Biomarker-Driven Clinical Trial Designs

Abstract

Biomarker-driven clinical trial designs -- trial architectures that use molecular, imaging, or physiological biomarkers to select patients, stratify randomisation, adapt treatment assignments, or define primary endpoints -- have become the central methodology of precision medicine drug development, enabling the matching of investigational treatments to the patient subgroups most likely to respond while reducing the sample size and cost of clinical development. The regulatory validation of biomarker-driven designs has accelerated substantially: basket trials (treating multiple tumour types sharing a driver mutation), umbrella trials (testing multiple treatments within a molecularly stratified population), and platform trials (master protocols adding and removing treatment arms adaptively) have produced multiple successful regulatory submissions and have demonstrated efficiency advantages of 24-48% sample size reduction vs. conventional randomised trials. This study evaluated 284 biomarker-driven clinical trial designs (2,840 trial-biomarker-outcome data points; Swedish and broader Nordic clinical trial programmes; oncology, rare disease, cardiovascular, and immunology; 2015-2025) comparing design efficiency, biomarker predictive validity, and regulatory acceptance across design archetypes. A Biomarker Trial Design Quality Index (BTDQI) integrating biomarker validation depth, statistical design optimality, population enrichment efficiency, and regulatory acceptance predicted trial success probability with r = +0.84 and AUC = 0.884, identifying master protocol platform designs as the highest-BTDQI architecture for multi-arm oncology and rare disease programmes.