Mechanism-aware and explainable artificial intelligence for drug release prediction in polymeric drug delivery systems

The study develops an interpretable artificial intelligence framework for predicting cumulative drug release from PLGA nanoparticle formulations prepared by nanoprecipitation. The research uses a curated dataset containing 4,909 formulation-release observations. The proposed framework achieved exceptional accuracy with R² = 0.9977 on training data and R² = 0.9975 on an independent test set. The framework combines formulation-informed feature engineering, temporal representation learning, symbolic regression, and explainable machine learning. Y-scrambling testing confirmed that the achieved accuracy is not due to chance. Despite high predictive accuracy, the authors emphasize that experimentally validated results and external datasets are necessary before practical deployment.