Improved AI-Based Segmentation of Apical and Basal Slices from Clinical Cine CMR

Abstract Background Mitral valve regurgitation (MR) is currently primarily assessed by a multiparametric approach with transthoracic echocardiography (TTE) that can be further completed by 2D Cardiac Magnetic Resonance (2D CMR) in case of doubt or poor acoustic window. TTE and 2D CMR have nevertheless imperfect agreement in terms of MR quantification. Time-resolved phase-contrast cardiac magnetic resonance imaging with flow-encoding in three spatial directions (4D Flow CMR) could help in MR quantification. Purpose Compare 4D Flow CMR quantification of MR with TTE using a multiparametric approach. Methods We conducted a monocentric, prospective study at the Institut Mutualiste Montsouris in Paris between November 2016 and 2017 including patients with chronic primitive MR. MR was evaluated with a multiparametric approach by two cardiologists with TTE and quantitatively by two radiologists with 4D Flow CMR. MR was classified as mild, moderate or severe and evaluated blindly with consensus in case of disagreement. 4D Flow CMR measurements consisted in quantifying MR regurgitant volume (RV) and MR regurgitant fraction (RF). 4D anterograde mitral flow was compared to left ventricular stroke volume using 2D-cine CMR. Results 33 patients were included. Inter-observer agreement was good in TTE (kappa= 0.75 95% CI [0.57- 0.92]) and excellent in 4D Flow CMR (ICC= 0.94 95% CI [0.79–0.95]). Agreement with TTE was excellent using optimized thresholds (Mild: RV≤20mL RF≤20%, Moderate: RV=21–39mL RF=21–36%, Severe: RV≥40mL RF≥37%): kappa= 0.93 95% CI [0.8–1] for RV and kappa= 0.90 95% CI [0.7–0.9] for RF. A validation cohort confirmed that the 4D flow thresholds as determined were accurate for MR grading. Agreement between 4D anterograde mitral flow and 2D-cine CMR left ventricular stroke volume was also excellent (ICC= 0.92 95% CI [0.85–0.96]). Conclusion 4D Flow CMR is a reliable tool for MR quantification. It provides direct quantitative evaluation of MR with low inter-observer variability. It may therefore be used as a gatekeeper before therapeutic decisions such as surgery.

Download Full-text

477Partial fusion of two aortic valve leaflets is related to alterations in ascending aorta flow: 4D flow CMR study

European Heart Journal ◽

10.1093/eurheartj/ehz747.0127 ◽

2019 ◽

Vol 40 (Supplement_1) ◽

Author(s):

A Guala ◽

L Galian ◽

G Teixido Tura ◽

L Dux-Santoy ◽

A Ruiz Munoz ◽

...

Keyword(s):

Aortic Valve ◽

Aortic Arch ◽

Ascending Aorta ◽

Aortic Dilation ◽

4D Flow ◽

4D Flow Cmr ◽

Flow Displacement ◽

Aortic Valve Leaflets ◽

Partial Fusion ◽

Cine Cmr

Abstract Introduction Bicuspid aortic valve (BAV) is the most common congenital valve defect. It consists in the fusion of two aortic valve leaflets, and it is associated with a high prevalence of proximal aorta dilation. Dilation is highly prevalent (around 30%) in BAV patient relatives with a tricuspid valve (TAV) identified by echocardiography. However, the presence of partial aortic valve leaflet fusion (also called mini-raphe or forme fruste BAV, see figure 1A) is easily missed by echocardiography. A recent study reported that 44% of patients from a small cohort of BAV patient relatives with aortic dilation followed by CT showed mini-raphe. Purpose We aimed to use 4D flow CMR to assess if the presence of mini-raphe is associated with aortic flow alterations, which may be concurs in the etiology of aortic dilation in BAV patient relatives. Methods Twenty BAV patients first-degree relatives with partial fusion (<50%) of aortic valve leaflets and proximal aorta dilation were identified by CT or cine CMR and prospectively included. One-hundred twenty-five BAV and 95 patients with TAV from our prospective dataset of 4D flow CMR were included for comparison. Propensity score matching was used throughout the study to correct the comparisons between mini-raphe and BAV and mini-raphe and TAV patients for differences in age, maximum aortic diameter, sex, height, weight, proximal aortic pulse wave velocity and, only for BAV, fusion pattern. The hemodynamic parameters previously related to aortic dilation were computed. They were jet angle, normalized flow displacement and systolic flow reversal ratio (SFRR, identifying through-plane vortexes) were computed and compared in the ascending aorta and in the aortic arch. Results The presence of mini-raphe was statistically-significantly associated with increase in jet angle (Figure 1B), flow displacement (Figure 1C) and vortexes (Figure 1D) in most of the ascending aorta and aortic arch when mini-raphe patients were compared with TAV patients. The severity of flow asymmetry found in mini-raphe patients was lower than the one characteristic of BAV patients, but vortexes were even higher in a small region at the distal ascending aorta. Figure 1 Conclusion Partial fusion of the aortic valve leaflets is related to increase in proximal aorta flow eccentricity and vorticity. These flow abnormalities are not as marked as those associated with BAV. Data regarding prevalence of mini-raphe as evaluated with CT or cine CMR are needed, especially in familiar of BAV patients. Acknowledgement/Funding European FP7/People 267128; Spanish Ministry of Economy and Competitiveness RTC-2016-5152-1 and Instituto de Salud Carlos III PI14/0106

Download Full-text