933-96 LV Mass Determination by 3D Echocardiography: Initial Clinical Experience and Comparison to Magnetic Resonance Imaging in Patients with Abnormal Ventricular Geometry

Abstract

Measurement of LV mass and its regression is important in dilated and congenitally deformed ventricles as well as in those with normal shape and concentric hypertrophy. Conventional echo methods for mass estimation use geometric and image plane positioning assumptions that may be invalid in patients with abnormally shaped ventricles. 3D echo eliminates these assumptions and has been previously validated for LV mass determination in normally shaped ventricles (r = 0.93, SEE = 9–11 g). In the same study Penn convention and 2D echo methods had standard errors greater than two times larger than 3D echo. To compare LV mass determination by 3D echo and MRI in patients with abnormal ventricular shape. 30 patients underwent 3D echo and short-axis gradient reversal MRI within 6 hours. 3D echo data sets of 8-10 short axis, non-parallel, non-intersecting short axis crosssections were obtained using an acoustic spatial locater and a line of intersection display for guidance. End-diastolic image data sets were obtained, boundaries were traced manually and mass computed. Etiologies of abnormal ventricular shape included ischemic, idiopathic, congenital, valvular and alcoholic cardiac disease. Results were compared by linear regression and limits of agreement analysis. Linear Regression: Range: 86.3-400.3 g; Mean = 205.7 ± 82.8 r = 0.82; SEE = 34.9 g; P < 0.001 3D ECHO = 0.59 MRI + 72.8 g Limits of Agreement: Bias (mean of differences) = -11.4 g Limits (2 SD of differences) = 96.4 g Left ventricular mass determination by 3D echo is clinically feasible. Initial clinical experience in unselected patients with abnormal ventricles demonstrates good correlation and standard error, small bias and acceptable limits of agreement with MRI. 3D echo offers a new means for clinical assessment of LV mass