Machine learning strengthened formulation design of pharmaceutical suspensions - A statistical and explainable ML screening approach

Q: 为什么需要对药物混悬剂进行稳定剂筛选与配方优化？

纳米/微米混悬剂的分散体系存在不稳定热力学，可能导致团聚或奥斯特瓦尔德熟化等失稳。稳定剂（表面活性剂和/或聚合物）通过静电排斥和/或空间位阻稳定来降低粒子间作用并提升长期稳定性。

Q: 本研究如何利用机器学习改进混悬剂配方设计？

研究通过全因子研磨实验系统考察混悬特征与配方参数之间的关系，并采用统计与机器学习策略构建配方稳定性分类模型。

Q: 哪些因素对配方稳定性和粒径指标影响更显著？

结果显示，稳定剂浓度是影响中位粒径 D50 的显著因素（p < 0.001）。可解释性分析（SHAP）进一步揭示稳定剂浓度与研磨珠粒径对配方稳定性具有突出影响。

Aug 5, 2026

Upload by Sage

Many formulation strategies target suboptimal management of long-term chronic conditions, including nano- and microsuspensions formulated as long-acting injectables to prolong drug release by particle-size regulation via milling. Surfactants and polymers stabilize these dispersions, yet milling studies often rely on expertise and trial-and-error, with interacting parameters complicating design. This work applies statistical and machine learning to full-factorial milling experiments, identifies stabilizer concentration as significant for D50, and builds an explainable ML stability classifier with high validation performance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Long-acting injectables and particle-size control

Wet bead media milling and stability challenges

Role of stabilizers and formulation screening

Research approach and findings

Full-factorial experiments and ML modeling

Feature importance and SHAP interpretation

Predictive performance and classification outcomes

International Journal of Pharmaceutics 668 (2025) 124967 Contents lists available at ScienceDirect International Journal of Pharmaceutics journal [homepage:](homepage: www.elsevier.com/locate/ijpharm)[ www.elsevier.com/locate/ijpharm](homepage: www.elsevier.com/locate/ijpharm) | Machine learning strengthened formulation design of pharmaceutical suspensions | | | | | --- | --- | --- | --- | | Nadina Zulbearia,1, Fanjin Wang b,1, Sibel Selyatinova Mustafova a, Maryam Parhizkar b, Ren´e Holm a,*
a Department of Physics, Chemistry, and Pharmacy, University of Southern Denmark, Campusvej 55, 5230 Odense, Denmark
b Deparment ofPharmaceutics, UCL School of Pharmacy, University College London, 29-39 Brunswick Square, WC1N 1AX London, United Kingdom | | | | | A R T I C L E I N F O | | A B S T R A C T | | | Keywords:
Stabilizers Formulation screening Nano Microsuspensions Machine learning | | Many different formulation strategies have been investigated to oppose suboptimal treatment of long-term or chronic conditions, one of which are the nano- and microsuspensions prepared as long-acting injectables to prolong the release of an active pharmaceutical compound for a defined period of time by regulating the size of particles by milling. Typically, surfactant and/or polymers are added in the dispersion medium of the suspension during processing for stabilization purposes. However, current formulation investigations with milling are heavily based on prior expertise and trial-and-error approaches. Various interacting parameters such as the milling bead size, stabilizer type and concentration have confounded the investigation of milling process. The present study systematically exploited statistical and machine learning (ML) strategies to understand the relationship between suspension characteristics and formulation parameters under full-factorial milling experiments. Stabilizer concentration was identified as a significant factor (p < 0.001) for median suspension diameter (D50). A formulation stability classification ML model with high prediction accuracy (0.91) and F1-score (0.91) under 10-fold cross-validation was constructed based on 72 formulation datapoints. Model interpretation through Shapley additive explanations (SHAP) revealed the prominent impact of stabilizer concentration and milling bead size on formulation stability. The present work demonstrated the potential to achieve a deeper understanding of the design and optimization of nano- and microsuspensions through explainable ML modelling on formulation screening data. | | 1. Introduction To overcome suboptimal treatment outcome of long-term or chronic conditions (e.g., schizophrenia or human immunodeficiency virus (HIV)) and poor patient compliance due to frequent administration, several formulation strategies have been investigated (Park et al., 2013; Okoli et al., 2022). Among these, long-acting injectables (LAIs) have clinically been shown to prolong the release of an active pharmaceutical compound for weeks to months by a single injection. LAIs are based on different formulation techniques, where aqueous suspensions containing crystalline drug particles can be engineered to release the active compound for a predefined plasma-concentration profile, within a safe therapeutic range, by controlling the sizes of the drug particles (Park et al., 2013; Owen & Rannard, 2016; Pacchiarotti et al., 2019; Nkanga et al., 2020; Bao et al., 2021; Okoli et al., 2022; Wilkinson et al., 2022; Bauer et al., 2023; Holm et al., 2023; Alidori et al., 2024). The manufacturing of nano-and microsuspensions is often prepared by the highly efficient wet bead media milling (i.e., top-down size reduction approach) which reduces the size of larger drug particles into smaller particles by the friction from mechanical forces generated by milling beads while suspended in an aqueous stabilizer vehicle (Verma et al., 2009; Nakach et al., 2014; Mishra et al., 2015; Hagedorn et

FAQs

为什么需要对药物混悬剂进行稳定剂筛选与配方优化？

纳米/微米混悬剂的分散体系存在不稳定热力学，可能导致团聚或奥斯特瓦尔德熟化等失稳。稳定剂（表面活性剂和/或聚合物）通过静电排斥和/或空间位阻稳定来降低粒子间作用并提升长期稳定性。

本研究如何利用机器学习改进混悬剂配方设计？

研究通过全因子研磨实验系统考察混悬特征与配方参数之间的关系，并采用统计与机器学习策略构建配方稳定性分类模型。

哪些因素对配方稳定性和粒径指标影响更显著？

结果显示，稳定剂浓度是影响中位粒径 D50 的显著因素（p < 0.001）。可解释性分析（SHAP）进一步揭示稳定剂浓度与研磨珠粒径对配方稳定性具有突出影响。

Aug 5, 2026

Upload by Sage