LEFT ATRIAL FLOW IN PATIENTS WITH PAROXYSMAL ATRIAL FIBRILLATION USING 4D PHASE CONTRAST MAGNETIC RESONANCE IMAGING

Abstract

Atrial fibrillation (AF) is a common arrhythmia associated with elevated morbidity and mortality from systemic thrombo-embolism. Patterns of left atrial (LA) flow may provide valuable insights for thrombogenic risk. However, such fluid dynamics are challenging to explore given 3-dimensional complexity and possible dependence on pulmonary vein inflow symmetry. This study aimed to explore patterns of LA flow in patients with paroxysmal AF and provide preliminary insights surrounding need to consider pulmonary vein inflow symmetry to understand such patterns. 30 subjects (20 with paroxysmal AF, and 10 controls) in sinus rhythm were scanned using a 3T MRI scanner with a standardized protocol inclusive of full heart 4D flow MRI (Fig. 1A-B). Analysis planes were prescribed at the posterior, middle, and distal third (close to mitral valve) of the left atrium (Fig. 1C) as well as in the ostium of each dominant pulmonary vein (Fig. 1D). Blood flow analyses for each plane included: peak velocity (PVel), net flow, forward flow, retrograde flow, and regurgitant fraction (%). Left atrial flow patterns were visually scored for vortex formation, rotation and fractionation in all subjects. Age was 53±17 years in AF subjects and 48±11 years in controls. At the basal, mid and distal planes LA PVel were significantly lower in subjects with AF versus controls (0.38±0.10 m/s vs. 0.57±0.18 m/s, p=0.003; 0.40±0.12 m/s vs. 0.53±0.22 m/s, p=0.043; 0.44±0.15 m/s vs. 0.66±0.32 m/s, p=0.017, respectively). Pulmonary vein PVel were similarly between AF and control subjects, both for individual veins and for mean patient values. Patients with AF showed significant asymmetry in PVel between the left superior (LS) and right superior (RS) pulmonary veins (0.51±0.16 m/s vs. 0.41±0.11 m/s, p=0.003, respectively). Vortical LA flow patterns were observed in all subjects and reliably followed a clockwise rotation pattern. Patients with AF demonstrated broader, more complex, and fractionated vortical flow patterns than control subjects (Fig. 1E). This study demonstrated that 4D flow imaging of LA inflow and vortex formation is clinically feasible and identifies observable differences in LA vortices between patients with paroxysmal AF and healthy controls. Loss of pulmonary vein inflow symmetry was observed in those with AF. Whether this is a cause or a result of altered LA flow vortex characteristics remains unknown. Future work in larger patient cohorts, inclusive of developing standardized measures of vortex behavior, is required to explore the prognostic role of LA vortex analysis in patients with atrial disease.