LAUSR

allows the calculation of the effective AVA, but the accuracy of this measurement has been criticized and is not part of core guidelines. Multidetector computed tomography (MDCT) supports AS diagnosis using valvular calcium scoring and MDCT can delineate aortic cusps, but anatomical AVA measure remains untested. Purpose: To compare the measurement of 4D-MDCT derived anatomic AVA, obtained with new, custom-made software, with effective AVA by transthoracic echocardiography (TTE) continuity equation. Methods: Twenty patients with severe AS and clinically indicated 4D-MDCT of the aortic valve were included. AVA was obtained using continuity equation for Doppler-Echocardiography. Using 4D-MDCT with contrast after imaging reregistration, custom semi-automated software allowed aortic cusp delineation and anatomical AVA measurement. With this software, a systolic 3D model of the valve is obtained after cusps’ profiling using 18 automatically generated long-axis planes (Figure, top panel). Then, orifice area (anatomic AVA) was automatically calculated using 3 different algorithms (Figure, bottom panel): by using smallest 2D-projection of aortic cusps profile (method A), by computing 2D area of cusps’ free margin (blue) (B), and by using any plane (yellow) with the smallest area between cusps (C). Results: In 18 out of 20 patients (80%) MDCT image quality allowed complete delineation of aortic cusps. AVA by Doppler-Echo was 82±15 mm2. Anatomic AVA measured 80±16 mm2 for method-A, 88±20 mm2 for method-B, 93±21 mm2 for method-C, and 87±19 mm2 when averaging over the three methods. Absolute differences between Echo and 4D-MDCT measurements were 7.7±4.6 mm2 (p for difference=0.3; r=0.85, p<0.0001) for method-A; 9.2±6.8 mm2 (p for difference= 0.074; r=0.86, p<0.0001) for methods-B; 11.9±9.1 mm2 (p for difference=0.0002; r=0.90, p<0.0001) for method-C, and 7.7±5.9 mm2 (p for difference=0.03; r=0.89, p<0.0001) for the average. Analysis of regression slopes >1 (echo lower than MDCT) was observed for methods-B, -C, and -average (1.26, 1.13, and 1.11 respectively) but was 0.93 for method-A.