A realistic dynamic cardiac phantom for evaluating radionuclide ventriculography: description and initial studies with the left ventricular chamber.

A phantom was devised to validate scintigraphically determined left ventricular ejection fractions (LVEFs) and cardiac chamber volumes in the following simulated cardiac situations: normal contraction, moderately impaired left ventricular contraction, severely impaired left ventricular contraction, mitral regurgitation, and cardiomyopathy. The phantom, assembled from anatomically realistic cardiac chambers, simulated contraction and expansion using individual chamber pumps coordinated by a microcomputer. Scintigraphic studies were performed by sequential imaging of [99mTc]pertechnetate introduced into each chamber. The images were analyzed like conventional clinical studies, using both automatic and manual techniques. Scintigraphic techniques correlated with chamber volumes that were determined by weight to yield the following regression formulae: LVEF (by automatic method 1) = 1.08 x LVEF (by weight) -5.11; LVEF (by automatic method 2) = 1.00 x LVEF (by weight) -3.15; and LVEF (by manual method) = 1.04 x LVEF (by weight) -5.08 ml (Correlation coefficients greater than 0.98). The absolute left ventricular volumes (LVVs), determined by scintigraphy, correlated well with LVVs determined by weight. These correlations were performed with separations between the center of the left ventricle and the collimator varying from 5 cm to 9 cm. The regression formulae for 5, 7, and 9 cm distances were: LVV (by counts) = 0.99 x LVV (by weight) + 0.13, LVV (by counts) = 1.04 x LVV (by weight) + 9.08, LVV (by counts) = 0.88 x LVV (by weight) + 15.25, respectively. At 9 cm, slight volumetric underestimation occurred, as predicted from the work of Fearnow et al., possibly because of oversubtraction of background. Thus, this phantom provides a useful tool for validating scintigraphic cardiac blood-pool studies simulating a wide range of clinically relevant situations.