Deep vectorised operators for pulsatile hemodynamics estimation in coronary arteries from a steady-state prior

2024-10-15

Comput. Methods Programs Biomed.

AI Summary

Accurate and efficient estimation of pulsatile coronary hemodynamics remains challenging, as conventional CFD simulations are computationally prohibitive for clinical deployment in non-invasive coronary artery disease assessment. Method: We propose a discretization-invariant deep vectorization operator framework that achieves functional-space generalization via permutation-equivariant modeling—replacing pointwise mapping with structured geometric learning. The framework integrates neural fields, message-passing networks, and self-attention mechanisms, and is conditioned on CCTA-derived vascular geometry and boundary conditions. Contribution/Results: Evaluated on 74 clinically acquired stenotic coronary cases, our method achieves pulsatile velocity and pressure prediction errors of 0.368 ± 0.079, significantly outperforming baseline methods (p < 0.05), while accelerating computation by two orders of magnitude. To our knowledge, this is the first work to enable high-fidelity pulsatile flow surrogate modeling guided by steady-state priors, establishing a new paradigm for real-time, physics-informed clinical hemodynamic evaluation.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Computer Vision: Learning & Optimization for CV

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

BACKGROUND AND OBJECTIVE Cardiovascular hemodynamic fields provide valuable medical decision markers for coronary artery disease. Computational fluid dynamics (CFD) is the gold standard for accurate, non-invasive evaluation of these quantities in silico. In this work, we propose a time-efficient surrogate model, powered by machine learning, for the estimation of pulsatile hemodynamics based on steady-state priors. METHODS We introduce deep vectorised operators, a modelling framework for discretisation-independent learning on infinite-dimensional function spaces. The underlying neural architecture is a neural field conditioned on hemodynamic boundary conditions. Importantly, we show how relaxing the requirement of point-wise action to permutation-equivariance leads to a family of models that can be parametrised by message passing and self-attention layers. We evaluate our approach on a dataset of 74 stenotic coronary arteries extracted from coronary computed tomography angiography (CCTA) with patient-specific pulsatile CFD simulations as ground truth. RESULTS We show that our model produces accurate estimates of the pulsatile velocity and pressure (approximation disparity 0.368 ± 0.079) while being agnostic (p<0.05 in a one-way ANOVA test) to re-sampling of the source domain, i.e. discretisation-independent. CONCLUSION This shows that deep vectorised operators are a powerful modelling tool for cardiovascular hemodynamics estimation in coronary arteries and beyond.

Problem

• Estimating pulsatile hemodynamics from steady-state priors
• Developing discretization-independent deep learning models
• Accurately predicting velocity and pressure in coronary arteries

Innovation

• Machine learning surrogate model for pulsatile hemodynamics
• Deep vectorised operators for discretisation-independent learning
• Neural field conditioned on hemodynamic boundary conditions