Statistical Methods for Biomarker Informed Personalized Medicine: Discovery and Translation

Abstract

Identifying biomarkers is key to the success of personalized medicine. The recent progress in high throughput genotyping and genomics technologies has enabled rapid discovery of biomarkers that could potentially be used for informing clinical decision making. Towards this direction, we developed summary statistics-based mixed effects score test statistics (sMiST) for testing the association of multiple genetically predicted mediators simultaneously and direct association of individual variants independent of mediators by using a random effects model. Extensive simulation and real data analyses demonstrate that sMiST recovers the results of MiST that is based on individual level data, but is computationally much faster. We applied our approach to a genome-wide association study of colorectal cancer and gene expression and identified several novel and secondary genetic loci.Medical decision making is often complex in that a treatment that improves clinical efficacy may also incur more medical costs, compared to standard care. In cost effectiveness analysis, incremental cost effectiveness ratio (ICER) is an important metric that measures the trade-off between the costs and health benefits of two medical programs. Individualized treatment regimes (ITRs) take into account patient heterogeneity and thus different ITRs may exhibit different health benefits and costs. Identifying a promising ITR that takes into account both efficacy and cost would be of great interest in practice. We construct ITRs that optimize the ICER, where we adopt Dinkelbach’s algorithm to translate a fractional program into an equivalent parametric program that is easy to handle. We conduct extensive simulation studies to show satisfactory performances of our methods, compared to ITRs that only optimizes one single outcome (benefits or costs). Lastly, we apply our method to Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy (MADIT-CRT) study, a randomized trial. Dynamic surveillance rules (DSRs) are sequential surveillance decision rules informing the monitoring schedules in clinical practice, which can adapt overtime according to a patient’s evolving characteristics. In many clinical applications, it is desirable to identify and implement optimal stabilized DSRs, where the parameters indexing the decision rules are shared across different decision points and estimated simultaneously. We propose a new criterion for DSRs that account for benefit-cost tradeoff. We develop two methods to estimate the stabilized DSRs optimizing the proposed criterion, which are easy to implement using slightly modified standard statistical software. We establish the asymptotic properties of the estimated coefficient parameters of biomarkers indexing the decision rules. Extensive simulation studies are conducted to demonstrate the superior performance of the proposed methods. The methods are further applied to the Canary Prostate Active Surveillance Study(PASS) study.