Progression-free survival (PFS) as a surrogate for overall survival (OS) in patients with advanced colorectal cancer: An analysis of 3159 patients randomized in 11 trials.

Abstract

The traditional endpoint for efficacy in advanced colorectal cancer is OS, but this endpoint requires prolonged follow-up and the impact of first-line therapy on OS may be confounded by the effect of second-line therapies. We investigate whether PFS can replace OS as the primary endpoint in randomized trials on patients with advanced colorectal cancer. Methods: Individual patient data were available from 11 randomized phase III trials comparing 5-fluorouracil (5FU) + folinic acid (LV) to either 5FU alone (8 trials, 1814 patients) or raltitrexed (3 trials, 1345 patients). The correlation between PFS and OS was estimated through the bivariate distribution of these endpoints over the entire time range, or using Kaplan-Meier estimates of 6-month PFS and 12- month OS within each trial. The correlation between the treatment effects on PFS and OS was estimated using the hazard ratios over the entire time range, or up until 6 months for PFS and 12 months for OS. Correlation coefficients were estimated with their 95% confidence intervals (CI). Results: Median follow-up of alive patients was 13 months, median OS 11 months, and median PFS 5.5 months. A total of 1688 (53%) patients had progressed at 6 months, and 1659 (52%) had died at 12 months. The correlation between PFS and OS was 0.64 (CI 0.62 - 0.65). The correlation between 6-month PFS and 12-month OS was 0.19 (CI -0.26 - 0.65). PFS and OS were significantly longer with 5FU + LV than with 5FU or raltitrexed alone, with hazard ratios in favor of the combination of 0.78 (CI 0.72 - 0.85, P<0.001) for PFS and 0.86 (CI 0.79 - 0.93, P<0.001) for OS. The correlation between the log hazard ratios was 0.99 (CI 0.94 - 1.03). The correlation between the log hazard ratios up until 6-month for PFS and 12 months for OS was 0.93 (CI 0.84 - 1.01), and therefore the treatment effect on PFS at 6 months reliably predicted the treatment effect on OS at 12 months. Conclusions: PFS is poorly correlated with OS in patients with advanced colorectal cancer. In contrast, the effects of treatment upon PFS is useful surrogate for OS in randomized trials on patients with advanced colorectal cancer.