Attempts to measure left ventricular stroke volume utilizing the Doppler aortic flow method have found varying correlations between invasive thermodilution and noninvasive Doppler methods. Because stroke volume is the product of the Doppler flow velocity integral (that is, the area under the flow velocity curve) and the cross-sectional area of the vessel through which blood flows, both variables are potential sources of error. Previous studies have shown that the Doppler flow velocity integral can be measured with acceptable reproducibility in the ascending aorta. Consequently, in this study an attempt was made to determine empirically the optimal method for measuring aortic diameter and area. The diameter of the ascending aorta was measured utilizing four Mmode and seven two-dimensional echocardiographic conventions. Doppler aortic flow velocity patterns were recorded with a 2.25 MHz M-mode echocardiographic transducer from the suprasternal notch by mapping the ascending aorta until aortic peak flow velocity was recorded. In 19 adult patients undergoing cardiac catheterization for clinical indications, Doppler stroke volume estimates utilizing the various echocardiographic conventions for measuring aortic root diameter and area were compared with simultaneous measurements of stroke volume by the thermodilution technique. The best correlation (r = 0.87) with thermodilution stroke volume was obtained by estimating aortic area from the two-dimensional parasternal long-axis images with the aortic dimension measured distal to the aortic sinuses from the inner to inner wall. The data were related by the equation: Thermodilution stroke volume=(0.73)×(two-dimensional Doppler stroke volume)+17cc. M-mode echographic measurements of aortic diameter resulted in Doppler estimates of left ventricular stroke volume that correlated poorly (r < 0.3) with thermodilution estimates. © 1985, American College of Cardiology Foundation. All rights reserved.