Matthew Psioda, PhD, GSK
Title: Bayesian Adaptive Basket Trial Design Using Model Averaging.
础产蝉迟谤补肠迟:听Matt Psioda is Head of Statistical Innovation for Oncology and Vaccines within GSK鈥檚 Statistics and Data Science Innovation Hub. In that role, he leads a small team of statistical consultants to support use of innovative study designs and advanced statistical methods in GSK studies. His group works on a variety of applied and methodological research problems. Examples include extrapolating information on treatment effectiveness from adult to adolescent/pediatric settings, design and analysis of clinical trials with hybrid or external control arms, and design and analysis of adaptive basket and/or platform trials. Prior to joining GSK, most recently Matt was on the faculty in the Department of Biostatistics at the University of North Carolina at Chapel Hill and was a Statistical Advisor to the United States Food and Drug Administration鈥檚 Center for Drug Evaluation and Research.
We discuss a Bayesian adaptive design methodology for oncology basket trials with binary endpoints using a Bayesian model averaging framework. Most existing methods seek to borrow information based on the degree of homogeneity of estimated response rates across all baskets. In reality, an investigational product may only demonstrate activity for a subset of baskets, and the degree of activity may vary across the subset. A key benefit of our Bayesian model averaging approach is that it explicitly accounts for the possibility that any subset of baskets may have similar activity and that some may not. Our proposed approach performs inference on the basket-specific response rates by averaging over the complete model space for the response rates, which can include thousands of models. We present results that demonstrate that this computationally feasible Bayesian approach performs favorably compared to existing state-of-the-art approaches, even when held to stringent requirements regarding false positive rates.
听