Abstract: Objective To evaluate the consistency and reproducibility of aortic root measurements by Anythink, a semi-automated preoperative CT analysis software, with those of 3mensio. Methods Sixty-seven patients undergoing transcatheter aortic valve replacement (TAVR) in the First Medical Center of Chinese PLA General Hospital from December, 2016 to February, 2022 were retrospectively enrolled in this study. A cardiology resident who completed his professional training used both the software Anythink and 3mensio (as the gold standard) to reconstruct the aortic root model and analyze the parameters of the aortic annulus and the surrounding structures. The correlation and consistency of the measurement results of two software were analyzed. Two independent residents also used Anythink software to repeat the measurements for the same patient for assessment of the reproducibility of Anythink measurements. The valve models were selected based on the measurements by Anythink and 3mensio, and similarities and differences of the two software in clinical valve selection were assessed. Results The measurements of the distances from the anulus plane to the left and right coronary ostium, average diameter of the anulus, anulus area, anulus perimeter, and the angle between the annulus and horizontal plane did not differ significantly between the two software (P>0.05), and their measurements showed positive correlations (r=0.884- 0.981, P<0.01). The intra-group and inter-group correlation coefficients of the anulus parameters measured by Anythink ranged from 0.894 to 0.992 and from 0.651 to 0.954, respectively. The Kappa-test values of valve models selected by Anythink and 3mensio based on the average diameter, area diameter and perimeter diameter were 0.886, 0.796 and 0.775, respectively. The intra-group Kappa values for the valve models selected based on Anythink measurements were 0.819, 0.841, and 0.795, and the inter-group Kappa values were 0.812, 0.812, and 0.768, respectively. Compared with the measurements by 3mensio, the recommended area diameter measured by Anythink was slightly greater in patients with postoperative paravalvular leakage, but slightly smaller in patients with postoperative new-onset conduction block. Conclusion Anythink has excellent measurement consistency and high reproducibility for aortic root measurements, and trained cardiologists can use Anythink to obtain accurate aortic root parameters before TAVR.

Key words: aortic valve disease; transcatheter aortic valve replacement; computed tomography; annulus sizing