DILI-Context: A Dose- and Exposure-Enriched Knowledge Base for Translational Liver Safety Assessment.

Area of Science:

• Pharmacology and Toxicology
• Drug Safety and Development
• Computational Chemistry

Background:

• Drug-induced liver injury (DILI) is a primary reason for drug development failure and market withdrawal.
• Current DILI datasets lack exposure-dependent risk, focusing instead on categorical hazard labels.
• This limits understanding to inherent hazard rather than context-dependent exposure risks.

Purpose of the Study:

• To develop DILI-Context, a knowledge base integrating dose and exposure data with hazard annotations.
• To contextualize hepatotoxic liability using therapeutic dose, duration, and preclinical safety thresholds.
• To advance exposure-aware DILI modeling and risk assessment.

Main Methods:

• Integrated DILI hazard annotations with regulatory and toxicological data, including therapeutic dose and duration.
• Analyzed relationships between dose, safety margins (Therapeutic Index, Margin of Safety), and DILI severity.
• Introduced Chronic Load Score, combining dose magnitude and treatment duration.
• Compared structure-only classifiers with those incorporating exposure-derived covariates.

Main Results:

• A monotonic relationship exists between therapeutic dose and DILI severity.
• Drugs with higher DILI risk have narrower therapeutic windows and reduced safety margins.
• Chronic no-effect thresholds (NOAEL) are more effective than lowest-effect endpoints (LOAEL) for DILI risk discrimination.
• Chronic Load Score effectively stratifies DILI concern and aligns with regulatory signals.
• Exposure-derived covariates provide complementary predictive signals to structural information.

Conclusions:

• Hepatotoxic liability is strongly associated with required systemic exposure and limited safety margins.
• DILI-Context provides a resource for exposure-aware DILI modeling.
• The findings support a shift towards exposure-dependent risk assessment in DILI studies.