Accuracy of radionuclide ventriculography for estimation of left ventricular volume changes and end-systolic pressure-volume relations

Estimation of left ventricular end-systolic pressure-volume relations depends on the accurate measurement of small changes in ventricular volume. To study the accuracy of radionuclide ventriculography, paired radionuclide and contrast ventriculograms were obtained in seven dogs during a control period and when blood pressure was increased in increments of 30 mm Hg by phenylephrine infusion. The heart rate was held constant by atropine infusion. The correlation between radionuclide and contrast ventriculography was excellent. In the individual animals, the average r value for left ventricular volume was 0.96 +/- 0.03 (+/- SD) (p = 0.001, n = 7) and the mean r value for end-systolic volume changes was 0.90 +/- 0.08 (n = 7, range 0.76 to 0.99). For the entire series, there were 33 end-systolic volume changes, and there was an equally strong radionuclide-contrast correlation (r = 0.89, p less than 0.001, n = 33), even though the volume changes averaged only 11.9 +/- 8.2 ml (range 0.3 to 38.1). The systolic pressure-volume relations were linear for both radionuclide and contrast ventriculography (r = 0.98 and 0.97, respectively, n = 7). The mean slope for radionuclide ventriculography (2.9 +/- 1.4) was lower than the mean slope for contrast ventriculography (4.8 +/- 1.7) (p = 0.004); however, the slopes correlated well (r = 0.81, n = 7, p = 0.026). The radionuclide-contrast volume relation was compared using background subtraction, attenuation correction, neither of these or both. By each method, radionuclide ventriculography was valid for measuring small changes in left ventricular volume and for defining end-systolic pressure-volume relations.