Keywords: basket, biomarker, Bayesian, subsets
Basket designs have become popular in oncology for evaluating a treatment which targets the same genomically deregulated gene or pathway in each of several types of cancer. By borrowing information among cancer type strata, basket designs can achieve a greater efficiency than performing an independent phase 2 trial for each cancer type. Basket designs can be of use in fields other than oncology for evaluating a treatment among related patient populations or in assessing multiple candidate predictive biomarkers.
I will review the several of the Bayesian Basket designs which have been developed. I will discuss some of the large basket studies that have been reported, extension of the design to randomized baskets and the role of the basket design in the regulatory pathway.