RESUMO
BACKGROUND: New postprocessing software facilitates 3-dimensional (3D) echocardiographic determination of mitral annular (MA) and neo-left ventricular outflow tract (neo-LVOT) dimensions in patients undergoing transcatheter mitral valve replacement (TMVR). OBJECTIVES: This study aims to test the accuracy of 3D echocardiographic analysis as compared to baseline computed tomography (CT). METHODS: A total of 105 consecutive patients who underwent TMVR at 2 tertiary care centers between October 2017 and May 2023 were retrospectively included. A virtual valve was projected in both baseline CT and 3D transesophageal echocardiography (TEE) using dedicated software. MA dimensions were measured in baseline images and neo-LVOT dimensions were measured in baseline and postprocedural images. All measurements were compared to baseline CT as a reference. The predicted neo-LVOT area was correlated with postprocedural peak LVOT gradients. RESULTS: There was no significant bias in baseline neo-LVOT prediction between both imaging modalities. TEE significantly underestimated MA area, perimeter, and medial-lateral dimension compared to CT. Both modalities significantly underestimated the actual neo-LVOT area (mean bias pre/post TEE: 25.6 mm2, limit of agreement: -92.2 mm2 to 143.3 mm2; P < 0.001; mean bias pre/post CT: 28.3 mm2, limit of agreement: -65.8 mm2 to 122.4 mm2; P = 0.046), driven by neo-LVOT underestimation in the group treated with dedicated mitral valve bioprosthesis. Both CT- and TEE-predicted-neo-LVOT areas exhibited an inverse correlation with postprocedural LVOT gradients (r2 = 0.481; P < 0.001 for TEE and r2 = 0.401; P < 0.001 for CT). CONCLUSIONS: TEE-derived analysis provides comparable results with CT-derived metrics in predicting the neo-LVOT area and peak gradient after TMVR.