Background: 3Y DFS is a validated SE for 5Y OS in adjuvant chemotherapy (AC) CC trials, while no SE have been established for radiation (RT) in RC trials. The objective of this study was to validate 3Y DFS as a SE for 5Y OS in RC and CC. Methods: Patients diagnosed with stage I-III RC and CC referred to any 1 of 5 regional cancer centers in British Columbia between 1990 and 2009 were followed by the GI Cancer Outcomes Unit. Clinical, pathological, and treatment characteristics were prospectively collected, with DFS and OS measured from the date of definitive surgery until relapse or death. RC patients who received RT and CC patients who received AC were propensity-score matched to those who did not, within each cancer center, using an absolute standard difference of ≤0.2. Individual level associations between DFS and 5Y OS were calculated using Spearman’s rho . Propensity-score adjusted treatment effects, by center were calculated using R². Results: Of 3153 patients with RC, 2564 received RT and 589 did not. Of 4638 patients with CC 2563 received AC and 2075 did not. Patients with AC and RT were younger and more likely to have grade 3, lymphovascular and perineural invasion, and stage III disease. RC patients also had higher CEA levels and received more CT. After matching, most patients at the extremes of age and with stage I-II disease were removed, resulting in 1544 RC and 2226 CC patients with no significant differences between treatment groups. There was a strong correlation between 2, 3 and 4Y DFS and 5Y OS by Spearman’s rho and R². Crude recurrence rates for both RC and CC were similar, peaking in the first year and declining to zero after eight years. Conclusions: These data support 3 year DFS as a SE for 5Y OS, for RT in RC and AC in CC. Correlations remained significant at both individual and across-centers for both RC and CC. The recurrence rates for both RC and CC are similar.