High-Dimensional Precision Medicine from Patient Derived Xenograft Data

• Activity Number: 561 - JASA Applications and Case Studies Invited Session
• Type: Invited
• Date/Time: Wednesday, August 2, 2017 : 2:00 PM to 3:50 PM
• Sponsor: JASA, Applications and Case Studies
• Author(s): Naim Rashid and Jingxiang Chen and Michael Lawson and Daniel Luckett and Longshaokan Wang and Eric Laber and Yufeng Liu and Jen Jen Yeh and Donglin Zeng and Michael R Kosorok*
• Companies: University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill and North Carolina State University and North Carolina State University and University of North Carolina and University of North Carolina at Chapel Hill and University of North Carolina and University of North Carolina at Chapel Hill
• Keywords: Biomarkers ; Deep Learning Autoencoders ; Machine Learning ; Outcome Weighted Learning ; Patient-Derived Xenografts ; Q-Learning

Abstract

Patient-derived xenograft (PDX) studies are a rich cancer research resource that allows for the examination of the effects of multiple cancer treatments within the same patient tumor as well as between different tumors. We analyze PDX data from multiple cancers to identify predictive biomarkers informative for developing personalized optimal treatment recommendations, given pre-treatment multi-platform genomic data and post-treatment tumor response measurements. Because the original tumors are human in origin, the biomarkers discovered are potentially applicable to treating cancer in humans. We estimate optimal Individualized Treatment Rules (ITRs) using variants of both indirect, regression based approaches such as Q-learning and direct approaches such as Outcome Weighted Learning. Because available genomic data is high-dimensional, we explore several novel methods for dimension reduction that are suitable to estimation of optimal ITRs. Our results indicate that PDX data is a valuable resource for precision medicine. Conclusions and methodological recommendations are also provided.