Explainable artificial intelligence for patient stratification and drug repositioning

Abstract

Enabling precision medicine requires developing robust patient stratification methods as well as drugs tailored to homogeneous subgroups of patients from a heterogeneous population. Developing de novo drugs is expensive and time consuming with an ultimately low FDA approval rate. These limitations make developing new drugs for a small portion of a disease population unfeasible. Therefore, drug repositioning is an essential alternative for developing new drugs for a disease subpopulation. There is a crucial need to develop data-driven approaches that find druggable homogeneous subgroups within the disease population and reposition the drugs for these subgroups. In this study, we developed an explainable AI approach for patient stratification and drug repositioning. Exploratory mining mimicking the trial recruitment process as well as network analysis were used to discover homogeneous subgroups within a disease population. For each subgroup, a biomedical network analysis was done to find the drugs that are most relevant to a given subgroup of patients. The set of candidate drugs for each subgroup was ranked using an aggregated drug score assigned to each drug. The method represents a human-in-the-loop framework, where medical experts use data-driven results to generate hypotheses and obtain insights into potential therapeutic candidates for patients who belong to a subgroup. To examine the validity of our method, we implemented our method on individual cancer types and on pan-cancer data to consider the inter- and intra-heterogeneity within a cancer type and among cancer types. Patients' phenotypic and genotypic data was utilized with a heterogeneous knowledge base because it gives a multi-view perspective for finding new indications for drugs outside of their original use. Our analysis of the top candidate drugs for the subgroups showed that most of these drugs are FDA-approved drugs for cancer, and others are non-cancer related, but have the potential to be repurposed for cancer. We have discovered novel cancer-related mechanisms that these drugs can target in different cancer types to reduce cancer treatment costs and improve patient survival. Further wet lab experiments to validate these findings are required prior to initiating clinical trials using these repurposed therapies.