Left Atrial 4-Dimensional Flow Magnetic Resonance Imaging: Stasis and Velocity Mapping in Patients With Atrial Fibrillation

Abstract

Objectives: Left atrial (LA) 4-dimensional flow magnetic resonance imaging (MRI) was used to derive anatomic maps of LA stasis, peak velocity, and time-to-peak (TTP) velocity in patients with atrial fibrillation (AF) and to identify relationships between LA flow with LA volume and patient characteristics.

Materials and methods: Four-dimensional flow MRI for the in vivo assessment of time-resolved 3-dimensional LA blood flow velocities was performed in 111 subjects: 42 patients with a history of AF and in sinus rhythm (AF-sinus), 39 patients with persistent AF (AF-afib), 10 young healthy volunteers (HVs), and 20 age-appropriate controls (CTRL). Data analysis included the 3-dimensional segmentation of the LA and the calculation of LA stasis, peak velocity, and TTP maps. Regional LA flow dynamics were quantified by calculating mean stasis, peak velocity, and TTP in the LA center region and the region adjacent to the LA wall.

Results: A sensitivity analysis identified thresholds for global LA stasis (<0.1 m/s) and peak velocity (top 5% LA velocities), which detected significant differences between AF patients and controls for global LA stasis (HV, 25% ± 5%; CTRL, 29% ± 10%; AF-sinus, 41% ± 13%; AF-afib, 52% ± 17%) and peak velocity (HV, 0.43 ± 0.02 m/s; CTRL, 0.37 ± 0.04 m/s; AF-sinus, 0.33 ± 0.05 m/s; AF-afib, 0.30 ± 0.05 m/s). Regional analysis revealed significantly increased stasis at both LA center and wall for AF patients compared with age-appropriate controls (29%-84% difference, P < 0.006) and for AF-afib versus AF-sinus patients (22%-30% difference, P < 0.004). In addition, stasis close to the LA wall was significantly elevated (P < 0.001) compared with the LA center for all subject groups. Multiple regressions revealed significant (RAdj = 0.45-0.50, P < 0.001) relationships between impaired global LA flow (reduced velocity and increased stasis) with age (|β| = 0.27-0.50, P < 0.002) and LA volume (|β| = 0.26-0.50, P < 0.003).

Conclusions: Atrial 4-dimensional flow MRI detected changes in global and regional LA flow dynamics associated with AF, age, and LA volume. Longitudinal studies are needed to test the diagnostic value of LA flow metrics as potential risk factors for thromboembolic events.

Figure 1

4D flow MRI data analysis in a 62 year old male patient with a history of AF (AF-sinus, subject #35) including A: 3D segmentation of the left atrium (LA) based on the 3D-PCMRA data (gray shaded iso-surface) and masking of velocities inside the segmented LA. B: Calculation of LA stasis maps for a velocity threshold of 0.1m/s (left), peak velocity maximum intensity projections (MIPs, mid), and time-to-peak velocity (TTP) maps. Atrial areas for the analysis regional velocities and stasis in the LA center and adjacent to the LA wall are delineated by black lines. The mean for all three metrics of regional LA flow dynamics are listed below each map. AAo = ascending aorta, DAo = descending aorta, PA = pulmonary artery.

Figure 2

Maps of left atrial flow parameters superimposed on the underlying anatomic 4D flow MRI data in oblique sagittal orientation. The individual images show stasis maps for a velocity threshold of 0.1m/s (left column), peak velocity maximum intensity projections (4D MIPs, mid column), and time-to-peak (TTP) velocity maps (right column) in four subjects representing each of the sub-groups (A-D). Color coding illustrates regions in the LA with high (red) and low (blue) stasis, peak velocity and TTP. Areas for the analysis of regional flow dynamics (LA center and LA wall) are delineated and the resulting regional LA stasis, peak velocities and TTP are listed below each map. AAo = ascending aorta, PA = pulmonary artery, LA = left atrium.

Figure 3

Sensitivity analysis for the identification of LA stasis and LA peak velocity thresholds. A: LA stasis (top) as a function of the LA velocity threshold (% LA velocities < LA velocity threshold, v_{stasis_thresh}) for all four groups (young volunteers, age appropriate controls, AF-sinus, AF-afib). Corresponding p-values for pair-wise comparisons of LA stasis between groups are shown below. B: Peak LA velocities (top) for different thresholds (% of top LA velocities, v_{peakvel_thresh}) and p-values for between group comparisons (bottom). The individual data points represent averaged over all subjects in each group. Error bars indicate inter-individual standard deviations of LA flow metrics. LA = left atrium.

Figure 4

Group wise comparisons of regional LA peak velocities (A) and LA stasis (B). Flow stasis adjacent to the LA wall was significantly elevated compared to the LA center for all subject groups. The individual box plots illustrate the median (central mark) and the 25th and 75th percentiles (edges), the whiskers extend to the most extreme data points not considered outliers, and outliers are plotted individually as ‘+’.

Figure 5

Bland Altman analysis of inter-observer variability for the quantification of mean LA velocity, peak LA velocity (v_{peakvel_thresh}=5%) and LA stasis (v_{stasis_thresh}=0.1m/s) in a subgroup of n=29 subjects.