All Times EDT
Keywords: companion diagnostic, CDx intended use population, Bayesian hierarchical model
Companion diagnostics (CDx) was considered a medical device to identify patients who might benefit from treatment. In CDx studies, a critical statistical question is how to estimate the efficacy in the CDx intended use population, especially when the CDx negative patients were not enrolled in clinical trials. Frequentist approaches have been widely used to estimate the efficacy in the CDx intended use population in CDx trials. In this study, the Bayesian hierarchical models were developed to assess the drug efficacy, and simulations were conducted to evaluate the model performance. In addition, general statistical considerations in CDx studies about acceptance criteria, sample size estimation, sensitivity analysis for missing values, and tissue agnostic CDx, will also be discussed.