CombPDX: a unified statistical framework for evaluating drug synergism in patient-derived xenografts

Abstract

AbstractMotivationAnticancer combination therapy has been developed to increase efficacy by enhancing synergy. Patient-derived xenografts (PDXs) have emerged as reliable preclinical models to develop effective treatments in translational cancer research. However, in most PDX combination experiments, PDXs are tested on single dose levels and dose-response surface methods are not applicable for testing synergism.ResultsWe propose a comprehensive statistical framework to assess joint action of drug combinations from PDX tumor growth curve data. We provide various metrics and robust statistical inference procedures that locally (at a fixed time) and globally (across time) access combination effects under classical drug interaction models. Integrating genomic and pharmacological profiles in non-small-cell lung cancer (NSCLC), we have shown the utilities of combPDX in discovering effective therapeutic combinations and relevant biological mechanisms.AvailabilityWe provide an interactive web server, combPDX (https://licaih.shinyapps.io/CombPDX/), to analyze PDX tumor growth curve data and perform power analyses.ContactMJHa@mdanderson.orgSupplementary informationSupplementary data are available at Bioinformatics online.