Echocardiographic automated border detection can provide on-line estimates of left ventricular cavity area by differentiating blood from tissue backscatter characteristics. The objective of this study was to assess the ability of short-axis measurements of left ventricular cavity area by automated border detection to determine left ventricular function by comparing these measurements to radionuclide measures of ejection fraction in the same patients. Eighty-eight consecutive patients, aged 53 ± 14 years, underwent automated border detection studies within 2 hours of radionuclide ventriculography. Short-axis imaging with automated border detection was attempted at basal, midpapillary muscle, and apical levels. Maximal left ventricular length was also measured from apical 4- and 2-chamber views by standard imaging. Fractional area change - (end-diastolic area-end-systolic area)/end-diastolic area - was determined at each short-axis level. Volumes and ejection fractions were calculated using: volume = 5 6 (midventricular area) · length. Simpson's rule for 3 short-axis measurements was calculated using: volume = (length/12) (5 · basal area + 2 · mid-area + 4 · apical area). Technically adequate automated border detection data could be obtained on 69 patients (78%) at basal and mid-levels, and at all 3 short-axis levels in 66 patients (75%). Correlations with radionuclide ejection fraction were as follows: midventricular fractional area change - R = 0.84, SEE =12%, y = 0.86 × - 7; area-length ejection fraction - R = 0.89, SEE = 9%, y = 0.96 × - 4; and Simpson's rule - R = 0.91, SEE = 8%, y = 0.89 × + 1. In conclusion, automated border detection data from short-axis views may be used as a method to quantify left ventricular function in selected patients.