A Meta-Analysis To Indirectly Compare The Safety And Efficacy Of Pd-1 And Pd-L1 Antibodies Across Solid Tumors Using A Bayesian Hierarchical Model.

3065 Background: Many PD-1/PD-L1 mAb are in development across a broad range of tumors. The expanded literature provided an opportunity to address a critical question: if PD-1 mAbs targeting the receptor have a different safety/efficacy profile than the PD-L1 mAbs targeting the ligand. Direct comparison of these mAb would be ideal but highly unlikely. Methods: We performed a meta-analysis using a Bayesian model to provide indirect comparison based on data in the public domain including PubMed and key medical conference abstracts (eg ASCO, ESMO etc.). Key inclusion/exclusion criteria include patients 1) with any specific solid tumor, 2) PD-L1 status if tested using one of the approved assays, 3) received one of the five approved mAbs as a monotherapy, and 4) with disclosure date between June 4th, 2013 and June 6th, 2017. The search resulted in about 70 clinical trials in 31 tumors and 12,025 patients. The endpoints include ORR, PFS for efficacy and the overall G3/4 AE rate for safety. Results: For efficacy, in all solid tumors except for HNSCC, ORR is slightly greater for PD-1 than PD-L1 mAb, with the largest odds ratio (OR) in 2L+ NSCLC favoring the PD-1 mAb [ORRPD-1= 19.8% vs ORRPD-L1= 16.5%, OR = 1.27, 95% CI = (0.96, 1.70)] and the smallest OR in 1L HNSCC favoring the PD-L1 mAb [ORRPD-1= 27.4% vs. ORRPD-L1= 28.7%, OR = 0.975, 95% CI = (0.38, 1.50)]. Overall, there is 81% posterior probability that ORRs are equivalent for PD-1 and PD-L1 mAb. For safety, the Gr 3/4 AE rate (e.g. pneumonitis) is numerically higher for PD-1 than PD-L1 mAb [OR = 1.48, 95% CI = (0.19, 12.39)] across tumors. Overall, there is 79% posterior probability that the safety profile in terms of G3/4 AE rate is equivalent for PD-1 and PD-L1 mAb. Conclusions: There was no significant difference between PD-1 and PD-L1 mAb across tumor types for ORR, PFS and Gr 3/4 AE rate. Moreover, when the information was analyzed across tumor types, the small magnitude of the difference relative to the variability across tumor types suggests strong interchangeability of efficacy and safety profile of the antibodies targeting either PD-1 or PD-L1. Ultimately, the variability within this class of antibodies is not likely to be clinically meaningful.