Evaluation of reference systems for quantitative wall motion analysis from three-dimensional endocardial surface reconstruction: an echocardiographic study in subjects with and without myocardial infarction.

Abstract

Six relevant computer-implemented reference systems for three-dimensional quantitative assessment of left ventricular wall motion abnormalities were compared with visual wall motion analysis of two-dimensional images. Endocardial borders were traced in three apical echocardiographic views at end-diastole and end-systole in 10 patients with myocardial infarction and 5 healthy subjects, and three-dimensional reconstruction of endocardial surfaces was performed. End-diastolic and end-systolic surfaces were aligned in a common axis system depending on the reference system, and systolic wall motion was assessed at 1,024 points on the endocardial surface. The localization of abnormal wall motion was displayed in bull's-eye maps, and the area was determined as a percentage of total endocardial area. For each reference system, the segmental concordance between three-dimensional computerized and visual assessment was determined. The best agreement between computerized and visual analysis was obtained with a reference system based on wall motion towards the major ventricular axis, whereas the poorest result was obtained using the center of left ventricular cavity as reference. Correlation between the estimated area of wall motion abnormality and visually determined wall motion score index was best using the aligned center of mitral valve plane as reference (r = .92).