Enabling Earlier Identification of Genomic Predictors in Drug Development
Paul Bernhardt, North Carolina State University  *Devan Mehrotra, Merck  Robin Mogg, Merck 

Keywords: genomewide association study, pharmacogenetics, pharmacogenomics, patient segmentation, personalized medicine

Pharmaceutical companies continue to seek innovative ways to explore whether a drug under development is likely to be suitable for all or only an identified subset of patients in the target population. The sooner this can be done during the clinical development process, the better it is for the company, as well as for patients, prescribers, and payers. Rapid advances in genotyping technologies have created an opportunity to make this timely exploration a reality via the use of genomewide association studies (GWAS). In this talk, we present a remarkably efficient statistical approach that enables the discovery of potentially “actionable” genomic predictors via a GWAS embedded within a routine phase II dose-ranging trial. We use an illustrative example and extensive simulations to contrast our proposed method with several alternative methods, focusing on trials with binary response endpoints. In addition, we emphasize the need to replicate any potential discoveries in phase III, and we discuss issues related to the timely development of a companion diagnostic to facilitate patient segmentation.