma. They were randomly assigned (1:1) by random number table method to either shorter dx5 schedule (Irinotecan 50mg/m²/d D_1-5, Vincristine 1.4mg/m² D₁) or protracted dx5x2 schedule (Irinotecan 20mg/m²/d D_1-5,8-12, Vincristine 1.4mg/m² D_1,8) together with a fixed dose of temozolomide (100mg/m²/d D_1-5) in both groups. Patients were treated every 3 weeks till progression or unacceptable toxic effects for up to eight cycles. The primary endpoint was objective response rate at 12 weeks (ORR_12w). The secondary endpoint was progression-free survival (PFS), overall survival (OS) and safety. The study was powered to detect a 30% improvement in the response rate from 20% of dx5 schedule to 50% of dx5x2 schedule (α = 0.1, 1-β = 0.9, one-sided test favoring dx5x2 schedule since the only difference of clinical importance was an improved response with the inconvenient schedule of dx5x2). A sample size of 30 patients per group (60 randomly assigned patients) was required to detect a significant improvement in ORR_12w. Results: Between May 21, 2020, and August 8, 2022, 24 patients were randomly assigned to dx5 schedule while 22 to dx5x2 schedule, respectively. Median follow-up was 10.7 months (IQR 9.7-13.9) in the dx5 group and 8.3 months (IQR 4.4-15.3) in the dx5x2 group. ORR_12w was lower in dx5 group than in dx5x2 group (5 [20.8%] of 24 patients vs 12 [54.5%] of 22; p = 0.019). There was no significant difference in PFS between the two groups (median PFS 2.3 months [95% CI 0.0-4.7] in dx5 group vs 4.3 months [2.7-6.0] in dx5x2 group; hazard ratio [HR] 0.956 [95% CI 0.84-1.09]; p = 0.434). Also, there was no significant difference in OS (median OS 14.8 months [95% CI 12.0-17.6] vs 12.8 months [95% CI 6.9-18.7]; HR 0.957 [95% CI 0.81-1.12]; p = 0.594). Patients in dx5 schedule reported more grade 3 and 4 adverse events (AEs) than dx5x2 schedule, including diarrhea/abdominal pain (23[23.7%] of 97 courses vs 8[8.4%] of 95 courses, p = 0.005), vomiting/nausea (6[6.2%] vs 1[1.15%], p = 0.035). Other common grade 3 and 4 AE ( > 5%) were similar between two schedules (p > 0.05), including fatigue (24[24.7%] vs 14[14.7%]), leukopenia (6[6.2%] vs 8[8.4%]), neutropenia (5[5.2%] vs8 [8.4%]), and anemia (5[5.2%] vs 7[7.4%]). Unfortunately, the isolation policy of COVID-19 significantly restricted the speed of recruiting and the sponsor stopped following support to our study. Conclusions: Protracted dx5x2 VIT schedule showed superior efficacy and favorable tolerability compared with the shorter dx5 VIT schedule in patients with relapsed or refractory Ewing sarcoma. Clinical trial information: NCT03359005.