Programmed death 1 (PD-1) inhibitors, such as nivolumab and pembrolizumab, have now been approved for various cancers based on results from pivotal randomized controlled trials (RCTs). These drugs are known for unconventional response patterns with varying effects on PFS and OS. We aimed to compare the correlation between PFS and OS and evaluate the differences in treatment size between PFS and OS for PD-1 inhibitors.
We carried out a systematic search on PubMed and conference abstracts for RCTs of nivolumab and pembrolizumab versus non-immunotherapy control and obtained data on median PFS, median OS for both arms and hazard ratio (HR) and confidence intervals (CIs) for PFS and OS. We evaluated the correlation between PFS and OS as well as between Delta (PFS) and Delta (OS). We also evaluated the ratio of HR of PFS to HR of OS for each trial (rHR) and obtained a summary rHR by random-effects meta-analysis across trials.
Of 52 studies identified, a total of 11 phase 3 RCTs met the eligibility criteria. However, 2 trials didn’t have data on OS. So our analysis includes 9 RCTs that had data on both PFS and OS (6 Nivolumab, 3 Pembrolizumab). There was no significant correlation between PFS and OS (r = 0.676, R2= 0.457, P = 0.095) or between Delta (PFS) and Delta (OS) (r = 0.474, R2= 0.225, P = 0.282). Using random-effects meta-analysis, treatment effects were in general 19% higher for OS than PFS (rHR 1.19, 95% CI 1.07 to 1.32, p = 0.001). There was no statistical evidence for lack of homogeneity (I2= 0.0%, p = 0.850) and thus, subgroup analysis were not conducted. PFS and OS were discordant for 5 RCTs (3 Nivolumab, 2 Pembrolizumab) and in all these 5 RCTs, OS was significant but PFS was not. All RCTs (n = 3) showing benefit for PFS also showed benefit for OS. Only one RCT was negative for OS.
Unlike targeted therapies where benefit in PFS mayn't translate to OS, treatment effect sizes in RCTs of PD-1 inhibitors were greater for OS than PFS. The benefit in OS was poorly captured by PFS. There was no correlation between PFS and OS. OS should remain the standard endpoint for PD-1 inhibitor RCTs unless better surrogate endpoints such as immune-criteria based PFS are introduced and validated.
Clinical trial identification
Legal entity responsible for the study
Y. Ando: Reports grants and personal fees from Taiho Pharmaceutical Co., Ltd., and personal fees from Merck Serono Co., Ltd., Ono Pharmaceutical Co., Ltd and Bristol-Myers Squibb. All other authors have declared no conflicts of interest.