A New Bayesian Dose-Finding Design for Drug Combination Trials

We consider the problem of finding maximum tolerated dose with targeted dose-limiting toxicity rate in drug combination trials. We propose a novel Bayesian adaptive design, which features adaptive local modeling and weighted learning along the search path. The method is robust in the sense that neither prespecifications of marginal toxicity probabilities nor subjective priors for model parameters are required. Extensive simulation studies show that the proposed method is comparable to some leading methods in terms of correct selection and superior to them when the priors are misspecified. Moreover, the proposed model can be extended to the case of trials with more than two drugs. Supplementary materials for this article are available online.