The ARPA-H ADvanced Analysis for Precision cancer Therapy (ADAPT) program invests $28M to launch a first-of-its-kind, 500-participant metastatic breast cancer trial led by UNC Lineberger and the Translational Breast Cancer Research Consortium. Dr. Rashid co-leads the statistical innovation arm, delivering a Bayesian adaptive design that learns from longitudinal ctDNA, imaging, and clinical endpoints as the study progresses.

Key elements of the platform:

• Adaptive learning phases. Three coordinated stages—warm-up, active learning, and confirmatory testing—allow biomarker hypotheses to be discovered, validated, and deployed without pausing enrollment.
• Posterior-guided decisions. Predictive probabilities drive interim treatment adaptations, arm graduation, and dynamic cohort expansion, ensuring patients access the regimens most likely to benefit them.
• Embedded real-time analytics. Serial multi-omic profiling is paired with cloud-based dashboards that feed machine-learning risk models for the care team.

The program is a flagship collaboration between Gillings biostatisticians, Lineberger clinicians, and national partners at Dana-Farber, Duke, Johns Hopkins, and MD Anderson. Methods emerging from ADAPT are open-sourced through adapt-bayes, with preprints, JCGS submissions, and ASA presentations available to consortia partners.