Commercial toxicology prediction systems: a regulatory perspective

Area of Science:

• Toxicology
• Computational Chemistry
• Regulatory Science

Background:

• Commercial toxicity prediction systems are increasingly used in regulatory contexts.
• Understanding the capabilities and limitations of these systems is crucial for their effective application.
• Current systems excel at hazard identification but are less adept at ruling out hazards.

Purpose of the Study:

• To evaluate the use of commercial toxicity prediction systems in regulatory settings.
• To highlight challenges and illustrate issues in applying these systems using real-world examples.
• To propose future directions for improving toxicity prediction technologies.

Main Methods:

• Analysis of the capabilities and limitations of commercial toxicity prediction systems.
• Review of regulatory applications, including testing prioritization and decision support.
• Examination of case studies involving the U.S. Environmental Protection Agency (EPA) and specific chemical assessments.

Main Results:

• Commercial systems are more effective for hazard identification than for hazard exclusion.
• Regulatory applications require careful consideration of prediction system performance and intended use.
• Case studies demonstrate practical challenges in integrating prediction systems into regulatory workflows.

Conclusions:

• Future advancements in toxicity prediction require improving technologies within data constraints.
• Optimal use of new test data and integration of quantitative structure-activity relationships (QSAR), empirical data, and mechanistic insights are essential.
• Enhanced integration of diverse data sources and modeling approaches will advance the goal of reliable toxicity prediction.