Fast ai predicts heart valve mechanics for disease assessment
Real-time Generalizable Heart Valve Mechanics for Clinical Disease Assessment via a Physics-Conditioned Neural Operator
Machine Learning
Summary
Heart valve problems affect many people but are often detected too late for easy treatment. The authors created a fast AI tool that can quickly predict how heart valve parts move and stretch under pressure. This can help doctors understand valve problems sooner and plan treatments better. Their tool works much faster than older computer simulations and still gives accurate results.
What this means in practice
- •For cardiac imaging teams: Provide rapid predictions of valve mechanics to assist timely and accurate diagnosis of heart valve diseases during clinical workflows.
- •For medical device designers: Use fast and generalizable valve mechanics predictions to optimize design and testing of repair devices under varied patient conditions.
Authors
Shawn Koohy, Wensi Wu, Matthew A Jolley, Paris Perdikaris
Abstract
Mitral regurgitation is the most common heart valve disorder worldwide, affecting over 2% of the global population, rising to at least 10% in adults over 75, and causing approximately 15% of valvular heart disease-related deaths. Yet only a minority of patients with severe disease undergo corrective surgery. Rapid assessment of valve mechanics could enable earlier, more precise intervention, but traditional finite element simulations remain too slow for clinical timelines and parameter sweeps. We introduce the Physics-Conditioned Neural Operator (PCNO), a transformer-based surrogate that predicts leaflet displacement, strain, and stress fields across mitral and tricuspid geometries, conditioned on systolic blood pressure and tissue properties. Trained on functional, regurgitated, and pathological valves, including tethering, P2 prolapse, and annular dilation, PCNO achieves up to a 15,260x speedup over fine mesh finite element simulations with comparable accuracy, identifies pathology class, and resolves diagnostic metrics within 3.5% error under out-of-distribution extrapolation.