Count-based scintigraphic method to calculate ventricular volumes in children: in vitro and clinical validation.

A "phantom" was used to validate 1) estimates of different depths of a constant radioactivity source, and 2) the calculation of different volumes using a constant depth and different attenuation coefficients. Using data from this in vitro study, scintigraphic estimates of right ventricular volume and ejection fraction were compared with those obtained by cineangiography in 36 children with either a normal right ventricle or various right ventricular diseases. The static program accurately estimates the distance from the radiation source to the collimator surface (r = 0.99). Radionuclide count methods best predict "phantom" volumes using attenuation coefficients between 0.11(-1) and 0.13(-1) cm. A coefficient of 0.10(-1) underestimates, whereas 0.15(-1) cm grossly overestimates actual volumes. In vivo data were therefore analyzed using an attenuation coefficient of 0.11(-1) with right ventricular counts corrected using either right ventricular or left ventricular background. Closest agreement between scintigraphic and cineangiographic volumes was obtained using right ventricular background, although end-diastolic volumes larger than 100 ml were substantially underestimated. On the basis of this study, the use of two different attenuation coefficients is suggested: the smaller 0.11(-1) cm to calculate end-systolic and end-diastolic volumes and the larger 0.15(-1) cm for volumes greater than 100 ml.