Accuracy of absolute left ventricular volumes and cardiac output determined by radionuclide cardiography

We determined left ventricular (LV) volumes and derived variables by gated equilibrium radionuclide imaging at rest and during exercise in 12 patients without valve disease or intracardiac shunts. LV volume was determined as the product of the background-corrected LV count rate and an individual attenuation correction factor divided by the count rate in peripheral blood. Attenuation correction was based on measurement of LV depth within the chest from an initial first pass study in the left lateral view and a linear attenuation coefficient of 0.156 cm-1 determined in phantom studies. The average LV depth was 8.0 cm (range 6.9-9.1) in agreement with an average depth measured by echocardiography of 8.2 cm (6.3-9.4), P much greater than 0.05. The correlation between radionuclide (RC) and simultaneous thermodilution (TD) measurements was for cardiac output (CO): r = 0.95; CO (RC) = 1.00 X CO (TD) + 0.10 1/min with a standard error of the estimate (SEE) of 0.79 1/min; for stroke volume (SV): r = 0.90; SV(RC) = 0.93 X SV (TD) + 5 ml; SEE = 8 ml; for end-diastolic volume (EDV): r = 0.96; EDV(RC) = 1.06 X EDV(TD) -14 ml; SEE = 27 ml; and for end-systolic volume (ESV): r = 0.98; ESV(RC) = 1.05 X ESV (TD) -6 ml; SEE = 20 ml. The interobserver variation, expressed as the coefficient of variation, was for cardiac output 6%, for stroke volume 6%, for end-diastolic volume 4%, and for end-systolic volume 5%. This method permits non-invasive determination of LV volume and total LV output per beat based exclusively on data obtained during radionuclide imaging.