Background: T is a HSV-1-based oncolytic virus designed to selectively replicate in tumors and produce GM-CSF to stimulate antitumor immune responses. I is an anti-CTLA-4 Ab that blocks inhibition of activated T cells. This is the first randomized study to evaluate the addition of an oncolytic virus to a checkpoint inhibitor. Methods: The 1^o endpoint was ORR by immune-related response criteria. Key 2^o endpoints: duration of response, disease control rate (DCR), PFS, OS, and safety. Prior treatment was allowed but not required. Pts had unresected stage IIIB-IV melanoma with measurable/injectable tumor(s) and no evidence of immunosuppression. T was given at approved dosing until no injectable tumors, disease progression (PD), or intolerance. I was started at w6 in T+I and at w1 in I at 3 mg/kg IV q3w x 4. Primary analysis occurred 6 m after last pt enrolled. Results: 198 pts were randomized: 98 T+I; 100 I. Characteristics were similar: 54% stage IIIB-IVM1a, 46% IVM1b/c. Median follow up time was 68 w (T+I) and 58 w (I). ORR was 38.8% (T+I) and 18.0% I, P = 0.002, Odds ratio (OR) 2.9. 89% T+I and 83% I pts remain in response. Unconfirmed visceral lesion response was 35.5% T+I vs 13.6% I. OS is immature. Of 190 pts (safety set: 95 T+I, 95 I), most common adverse events (AEs) for T+I, I (%) were fatigue (59, 42), chills (53, 3), and diarrhea (42, 35). 28% T+I and 18% I pts had gr ≥3 tx-related AE. There were 3 deaths (all unrelated) in T+I: 1 myocardial infarction and 2 PD. Conclusions: The study met the 1^o endpoint. ORR was significantly higher for T+I vs I; responses were not limited to injected lesions. Toxicity of T+I combination was tolerable with no unexpected AEs. Clinical trial information: NCT01740297

^aCR/PR required confirmation ≥ 4 w apart ^bDurable response rate = response ≥ 6 m; descriptive P= 0.007^ccChi-square test with continuity correction