OBJECTIVES: Among surrogate endpoints for overall survival (OS) in oncology trials, progression-free survival (PFS) is increasingly taking the lead. Although there have been some empirical investigations on inter-dependence of OS and PFS in different tumor types, new ways to model and interpret this inter-dependence are scarce, and only limited evidence is available for metastatic renal cell carcinoma (mRCC). METHODS: We assessed the relationship between PFS (primary endpoint) and OS in 750 patients with treatment-naïve mRCC randomized 1:1 to receive sunitinib (SU) or interferon-alfa (IFN) in a pivotal phase III study, pooling data for all available patients across treatment arms. A Weibull parametric model for failure-time data was applied to the dataset. The difference between OS and PFS was used as the outcome to remove inherent dependencies between PFS and OS. By excluding PFS time from OS time we obtain a distinct measure of survival beyond PFS: post-progression survival (PPS). RESULTS: The model demonstrated that longer PFS was significantly predictive of longer PPS (P<0.001). Estimated median PPS time was linked to a particular PFS time. For example, for PFS of 20 weeks, the median PPS was 43.9 weeks (95% confidence interval [CI]: 40.1, 48.1); for PFS of 60 weeks, the median PPS was 57.9 weeks (95% CI: 50.3, 66.7). A non-parametric Kaplan-Meier approach supported these results. CONCLUSIONS: For patients with mRCC randomized to either sunitinib or IFN, a distinct and quantifiable relationship was found between PFS and PPS. This suggests that PFS can be used as a surrogate measure for OS in mRCC, although more research is needed to generalize this finding beyond this particular study. This novel statistical approach using the Weibull parametric model can enrich the interpretation and understanding of that relationship, with potential implications for clinical trial design.