An Integrated AI-PBPK Platform for Predicting Drug In Vivo Fate and Tissue Distribution in Human and Inter-Species...

Area of Science:

• Pharmacology
• Computational Chemistry
• Drug Development

Background:

• Traditional pharmacokinetic (PK) estimation is costly, time-consuming, and limited in evaluating synergistic drug properties.
• Early drug development requires robust methods for predicting in vivo drug fate and tissue distribution.
• Optimizing PK profiles is crucial for successful clinical trials.

Purpose of the Study:

• To develop an integrated artificial intelligence (AI) and physiologically based pharmacokinetic (PBPK) platform for rapid PK estimation from molecular structures.
• To predict key drug properties and forecast PK curves without additional training.
• To validate the AI-PBPK model against extensive human PK data.

Main Methods:

• AI models were trained to predict eight critical physicochemical and ADME properties.
• These predictions fed into a PBPK model to forecast PK curves.
• The AI-PBPK approach was validated using 71 IV and 606 oral human PK datasets from the PK-DB database.

Main Results:

• The AI-PBPK model demonstrated robust predictions, with most AUC values within 2-3 fold error ranges.
• Accurate prediction of drug organ selectivity was achieved.
• Inter-species extrapolation optimized predictions for high plasma clearance drugs.

Conclusions:

• The developed AI-PBPK strategy effectively addresses PK challenges in drug discovery.
• This integrated platform enhances drug development efficiency by guiding candidate selection.
• The system facilitates advancing drugs with favorable PK profiles into clinical trials.