Accelerating Polymorph Screening with AI & ML: A New Era in Drug Development

This is the authors' abstract. We don't add key points for this paper.

Polymorph screening plays a crucial role in pharmaceutical development, influencing the solubility, stability, and bioavailability of active pharmaceutical ingredients (APIs). Traditional screening methods are time-consuming, labor-intensive, and often yield unpredictable results. Recent advancements in artificial intelligence (AI), machine learning (ML), and computational modeling such as molecular dynamics (MD) and density functional theory (DFT) have revolutionized this process, enabling faster and more accurate predictions of polymorphs, solvates, hydrates, cocrystals, and salts. The integration of AI-driven and computational models in polymorph screening is examined in this paper, with an emphasis on their potential applications in predicting the stability of amorphous solids, solubility of APIs, and the solvates, hydrates and cocrystals for the drug development. We also discussed how thermodynamic viability of solvate and hydrate formation and desolvation kinetics can be analyzed using computational techniques. AI and ML provides increase polymorph screening's success rate, which will impact the drug polymorphs selection for the manufacturing and regulatory compliance. The most recent advancements, challenges, and contemporary approaches in the use of AI/ML, MD and DFT in solid-state drug development are the main topics of this review.