Bayesian Hierarchical Modeling and Biomarker Cutoff Identification in Basket Trials
Guosheng Yin, Zhao Yang, Motoi Odani, Satoru Fukimbara
Abstract
Patients’ heterogeneity poses a fundamental problem in the rapidly developing field of precision medicine. Based on a prespecified cutoff, biomarker-based designs provide a flexible approach to selecting a subset of biomarker-positive patients who are most likely to benefit from the new therapeutics. However, a natural question is how to determine the biomarker cutoff that distinguishes biomarker-positive patients from the negatives, and then evaluate the efficacy of the new therapeutics in one trial. We propose a Phase II basket biomarker cutoff (BBC) design where a biomarker for identifying the sensitive patients is measured on a continuous scale. The proposed BBC design incorporates the biomarker cutoff identification procedure into a basket trial via Bayesian hierarchical modeling. We verify its feasibility and practicability via real trial examples, extensive simulation studies, and sensitivity analyses. The simulation studies show that the BBC design can select biomarker-positive patients accurately and may exhibit competitive improvement in regards to the overall Type I error, power, and average sample number.