Integrated QSAR-ML and QST Modeling for Early Mechanistic Prediction of Clinical Hepatotoxicity Across Multiple Drug Classes

Authors: Mohd H, Battista C, Yang K, Shoda LKM, Lawless M, Beaudoin J
Conference: ASCPT
Keywords: drug-induced livery injury (DILI), hepatotoxicity, machine learning (ML) models, quantitative structure–activity relationship (QSAR)
Software: DILIsym®

Background

Drug-induced liver injury (DILI) is a major cause of drug attrition, often undetected until latestage clinical trials. Early hepatotoxicity prediction reduces cost and risk in development. To address this, we developed quantitative structure-activity relationship (QSAR)-machine learning (ML) models to provide mechanistic hepatotoxicity inputs for DILIsym, a quantitative systems toxicology (QST) model of DILI. This integrated QSARML+ QST workflow enables prospective hepatotoxicity screening in early development.

By Hana Mohd, Christina Battista, Kyunghee Yang, Lisl K.M. Shoda, Michael S. Lawless, James J. Beaudoin

ASCPT 2026 Annual Meeting, March 4-6 in Denver, CO