Propensity weighted indirect treatment comparison of nivolumab (NIVO) versus placebo (PBO) as adjuvant therapy for resected melanoma: A number needed to treat and overall survival analysis

9572

Background: The CheckMate 238 trial demonstrated that NIVO improved recurrence free survival (RFS) vs. ipilimumab (IPI). The EORTC 18071 trial demonstrated that IPI improved RFS and overall survival (OS) vs. PBO. The current study pooled data from these two trials to indirectly assess the RFS and OS of NIVO vs. PBO and the numbers needed to treat (NNTs) for one additional recurrence-free survivor and survivor over 4 years. Methods: Patients with resected AJCC 7th edition stage IIIB/C cutaneous melanoma from CheckMate 238 (NIVO vs. IPI) and EORTC 18071 (IPI vs. PBO) were pooled together with inverse probability weighting to balance between-trial differences in baseline characteristics. NNTs were calculated for RFS and OS comparing NIVO vs. IPI and PBO over 4 years. To account for improved post-recurrence survival over time, a sensitivity analysis that adjusted for post-recurrence survival in the PBO arm of EORTC 18071 was performed. Results: A total of 278, 643, and 365 patients treated with NIVO, IPI, and PBO, respectively, were included. In the weighted samples, patients treated with NIVO had consistently higher RFS rates than those treated with IPI (HR [95% CI]: 0.69 [0.56, 0.85]) and PBO (HR: 0.49 [0.39, 0.61]). NIVO was associated with similar OS as IPI (HR: 0.80 [0.60, 1.08]) and superior OS compared to PBO (HR: 0.45 [0.33, 0.60]). At 4 years, the weighted RFS rate was 53.1% for NIVO, 41.8% for IPI, and 29.1% for PBO. The NNT to achieve one additional recurrence-free survivor was 4.2 for NIVO vs. PBO and 8.9 for NIVO vs. IPI. The NNT to obtain one additional survivor was 4.8 for NIVO vs. PBO and 22.2 for NIVO vs. IPI. The OS rate for PBO after adjusting for differences in post-recurrence treatments at 4 years was 64.1%, and the corresponding NNT of OS comparing NIVO vs. adjusted PBO was 8.5. Conclusions: In patients with resected AJCC 7th edition stage IIIB/C cutaneous melanoma, this indirect comparison showed that NIVO improved RFS and OS vs placebo, with OS improvement maintained after adjustment for post-recurrence therapy.[Table: see text]