AimsTranscatheter aortic valve replacement (TAVR) has become the standard of care for selected patients with severe aortic stenosis. Multidetector computed tomography (MDCT) and transoesophageal 2D/3D (two-dimensional/three-dimensional) echocardiography (ECHO) are used for aortic annulus (AA) sizing. The aim of this study was to compare the accuracy of AA sizing by ECHO versus MDCT for Edwards Sapien balloon expandable valve in a single center. Methods and resultsData from 145 consecutive patients with TAVR (Sapien XT or Sapien S3) were analyzed retrospectively. A total of 139 (96%) patients had favorable outcomes after TAVR (at most mild aortic regurgitation and only one valve implanted). The 3D ECHO AA area and area-derived diameter were smaller than the corresponding MDCT parameters (464 & PLUSMN; 99 vs. 479 & PLUSMN; 88 mm(2), p < .001, and 24.2 & PLUSMN; 2.7 vs. 25.0 & PLUSMN; 5.5 mm, p = .002, respectively). The 2D ECHO annulus measurement was smaller than both the MDCT and 3D ECHO area-derived diameters (22.6 & PLUSMN; 2.9 vs. 25.0 & PLUSMN; 5.5 mm, p = .013, and 22.6 & PLUSMN; 2.9 vs. 24.2 & PLUSMN; 2.7 mm, p < .001, respectively) but larger than the minor axis diameter of the AA derived from MDCT and 3D ECHO by multiplanar reconstruction (p < .001). The 3D ECHO circumference-derived diameter was also smaller than the MDCT circumference-derived diameter (24.3 & PLUSMN; 2.5 vs. 25.0 & PLUSMN; 2.3, p = .007). The sphericity index by 3D ECHO was smaller than that by MDCT (1.2 & PLUSMN; .1 vs. 1.3 & PLUSMN; .1, p < .001). In up to 1/3 of the patients, 3D ECHO measurements would have predicted different (generally smaller) valve size than was the valve size implanted with favorable result. The concordance of the implanted valve size with the recommended size based on preprocedural MDCT and 3D ECHO AA area was 79.4% versus 61% (p = .001), and for the area-derived diameter, the concordance was 80.1% versus 61.7% (p = .001). 2D ECHO diameter concordance was similar to MDCT (78.7%); however, MDCT AA area derived diameter extrapolated from 2D ECHO measurement would have led to the different valve size selection in additional 33% of the patients. Conclusions3D ECHO AA measurements are smaller than MDCT measurements. If 3D ECHO-based parameters alone are used to size the Edwards Sapien balloon expandable valve, then the selected valve size would have been smaller than the valve size implanted with favorable result in 1/3 of the patients. MDCT preprocedural TAVR assessment should be the preferred method over 3D ECHO in routine clinical practice to determine Edwards Sapien valve size.