Optimality Considerations in Two‐Stage Biomarker‐Stratified Trials With Adaptive Enrichment

ABSTRACT We consider biomarker‐stratified trials that test the treatment effect in a targeted patient group, the biomarker positive subgroup, and in the overall population. If the biomarker negative subgroup is stopped for futility during the interim analysis the number of participants assigned can be increased. Optimal two‐stage designs are obtained using several optimality criteria. Recommendations are given on the best interim analysis timing, futility boundaries, α allocation between the two hypotheses and stage 2 sample size for the biomarker positive subgroup in case its complement, the biomarker negative subgroup, is stopped for futility in the interim analysis. This work is motivated by PrecISE, a two‐stage biomarker‐stratified clinical trial in patients with severe asthma.