2D cine vs. 3D self-navigated free-breathing high-resolution whole heart cardiovascular magnetic resonance for aortic root measurements in congenital heart disease

Abstract

Background: Cardiovascular magnetic resonance (CMR) is considered the method of choice for evaluation of aortic root dilatation in congenital heart disease. Usually, a cross-sectional 2D cine stack is acquired perpendicular to the vessel's axis. However, this method requires a considerable patient collaboration and precise planning of image planes. The present study compares a recently introduced 3D self-navigated free-breathing high-resolution whole heart CMR sequence (3D self nav) allowing a multiplanar retrospective reconstruction of the aortic root as an alternative to the 2D cine technique for determination of aortic root diameters.

Methods: A total of 6 cusp-commissure (CuCo) and cusp-cusp (CuCu) enddiastolic diameters were measured by two observers on 2D cine and 3D self nav cross-sectional planes of the aortic root acquired on a 1.5 T CMR scanner. Asymmetry of the aortic root was evaluated by the ratio of the minimal to the maximum 3D self nav CuCu diameter. CuCu diameters were compared to standard transthoracic echocardiographic (TTE) aortic root diameters.

Results: Sixty-five exams in 58 patients (32 ± 15 years) were included. Typically, 2D cine and 3D self nav spatial resolution was 1.1-1.5² × 4.5-7 mm and 0.9-1.15³ mm, respectively. 3D self nav yielded larger maximum diameters than 2D cine: CuCo 37.2 ± 6.4 vs. 36.2 ± 7.0 mm (p = 0.006), CuCu 39.7 ± 6.3 vs. 38.5 ± 6.5 mm (p < 0.001). CuCu diameters were significantly larger (2.3-3.9 mm, p < 0.001) than CuCo and TTE diameters on both 2D cine and 3D self nav. Intra- and interobserver variabilities were excellent for both techniques with bias of -0.5 to 1.0 mm. Intra-observer variability of the more experienced observer was better for 3D self nav (F-test p < 0.05). Aortic root asymmetry was more pronounced in patients with bicuspid aortic valve (BAV: 0.73 (interquartile (IQ) 0.69; 0.78) vs. 0.93 (IQ 0.9; 0.96), p < 0.001), which was associated to a larger difference of maximum CuCu to TTE diameters: 5.5 ± 3.3 vs. 3.3 ± 3.8 mm, p = 0.033.

Conclusion: Both, the 3D self nav and 2D cine CMR techniques allow reliable determination of aortic root diameters. However, we propose to privilege the 3D self nav technique and measurement of CuCu diameters to avoid underestimation of the maximum diameter, particularly in patients with asymmetric aortic roots and/or BAV.

Keywords: 3D self-navigation whole heart; Aortic root dilatation; Bicuspid aortic valve; Congenital heart disease.

Fig. 1

The doubly-oblique cross-sectional stack of 2D cines was planned on two orthogonal cines of the left ventricular outflow tract (LVOT)  (a). For 3D self nav, a multiplanar reconstruction of the aortic root plane was performed (b). Cusp to the opposite commissure (CuCo) (c) and cusp to cusp (CuCu, d) inner to inner edge diameters were determined. For bicuspid aortic valve (BAV) patients in whom identification of three cusps was not possible, the maximum diameter and a second, orthogonal diameter were obtained (e)

Fig. 2

Comparison of image quality of 2D cine (red bars) and 3D self nav (blue bars). There were no differences of image quality neither between techniques nor between observers (p > 0.05)

Fig. 3

Upper panels: Bland–Altman plots for the comparison of the maximum cusp to commissure (CuCo) (a) and cusp to cusp (CuCu) (b) diameters between 2D cine vs. 3D self nav. Lower panels: Comparison maximum CuCo to CuCu diameters measured on 2D cine (c) and 3D self nav (d). The blue line represents the mean bias and the doted red lines indicate the 95% confidence interval. CuCo cusp to commissure, CuCu cusp to cusp