2023 FDA Science Forum
Using Various Machine Learning Algorithms to Determine the Best Method for Predicting Population Physiologically Based Pharmacokinetic Model Plasma Profiles
- Authors:
- Center:
-
Contributing OfficeNational Center for Toxicological Research
Abstract
The use of artificial intelligence (AI) for predicting drug pharmacokinetics (PK) is a relatively new effort. Several papers have been published covering various methods for using AI machine learning (ML) to predict PK. The objective of this study was to determine which ML method would best predict the plasma concentration of a simulated subject taking various doses of labetalol. In the following experiment, we used various machine learning to determine which algorithm best predicted the data in men and women. . Population plasma PK data were generated with Berkeley Madonna using a physiologically based pharmacokinetic model with 10-30% variation in parameters. The time since the first dose was given, time since last dose, dose frequency, dose amount, study number, patient number, and dosing cycle were also incorporated into the datafile for each patient. The data for each patient were appended together to represent a single study. Fifty subjects were simulated for each study. Multiple scenarios in which sex (male or female), dose (200-1000mg), and frequency (6-12 hours) were varied for each study. The data file was randomized for training, validation, and test sets for the AI models. Once the data set was split using python-based software, the performance was evaluated using the coefficient of determination (R2) for the regression models and the root mean squared error (RMSE) for predictive error. The ML models that performed better than the others tested had an R2 value of greater than 0.6. However, the RMSE overall needs improvement. The results will guide exploration of other model features for the best performing ML methods and algorithms.
