[Optimized calculation of left ventricular enddiastolic volume by equilibrium radionuclide ventriculography (MUGA)--the influence of background correction and absorption (author's transl)].

Left ventricular (LV) enddiastolic volume (EDV) was calculated in 23 patients from background-corrected enddiastolic count rates (ROI technique). The calculation in volume units (ml) was based on the measurement of a well-defined amount of patients' venous blood (50 ml in a syringe) counted by the gamma camera. Decrease in count rate due to tissue attenuation was corrected for by calculation of soft-tissue absorption in the thoracic and LV wall as well as the blood self-absorption within the LV, applying a rotational ellipsoid with an assumed LV depth of 1.5 times that of the LV width within the LV ROI. Furthermore, the influence of various models of LV background superimposition (homogeneous, parabolic and no background) were tested correlatively for calculations of volume. Best correlation with cineventriculographically measured volumes (r = 0.92) was found with a parabolically weighed background correction, determining an individual value of background superimposition for each LV pixel. This approach was chosen because a homogeneous superimposition is an inadequate model to correct for background due to the geometric extension of the LV. Volumes, parabolically corrected were found to correlate most closely to cineventriculography (CVG) (VMUGA = 0.997 X VCVG - 12.6 ml).