BackgroundThe administration of second-line or subsequent immune checkpoint inhibitors (ICIs) in previously treated patients with advanced solid cancers has been clinically investigated. However, previous clinical trials lacked an appropriate primary endpoint for efficacy assessment. This systematic review aimed to explore the most optimal early efficacy endpoint for such trials.MethodsPhase 2 or 3 clinical trials involving patients with advanced solid cancers with disease progression following standard first-line therapy receiving second-line or subsequent ICI administration, with adequate survival outcome data, were included from PubMed, Embase, Web of Science, and Cochrane Library databases before February 2023. Quality assessment was conducted using the Cochrane tool and Newcastle-Ottawa Quality Assessment Scale for Cohort Studies for randomized controlled trials (RCTs) and non-randomized trials, respectively. Objective response rate (ORR) and progression-free survival (PFS) at 3, 6, and 9 months were investigated as potential early efficacy endpoint candidates for 12-month overall survival (OS), with a strong correlation defined as Pearson's correlation coefficient r >= 0.8.ResultsA total of 64 RCTs comprising 22,725 patients and 106 non-randomized prospective trials involving 10,608 participants were eligible for modeling and external validation, respectively. RCTs examined 15 different cancer types, predominantly non-small-cell lung cancer (NSCLC) (17, 28%), melanoma (9, 14%), and esophageal squamous cell carcinoma (5, 8%). The median sample size of RCTs was 124 patients, and the median follow-up time was 3.2-57.7 months. The ORR (r = 0.38; 95% confidence interval [CI], 0.18-0.54) and PFS (r = 0.42; 95% CI, 0.14-0.64) exhibited weak trial-level correlations with OS. Within ICI treatment arms, the r values of ORR and 3-, 6-, and 9-month PFS with 12-month OS were 0.61 (95% CI, 0.37-0.79), 0.78 (95% CI, 0.62-0.88), 0.84 (95% CI, 0.77-0.90), and 0.86 (95% CI, 0.79-0.90), respectively. External validation of 6-month PFS indicated an acceptable discrepancy between actual and predicted 12-month OS.ConclusionsIn non-randomized phase 2 trials on second-line or subsequent ICI therapy in patients with advanced solid cancers, 6-month PFS could serve as an early efficacy endpoint. However, early efficacy endpoints are not recommended in RCTs to replace OS.