AI-powered multimodal modeling of personalized hemodynamics in aortic stenosis

• DOI: 10.48550/arxiv.2407.00535
• Publication Date: 2024-06-29
• Authors: Çağlar Öztürk; Daniel H. Pak; Luca Rosalia; Debkalpa Goswami; Mary Robakowski; Raymond G. McKay; Christopher Nguyen; James S. Duncan; Ellen T. Roche
• Keywords: FOS: Computer and information sciences; Science; Computer Vision and Pattern Recognition (cs.CV); Q; Computer Science - Computer Vision and Pattern Recognition; aortic stenosis; deep learning; fluid‐structure interaction; computational fluid dynamics; heart meshing; Computational Engineering, Finance, and Science (cs.CE); 03 medical and health sciences; 0302 clinical medicine; multimodal modeling; Computer Science - Computational Engineering, Finance, and Science; Research Article

Aortic stenosis (AS) is the most common valvular heart disease in developed countries. High-fidelity preclinical models can improve AS management by enabling therapeutic innovation, early diagnosis, and tailored treatment planning. However, their use currently limited complex workflows necessitating lengthy expert-driven manual operations. Here, we propose an AI-powered computational framework for accelerated democratized patient-specific modeling of hemodynamics from computed tomography. First, demonstrate that our automated meshing algorithms generate task-ready geometries both benchtop simulations with higher accuracy 100 times faster than existing approaches. Then, show approach be integrated fluid-structure interaction soft robotics to accurately recapitulate a broad spectrum clinical hemodynamic measurements diverse patients. The efficiency reliability these make them ideal complementary tool personalized high-fidelity biomechanics, hemodynamics,