Predicting the viability of pharmaceutical formulations for continuous direct compression using machine learning approaches

Tools

Diaz, Laura Pereira and Marchal, Stéphanie and Kroll, Paul and Hofstetter, Albert and Lang, Moritz and Piccione, Patrick M. and Brown, Cameron J. and Salehian, Mohammad and Florence, Alastair J. (2026) Predicting the viability of pharmaceutical formulations for continuous direct compression using machine learning approaches. International Journal of Pharmaceutics, 698. 126911. ISSN 1873-3476 (https://doi.org/10.1016/j.ijpharm.2026.126911)

[thumbnail of Diaz-etal-IJP-2026-Predicting-the-viability-of-pharmaceutical-formulations-for-continuous-direct-compression]
Preview
 Text. Filename: Diaz-etal-IJP-2026-Predicting-the-viability-of-pharmaceutical-formulations-for-continuous-direct-compression.pdf
Final Published Version
License: Creative Commons Attribution-NonCommercial 4.0 logo
Download (3MB)| Preview

Abstract

Pharmaceutical formulation is the activity in which the chemical substances that form a final medicinal product are combined, including the active pharmaceutical ingredient and excipients. Changes in formulation from variations in excipients, their composition, or variations in drug loading can impact bulk properties such as powder flowability. Such properties, in turn, may impact subsequent manufacturing processes adversely. More subtle changes, for instance in the physical properties of APIs, such as particle size and shape, can also influence the manufacturability of the drug product. It is therefore important to use state-of-the-art techniques to predict formulation properties, in particular for manufacturability. In this context, Artificial intelligence and Machine Learning (ML) have emerged as potential tools to optimise the transition from formulation development to manufacturing and thus, the use of digital design and data-driven models provides the prospect to accelerate these important development steps. This paper presents three complementary ML models that, when used together, support early assessment of the viability of pharmaceutical formulations for continuous direct compression (cDC). The combined modelling approach provides a practical framework for predictive screening of formulation viability and for supporting more informed decision-making during formulation development.

ORCID iDs

Diaz, Laura Pereira, Marchal, Stéphanie, Kroll, Paul, Hofstetter, Albert, Lang, Moritz, Piccione, Patrick M., Brown, Cameron J. ORCID logoORCID: https://orcid.org/0000-0001-7091-1721, Salehian, Mohammad ORCID logoORCID: https://orcid.org/0000-0003-4073-292X and Florence, Alastair J. ORCID logoORCID: https://orcid.org/0000-0002-9706-8364;
CORE (COnnecting REpositories)