Predicting Early and Complete Drug Release from Long-Acting Injectables Using Explainable Machine Learning

TL;DR

This work develops a time-independent, explainable ML framework to predict drug release dynamics from polymer-based long-acting injectables (LAIs) using 321 PLGA microparticle formulations. It tackles three tasks: predicting fractional release at $24$, $48$, and $72$ h, classifying release profile types via AUC-based labels, and forecasting complete release curves without time inputs, aided by SHAP-based explainability. Key findings show strong predictive power for early release ($R$ around $0.63$–$0.67$) and profile-type classification ($\approx$80% accuracy, $\approx$0.87 F1), with a novel RNN-based model capable of predicting entire release curves from static material features. The results offer concrete material-characteristic descriptors (e.g., encapsulation efficiency, polymer MW, particle size, drug TPSA, LogP, loading capacity) that inform LAI design and optimization, and the open-source code supports reproducibility and practical application in formulation science.

Abstract

Polymer-based long-acting injectables (LAIs) have transformed the treatment of chronic diseases by enabling controlled drug delivery, thus reducing dosing frequency and extending therapeutic duration. Achieving controlled drug release from LAIs requires extensive optimization of the complex underlying physicochemical properties. Machine learning (ML) can accelerate LAI development by modeling the complex relationships between LAI properties and drug release. However, recent ML studies have provided limited information on key properties that modulate drug release, due to the lack of custom modeling and analysis tailored to LAI data. This paper presents a novel data transformation and explainable ML approach to synthesize actionable information from 321 LAI formulations by predicting early drug release at 24, 48, and 72 hours, classification of release profile types, and prediction of complete release profiles. These three experiments investigate the contribution and control of LAI material characteristics in early and complete drug release profiles. A strong correlation (>0.65) is observed between the true and predicted drug release in 72 hours, while a 0.87 F1-score is obtained in classifying release profile types. A time-independent ML framework predicts delayed biphasic and triphasic curves with better performance than current time-dependent approaches. Shapley additive explanations reveal the relative influence of material characteristics during early and for complete release which fill several gaps in previous in-vitro and ML-based studies. The novel approach and findings can provide a quantitative strategy and recommendations for scientists to optimize the drug-release dynamics of LAI. The source code for the model implementation is publicly available.

Figures (8)

• Figure 1: Overview of standard approaches to predicting drug release and methodological contributions of this article.
• Figure 2: Two distinct types of drug release profiles with a) AUC $\leq$ 0.5 and (b) AUC $>$ 0.5. (c) The sample distribution based on the area under the curve (AUC) of release profiles. Blue for AUC $\leq$ 0.5 and red for AUC $>$ 0.5 (red) profile types.
• Figure 3: Customized recurrent neural network framework used for complete drug release profile prediction.
• Figure 4: Model validation and selection strategy using 10x2 nested cross-validation.
• Figure 5: Prediction of the drug release magnitude after (a) 24 hours, (b) 48 hours, and (c) 72 hours following the onset of the drug release using the best regression model (XGB).