On Enrichment Strategies for Biomarker Stratified Clinical Trials.

摘要: In the era of precision medicine, drugs are increasingly developed to target subgroups patients with certain biomarkers. large all-comer trials using a biomarker stratified design, cost treating and following for clinical outcomes may be prohibitive. With fixed number randomized patients, efficiency testing treatments parameters, including treatment effect among biomarker-positive interaction between biomarker, can improved by increasing proportion positives on study, especially when prevalence rate is low in underlying patient population. When assessing true prohibitive, one further improve study oversampling cheaper auxiliary variable or surrogate that correlates biomarker. To reduce cost, we adopt an enrichment strategy both scenarios concentrating contain more information about specific parameters primary interest investigators. first scenario, enriched design enriches cohort directly relevant while second auxiliary-variable-enriched based inexpensive variable, thereby avoiding all screened reducing waiting time. For designs, discuss how choose optimal single hypothesis two hypotheses simultaneously. At requisite power, compare new designs BSD terms trial under mimicking real trials. The illustrated hypothetical examples designing biomarker-driven cancer