Three-dimensional echocardiography: optimizing the amount of data acquired for determination of left ventricular ejection fraction

Abstract

Three-dimensional (3D) echocardiography has been validated as an accurate and reproducible means of measuring left ventricular volume in normal subjects and patients. The authors have recently completed a comparison of 3D echocardiography to gated radionuclide angiography for determination of left ventricular ejection fraction in unselected patients. The results showed that 3D echocardiography provided results comparable to radionuclide angiography (n=51, range 15-75%, r=0.94-0.97, SEE=3.6-5.4%). The limits of agreement analysis showed no bias for 3D echocardiography compared to radionuclide angiography and indicated that the results of the 2 tests would be within 10-13% (2 standard deviations) 95% of the time. Various two-dimensional (2D) echocardiography techniques were shown to have limits of agreement of 19-21%, or more than 50% greater. The authors thus conclude that 3D echocardiography is clearly superior to 2D echocardiography for measuring left ventricular ejection fraction when compared to equilibrium radionuclide angiography. Data acquisition for 3D echocardiography in this study involved using separate end-diastolic and end-systolic long axis reference images to acquire separate short axis image data sets for end-diastole and end-systole that account for systolic translation of the ventricle. The time currently required for acquisition of both sets is about 15-20 minutes. With the objective of reducing this time requirement, the authors hypothesized that a single set of data analyzed for both systole and diastole would not produce a significantly different result, and that separate data sets are unnecessary. To test this hypothesis the end-diastolic data sets of 30 patients were determined to be satisfactory for tracing at end-systole and were re-traced for both systole and diastole. >