Truace
Subscribe

A Multiparametric in vitro Strategy for Small Molecule Drug-Induced Liver Injury Risk Mitigation in Early Drug Development

Drug-induced liver injury (DILI) remains a major cause of preclinical and clinical attrition, reflecting persistent gaps in the predictive translation of preclinical safety data to human outcomes. To address these gaps, we established a refined multiparametric framework for early DILI risk prediction using a balanced reference set of 170 drugs with established clinical outcomes. Thirty physicochemical, mechanistic endpoints and exposure-related features were systematically evaluated to identify high-specificity…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1295· cite & verify
A Multiparametric in vitro Strategy for Small Molecule Drug-Induced Liver Injury Risk Mitigation in Early Drug Development
Hepatocyte Culture by Lauren Franza. CC BY-SA 3.0 · https://creativecommons.org/licenses/by-sa/3.0

In brief

By the publication date of 2026-10-03, researchers reported a refined multiparametric in vitro strategy for early DILI risk mitigation built on 170 reference drugs with known clinical outcomes. They combined cytotoxicity, mitotoxicity and transporter inhibition assays with exposure data and machine learning, showing that a >=2 flag system and Random Forest models improved specificity and balanced accuracy for early compound triage.

The work matters because DILI remains a major cause of preclinical and clinical attrition, and a high-specificity, interpretable platform could reduce late-stage failures and guide safer design. Uncertainty remains about how well the 98% specificity and 0.72 balanced accuracy observed in this reference set and Genentech portfolio will translate to diverse chemistries, patient populations, and in vivo outcomes.

Main points

  1. Study used a balanced reference set of 170 drugs with established clinical outcomes and evaluated 30 physicochemical, mechanistic and exposure-related features.
  2. Top predictive features were cytotoxicity in primary human hepatocytes and human liver microtissues, mitochondrial toxicity and bile salt export pump inhibition.
  3. Incorporation of clinical exposure data to derive margins of safety substantially improved assay performance.
  4. Exposure-informed Random Forest models achieved a balanced accuracy of 0.72 with exposure, cLogD, mitotoxicity, cytotoxicity and hepatic clearance as major contributors.

The gain

Random Forest and flag-based models integrating exposure, lipophilicity, mitotoxicity and cytotoxicity improved early DILI risk prediction with high specificity, supporting safer molecule triage to reduce hepatotoxicity-related attrition.

The rundown

Researchers systematically evaluated 30 endpoints including cytotoxicity in primary human hepatocytes and human liver microtissues, mitochondrial toxicity and bile salt export pump inhibition, and found that exposure normalization via margins of safety substantially improved prediction.

A flag-based hazard framework integrating >=2 hazard flags reached 98% specificity and positive likelihood ratio of 18, while Random Forest models identified exposure, predicted lipophilicity (e.g., cLogD), mitotoxicity, cytotoxicity and hepatic clearance as key drivers, with translational testing on Genentech portfolio molecules to refine borderline 0-1 flag compounds.

Sources

  1. Peer-reviewedToxicological Sciences2026-10-03

The debate