Myers R W, Bails R P, Reed V R, Price J M, Drasin E, Hill J D, Saphir J R
Clin Nucl Med. 1982 Apr;7(4):151-6. doi: 10.1097/00003072-198204000-00002.
Determination of left ventricular function with tomographic radionuclide angiocardiography using a seven-pinhole collimator was evaluated by comparing results of invasive contrast studies and planar multigated blood pool imaging to the tomographic study of 25 patients. LAO, seven-pinhole multigated blood pool acquisition was reconstructed to produce eight slices from apex to base in each of eight segments of the cardiac cycle. After applying an edge detection routine, three-dimensional reconstruction of perimeters allowed cyclic viewing of the left ventricular angiogram in any projection. When planar and tomographic radionuclide techniques were compared to contrast studies, sensitivity and specificity for identification of segmental wall motion abnormalities were not different (93% and 90% for both). Ejection fraction was determined from tomography by integration of slices to produce a noncalibrated volume and from planar blood pool imaging and contrast ventriculograms by standard techniques. Ejection fraction as compared to contrast studies was accurately determined on the planar angiogram (R = .87, P less than .001) with tomographic analysis showing similar significant correlation (R = .79, P less than .001). Methodologic evaluation indicated that the tomographic study was readily positioned and acquired and provided the advantage of requiring little operator interaction. It suffers, however, from the disadvantage of long reconstruction times with current software and occasional difficulties in defining the superior extent of the left ventricle.