Quantification of pulmonary regurgitation in patients with repaired Tetralogy of Fallot by 2D phase-contrast MRI: Differences between the standard method of velocity averaging and a pixel-wise analysis

Abstract

Objectives: To compare the values of pulmonary regurgitation in patients with repaired Tetralogy of Fallot quantified from two-dimensional phase-contrast data, by using a new pixel-wise analysis and the standard velocity-averaging method.

Design: Quantitative in silico and in vivo analysis.

Setting: Hospital Sótero del Río. The magnetic resonance images were acquired using a Philips Achieva 1.5T scanner.

Participants: Twenty-five patients with repaired Tetralogy of Fallot who underwent cardiovascular magnetic resonance imaging requested by their referring physicians were included in this study.

Main outcome measures: Using a computational fluid dynamics simulation, we validated our pixel-wise method, quantifying the error of our method in comparison with the standard method. The patients underwent a standard two-dimensional phase-contrast magnetic resonance imaging acquisition for quantifying pulmonary artery flow. Pulmonary regurgitation fraction was estimated by using our pixel-wise and the standard method. The two-dimensional flow profiles were inspected looking for simultaneous antegrade and retrograde flows in the same cardiac phase. Statistical analysis was performed with t-test for related samples, Bland-Altman plots, and Pearson correlation coefficient.

Results: Estimation of pulmonary regurgitation fraction using the pixel-wise analysis revealed higher values compared with the standard method (39 ± 16% vs. 30 ± 22%, p-value <0.01). Eight patients (32%) had a difference of more than 10% between methods. Analysis of two-dimensional flow profiles in these patients revealed simultaneous antegrade and retrograde flows through the pulmonary artery during systole-early diastole.

Conclusion: Quantification of pulmonary regurgitation fraction in patients with repaired Tetralogy of Fallot through a pixel-wise analysis yields higher values of pulmonary regurgitation compared with the standard velocity-averaging method.

Keywords: Tetralogy of Fallot; magnetic resonance imaging; pulmonary valve insufficiency; two-dimensional flow.

Figure 1.

Quantification of pulmonary regurgitation. (a, b) Two perspectives of a flow curve where simultaneous antegrade (light gray area) and retrograde (dark gray area) flows are present from cardiac phases 8 to 16 (at time 0.23 to 0.47, respectively). Regurgitation fraction is the ratio of retrograde net flow (dark gray area) to antegrade net flow (light gray area). (a) The standard velocity-averaging method masks the retrograde flow from phases 8 to 10, because positive blood velocities have higher magnitudes than negative blood velocities. Regurgitation fraction is 10.7% in this case. (b) Since the pixel-wise analysis estimates antegrade and retrograde flows individually, retrograde flow from phases 8 to 10 is still quantified, although positive blood velocities have higher values than negative velocities. Regurgitation fraction is 19.2% in this case. (c) A cross-sectional example from ninth cardiac phase (green dashed lines in (a) and (b)) showing how blood flow is calculated by both methods. Light gray squares represent pixels with positive blood velocities and dark gray squares the negative blood velocities. In this example, the averaging process of the standard method determines that the positive velocities mask negative ones. This does not occur with the pixel-wise analysis. Note that the net flow is equal between methods, because we assumed a contour area that is equal to the sum of the pixel areas.

Figure 2.

3D model and velocity magnitude of CFD pulsatile simulation. (a) The geometry of the 3D Retrograde Flow Model. In (b) we show the velocity magnitude in systolic phase and additionally the reformatted 2D section that we transform in 2D images, in which we compared the regurgitation fraction obtained by our method with the standard method. We also show in (b) the velocity wave setting in the inlet face of the model.

Figure 3.

Flow wave curves. (a) Flow measurement with the standard method and (b) flow measurement with our pixel-wise method. Additionally, we show the volume of the forward and backward flow curve measurements for both methods. For (a) we calculated a pulmonary regurgitation fraction of 9.73%. For (b) we calculate a pulmonary regurgitation fraction of 28.67%.

Figure 4.

(a) Bland–Altman plot showing the mean difference between values of pulmonary regurgitation obtained by both methods. The graph demonstrates a systematic bias, with greater values of pulmonary regurgitation obtained by the pixel-wise analysis. (b, c) 2D flow profiles at 20% of the cardiac cycle from patients whose PR fraction showed a difference of more than 10% between methods. Images show simultaneous antegrade flow (arrow) and retrograde flow (arrow head) in the same cardiac phase. (b) 2D flow profile from a 14-year-old female (difference between methods = 23.5%). (c) 2D flow profile from a 20-year-old male (difference between methods = 37.7%).