A randomized phase II study of nivolumab monotherapy versus nivolumab combined with ipilimumab in advanced gastrointestinal stromal tumor (GIST).

Abstract

11017 Background: Most GISTs are driven by mutations in KIT and PDGFRa and secondary mutations are felt to confer resistance to TKIs. In advanced/metastatic GIST, the benefit of second line TKIs and beyond is progressively less after imatinib failure. As such, novel non-TKI approaches are important to explore. Here we report interim analyses of safety and efficacy in advanced GIST patients treated with immunotherapy. Methods: Patients with advanced/metastatic GIST refractory to at least imatinib were enrolled on a randomized, parallel group, unblinded Phase 2 trial of either nivo (240 mg Q2wks) or nivo (240 mg Q2wks) with ipi (1mg/kg Q6wks) for up to 2 years. The primary endpoint was the objective response rate(ORR) of nivo alone or nivo + ipi by RECIST 1.1 criteria. Imaging was assessed by investigator and 3 independent radiologists. Patients were randomized 1:1 and were restaged every 8 weeks. With a sample size of 20 per group, an exact binomial test with a nominal 0.050 one-sided significance level will have 82% power to detect the difference between the null hypothesis response rate 1.5% and the alternative response rate of 15%. Secondary objectives are to ascertain the PFS, CBR, RR by Choi criteria and safety. Blood and biopsies are also being collected. Results: At cutoff, 29 patients (27 evaluable) with a median of 3 (1-7) lines of prior therapies have started on trial. In the nivo only arm, 7/15 pts had a best response of SD for a CBR of 46.7% with the median PFS being 8.57 wks. In the nivo + ipi arm, 1/12 patients had a PR and 2/12 have SD for a CBR of 25.0% (95% exact C.I. 5.5%-57.2%) with a median PFS of 9.1 wks. 8 patients have been on therapy for more than 6 months and two patients with a KIT Exon 17 mutation had radiographic disease shrinkage. Most AEs were grades 1-2 with fatigue (37%) being the most common. 4 Grade 3/4 AEs occurred in the nivo and ipi arm (hyperglycemia, weakness, diarrhea x 2) and 4 grade 3/4AEs occurred in the nivo arm (DKA, hyperglycemia, rash, fatigue). Pretreatment biopsies have been obtained in all patients and blood has been collected on all patients for correlative analysis. Conclusions: In a heavily pretreated GIST population, responses and disease control with both nivo and nivo + ipi were observed. To date, the drugs have been well tolerated and no new safety signals have been observed in this disease state. Clinical trial information: NCT02880020.