Correlations between disease-free survival (DFS) and overall survival (OS) in patients (pts) with renal cell carcinoma (RCC) at high risk for recurrence: Results from S-TRAC trial

Background: S-TRAC, a prospective phase 3 randomized trial in pts with resected RCC at high risk for recurrence, demonstrated a statistically significant improvement in DFS with sunitinib (SU) vs placebo; HR 0.76; 95% CI: 0.59, 0.98; P = 0.03. We hypothesized that DFS is a surrogate for OS and evaluated the association of DFS events with OS. Methods: DFS (by blinded independent review) was defined from randomization to first evidence of recurrence, second primary malignancy, or death, whichever occurred first. Pts were categorized as having DFS or censored ≤2 years vs DFS >2 years, and as having OS or censored ≤5 years vs OS > 5 years. The odds ratio (OR), positive predictive values (PPV) and negative PV (NPV) were reported. To determine if DFS can be a surrogate for OS, 2 conditions need to be tested: 1) DFS and OS are strongly correlated, and 2) the treatment effect on DFS is sufficiently correlated with the effect on OS. The validity of the surrogate is reflected by the strength of these correlations. The correlation between DFS and OS is tested at the pt level through copula models to estimate the correlation coefficient (Kendall τ). Trial level correlations have been explored considering region as the trial unit. Results: Median follow-up for OS was approximately 6.5 years with 141 (23%) observed deaths. There were 257 (42%) DFS events. Of 261 pts with DFS or censored ≤2 years of enrollment, 97 were alive and in follow-up >5 years (37%). In 354 pts with DFS >2 years, 318 pts were alive >5 years (90%). The OR, PPV and NPV were 14.9, 0.9 and 0.63, respectively. Kendall’s τ ranged from 0.51 to 0.88 using the Hougaard, Clayton, and Plackett copula, suggesting a moderate correlation at the individual pt level. Similar results were observed with investigator assessed DFS. Analyses of trial level correlations also suggest a moderate correlation. Further analyses are being explored. Conclusions: A moderate correlation between DFS and OS was observed in S-TRAC despite immature OS data. Additional analyses across completed trials are warranted to further assess the relationship between DFS and OS. Clinical trial identification: NCT00375674. Editorial acknowledgement: Editorial support was provided by Vardit Dror, PhD, of Engage Scientific Solutions, and funded by Pfizer. Legal entity responsible for the study: Pfizer. Funding: Pfizer. Disclosure: D.J. George: Honoraria and consulting: Sanofi, Exelixis, Bayer; Consulting: Merck, Sanofi; Grants: Genentech/Roche, Novartis, Astellas, Celldex, Acerta; Grants & consulting: Exelixis, Janssen, Pfizer, Innocrin Pharma, BMS. A.J. Pantuck: Consulting fees: Pfizer. R. Figlin: Research funding: Peloton, BMS, Argos Therapeutics, Exilexis, and Cerulean; Consulting fees: Novartis, Pfizer, Nektar, Peloton, Calithera, Acceleron. B. Escudier: Consulting fees: Bayer, Pfizer, Novartis; Honoraria: Bayer, Roche, Pfizer, Genentech, Novartis, Aveo. S. Halabi: Data and Safety Monitoring Board member: Eisai. M. Casey, X. Lin, L. Serfass, M.J. Lechuga Frean: Employee of and own stock in Pfizer. A. Ravaud: Advisory boards: Pfizer, Novartis, GSK, Roche, BMS; Institutional support grants: Pfizer, Novartis; Housing and transportation for meetings and speeches: Pfizer, Novartis, BMS.