Assessing the Viability of a Surrogacy Relationship between Disease-Free Survival (DFS) and Overall Survival (OS) for the Patients in Clinical Trials for Adjuvant Treatment of Muscle-Invasive Urothelial Carcinoma (MIUC) Following Radical Cy ...
Author(s)
Barkema AE1, Simons C2, Kurt M3, Broughton E4, Teitsson S5, May JR5, Kroep S6
1OPEN Health Group, Rotterdam, Netherlands, 2OPEN Health Group, York, NYK, UK, 3Bristol Myers Squibb, Lawrenceville, NJ, USA, 4Bristol Myers Squibb, Princeton, NJ, USA, 5Bristol Myers Squibb, Uxbridge, UK, 6OPEN Health, Rotterdam, Netherlands
Presentation Documents
OBJECTIVES To assess the strength of the association between the treatment effects on DFS and OS after radical cystectomy in MIUC using published literature. METHODS Eight randomized controlled trials identified from the SLR, published between 1991 and 2020, were deemed eligible for inclusion in the analysis. For the sensitivity analysis the evidence base was expanded by two trials. For all but one of the included studies, treatment proportionality was verified. First, the strength of the evidence base and Pearson’s ρ between the hazard ratio of DFS (DFSHR) and OS (OSHR) in the selected chemotherapy trials were assessed. Second, the surrogacy equations were generated via weighted frequentist meta-regression, univariate Bayesian meta-regression and bivariate Bayesian meta-analysis. The accuracy of our models was tested by validating the predicted OSHR with the reported OSHR in the IMvigor-010 trial. RESULTS The estimated correlation between DFSHR and OSHR was 0.84 (95% CI: -0.18, 0.99) and 0.93 (95% CI: 0.57, 0.99) for the base case and sensitivity analyses. These are considered moderate strength correlations and do not independently prove a surrogacy relationship. Due to limited number of studies in the evidence base and their relatively low sample sizes, prediction intervals around the estimated surrogacy equations were wide. In the base case, the slope coefficient of the resulting equation for the univariate Bayesian meta-regression was found to be significant. This model overestimated the known OSHR of IMvigor-010 by 0.14. For the sensitivity analysis, the slope coefficient was found to be significant for all models, and the OSHR was overestimated with errors in the range of [0.14, 0.21]. CONCLUSIONS OSHR predictions from the surrogacy equations using DFSHR for adjuvant treatment of MIUC should be interpreted with a high degree of caution due to the limited evidence base and moderate association between DFSHR and OSHR in this study.
Conference/Value in Health Info
2021-11, ISPOR Europe 2021, Copenhagen, Denmark
Value in Health, Volume 24, Issue 12, S2 (December 2021)
Code
POSB302
Topic
Clinical Outcomes
Topic Subcategory
Relating Intermediate to Long-term Outcomes
Disease
Drugs, Oncology