Assessment of altered three-dimensional blood characteristics in aortic disease by velocity distribution analysis

Abstract

Purpose: To test the feasibility of velocity distribution analysis for identifying altered three-dimensional (3D) flow characteristics in patients with aortic disease based on 4D flow MRI volumetric analysis.

Methods: Forty patients with aortic (Ao) dilation (mid ascending aortic diameter MAA = 40 ± 7 mm, age = 56 ± 17 years, 11 females) underwent cardiovascular MRI. Four groups were retrospectively defined: mild Ao dilation (n = 10; MAA < 35 mm); moderate Ao dilation (n = 10; 35 < MAA < 45 mm); severe Ao dilation (n = 10; MAA > 45 mm); Ao dilation+aortic stenosis AS (n = 10; MAA > 35 mm and peak velocity > 2.5 m/s). The 3D PC-MR angiograms were computed and used to obtain a 3D segmentation of the aorta which was divided into four segments: root, ascending aorta, arch, descending aorta. Radial chart displays were used to visualize multiple parameters representing segmental changes in the 3D velocity distribution associated with aortic disease.

Results: Changes in the velocity field and geometry between cohorts resulted in distinct hemodynamic patterns for each aortic segment. Disease progression from mild to Ao dilation + AS resulted in significant differences (P < 0.05) in flow parameters across cohorts and increased radial chart size for root and ascending aorta segments by 146% and 99%, respectively.

Conclusion: Volumetric 4D velocity distribution analysis has the potential to identify characteristic changes in regional blood flow patterns in patients with aortic disease.

Keywords: 4D flow MRI; aortic dilation.

FIGURE 1. Workflow of hemodynamic velocity distribution analysis

4D flow MRI data (time-resolved magnitude images and velocity vector components along (X, Y, Z)) were used to generate a 3D PC-MR angiogram and a 3D segmentation of the aorta which was divided into four segments: Segment 1 (aortic root in yellow) left ventricular outflow tract to the sino-tubular junction; Segment 2 (ascending aorta in orange) sino-tubular junction to the aortic arch; Segment 3 (aortic arch in magenta) aortic arch; Segment 4 (descending aorta in green) proximal to distal descending aorta. Segments were used to mask 4D flow velocity field in order to perform volumetric velocity distribution analysis. An example of the velocity distribution analysis for Segment 2 is shown in a patient with severe aortic dilation (patient #4, Fig. 2). Segmental aortic volume and velocity histogram characteristics are used to generate a radial chart display. sd: standard deviation; kur: kurtosis; ske: skewness.

FIGURE 2. Systolic velocity maximum intensity projection

The masked aorta velocity fields of three highest systolic velocity frames were used to generate a maximum intensity projection (MIP) in an oblique sagittal plane for all n=40 subjects. AS: aortic stenosis.

FIGURE 3. Cohort averaged histograms for Segment 1

Illustration of flow characteristic differences between the four patient cohorts is provided in panel A-D. Panels A-C show the cohort averaged histogram for the entire cohort for mild Ao dilation, moderate, and severe Ao dilation, respectively. Panel D shows the cohort averaged histogram for the entire cohort for Ao dilation + AS. The red line represents the median velocity, the green line represents the mean velocity, and the pink line represents the incidence threshold set at 1 m/s. Ao: aortic; AS: aortic stenosis; sd: standard deviation; kur: kurtosis; ske: skewness.

FIGURE 4. Radial chart display of the histogram characteristics

Radial chart displays provide a visual impression of the histogram characteristics for each aortic segment (1-4) using averaged values for each cohort and standard deviation (SD). Charts are in arbitrary units (AU). Blue band: SD for mild Ao Dilation; Green band: SD for moderate Ao Dilation; Yellow band: SD for severe Ao Dilation; Red band: SD for Ao Dilation + AS; Ao: aortic; AS: aortic stenosis. *: P<0.05 between groups using ANOVA.

FIGURE 5. Bland-Altman for intra-observer variability for all parameters included in the radial chart display

Panel A-G show the plots for segment volume, median velocity, mean velocity, standard deviation velocity, incidence above 1 m/s, skewness and kurtosis, respectively.

FIGURE 6. Bland-Altman for inter-observer variability for all parameters included in the radial chart display

Panel A-G show the plots for segment volume, median velocity, mean velocity, standard deviation velocity, incidence above 1 m/s, skewness and kurtosis, respectively.