Disease-free survival (DFS) vs. overall survival (OS) as a primary endpoint for adjuvant colon cancer studies: Individual patient data from 12,915 patients on 15 randomized trials.

3502 Background: The traditional endpoint for colon adjuvant clinical trials (CT) is OS, with 5 years (yrs) used to demonstrate adequacy of follow-up (f-u). A shorter-term endpoint providing both clinically and statistically convincing evidence to allow a comparison of treatments could significantly speed both the CT process and the translation of advances into clinical practice. METHODS Individual patient data were pooled from 15 large randomized phase III colon adjuvant CT. Trials included 33 arms: 9 no chemotherapy control arms, 24 5-FU based chemotherapy arms, with a pooled sample size of 12,915 patients (pts). The primary hypothesis was that DFS, with 3 yrs f-u, is an appropriate primary endpoint to replace OS with 5 yrs f-u in phase III colon adjuvant CT. Median f-u was 8 yrs; 5 yr f-u was available on 93% of patients (pts). RESULTS The recurrence rate for yrs 1-5 was 11%, 13%, 7%, 4%, and 3%. Median time from recurrence to death was 14 months. 95% of pts alive and disease-free at 3 yr were alive at 5 yrs; 86% of pts with recurrence by 3 yrs died before 5 yrs. Overall per-pt concordance between 2, 3, and 4 yr DFS and 5 yr OS was 88%, 92%, and 92% respectively. Across all 33 study arms, correlation between 3 yr DFS and 5 yr OS was 0.94. Correlation of the difference in 3 yr DFS and 5 yr OS comparing control vs experimental arms within each trial was 0.88. Within trial log-rank testing using DFS with 3 yrs of f-u versus OS with 5 yrs f-u resulted in the same conclusion in 15 of 18 cases (83%); all three discordant results were only marginally significant for DFS (0.03 < p < 0.05). Differences in 5 yr OS retained 92% of the difference in 3 yr DFS (95% CI 67%-117%). CONCLUSIONS In pts treated on phase III adjuvant colon CT, 3 yr DFS and 5 yr OS are highly correlated, both within pts and across trials. Within-trial comparisons of 3 yr DFS predict 5 yr OS comparisons with high concordance. These results suggest that DFS after 3 years f-u is an appropriate endpoint for adjuvant colon CTs of 5-FU-based regimens, although marginally significant 3 yr DFS improvements may not translate into significant 5 yr OS benefits. No significant financial relationships to disclose.