Periodic Reporting for period 1 - HiDDyn-AF (High-resolution Decryption of Left Atrial Haemodynamics in Atrial Fibrillation (HiDDyn-AF))
Período documentado: 2023-10-01 hasta 2025-09-30
The project aimed to improve understanding of left atrial (LA) and left atrial appendage (LAA) haemodynamics in atrial fibrillation (AF), a major contributor to cardioembolic stroke. Current clinical imaging modalities cannot fully resolve intracardiac flow structures or wall shear stress patterns, limiting risk stratification and personalised treatment planning.
The overarching objective was to develop and validate multi-modal, patient-specific computational models that combine high-resolution CT anatomy with in vivo 4D flow MRI–derived motion and boundary conditions. This integrated approach sought to provide more accurate simulations of LA haemodynamics and identify biomechanical markers linked to thrombus formation.
The overarching objective was to develop and validate multi-modal, patient-specific computational models that combine high-resolution CT anatomy with in vivo 4D flow MRI–derived motion and boundary conditions. This integrated approach sought to provide more accurate simulations of LA haemodynamics and identify biomechanical markers linked to thrombus formation.
During the fellowship, I developed a complete image-based workflow for personalised LA modelling, including CT-based anatomical reconstruction, motion extraction from cardiac MRI, boundary-condition derivation from 4D flow, and fully transient CFD simulation.
The work resulted in multiple peer-reviewed publications demonstrating (i) the feasibility of combining multi-modal clinical imaging into a single computational haemodynamic model and (ii) rigorous validation of simulated velocity fields against 4D flow MRI (iii) characterisation of LA/LAA haemodynamics (iv) investigation of LA/LAA haemodynamics with fibrosis and ablation status.
Major achievements include: establishing an end-to-end pipeline for moving-wall cardiac CFD; quantifying LA and LAA shear stress, oscillatory shear, and endothelial activation markers; and identifying strong associations between morphology, haemodynamics, and thrombogenic potential. The work also included four conference presentations targeting both clinical and scientific audiences.
The work resulted in multiple peer-reviewed publications demonstrating (i) the feasibility of combining multi-modal clinical imaging into a single computational haemodynamic model and (ii) rigorous validation of simulated velocity fields against 4D flow MRI (iii) characterisation of LA/LAA haemodynamics (iv) investigation of LA/LAA haemodynamics with fibrosis and ablation status.
Major achievements include: establishing an end-to-end pipeline for moving-wall cardiac CFD; quantifying LA and LAA shear stress, oscillatory shear, and endothelial activation markers; and identifying strong associations between morphology, haemodynamics, and thrombogenic potential. The work also included four conference presentations targeting both clinical and scientific audiences.
The project delivered one of the first frameworks capable of integrating high-resolution CT, in vivo LA wall motion, and patient-specific pulmonary vein inflow into a unified CFD model of the left atrium. The number of cases analysed well exceeds any group previously considered.
This goes substantially beyond typical LA CFD studies, which rely on rigid walls and simplified or artificial flow conditions. The moving-wall, multi-modal approach produced more physiologically realistic haemodynamics and revealed systematic underestimation of shear-related biomarkers in rigid-wall models.
The correlations identified between LA/LAA geometry and haemodynamic risk markers advance mechanistic understanding of thrombus formation pathways and provide a foundation for future personalised stroke-risk assessment tools.
This goes substantially beyond typical LA CFD studies, which rely on rigid walls and simplified or artificial flow conditions. The moving-wall, multi-modal approach produced more physiologically realistic haemodynamics and revealed systematic underestimation of shear-related biomarkers in rigid-wall models.
The correlations identified between LA/LAA geometry and haemodynamic risk markers advance mechanistic understanding of thrombus formation pathways and provide a foundation for future personalised stroke-risk assessment tools.