Year: 2025
Type: book-chapter
Abstract: A novel approach for predicting stroke risk using graph neural networks (GNNs) from computational fluid dynamics (CFD) simulations of blood flow in patientbased arterial vessel geometries is proposed.Atherogenesis is the build-up of plaque along arterial vessel walls.It is influenced by the geometrical configuration of the vessel, typically in the vicinity of a bifurcation.Furthermore, some geometric features of the bifurcation such as larger curvature can cause intense perturbations in the flow.Using simulations of blood flow, it is possible to obtain flow parameters such as the distribution of wall shear stress, which would otherwise be impossible to measure.GNNs enable learning of the complex relationships linked to the geometry of the blood vessel and the development of atherosclerosis.Our approach provides new insights into the relationship between blood flow patterns and stroke risk, potentially enabling more personalized prevention and treatment strategies.
Source: De Gruyter eBooks
Parent work: Mathematical Optimization for Machine Learning
Authors: Rohit Pochampalli, Nicolas R. Gauger
Language: English
Subfield: Statistical and Nonlinear Physics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4409704448
