|
Abstract:
|
In early phase oncology drug development, single arm expansion cohorts are increasingly being used as proof-of-concept (POC) studies to drive the decisions for future development of the drug. Decision-makings based on such POC studies with limited data and early efficacy endpoints are extremely challenging. As a result, it is crucial to take into account the prior knowledge of the drug and uncertainties of the observed data. To this end, we have proposed a Bayesian approach to facilitate the decision-making process. With pre-specified Bayesian decision criteria, decisions regarding the future of the drug can be made in terms of terminating the drug, accelerating the drug or initiating further investigations. The proposed decision framework can be used to extend the design of the study to a two-stage design by incorporating interim analyses that allow for early termination. We illustrate the Bayesian approach through examples and compare the properties of the proposed design with that from the well-established frequentist approaches based designs, such as Simon’s and Fleming’s two stage designs, through simulations.
|
