Surrogate end points and postprogression survival in renal cell carcinoma: an analysis of first-line trials with...

Our end point was to determine the correlations between progression-free survival (PFS), postprogression survival (PPS), response rate (RR), and disease control rate (DCR) (RR + stable disease) and overall survival (OS) in first-line trials of renal cell carcinoma (RCC) treated with targeted therapies and to identify a potential surrogate for OS. Data were collected from first-line phase III randomized trials in RCC. Linear regression was undertaken to evaluate the correlations between end points and a potential surrogate end point for OS. Six randomized trials were identified containing a total of 7 treatment arms. The nonparametric Spearman rank correlation coefficients (r(s)) between PFS, PPS, and RR/DCR and OS are 0.869, and 1, 0.96/1 (all P < .0001), respectively. There is a strong relationship between differences (Δ) in DCR and ΔOS (r(s) = 1). The slope of the regression line is 0.3963 ± 0.0019, indicating that a novel drug producing a 10% increase for DCR will yield an estimated absolute 3.9% increase in OS. In first-line trials including novel targeted agents for RCC, PFS is a relatively flawed surrogate end point because of PPS influence. Improvement in DCR is strongly associated with improvement in median OS. In this population, DCR may be an appropriate surrogate for OS.