Artificial intelligence-augmented mixture toxicology: reframing unresolved risk at Camp Lejeune

Area of Science:

• Environmental toxicology
• Computational toxicology
• Risk assessment

Background:

• Evaluating health risks from complex chemical mixtures in diverse populations is challenging.
• Traditional toxicology methods struggle with multiple volatile organic compounds (VOCs) that share metabolic pathways.
• Camp Lejeune's contaminated water poses persistent health risk uncertainty due to these limitations.

Purpose of the Study:

• To demonstrate how AI-enabled computational toxicology can enhance mixture risk assessment.
• To apply AI approaches to the Camp Lejeune VOC mixture (trichloroethylene, tetrachloroethylene, benzene, vinyl chloride) as a case study.
• To address the unresolved issue of disproportionate effects on susceptible subpopulations from chronic low-level mixture exposure.

Main Methods:

• Integrating toxicokinetic, toxicodynamic, and toxicogenomic data using AI.
• Developing mechanistically coherent, testable computational models.
• Utilizing AI to identify interaction mechanisms and quantify genotype-dependent variability in internal dose.

Main Results:

• AI-augmented approaches can identify plausible interaction mechanisms within chemical mixtures.
• Quantification of genotype-dependent variability in internal dose is feasible with AI.
• Probabilistic risk estimates can be generated for complex mixture exposures.

Conclusions:

• AI-driven computational toxicology offers a tractable framework for assessing risks of chemical mixtures.
• These approaches can clarify the interface between mechanistic toxicology and public health decision-making.
• Addressing susceptible subpopulations is a key application for AI in environmental health risk assessment.