FDA Accepts In Silico DDT to Predict DILI

Per the U.S. Food and Drug Administration, the FDA's Center for Drug Evaluation and Research (CDER) has accepted the first Letter of Intent (LOI) for an in silico drug development tool (DDT) into the Innovative Science and Technology Approaches for New Drugs (ISTAND) DDT Qualification Program. The submission concerns an AI-Driven Digital Liver Model intended to support prediction of drug-induced liver injury (DILI) for small-molecule new drug candidates. The FDA states this LOI acceptance is the first step in a three-step DDT qualification process. "New technologies are showing incredible promise in helping improve and streamline drug development, with the ultimate goal of enhancing patient care," said Michael Davis, MD, PhD, Acting Director of CDER. "The AI-Driven Digital Liver Model shows promise in assessing the risk of hepatotoxicity during preclinical phases of drug development," said Jeffrey Siegel, MD, Director of the Office of Drug Evaluation Sciences.

Per the FDA announcement, the DDT is described as a New Approach Methodology (NAM) that leverages artificial intelligence to compare the chemical structures of new drug candidates against historical reference drugs with known DILI risk. The agency frames the tool as intended to complement other methods of DILI risk assessment as part of a weight-of-evidence approach. The FDA also notes that current modelling does not accurately identify DILI risk in humans and that DILI is a leading cause of clinical-trial termination and drug attrition during the Investigational New Drug process.

Industry observers note that regulatory qualification pathways, such as ISTAND, function as formal interfaces between computational tool developers and regulators. Such acceptance of an LOI typically signals that a tool will undergo structured evaluation against validation, reproducibility, and context-of-use criteria, which can reduce uncertainty for sponsor submissions and external validators.

Observers should track progression through the ISTAND qualification steps, publication or independent benchmarking of the model's predictive performance, whether the FDA issues context-of-use language if the DDT becomes qualified, and subsequent uptake by sponsors in preclinical packages. Per the FDA, no qualification is final at the LOI stage; additional data and review steps remain.