Integrating Machine Learning with Computational Fluid Dynamics Models of Orally Inhaled Drug Products (U01) Clinical Trials Not Allowed

Computational fluid dynamics (CFD) has played a crucial role in providing an alternative bioequivalence (BE) approach for generic orally inhaled drug products (OIDPs), in addition to comparative clinical endpoint or pharmacodynamic BE studies, as a relatively cost- and time-efficient complement to benchtop and clinical experiments that has been widely used in developing and assessing generic inhaler devices. However, despite the advances in the power of modern computers, there are still some bottlenecks in using CFD due to computational time, limited grid resolution, pre- and post-processing of large simulation data sets, model parameter estimations, and uncertainty quantifications. Machine learning (ML) has been gaining more attention as a potential tool to alleviate such limitations that arise in CFD. The purpose of this grant is to develop a methodology to integrate ML with CFD models of OIDPs to promote alternative BE studies to enhance and accelerate the development and approval of generic OIDPs.