Request Copyright Permission

Machine Learning and Artificial Intelligence in Generic Drug Development of Orally Inhaled Drug Products

Chopski S, Walenga RL, Hu M, Babiskin A, Fang L.

Respiratory Drug Delivery 2026. Volume 1, 2026: 9-0.

Abstract:

The U.S. Food and Drug Administration (FDA) publishes bioequivalence (BE) recommendations for orally inhaled drug products (OIDPs) which may include a combination of in vitro or in vivo studies. For the development of more efficient BE methods for OIDPs, FDA has employed quantitative medicine (QM) approaches such as computational fluid dynamics (CFD) and physiologically based pharmacokinetic (PBPK) modeling. Artificial intelligence (AI) and machine learning (ML) can reduce the burden of computationally expensive CFD simulations such as using machine learning-based reduced order models (ROMs) to more rapidly predict powder aerodynamic particle size distributions from dry powder inhalers (DPIs) and accelerate CFD simulation times. Besides the impact of AI/ML on enhancement of QM approaches such as CFD, FDA has embraced the use of AI/ML through the adoption of AI/ML for regulatory review activities and as evidenced by the AI guidance as a risk-based framework. The Office of Generic Drugs (OGD) is employing AI in three key areas of research and development which are AI Data Infrastructure, AI-Powered Workflow Automation, and AI-Enhanced Quantitative Medicine. Altogether, AI/ML has already been used to augment QM approaches and accelerate regulatory review activities, and its impact is expected to only increase in scope over time.