Abstract:
|
Basket trials evaluate a single drug targeting a single biomarker mutation in multiple cancer cohorts. Empirical findings from published studies suggest that we can expect heterogeneity in efficacy across baskets. Nowadays, most of the novel approaches for basket design use Bayesian methods. These methods require a prior specification of at least one parameter that permits sharing of information across baskets. In this research, we endeavor to provide recommendations on selecting a prior for the scale parameter in an adaptive basket trial design using Bayesian hierarchical modeling, and address the essential of the Bayesian hierarchical approach, the exchangeability among hyperparameters due to the possibility of the high heterogeneity across baskets. Thus, we allowed each stratum-specific parameter to be exchangeable with other similar strata parameters or nonexchangeable with any of them through the observed data during a sequence of interim analysis. In simulation study, we evaluate the overall performance of our design based on statistical power and false-positive rates. We believe that our research is contributed to understand some properties of Bayesian basket design.
|