AUTOMATED LEFT-VENTRICULAR ENDOCARDIAL BORDER DETECTION USING ACOUSTIC QUANTIFICATION IN CHILDREN

Objectives: The purpose of this study was to determine the reliability and accuracy of automated border detection using acoustic quantificat ion in children. Background: Acoustic quantification has shown promise in adult patients as a method for on-line estimation of left ventricu lar size and function. However, in children, the smaller ventricular s ize might magnify the importance of measurement error. Methods: We com pared the cross-sectional area and fractional area change of the left ventricle as measured on line by acoustic quantification with the area and fractional area change derived by hand-digitizing the endocardial border of the left ventricle off Line, both with and without the papi llary muscles included in the left ventricular cavity. Results: The ar eas and area change fractions from the true methods were highly correl ated, both with inclusion and exclusion of the papillary muscles for o ff-line analysis. However, the regression slope was closer to unity wh en the papillary muscles were excluded from the left ventricular cavit y during off-line digitization of the endocardial border. Analysis of agreement between the two methods showed good agreement for area measu rements and fair agreement for function measurements. The magnitude of the difference between the two methods for area measurement was direc tly proportional to the size of the ventricle. That is, the larger the ventricle the larger the difference between the area measurements by the two methods. Discussion: Automatic border detection using acoustic quantification appears to be an acceptable method for estimating the cross-sectional area and fractional area change of the left ventricle in children.