Overall Survival in Patients With Advanced Melanoma Who Received Nivolumab Versus Investigator's Choice Chemotherapy in CheckMate 037: A Randomized, Controlled, Open-Label Phase III Trial

Abstract

Purpose Until recently, limited options existed for patients with advanced melanoma who experienced disease progression while receiving treatment with ipilimumab. Here, we report the coprimary overall survival (OS) end point of CheckMate 037, which has previously shown that nivolumab resulted in more patients achieving an objective response compared with chemotherapy regimens in ipilimumab-refractory patients with advanced melanoma. Patients and Methods Patients were stratified by programmed death-ligand 1 expression, BRAF status, and best prior cytotoxic T-lymphocyte antigen-4 therapy response, then randomly assigned 2:1 to nivolumab 3 mg/kg intravenously every 2 weeks or investigator's choice chemotherapy (ICC; dacarbazine 1,000 mg/m² every 3 weeks or carboplatin area under the curve 6 plus paclitaxel 175 mg/m² every 3 weeks). Patients were treated until they experienced progression or unacceptable toxicity, with follow-up of approximately 2 years. Results Two hundred seventy-two patients were randomly assigned to nivolumab (99% treated) and 133 to ICC (77% treated). More nivolumab-treated patients had brain metastases (20% v 14%) and increased lactate dehydrogenase levels (52% v 38%) at baseline; 41% of patients treated with ICC versus 11% of patients treated with nivolumab received anti-programmed death 1 agents after randomly assigned therapy. Median OS was 16 months for nivolumab versus 14 months for ICC (hazard ratio, 0.95; 95.54% CI, 0.73 to 1.24); median progression-free survival was 3.1 months versus 3.7 months, respectively (hazard ratio, 1.0; 95.1% CI, 0.78 to 1.436). Overall response rate (27% v 10%) and median duration of response (32 months v 13 months) were notably higher for nivolumab versus ICC. Fewer grade 3 and 4 treatment-related adverse events were observed in patients on nivolumab (14% v 34%). Conclusion Nivolumab demonstrated higher, more durable responses but no difference in survival compared with ICC. OS should be interpreted with caution as it was likely impacted by an increased dropout rate before treatment, which led to crossover therapy in the ICC group, and by an increased proportion of patients in the nivolumab group with poor prognostic factors.

Trial registration: ClinicalTrials.gov NCT01721746.

Fig 1.

Trial design. ICC, investigator’s choice chemotherapy.

Fig 2.

(A and B) Overall survival (OS) in all randomly assigned patients and OS censoring at the start of programmed death 1 (PD-1) or programmed death-ligand 1 (PD-L1) agent after assigned therapy in investigator’s choice chemotherapy (ICC). (A) Kaplan-Meier curves for OS in all randomly assigned patients. Median OS was 15.7 months (95% CI, 12.9 to 19.9) in the nivolumab (NIVO) group and 14.4 months (95% CI, 11.7 to 18.2) in the ICC group (hazard ratio for death, 0.95; 95.54% CI, 0.73 to 1.24; P = .716). (B) Kaplan-Meier curves for OS in all treated patients censoring at the start of PD-1 or PD-L1 agent after assigned therapy in ICC. Median OS was 16.4 months (95% CI, 12.9 to 20.3) in the NIVO group and 11.8 months (95% CI, 9.9 to 14.4) in the ICC group (hazard ratio for death, 0.81; 95.54% CI, 0.59 to 1.11).

Fig 3.

Duration of response per independent radiologic review committee. Swimmer plots show time to first response and duration of response, as defined by RECIST v1.1, for responders who received nivolumab (NIVO) or investigator’s choice chemotherapy (ICC).

Fig 4.

Progression-free survival (PFS) by independent radiologic review committee (IRRC) assessment. Kaplan-Meier curves for PFS in all randomly assigned patients by IRRC assessment. Median PFS was 3.1 months (95% CI, 2.3 to 3.5) in the nivolumab (NIVO) group and 3.7 (95% CI, 2.3 to 5.3) in the investigator’s choice chemotherapy (ICC) group (hazard ratio for death or disease progression, 1.03; 95.1% CI, 0.78 to 1.436).