|
Abstract:
|
In order to accelerate clinical development, seamless 2/3 adaptive design is a popular strategy to combine phase 2 dose selection with phase 3 confirmatory objectives. For therapeutic area with high unmet medical need such as oncology, this is especially relevant. As the regulatory requirement for phase 3 optimal dose shifted from MTD to MED for oncology drug, it’s increasingly important to gather more data on multiple candidate doses to inform phase 3 dose selection. Using the seamless 2/3 design, phase 3 dose may be selected based on phase 2 portion and carried forward for the phase 3 portion of the design. For final inference, data from the phase 2 portion and phase 3 portion will be combined. For many oncology trials, the phase 2 decision is based on early biomarkers correlated with the phase 3 clinical efficacy endpoint. As discussed in Li et al. (2015) this will create type I error inflation when the phase 2 biomarker are positively correlated with the phase 3 efficacy endpoint. Sidak adjustment may be used to adjust phase 2 p-values for the combined p-value test to control the overall type I error rate. However,
|
