ABSTRACT
Often, in Proof-of-Concept (PoC) studies, a statistician is asked to fill out the operational characteristics of several design options for biomarker subgroups and the overall population. In addition to the cutoff for the biomarker positivity, the chapter shows that the prevalence of the marker-positive population is known at the time of designing the clinical trial. While it is generally assumed that treatment effect in BM+ group is higher, clear evidence does not exist that the agent is not active in the biomarker-negative subgroup. Hence, the justification for doing a confirmatory trial in the marker-positive subgroup needs to be established. Alternatively, the PoC study will guide us to do the confirmatory trial in all subjects. The BM+ subgroup power curves are on the right; the overall population curves are on the left; for each population, the different curves account for varying number of events.
