OBJECTIVES: Initial results from CheckMate 032, a Phase 1/2 clinical trial, showed durable responses and encouraging survival with nivolumab-based regimens in SCLC patients after previous platinum-containing regimens. Response-based modeling has been proposed to capture extended OS associated with I-O treatment. We examine the relationship between patients’ response status at a landmark and OS by treatment arm.

METHODS: Analysis was conducted on the November 2017 database lock of CheckMate 032. A Kaplan-Meier analysis of OS was performed by response status at a 4 month landmark for each study arm. Additional landmark time points were explored in sensitivity analysis. Patients with a complete/partial response as measured by RECIST 1.1 were classified as responders. Patients with stable or progressed disease were classified as non-responders.

RESULTS: Overall response rate was 11.8% (29/245) for nivolumab and 22.3% (35/157) for nivolumab plus ipilimumab. At 4 months there were 24 and 29 responders among nivolumab and nivolumab plus ipilimumab patients, respectively. The number of non-responders at 4 months was 101 and 59 among nivolumab and nivolumab plus ipilimumab patients, respectively. For nivolumab, median post-landmark OS was 6.8 months (95% CI: 5.9 - 9.4) for non-responders; the median was not reached for responders. For nivolumab plus ipilimumab, median post-landmark OS was 9.1 months (95% CI: 5.9 – 19.6) for non-responders and 26.2 months (95% CI: 14.0-NR) for responders. The hazard ratio between responders and non-responders was 0.13 (95% CI: 0.05-0.32) in the nivolumab arm and 0.47 (95% CI: 0.25-0.89) in the nivolumab plus ipilimumab arm.

CONCLUSIONS: In CheckMate 032, response at the 4-month landmark was associated with longer OS in both treatment arms. Caution should be taken when making inferences from the results due to the limited sample size. These findings support incorporating separate response-based survival curves in economic models.