How does two-dimensional echocardiography compare to cineventriculography in measuring left ventricular volumes and ejection fraction?
Two-dimensional echocardiography systematically underestimates left ventricular volumes compared to cineventriculography due to tangential cuts, slice-thickness artifacts, and differences in contour outlining.
Two-dimensional echocardiography underestimates left ventricular volume compared with cineventriculography. To exclude the influence of difference in heart rate, blood pressure, respiration phases and any effect of the contrast material on left ventricular function, simultaneous studies of two-dimensional echocardiography and cineventriculography-echoventriculography were performed in 46 patients. Apical two-dimensional echocardiograms in the right anterior oblique (RAO) equivalent view were recorded before and during cineventriculography in the 30 degrees RAO projection. End-diastolic and end-systolic volumes (EDV and ESV) were calculated using a disc method with a semiautomatic computer system. The echo transducer position relative to the left ventricular apex and long axis was analyzed. For EDV determined by two-dimensional echocardiography and cineventriculography, the linear regression equation was y = 0.659x + 0.8, SEE = +/- 26.2 ml, r = 0.907. For ESV, the regression equation was y = 0.571x + 17.8, r = 0.938, SEE = +/- 18.6 ml, and for ejection fraction (EF) it was y = 0.606x + 13.0, r = 0.803, SEE = +/- 9.1%. Injection of contrast material resulted in only a small increase of stroke volume, caused by an increase of EDV as analyzed by echoventriculography. In all but two patients, the transducer position was found to be anterior and superior to the left ventricular anatomic apex, as evaluated by filming the echo transducer position during cineventriculography in 46 patients in the 30 degrees RAO projection and in 15 patients consecutively in the 60 degrees left anterior oblique and 30-40 degrees cranial projections. Thus, tangential cuts of the ventricle resulted in underestimation of diameters, long axis and ventricular volumes. These methodologic problems are exacerbated by slice-thickness artifacts. Furthermore, different outlining of left ventricular contour -- outer border of ventricular trabeculae for cine ventriculography and inner border for two-dimensional echocardiography -- seemed to result in underestimation of volume by echocardiography.
Erbel et al. (1983) studied this question.