Domain adaptation for cancer drug response using an adversarial network

Abstract

Pharmacogenomics aims to provide personalized therapy to patients based on their genetic variability. However, accurate prediction of cancer drug response (CDR) is challenging due to genetic heterogeneity. Since the number of clinical data is limited, most studies to predict drug response use pre-clinical data to train models. However, such models might not be generalizable to external clinical data due to differences between the pre-clinical and clinical datasets. In this study, we propose a Precision medicine prediction using an Adversarial Network for Cancer Drug Response (PANCDR). PANCDR consists of two sub-models, an adversarial model and a CDR prediction model. The adversarial model reduces the gap between the preclinical and clinical datasets, while the CDR prediction model extracts features and predicts responses. PANCDR was trained using both pre-clinical data and unlabeled clinical data. Subsequently, we tested it on external clinical data, including The Cancer Genome Atlas (TCGA) and brain tumor patients. PANCDR outperformed other machine learning models at predicting external test data. Our results demonstrate the robustness of PANCDR and its potential in precision medicine by recommending patient-specific drug candidates.