Abstract
Introduction:
Clinical scores used for stroke risk estimation in atrial fibrillation (AF) have limited predictive accuracy. 4D Flow MRI can acquire blood flow in a volume data set encompassing the entire left atrium (LA) with quantification of blood velocity in 3 orthogonal directions for every voxel at multiple time points throughout the cardiac cycle (4D = 3D coverage + time).
Hypothesis:
4D Flow MRI can assess 3D LA flow abnormalities in AF patients, which may potentially be linked to stroke.
Methods:
We performed 4D flow MRI in 30 volunteers (10 young, 20 older) and 70 patients with a history of AF: 40 in sinus rhythm (AF-sinus) and 30 in AF during MRI scan (AF-afib). The typical LA contained 3000 voxels and 18 phases per voxel. LA flow velocity histograms for each patient were quantified by mean LA velocity, median LA velocity, and the percentage of LA velocities < 0.2 m/s (%stasis). CHA2DS2-VASc scores were calculated for each patient.
Results:
Mean LA velocity, median LA velocity, and %stasis were significantly different between groups: young volunteers (0.26±0.02 m/s, 0.23±0.02 m/s, 37.3±7.6%), older volunteers (0.21±0.03 m/s, 0.20±0.03 m/s, 52.2±16.0%), AF-sinus (0.18±0.03 m/s, 0.17±0.03 m/s, 67.5±16.9%), and AF-afib (0.16±0.01 m/s, 0.14±0.03 m/s, 78.2±12.9%), p < 0.0125 for all comparisons. CHA2DS2-VASc had moderate but significant correlations with mean LA velocity (R
2
=0.27, p<0.001), median LA velocity (R
2
=0.23, p<0.001), and stasis% (R
2
=0.26, p<0.001). LA flow indices also correlated significantly with age, LA volume, and left ventricular ejection fraction (LVEF) as seen in the Table.
Conclusions:
Left atrial 4D flow MRI is a novel approach to identify patients with reduced LA blood velocities and increased LA stasis. Further study is needed to determine whether these measures can improve upon the CHA2DS2-VASc score for stroke risk prediction and enhance individual decisions on anticoagulation in patients with AF.