[Bayesian Regression Modeling of the Correlation between Post-Progression Survival and Overall Survival in Immune Checkpoint Inhibitor Therapy].

Abstract

Progression-free survival(PFS)and overall survival(OS)are key indicators commonly used to evaluate the effectiveness of treatments for advanced or recurrent cancer. However, in immune checkpoint inhibitor(ICI)therapy, improvements in PFS do not always correlate directly with extended OS. This study analyzed PFS, OS, and post-progression survival(PPS)using Bayesian statistics in cases of advanced non-small cell lung cancer treated with atezolizumab and pembrolizumab. The findings indicate that PPS extension significantly contributed to prolonged OS for both drugs. Notably, in the pembrolizumab group, PPS demonstrated high predictive accuracy for OS, highlighting its potential as a stable metric. Additionally, ICI-associated delayed responses and pseudoprogression can affect PFS assessment, underscoring the importance of PPS evaluation in capturing the prolonged effects of ICI therapy. Future studies should further explore these relationships, potentially incorporating nonlinear models to enhance predictive accuracy.