Analyzing Basket Trials under Multisource Exchangeability Assumptions

Abstract

Basket designs are prospective clinical trials that are devised with the hypothesis that the presence of selected molecular features determine a patient's subsequent response to a particular "targeted" treatment strategy. Basket trials are designed to enroll multiple clinical subpopulations to which it is assumed that the therapy in question offers beneficial efficacy in the presence of the targeted molecular profile. The treatment, however, may not offer acceptable efficacy to all subpopulations enrolled. Moreover, for rare disease settings, such as oncology wherein these trials have become popular, marginal measures of statistical evidence are difficult to interpret for sparsely enrolled subpopulations. Consequently, basket trials pose challenges to the traditional paradigm for trial design, which assumes inter-patient exchangeability. The R-package basket facilitates the analysis of basket trials by implementing multi-source exchangeability models. By evaluating all possible pairwise exchangeability relationships, this hierarchical modeling framework facilitates Bayesian posterior shrinkage among a collection of discrete and pre-specified subpopulations. Analysis functions are provided to implement posterior inference of the response rates and all possible exchangeability relationships between subpopulations. In addition, the package can identify "poolable" subsets of and report their response characteristics. The functionality of the package is demonstrated using data from an oncology study with subpopulations defined by tumor histology.

Bayesian local exchangeability design for phase II basket trials

“…Previous works in MEM assume information sharing can occur between any two baskets and the amount of information sharing is proportional to the pairwise similarity between baskets. 14,[19][20][21] We refer to these approaches as "global-MEM." Under global-MEM, the posterior distribution for the response probability of basket b is…”

A comparison of priors for variance parameters in Bayesian basket trials

“…However straightforward, collections of separate studies do not account for any possible similarities in the response between tumors. This approach is known to alleviate possible bias, but at the same time may lead to loss of power especially in low sample size cases, which are frequent for cancer molecular characterization (Kane et al, 2019) In this view, Thall et al (2003) first proposed a Bayesian hierarchical modelling (BHM) approach for a phase II sarcoma trial with multiple subtypes, each corresponding to one arm, which allows borrowing strength of information (i.e., pooling data) between the different arms. This approach introduces a set of random effects to capture arm-specific drug responses and model them as independent random variables following a Gaussian distribution with mean µ and standard deviation σ.…”