A Randomized Phase II Study of Nivolumab Monotherapy or Nivolumab Combined with Ipilimumab in Patients with Advanced Gastrointestinal Stromal Tumors
Arun S. Singh, J. Randolph Hecht, Lee Rosen, Zev A. Wainberg, Xiaoyan Wang, Michael Douek, Anahis Hagopian, Rachel Andes, Lauren Sauer, Sandra R. Brackert, Warren Chow, Ronald DeMatteo, Fritz Christian Eilber, John A. Glaspy, Bartosz Chmielowski
Abstract
PURPOSE: Most gastrointestinal stromal tumors (GIST) are driven by KIT/PDGFRa mutations. Tyrosine kinase inhibitor benefit is progressively less after imatinib failure. This phase II trial analyzed the efficacy of nivolumab (N) or nivolumab + ipilimumab (N + I) in patients with refractory GIST. PATIENTS AND METHODS: Patients with advanced/metastatic GIST refractory to at least imatinib were randomized 1:1 in a noncomparative, parallel group, unblinded phase II trial of N (240 mg every 2 weeks) or N + I (240 mg every 2 weeks + 1 mg/kg every 6 weeks). The primary endpoint was the objective response rate of N alone or N+I by RECIST 1.1 in the intent-to-treat population. RESULTS: A total of 36 patients with a median of 3 (1-6) prior lines of therapies were enrolled. Ten of 19 (52.6%) patients had stable disease (SD) for a clinical benefit rate (CBR) of 52.6% in the N arm and the median progression-free survival (PFS) was 11.7 weeks [95% confidence interval (CI), 7.0-17.4]. In the N+I arm, 1 of 16 (6.7%) patients had a complete response (CR) and 4/16 (25.0%) had SD for a CBR of 31.3% and a median PFS of 8.3 weeks (95% CI, 5.6-22.2). The 4- and 6-month PFS were 42.1% and 26.3%, respectively for N, and 31.3% and 18.8%, respectively for N+I. The most common adverse events (AE) attributed to N and N+I were fatigue: 13.9% and 22.2%, respectively. There were nine total attributable grade 3-4 AEs. CONCLUSIONS: The primary endpoint of response rate > 15% was not observed for N or N + I. In a heavily pretreated GIST population, responses and long-term disease control with both N and N+I were observed. No new safety signals have been observed.