Leveraging In Silico and Artificial Intelligence Models to Advance Drug Disposition and Response Predictions Across...

Area of Science:

• Pharmacology and Drug Development
• Computational Biology
• Systems Toxicology

Background:

• Ensuring safe and effective drug use requires understanding inter-individual differences in drug disposition and response.
• Clinical trials often underrepresent diverse populations (children, elderly, pregnant, etc.), necessitating advanced evaluation tools.
• Physiologically based pharmacokinetic (PBPK) and quantitative systems pharmacology/toxicology (QSP/QST) models offer solutions for virtual population analysis.

Purpose of the Study:

• To review the application of in silico and AI models in predicting drug exposure and responses across the lifespan.
• To highlight the use of virtual populations within PBPK and QSP/QST frameworks.
• To discuss opportunities and challenges of AI in modeling drug dosing for diverse age groups.

Main Methods:

• Literature review of in silico modeling techniques (PBPK, QSP/QST).
• Integration of machine learning (ML) and artificial intelligence (AI) for data analysis.
• Case study using QST modeling for drug-induced liver injury (DILI) in postmenopausal women.

Main Results:

• In silico and AI models enable the creation of virtual populations reflecting diverse physiological states.
• These models integrate clinical trial and real-world data (RWD) for improved prediction of drug efficacy and safety.
• AI tools can identify key physiological factors influencing drug response across different age groups.

Conclusions:

• In silico and AI modeling are crucial for advancing drug development by accounting for inter-individual variability.
• Virtual populations and AI-driven analysis enhance the prediction of drug disposition and response across the lifespan.
• Further development and application of these models are needed to address challenges in drug dosing for all patient populations.