Mapping the incidentalome: estimating incidental findings generated through clinical pharmacogenomics testing.

Abstract

Greater clinical validity and economic feasibility are driving the more widespread use of multiplex genetic technologies in routine clinical care, especially for applications in pharmacogenomics. Empirical data on the numbers and types of incidental findings generated through such testing are needed to develop policies and practices related to their clinical use. Of particular importance are disparities in findings relevant to different ancestry groups. The Pharmacogenomic Resource for Enhanced Decisions in Care and Treatment Resource, or PREDICT, is an institutional program to implement prospective clinical genotyping of 34 pharmacogenomic-related genes to guide drug selection and dosing. We curated 5,566 journal articles to quantify and characterize the incidental, non-pharmacogenomic genotype-phenotype associations that could be generated through this clinical genotyping project. We identified 372 putative incidental genotype-phenotype associations that might be revealed in patients undergoing clinical genotyping for pharmacogenomic purposes. Of these, 287 associations were supported by at least one study demonstrating an odds ratio ≥2.0 or ≤0.5. Numbers of potentially relevant findings varied widely by ancestry group. Rigorous clinical policies for the clinical management of incidental findings are needed because the sheer number of significant findings could prove overwhelming to health-care institutions, providers, and patients.