AI tool aims to predict drug-related liver injury before human trials

The FDA's Center for Drug Evaluation and Research has accepted a Letter of Intent for an AI-driven digital liver model designed to help predict drug-induced liver injury, also known as DILI. The tool is being reviewed through the FDA’s ISTAND Drug Development Tool Qualification Program, which supports innovative technologies that may be used in regulatory drug development.

Drug-induced liver injury remains one of the most difficult safety risks to predict before a medicine reaches human trials. While traditional preclinical models provide important safety information, they do not always translate reliably to human liver toxicity. This gap can lead to late-stage clinical trial failures, delayed development timelines, and in some cases, serious patient safety concerns.

The AI-based tool is designed to assess DILI risk in small-molecule drug candidates by comparing the chemical structure of new compounds with historical reference drugs that have known liver injury profiles. According to the FDA, the model would not replace existing toxicology methods, but would complement them as part of a broader weight-of-evidence approach.

If qualified, the tool could help pharmaceutical companies make more informed decisions before entering phase I clinical trials. This may support earlier identification of high-risk compounds, reduce unnecessary development costs, and potentially decrease reliance on animal testing by improving the predictive value of nonclinical assessments.

For physicians and clinical researchers, the review reflects a wider shift in how AI may be integrated into drug development. Rather than being limited to diagnosis or clinical workflow automation, AI is increasingly being explored as a way to strengthen safety evaluation before a drug reaches patients. In the context of hepatotoxicity, where risk can be complex, unpredictable, and clinically serious, better early prediction could have meaningful implications for both trial design and patient protection.

However, the FDA’s acceptance of the Letter of Intent is only the first step in a three-step qualification process. The company must still submit a Qualification Plan and then a full qualification package before the tool can be formally qualified for use in drug development programs.

This development does not mean AI is ready to independently determine drug safety. It does suggest that regulators are increasingly open to evaluating AI-based models when they are transparent, scientifically validated, and used within a clearly defined context. For drug-induced liver injury, where current prediction methods remain imperfect, this could represent an important step toward safer and more efficient drug development.

G-Med excels in HCP marketing by blending digital innovation with data-driven insights, creating an effective platform for reaching healthcare professionals, offering various advertising solutions. By using G-Med to engage HCPs, share data reports, and explore innovative channels, marketers can deliver targeted, impactful messages that foster strong connections. G-Med’s approach ensures that each campaign is tailored, scientifically rigorous, and effective, aligning perfectly with the best practices for successful HCP marketing.   

Contact us today to learn more: Contact@g-med.com