Statistics in Clinical Cancer Research Evaluation of Progression-Free Survival as a Surrogate Endpoint for Survival in Chemotherapy and Targeted Agent Metastatic Colorectal Cancer Trials

Pooled analyses of chemotherapy trials in metastatic colorectal cancer (mCRC) have suggested that progression-free survival (PFS) is a surrogate endpoint for overall survival (OS). However, this has not been evaluated under current standard-of-care regimens of chemotherapy in combination with targeted therapies. We conducted an analysis of published mCRC trials of chemotherapy and targeted therapies from 2000 to evaluate the surrogacy of PFS and response rate (RR) for OS. Study-level data was pooled from 24 randomized mCRC trials that evaluated fluoropyrimidine-based regimens and included trials conducted with targeted agents (panitumumab, cetuximab, bevacizumab, andaflibercept). A total of 69 treatment arms with a sample size of 20,438 patients was included. Linear regression analysis was carried out to estimate the correlation of PFS and RRwithOS. The correlation coefficient between PFSHRs andOSHRswas 0.86 for all trials, 0.89 for 12 phase III trials of targeted agents in combination with chemotherapy, 0.95 for 8 first-line phase III trials of targeted agents, and 0.83 for 9 trials of anti-EGFR-targeted agents. In all cases, correlation coefficients between RR and OS HRs were lower than those between PFS HRs and OS HRs (range, 0.42– 0.81). In this study-level analysis of randomized mCRC trials of chemotherapy and targeted agents, improvements in PFS are strongly correlated with improvements in OS. This suggests that PFS remains a valid surrogate endpoint for OS with current treatment regimens in the mCRC setting. Clin Cancer Res; 19(5); 969–76. 2012 AACR.