Two-dimensional Echocardiographic Automated Border Detection Accurately Reflects Changes in Left Ventricular Volume

Abstract

The objective of this study was to determine the relationship of on-line measurements of left ventricular cavity area generated by echocardiographic automated border detection to true volume measured by an intraventricular balloon in an isovolumically contracting isolated canine heart preparation. Seven excised dog hearts had placement of an intraventricular balloon and were perfused in an ex vivo apparatus. Left ventricular area data from the midventricular short-axis plane and pressure data were recorded on a computer through a customized hardware and software interface with the ultrasound system. Left ventricular volumes were varied from 5 ml to maximal values (30 to 40 ml) at 1-milliliter increments. Three increasing and decreasing volume ramps were analyzed on each of seven hearts for a total of 1260 simultaneous measurements. Linear regression analysis correlated mean automated border detection area with absolute volume from each preparation. A predominantly linear relationship was observed with an average correlation of r = 0.97 (y = 0.16x-0.69, SEE = 0.31 cm2, p < 0.01). Left ventricular area measures for six of seven dogs varied little during isovolumic contraction (< 0.4 cm2) but did show a systematic cardiac cycle-related variability in one dog (28% change in area, maximum to minimum, over all volumes). In conclusion, the relationship between cross-sectional area and left ventricular volume was predominantly linear and varied little during isovolumic contractions in the normal canine left ventricle. Echocardiographic automated border detection appears to be a promising method to reflect changes in left ventricular volume.