Comparing Left Atrial Spontaneous Echo Contrast Intensity with Gaussian Process Emulator Predictions

Funding Sponsor

King's College London

Third Author's Department

Mechanical Engineering Department

Find in your Library

https://doi.org/10.1007/978-3-031-87756-8_43

All Authors

Paolo Melidoro Malak Sabry Abdel Rahman Amr Sultan Ahmed Qureshi Gregory Y.H. Lip Natalie Montarello Nishant Lahoti Ronak Rajani Magdalena Klis Steven E. Williams Oleg Aslanidi Adelaide De Vecchi

Document Type

Research Article

Publication Title

Lecture Notes in Computer Science

Publication Date

1-1-2025

doi

10.1007/978-3-031-87756-8_43

Abstract

Atrial Fibrillation (AF) significantly increases the risk of ischemic stroke due to blood stasis and hypercoagulability in the left atrium (LA). Effective stroke risk stratification is crucial for identifying AF patients who require anticoagulation therapy. Spontaneous echo contrast (SEC), a phenomenon arising from blood stasis and fibrinogen-mediated red blood cell aggregation, serves as a strong predictor of stroke and can be observed via transesophageal echocardiograms (TEEs). This study employs Gaussian Process Emulators (GPEs) trained on non-Newtonian Computational Fluid Dynamics (CFD) simulations to predict the occurrence and intensity of LA SEC. Using haematocrit and fibrinogen concentration as inputs, the GPEs compute the average blood viscosity in the left atrial appendage (LAA). We also quantified geometric and functional parameters of the LAA to assess their relative impact on SEC intensity and compared these findings with those in patients without SEC. Our results indicate that both LAA motility and GPE-predicted LAA viscosity can distinguish between SEC and non-SEC groups (p > 0.001). In the SEC cohort, significant correlations were observed between grayscale intensity and LAA motility (r = 0.8), LAA orifice diameter (r = 0.71), and predicted LAA viscosity (r = 0.57). This study demonstrates the potential utility of GPEs for predicting LA SEC, thereby enhancing stroke risk stratification in AF patients. Additionally, we identify key geometric and functional features of the LAA that influence SEC formation.

First Page

443

Last Page

452

Link to Full Text
Find this item online

Share

COinS