Assessment of diastolic left ventricular filling by echocardiographic automated border detection and comparison with radionuclide ventriculography.

To determine whether indexes obtained from a newly developed echocardiographic automated border detection (ABD) technology provide a reliable estimate of left ventricular (LV) diastolic filling, ABD variables of LV filling were compared with volumetric measurements determined by radionuclide angiography. Forty-two patients with a variety of heart diseases (age range, 11 to 76 years) underwent ABD echocardiographic studies on the same day as the radionuclide examination. Technically adequate ABD data could be obtained in 31 patients (74%). Nineteen healthy subjects served as normal controls. Area-time and volume-time waveforms for echocardiographic measurements were obtained from LV short-axis views at the level of the papillary muscles and four-chamber apical views. Both the diastolic indexes derived from the waveform of area change (short-axis view) and volume change (four-chamber apical view) correlated with radionuclide variables. Values measured from the ABD area-time waveform showed the following correlations: peak filling rate (r = 0.86; standard error of the estimate [SEE] = 0.62), time to peak filling rate (r = 0.85; SEE = 23.11), rapid filling phase fractional change (r = 0.79; SEE = 5.51), and atrial filling phase fractional change (r = 0.71; SEE = 5.82). Correlations of indexes derived from the ABD volume-time waveform were as follows: peak filling rate (r = 0.87; SEE = 0.50), time to peak filling rate (r = 0.90; SEE = 22.03), rapid filling fractional change (r = 0.83; SEE = 5.33), and atrial filling fractional change (r = 0.77; SEE = 4.68). ABD LV filling parameters in patients with heart disease and normal control subjects were significantly different. Thus ABD data from short-axis and apical views have a strong linear relation with radionuclide ventriculographic measurements and may be used as a method to assess LV diastolic filling.