Left atrial wall shear stress correlates with fibrosis in patients with atrial fibrillation

Figures

• 

  Fig. 1. Graphical overview of the methodology. The diagram outlines the methodological steps, integrating data from the electroanatomical maps (left atrial geometry and BV), CFD analysis and CMR imaging with gadolinium contrast.

• 

  Fig. 2. CFD and postprocessing methodological overview. a, Derivation of patient-specific mitral flow velocity from phase-contrast CMR imaging with adaptation to an AF mitral velocity waveform (dashed line) by removing the A-wave, if required. b, Example of the boundary conditions and flow extensions on a 3D mesh. c, Methodological steps of the unfolding procedure to generate the 2D maps through a standardized atrial atlas. d, Regions of the 2D atrial maps. PC-MRI, phase-contrast MRI; V(t), time-dependent flow velocity; P, pressure; LAA, left atrial appendage; RSPV, right superior pulmonary vein; RIPV, right inferior pulmonary vein; LSPV, left superior pulmonary vein; LIPV, left inferior pulmonary vein.

• 

  Fig. 3. Three-dimensional results in an exemplary case. a–d, Distribution of wall shear and fibrosis on the 3D atrial model of participant 10 in anteroposterior (left) and posteroanterior (right) views. The colors represent the TAWSS (a); the BV (b); the IIR (c) in regions with fibrosis (red), interstitial fibrosis (white) and no fibrosis (gray); and the BA (d). Source data

• 

  Fig. 4. Combined 2D maps of four representative cases of the primary variables: TAWSS, BA, BV and IIR. Each panel corresponds to one participant denoted by the corresponding participant number.

• 

  Fig. 5. Participant-specific interplay between TAWSS, BA and fibrosis as measured using BV and IIR. The TAWSS is divided into quartiles (Q1–Q4). Above each chart is the corresponding two-sided P value from the Kruskal–Wallis test between the quartiles of TAWSS and BV, IIR or BA and the Pearson correlation coefficient (r) between all data points of the same variable pairs. The boxes correspond to the interquartile range, the notches to the 95% confidence interval of the median and the whiskers to the s.d. LGE-CMR acquisitions were not available for participants 13–15. Sample sizes: 1: 29,492, 2: 27,780, 3: 37,670, 4: 23,899, 5: 28,750, 6: 28,371, 7: 18,797, 8: 19,376, 9: 23,103, 10: 20,840, 11: 20,444, 12: 22,615, 13: 30,219, 14: 21,990 and 15: 35,771.

• 

  Fig. 6. Relationship between fibrosis distribution and TAWSS. CMR-derived fibrosis (left; IIR ≥ 1.2, n = 301,137) and electrical scar (right; BV < 0.5 mV, n = 389,117) data are shown according to the TAWSS quartiles for the total study population, with the corresponding two-sided P value from the chi-square test.

• 

  Fig. 7. Distribution of BA in the left atrial regions. Data show the normalized BA according to the left atrial regions of the total study population. The P value corresponds to the result of the two-sided Kruskal–Wallis test (n = 60, for the pulmonary veins, n = 15 for all other regions). The boxes correspond to the median and the interquartile range, the whiskers to the s.d. and the square to the mean value. PV, pulmonary vein. Source data

• 

  Extended Data Fig. 1. The connection between electrical scar and hemodynamic indices. Pearson’s correlation coefficient heatmap between the bipolar voltage and the calculated hemodynamic indices of each subject. The 3D electro-anatomical maps of subjects 2, 3, 5, 9, 10, 11 and 14 were obtained in atrial fibrillation. Source data

• 

  Extended Data Fig. 2. The connection between fibrosis and hemodynamic indices. Pearson’s correlation coefficient heatmap between the image intensity ratio and the calculated hemodynamic indices of each subject. Source data

• 

  Extended Data Fig. 3. The 2D maps with the results of all subjects. The 2D maps of the distribution of the time-averaged wall shear stress (TAWSS), bipolar voltage (BV), blood age (BA) and Image Intensity Ratio (IIR) for each subject. The colormap of IIR corresponds to healthy regions (black-gray), regions with interstitial fibrosis (white) and fibrotic regions (red). LGE-CMR acquisitions were not available for subjects 13–15.