All Times EDT
Keywords: dose-finding, drug combination, dose-response, dose interpolation
Combination of multiple investigational drugs is always a promising direction in the development of anti-tumor treatment. The design of dose-finding study for drug combination is therefore worth to discuss. Such design are usually determined by three elements: the dose-response curve model, the next-dose-finding algorithm, and the stopping rule. Comparing to the case of a single drug, the first two elements can be quite different in a drug combination setting, especially when efficacy response, in addition to toxicity response, are taken in to consideration for choosing RP2D. In this poster we summarized three most novel dose-finding designs for drug combination from literature, including parametric method, non-parametric method, and adaptive modeling method. Further more, in the case of all pre-specified doses are found not ideal during the study, we proposed a modification to the new dose interpolation algorithm described in literature. Comparing to the original method in literature, our modified algorithm is operationally practical, yet provide comparable dosing-finding accuracy in multiple scenarios.