Early completion based on multiple dosages to accelerate maximum tolerated dose-finding
share
Background: Phase I trials desire to identify the maximum tolerated dose (MTD) early and proceed quickly to an expansion cohort or phase II trial for efficacy. We propose an early completion method based on multiple dosages to accelerate the identification of MTD for model-assisted designs. Methods: The early completion performs based on each dose-assignment probability for multiple dosages. The formulas of probabilities are simple to calculate. We evaluated the early completion for an actual trial. In addition, a simulation study performed the percentage of correct MTD selection and early completion. Results: In the actual trial, the early completion could perform, and the MTD estimation phase could be shortened. In the simulation study, the change in the percentage of correct MTD selection from non-early completion version was 2.2 at the maximum, indicating that there was almost no decrease. In addition, depending on the toxicity rate, the percentage of correct MTD selection improved compared to the non-early completion version. The early completion performed from 18.1 method with not only little reduction in accuracy but also improving the percentage of correct MTD selection. We conclude the early completion can perform unproblematically for model-assisted design.
READ FULL TEXT