The relationship between response rate and survival benefits in randomized immunotherapy studies

Understanding the relationship between the Objective Response Rate (ORR) and survival outcomes, notably Progression-Free Survival (PFS) and Overall Survival (OS), is relevant for assessing the efficacy of regimens in oncology. We evaluate the relationship between ORR, PFS and OS in immuno-oncology (IO) trials. Data from 68 clinical trials submitted to the FDA were evaluated, examining immunotherapy regimens, notably immune checkpoint inhibitors such as anti-programmed death (ligand)-1 [anti-PD-(L)1], cytotoxic T-lymphocyte-associated protein-4 (CTLA-4) inhibitors and combination therapies [e.g., IO + IO, anti-PD-L1 + chemotherapy, anti-PD-L1 + CTLA-4, anti-PD-L1 + TKI (tyrosine kinase inhibitors)]. Studies were included based on their reporting of ORR, PFS, and OS. Of the 68 clinical trials reviewed, 55 were included in the analysis. The correlation between ORR and PFS was moderate across most immunotherapy regimens, indicating that ORR can serve as a useful predictor of short-term disease control. However, the correlation between ORR and OS was weaker, especially in trials including combination therapies, indicating that ORR alone may not reliably predict long-term survival outcomes. ORR predicts PFS better in first-line treatment but declines in later lines and remains a weak OS predictor overall. Differing degrees of correlation between ORR and survival metrics, particularly across treatment lines and combinations, are observed. While ORR can serve as a surrogate marker for PFS in IO trials, its utility in predicting OS is restricted and the interpretation of the relationship between ORR and PFS or OS is a key limitation. Rather, a decline in PFS with increasing ORR may reflect trial differences rather than a direct relationship. Future analyses should adopt better methodologies to capture these dynamics and focus on improving surrogate endpoints for immunotherapy to improve clinical trial design and patient outcomes.