All Times EDT
Keywords: Surrogate, bayesian, prior, interim analysis
In randomized-controlled trials, interim analyses are often planned for possible early trial termination to claim superiority or futility of a new therapy. Surrogate endpoint enables interim analyses to be conducted earlier, allowing to speed up the decision-making process during the conduct of the trial. We propose a Bayesian methodology for interim analysis predicting the success of the trial from both surrogate and primary endpoints in a way that makes the best use of all the available evidence. The interim results from the surrogate endpoint and the correlation between surrogate and primary endpoint established from historical data, are used to derive an informative prior for the primary endpoint, called surrogate prior. We investigate the operating characteristics of the methodology through a comprehensive simulation plan, focusing on an application to a phase III clinical trial in metastatic colorectal cancer, and considering potential prior data conflicts between the surrogate prior and the available evidence on the primary endpoint.