Differences in Two- and Three-Dimensional Assessment of the Mitral Valve by Novices and Experts, Illustrated Using Anterior Mitral Valve Leaflet Length

Abstract

In this measurement validation study, the authors evaluated agreement between 2-dimensional (2D) and three-dimensional (3D) transesophageal echocardiography (TEE), measuring anterior mitral valve leaflet length by both novice and experienced echocardiographers. This was a retrospective, observational study. Single university hospital. Analyses on datasets from 44 patients. None. Fifty datasets from 44 patients with mitral regurgitation were analyzed by 4 observers (2 novices, 2 experts). All observers measured the anterior mitral valve leaflet length from end-systolic 2D TEE images from the midesophageal longitudinal axis view and 3D software-augmented TEE images. The overall mean anterior mitral valve leaflet length was significantly shorter with 3D versus 2D TEE measurements (24.6 ± 4.5 mm v 26.2 ± 5.3 mm; p < 0.001), with novices measuring shorter leaflets than experts for both techniques (p < 0.001 and p = 0.005, respectively). Bland-Altman plots of 3D and 2D TEE measurements showed mean biases (95% limits of agreement) of -1.6 mm (-9.0 to 5.9 mm), -1.8 mm (-9.6 to 6.0 mm), and -1.3 mm (-8.4 to 5.7 mm) for all observers, novices, and experts, respectively. For 2D measurements, interobserver reliability was very strong among experts and strong among novices (Pearson's r = 0.83 v 0.66; p = 0.055). For 3D measurements, interobserver reliability was strong in experts and moderate in novices (Pearson's r = 0.69 v 0.51; p = 0.168). For both novices and experts, 3D TEE measurements of the anterior mitral valve leaflet were significantly shorter than 2D measurements. Interobserver reliability was lowest for novices making 3D TEE measurements, indicating that reliable, quantitative evaluation of 3D TEE may require a greater amount of practice.