Evaluation of objective response, disease control and progression-free survival as surrogate end-points for overall survival in anti–programmed death-1 and anti–programmed death ligand 1 trials

We aimed to assess whether the Response Evaluation Criteria in Solid Tumors (RECIST) criteria-based objective response rate (ORR), disease control rate (DCR) and progression-free survival (PFS) could be valid surrogate end-points for overall survival (OS) in anti-programmed death-1 (PD-1)/programmed death ligand 1 (PD-L1) trials.We systematically reviewed phase 2 and phase 3 trials of anti-PD-1/PD-L1 drug trials of advanced or recurrent solid tumours that reported OS and at least one of the RECIST criteria-based end-points. We used Spearman rank correlation to evaluate the strength of the association between these end-points and OS and a linear regression model, weighted by the sample size, to assess the association between the treatment effect on these end-points and OS. We also performed sensitivity analyses and a leave-one-out cross-validation approach to evaluate the robustness of our findings.Forty-three qualifying trails comprising 15,088 patients were eligible. PFS showed good correlation with OS (squared Spearman rank correlation coefficient [rs2] = 0.54; P < 0.001), while ORR and DCR illustrated moderate association with OS (rs2 = 0.29 and 0.28, respectively; both P < 0.001). The correlation was moderate between the treatment effects on PFS and OS (coefficient of determination [R2] = 0.37, P < 0.001) and poor among ORR, DCR and OS (R2 = 0.10 and 0.08, respectively); these were confirmed by sensitivity analyses (all R2 < 0.75) and the leave-one-out cross-validation approach.No RECIST criteria-based end-points could be a valid surrogate for OS. At present, we proposed to set OS as the primary end-point in anti-PD-1/PD-L1 drug trials of advanced or recurrent solid tumours.