作者: Shigeyuki Matsui , John Crowley
DOI: 10.1158/1078-0432.CCR-17-1552
关键词: Multiple comparisons problem 、 Sequential analysis 、 Clinical trial 、 Biomarker (medicine) 、 Research design 、 Internal medicine 、 Cetuximab 、 Medicine 、 Oncology 、 Statistical power 、 Phases of clinical research
摘要: Among various design approaches to phase III clinical trials with a predictive biomarker, the marker-stratified all-comers is advantageous because it allows for establishing utility of both treatment and but often criticized requiring large sample sizes, as includes marker-positive marker-negative patients. In this article, we propose simple flexible subgroup-focused that allow sequential assessment across marker-defined subgroups adaptive subgroup selection while retaining an using entire patient cohort at final analysis stage, possibly established marker-based multiple testing procedures. Numerical evaluations indicate proposed has robustness property in preserving statistical power detecting profiles effects effectively reducing number randomized patients presumably limited efficacy. contrast, traditional enrichment designs could suffer from low some possible effects. The latter also needs long study durations marker-screened We provide application SWOG S0819, trial assess role cetuximab treating non-small cell lung cancers. These approach can enhance efficiency definitive evaluation biomarker trials. Clin Cancer Res; 24(5); 994-1001. ©2017 AACR.