Complas 2023

COMPLAS 2023

Deep Learning Approach for Coronary In-stent Restenosis Using Physics-informed Neural Networks

Shi, Jianye (Institute of Applied Mechanics, RWTH Aachen)
Manjunatha, Kiran (Institute of Applied Mechanics, RWTH Aachen)
Reese, Stefanie (Institute of Applied Mechanics, RWTH Aachen)

In session: TUa-900a - - CT900a - Data Science, Machine Learning and Artificial Intelligence

Please login to view abstract download link

Coronary artery disease (CAD) is one of the largest causes of death worldwide. Percutaneous coronary intervention (PCI) is one of the minimally invasive procedures used to overcome CAD by restoring blood flow in clogged coronary arteries. Unfortunately, PCI is associated with several risk factors including in-stent restenosis and stent thrombosis. Drug-eluting stents have been developed to counteract the severe restenosis observed after bare-metal stent implantation. However, the risk of restenosis still prevails due to the inhibitory effect of the drug on endothelial healing. The current work focuses on developing a multiphysics model using a deep learning framework to include the effect of anti-inflammatory drugs embedded in the drug-eluting stents. The highly resolved multiphysics model is based on a set of coupled partial differential equations (PDEs), which govern the mechanism of neointimal hyperplasia by capturing the effects of platelet aggregation, growth-factor release, cellular motility, endothelial barrier function and drug deposition.