The nature of the background in radionuclide ventriculography: deductions from the mathematical behaviour of the ejection fraction.

The relationship between the ejection fractions calculated from 'uncorrected' radionuclide time activity curves (UEF) and angiographic ejection fractions (AEF) in 200 catheterized patients yielded the regression equation AEF = 1.74 UEF + 0.21. It follows from this linear relationship that the left ventricular ejection fraction can be estimated by linear regression without explicit background correction: RREF = 1.74 UEF + 0.21, where RREF is the radionuclide regression ejection fraction. We first investigated the possibility that changes in photon self-attenuation within the cardiac chambers cause the observed mathematical characteristics of the cardiac background, B. Self-attenuation was calculated for cylindrical and spherical ventricular models. The results were insensitive to the particular geometry and would have only a small effect on the observed EF. Alternatively, the 'background' may result from extra-ventricular radiation scattering from the heart into the detector. If we assume that B should be proportional to the ventricular scattering volume, Bd = Kd EDC for diastole and Bs = Ks ESC for systole, the background corrected ejection fraction will be BCEF = K UEF + (1-K) where K = (1-Ks)/(1-Kd). This agrees with the form of the empirical regression equation.