Effect of additional dimensions and views in the echocardiographic determination of 3-dimensional left ventricular volume in myxomatous mitral valve disease in dogs

Background

Left ventricular (LV) volumes can be calculated from various linear, monoplane, and multiplane echocardiographic methods, and the same method can be applied to different imaging views. However, these methods and their variations have not been comprehensively evaluated against real-time 3-dimensional echocardiography (RT3D).

Hypothesis/Objectives

To identify the LV volumetric approaches that produce the least bias and the best agreement with RT3D, and to assess interoperator reproducibility between an experienced and an inexperienced operator.

Animals

Fifty-nine client-owned dogs with myxomatous mitral valve disease (38 Stage B1, 13 Stage B2, 8 Stages C/D) received echocardiograms, with a subset of 28 dogs (14 Stage B1, 10 Stage B2, 4 Stages C/D) imaged by 2 operators.

Methods

Prospective method comparison study. Body weight-indexed end-diastolic and end-systolic LV volumes using linear methods in long- and short-axis views (Teichholz, cube, modified cube), monoplane methods in right parasternal and left apical views (area-length and Simpson's method of discs), biplane Simpson's method of discs, and real-time triplane (RT3P) were compared against RT3D.

Results

The RT3P method exhibited no bias and demonstrated the highest agreement with RT3D. The linear methods showed significant bias and lower agreements for end-diastolic volumes, end-systolic volumes, or both. Volumes derived from different imaging views using the same method showed poor agreement. Both RT3P and RT3D methods demonstrated poor interoperator reproducibility.

Conclusions and Clinical Importance

Incorporating additional dimensions improves bias and agreement in LV volume quantification, but comprehensive clinical experience with RT3P and RT3D is needed to improve consistency across all operators.