A new method for determining ejection fraction by two-dimensional echocardiography was assessed in 60 patients undergoing angiography. In method A, the left ventricular minor axis was measured at the midventricular cavity level in end-systole and end-diastole using the apical four chamber view in the 60 patients. The left ventricular major axis was also measured from the left ventricular apex to the base of the mitral valve at end-systole and end-diastole. The ejection fraction was determined using a modified cylinder-ellipse algorithm. In method B, measurements of the left ventricular minor axis were made in 40 consecutive patients, at the upper, middle and lower thirds of the left ventricular cavity at end-systole and end-diastole of the same cardiac cycle and left ventricular major axis was measured as in method A. With use of the same algorithm, three regional ejection fractions were determined and averaged to yield the total ejection fraction. The two echocardiographic methods were compared with single plane cineangiography in all patients and with gated nuclear scanning in 14 patients. Reproducibility was assessed by interobserver comparison. Correlation was determined in all patients and then separately for those with echocardiographic wall motion abnormalities. The correlation coefficient for all patients was 0.79 (probability [p] < 0.001) for method A and 0.90 (p < 0.001) for method B. For patients with wall motion abnormalities, method A had a correlation coefficient of 0.38 (p < 0.1) and method B showed much higher correlation with r = 0.82 (p < 0.001). Corresponding values for methods A and B in patients without wall motion abnormality were 0.85 (p < 0.001) and 0.88 (p < 0.001), respectively. Unlike a previous study, this method directly measures fractional shortening of left ventricular major axis and ejection fraction values are not arbitrarily modified by type of wall motion abnormality. With this method, accurate measurement of ejection fraction can be made by two-dimensional echocardiography without planimetry. In the absence of echocardiographic wall motion abnormalities, a very simple method A suffices. If wall motion abnormalities are present, the regional ejection fraction method B provides excellent results. © 1983, American College of Cardiology Foundation. All rights reserved.