Surrogate end points for median overall survival in metastatic colorectal cancer: literature-based analysis from 39 randomized controlled trials of first-line chemotherapy.

Abstract

Our aims were to determine the correlations between progression-free survival (PFS), time to progression (TTP), and response rate (RR) with overall survival (OS) in the first-line treatment of metastatic colorectal cancer (MCRC), and to identify a potential surrogate for OS. Randomized trials of first-line chemotherapy in MCRC were identified, and statistical analyses were undertaken to evaluate the correlations between the end points. Thirty-nine randomized controlled trials were identified containing a total of 87 treatment arms. Among trials, the nonparametric Spearman rank correlation coefficient (r(s)) between differences (Delta) in surrogate end points (DeltaPFS, DeltaTTP, and DeltaRR) and DeltaOS were 0.74 (95% CI, 0.47 to 0.88), 0.52 (95% CI, 0.004 to 0.81), 0.39 (95% CI, 0.08 to 0.63), respectively. The r(s) for DeltaPFS was not significantly different from the r(s) DeltaTTP (P = .28). Linear regression analysis was performed using hazard ratios for PFS and OS. There was a strong relationship between hazard ratios for PFS and OS; the slope of the regression line was 0.54 +/- 0.10, indicating that a novel therapy producing a 10% risk reduction for PFS will yield an estimated 5.4% +/- 1% risk reduction for OS. In first-line chemotherapy trials for MCRC, improvements in PFS are strongly associated with improvements in OS. In this patient population, PFS may be an appropriate surrogate for OS. As a clinical end point, PFS offers increased statistical power at a given time of analysis and a significant lead time advantage compared with OS.