Count-based left ventricular volume determination utilizing a left posterior oblique view for attenuation correction.

This study aimed to determine the inherent error of the left ventricular volume measurement from the gated equilibrium blood pool scintigram utilizing the count-based technique. The study population consisted of 26 patients who had undergone biplane contrast ventriculography. The patients were imaged with a parallel-hole collimator in the left anterior oblique position showing the septum to best advantage. A reference blood sample was counted and radionuclide volumes calculated without correction for attenuation. Attenuation corrected volumes were derived with the factor 1/e-mud, where d = distance from skin marker to center of the left ventricle in the orthogonal left posterior oblique view and mu = linear attenuation coefficient. A series of mu values from 0.08 to 0.15 cm-1 was evaluated. There was a trend toward improvement in the r value and lowering of the SEE with attenuation correction. The tightest 95% confidence limits achieved for an end-diastolic 150-ml ventricle were +/- 44 ml, and for an end-systolic 75-ml ventricle +/- 32 ml. In view of the magnitude of inherent error, the count-based volume measurement may be more suitable for group analyses and in cases in which an individual patient serves as his own control.