Statistical analysis for two-stage seamless design with different study endpoints

In the pharmaceutical industry, it is desirable to apply an adaptive seamless trial design to combine two separate clinical studies that are normally conducted for achieving separate objectives such as a Phase II study for dose finding and a Phase III confirmatory study for efficacy. As a result, an adaptive seamless Phase II/III trial design consisting of two phases, namely a learning phase and a confirmatory phase, is commonly considered in pharmaceutical development. In some cases, however, the study endpoints for the two separate studies may be different due to long treatment duration. In this case, test statistics for the final analysis based on the combined data are necessary developed. In this paper, a test statistic utilizing data collected from both phases is proposed assuming that there is a well established relationship between the two different study endpoints. Formula for sample size calculation based on the proposed test statistic is derived. Sample size allocation at the two phases is also discussed.