2026-09-20T07:40:14Zhttps://digital.lib.washington.edu/server/oai/request

oai:digital.lib.washington.edu:1773/538942026-04-15T10:00:52Zcom_1773_4888col_1773_4889

Dabiri, Dana Gupta, Akshay 2025-10-02T16:03:34Z 2025-10-02 2025 Gupta_washington_0250O_28869.pdf https://hdl.handle.net/1773/53894 Thesis (Master's)--University of Washington, 2025 Atrial fibrillation (AF), the most common persistent arrhythmia, increases stroke risk through altered left atrial (LA) wall motion and blood stagnation in the left atrial appendage (LAA). Standard scores like CHA₂DS₂-VASc overlook patient-specific motion patterns, while AF's episodic nature and anatomical variability hinder consistent analysis. This thesis presents a standardized framework for LA wall motion analysis using 4D cardiac CT in both sinus rhythm (SR) and AF. LA geometries were segmented, temporally aligned via Coherent Point Drift (CPD) registration, and mapped to a 2D Universal Atrial Coordinate (UAC) system. Wall motion, quantified from Signed Distance Fields (SDF) and decomposed via fast Fourier transform (FFT), showed coordinated low-frequency contraction in SR and reduced amplitude with higher-frequency content in AF. The framework enables anatomy-independent motion comparison and has the potential to predict AF signatures using SR data alone. application/pdf en_US CC BY Arrythmia Atrial Fibrillaiton Computational Fluid Dynamics Left Atrium Signed Distance Field Sinus Rhythm Mechanical engineering Aerospace engineering Biomedical engineering Aeronautics and astronautics Universal Atrial Coordinate (UAC) for Wall Motion Deep Learning Thesis Restrict to UW for 1 year -- then make Open Access 2026-10-02T16:03:34Z embargo