AIM 
Aortic root dilatation is associated with a high risk of mortality. Transthoracic echocardiography (TTE) is the first-line imaging method for measurements of the aortic root. However, cardiac magnetic resonance imaging (CMR) has become frequently available and allows for accurate measurements of the aortic root as well. We aimed to evaluate the comparability of aortic root measurements by standardized TTE and CMR imaging in a large, representative sample of the general population.

METHODS AND RESULTS          
Our study included 744 subjects (mean age 63.5 ± 8 years, 43.7% female) from the prospective Hamburg City Health Study (HCHS). All subjects underwent balanced steady-state free precession long-axis 3-chamber left ventricular (3-CH-LAX) and left ventricular outflow tract (LVOT) cine-view CMR and parasternal long-axis TTE (Figure). Aortic root measurements were performed in end-diastole at three levels: aortic annulus, Sinuses of Valsalva (SoV) and sinotubular junction (STJ). TTE measurements were performed from leading-edge to leading-edge (LL) and from inner-edge to inner-edge (II) while CMR measurements were performed II. Statistics included median and interquartile range of the diameter to show the distribution, spearman correlation coefficients (r), Bland-Altman plots (mean difference ± 2.576 SD [99%-CI]) to evaluate the relationship between CMR and TTE and intraclass correlation coefficients (ICC [95%-CI]) to analyze the inter-observer agreement.
CMR II measurements showed a stronger correlation when compared to TTE II vs TTE LL measurements (e.g. SoV TTE II and CMR LAX II vs TTE LL and CMR LAX II: r =0.83, p < 0.001 vs r = 0.79; p < 0.001). TTE LL vs II measurements overestimated aortic root dimensions when compared to CMR II measurements (SoV CMR LAX II vs. TTE, LL: bias= 1.5 ± 6.1 mm; II: bias = 0.1 ± 6.1 mm).  While the SoV and STJ showed a high degree of agreement with only minor bias between CMR and TTE, the agreement for the aortic annulus was fair (TTE II vs. CMR LAX II: STJ: r = 0.769; p < 0.001; bias = 0.8 ± 5.9; Annulus: r = 0.53, p<0.001; bias = -1.3 ± 6.6). 
ICCs of ≥ 0.9 demonstrated appropriate inter- and intraobserver reproducibility for all CMR and TTE measurements of the SoV and STJ (SoV II interobserver:TTE ICC = 0.96 [0.94, 0.98]; CMR LAX ICC = 0.98 [0.97;0.99] STJ II interobserver: TTE ICC = 0.96 [0.89, 0.99]; CMR LAX ICC = 0.78 [0.23;0.95]). In opposite, lower ICCs (Annulus II interobserver: TTE ICC = 0.76 [0.63;0.84]; CMR LAX ICC = 0.77 [0.22; 0.95]) in both TTE and CMR indicate lower reliability and reproducibility of measurements of the aortic annulus.

CONCLUSION  Aortic root measurements in standardized CMR sequences are comparable to standard-of-care TTE derived diameters. The highest agreement was at the level of the SoV and STJ, while the agreement at the aortic annulus was only fair. Comparability improved if II instead of LL TTE measurements were performed challenging the standard LL-convention in aortic root echocardiography. These results might support a reduction of redundant multimodality imaging in patients with suspected aortic root diseases.