End points for colon cancer adjuvant trials: observations and recommendations based on individual patient data from 20,898 patients enrolled onto 18 randomized trials from the ACCENT Group.

PURPOSE: The traditional end point for colon adjuvant clinical trials is overall survival (OS). We previously validated disease-free survival (DFS) after 3-year follow-up as an excellent predictor of 5-year OS results. Here we explore shorter term DFS and OS end points, as well as stage dependency. METHODS: Individual patient data from 18 phase III trials including 43 arms and 20,898 patients were pooled. Association measures included correlation of event rates within arms, correlation of hazard ratios (HRs) between arms, trial level significance comparisons (via log-rank testing), and a formal surrogacy model. RESULTS: DFS at earlier times was less accurate in predicting OS than 3-year DFS, but 2-year DFS remained a strong predictor. DFS with 1-year minimum follow-up demonstrated perfect negative predicted value; all trials negative at 1 year for DFS were negative for 5-year OS. OS with 3-year minimum follow-up was also an excellent predictor for 5-year OS; OS at earlier time points provided inaccurate prediction. The association between 3-year DFS and 5-year OS was greater for stage III patients; correlation of HR within trials was 0.92 (95% CI, 0.85 to 0.95) for stage III patients and 0.70 (95% CI, 0.44 to 0.80) for stage II patients. CONCLUSION: DFS outcomes after 2- or 3-year median follow-up are excellent predictors of 5-year OS. DFS outcomes are appropriate for trials in which the majority of patients are stage III. DFS after 2- or 3-year median follow-up should be considered as the primary end point in future colon adjuvant trials.