🤖 AI Summary
Predicting rupture risk in thoracic aortic aneurysms (TAAs) faces dual challenges: full-order computational fluid dynamics (FOM-CFD) models are computationally prohibitive, while conventional reduced-order models (ROMs) suffer from poor generalizability across diverse geometries and growth stages.
Method: We propose a graph neural network (GNN)-enhanced, geometry-aware ROM that directly leverages the intrinsic graph structure of finite-volume (FV) meshes—embedding geometric features and FOM-CFD data jointly to predict wall shear stress (WSS) and oscillatory shear index (OSI) efficiently across multiple aneurysm growth stages.
Contribution/Results: Unlike traditional ROMs, our approach eliminates dependence on fixed mesh topology and high-dimensional parametrizations, ensuring spatially localized physical consistency and computational scalability. Experiments demonstrate accuracy comparable to FOM-CFD, with over 100× speedup—enabling real-time, patient-specific TAA risk assessment deployable in clinical settings.
📝 Abstract
The complexity of the cardiovascular system needs to be accurately reproduced in order to promptly acknowledge health conditions; to this aim, advanced multifidelity and multiphysics numerical models are crucial. On one side, Full Order Models (FOMs) deliver accurate hemodynamic assessments, but their high computational demands hinder their real-time clinical application. In contrast, ROMs provide more efficient yet accurate solutions, essential for personalized healthcare and timely clinical decision-making. In this work, we explore the application of computational fluid dynamics (CFD) in cardiovascular medicine by integrating FOMs with ROMs for predicting the risk of aortic aneurysm growth and rupture. Wall Shear Stress (WSS) and the Oscillatory Shear Index (OSI), sampled at different growth stages of the thoracic aortic aneurysm, are predicted by means of Graph Neural Networks (GNNs). GNNs exploit the natural graph structure of the mesh obtained by the Finite Volume (FV) discretization, taking into account the spatial local information, regardless of the dimension of the input graph. Our experimental validation framework yields promising results, confirming our method as a valid alternative that overcomes the curse of dimensionality.