Machine learning applications in toxicology

Toxicology is a broad field overlapping with biology, chemistry, pharmacology, forensics, and medicine that seeks to understand the adverse effects of poisons on living organisms. Multiple data from disparate sources must be integrated in medical toxicology to predict the clinical course of a given toxin which is difficult. Artificially intelligent (AI) methods of analysis (including machine learning approaches) are of increasing importance in toxicology to mitigate these challenges. This chapter focus on experimental and emerging applications in clinical toxicology, where AI has been applied for the detection and treatment of substance use disorder (SUD), supporting clinical diagnosis for toxidromes, and surveillance for emerging recreational drugs. The use of AI to analyze data from wearable sensors capable of tracking digital biomarkers that enable prediction of a user’s state of health is discussed, and their application in terms of the detection and treatment of substance use disorders is summarized. The current limited state of AI implementation for diagnostic application in clinical toxicology due to its complexity is described. The barriers to the widespread implementation of machine learning in toxicology practice are reviewed and the requirements to overcome these challenges are highlighted.