Adaptive enrichment with subpopulation selection at interim: methodologies, applications and design considerations

Abstract

There is a growing interest in pursuing adaptive enrichment for drug development because of its potential to achieve the goal of personalized medicine. There are many versions of adaptive enrichment proposed across many disease indications. Some are exploratory adaptive enrichment and others aim at confirmatory adaptive enrichment. In this paper, we give a brief overview on adaptive enrichment and the methodologies that are growing in statistical literature. A case example is provided to illustrate a regulatory experience that led to drug approval. There were two design elements used for adaptation in this case example: population adaptation and statistical information adaptation. We articulate the challenges in the implementation of a confirmatory adaptive enrichment trial. The challenges include logistic aspects on the appropriate choice of study population for adaptation and the ability to follow the pre-specified rules for statistical information or sample size adaptation. We assess the consistency of treatment effect before and after adaptation using the approach laid out in Wang et al. (2013). We provide the rationales for what would be an appropriate treatment effect estimate for reporting in the drug label. We discuss and articulate design considerations for adaptive enrichment among a dual-composite null hypothesis, a flexible dual-independent null hypothesis and a rigorous dual-independent null hypothesis.

Keywords: Biomarker; Dual-composite null hypothesis; Flexible dual-independent null hypothesis; Pharmacogenomics; Rigorous dual-independent null hypothesis.