Robust Prediction of Patient-Specific Clinical Response to Unseen Drugs From in vitro Screens Using Context-aware Deconfounding Autoencoder

Abstract

Accurate and robust prediction of patient-specific responses to drug treatments is critical for drug development and personalized medicine. However, patient data are often too scarce to train a generalized machine learning model. Although many methods have been developed to utilize cell line data, few of them can reliably predict individual patient clinical responses to new drugs due to data distribution shift and confounding factors. We develop a novel Context-aware Deconfounding Autoencoder (CODE-AE) that can extract common biological signals masked by context-specific patterns and confounding factors. Extensive studies demonstrate that CODE-AE effectively alleviates the out-of-distribution problem for the model generalization, significantly improves accuracy and robustness over state-of-the-art methods in both predicting patient-specific ex vivo and in vivo drug responses purely from in vitro screens and disentangling intrinsic biological signals from confounding factors. Using CODE-AE, we screened 50 drugs for 9,808 cancer patients and discovered novel personalized anti-cancer therapies and drug-response biomarkers. Contact: lxie@iscb.org