Keywords: basket trials; targeted therapy; information sharing
Basket trials are clinical trials where we would like to test a new therapy across many disease types simultaneously where all patients have a common mutation or a particular genomic alteration. This new class of clinical trials where the drug is tested simultaneously across the different subgroups of different tumor types (or baskets) has raised numerous considerations and implications for clinical trial design. In this trial setting, we need to answer multiple questions simultaneously including “Does the drug work at all?” and “Does the response/outcome differ between disease types?” The activity of the drug may have a similar response rates across all disease types, or the response rates may be similar in some disease types and not so similar in others. In the basket trial design, it would be advantageous to be able to share or borrow information across the subgroups or baskets. In this talk, we will delineate the definitions of basket, umbrella and platform protocols, clarify the different objectives and clinical settings and describe how efficiencies can be obtained by quantifying various statistical design parameters such as sample size, trial duration and type I and type II errors. This is joint work with Colin Begg, Mithat Gonen, Ronglai Shen and Kristen Cunanan.