Tebentafusp in Combination With Durvalumab And/or Tremelimumab in Patients With Metastatic Cutaneous Melanoma: A Phase 1 Study

BACKGROUND: Immune checkpoint inhibitors have significantly improved outcomes in first line cutaneous melanoma. However, there is a high unmet need for patients who progress on these therapies and combination therapies are being explored to improve outcomes. Tebentafusp is a first-in-class gp100×CD3 ImmTAC bispecific that demonstrated overall survival (OS) benefit (HR 0.51) in metastatic uveal melanoma despite a modest overall response rate of 9%. This phase 1b trial evaluated the safety and initial efficacy of tebentafusp in combination with durvalumab (anti-programmed death ligand 1 (PDL1)) and/or tremelimumab (anti-cytotoxic T lymphocyte-associated antigen 4) in patients with metastatic cutaneous melanoma (mCM), the majority of whom progressed on prior checkpoint inhibitors. METHODS: In this open-label, multicenter, phase 1b, dose-escalation trial, HLA-A*02:01-positive patients with mCM received weekly intravenous tebentafusp with increasing monthly doses of durvalumab and/or tremelimumab starting day 15 of each cycle. The primary objective was to identify the maximum tolerated dose (MTD) or recommended phase 2 dose for each combination. Efficacy analyses were performed in all tebentafusp with durvalumab±tremelimumab treated patients with a sensitivity analysis in those who progressed on prior anti-PD(L)1 therapy. RESULTS: 85 patients were assigned to receive tebentafusp in combination with durvalumab (n=43), tremelimumab (n=13), or durvalumab and tremelimumab (n=29). Patients were heavily pretreated with a median of 3 prior lines of therapy, including 76 (89%) who received prior anti-PD(L)1. Maximum target doses of tebentafusp (68 mcg) alone or in combination with durvalumab (20 mg/kg) and tremelimumab (1 mg/kg) were tolerated; MTD was not formally identified for any arm. Adverse event profile was consistent with each individual therapy and there were no new safety signals nor treatment-related deaths. In the efficacy subset (n=72), the response rate was 14%, tu