Accurate estimates of absolute left ventricular volumes from equilibrium radionuclide angiographic count data using a simple geometric attenuation correction.

To simplify and clarify the methods of obtaining attenuation-corrected equilibrium radionuclide angiographic estimates of absolute left ventricular volumes, 27 patients who also had biplane contrast cineangiography were evaluated. Background-corrected left ventricular end-diastolic and end-systolic counts were obtained by semiautomated variable and hand-drawn regions of interest and were normalized to cardiac cycles processed, frame rate and blood sample counts. Blood sample counts were acquired on (d degree) and at a distance (d') from the collimator. A simple geometric attenuation correction was performed to obtain absolute left ventricular volume estimates. Using blood sample counts obtained at d degree or d', the attentuation-corrected radionuclide left ventricular end-diastolic volume estimates using both region of interest selection methods correlated with the cineangiographic end-diastolic volumes (r = 0.95 to 0.96). However, both mean radionuclide semiautomated variable left ventricular end-diastolic volumes (179 +/- 100 [+/- 1 standard deviation] and 185 +/- 102 ml, p less than 0.001) were smaller than the average cineangiographic end-diastolic volume (217 +/- 102 ml), and both mean hand-drawn left ventricular end-diastolic volumes (212 +/- 104 and 220 +/- 106 ml) did not differ from the average cineangiographic end-diastolic volume. Using the blood sample counts obtained at d degree or d', the attenuation-corrected radionuclide left ventricular end-systolic volume estimates using both region of interest selection methods correlated with the cineangiographic end-systolic volumes (r = 0.96 to 0.98). Also, using blood sample counts at d degree, the mean radionuclide semiautomated variable left ventricular end-systolic volume (116 +/- 98 ml, p less than 0.05) was less than the average cineangiographic end-systolic volume (128 +/- 98 ml), and the other radionuclide end-systolic volumes did not differ from the average cineangiographic end-systolic volume. Therefore, it is concluded that: 1) a simple geometric attenuation-correction of radionuclide left ventricular end-diastolic and end-systolic count data provides accurate estimates of biplane cineangiographic end-diastolic and end-systolic volumes; and 2) the hand-drawn region of interest selection method, unlike the semiautomated variable method that underestimates end-diastolic and end-systolic volumes, provides more accurate estimates of biplane cineangiographic left ventricular volumes irrespective of the distance blood sample counts are acquired from the collimator.