Left ventricular (LV) volume, and not only ejection fraction (EF), is a crucial parameter for assessing the severity of cardiac disease and determining the patient's prognosis. The purpose of this study was to compare LV volumes and EF computed automatically from gated blood pool tomography (gBPT), using QUBE, and from gated myocardial perfusion tomography (gMPT), using QGS, in the same patients with a known history of myocardial infarction. The effects of the extent and severity of the myocardial perfusion defects were investigated. Thirty-seven patients were injected in a random sequence with 740 MBq of technetium-99m human serum albumin and 925 MBq of Tc-99m-tetrofosmin, within an interval of 2 days. gBPT and gMPT were acquired on the same triple-head gamma camera using the following acquisition parameters: 360degrees step-and shoot rotation, 32 stops (96 projections), 30 s per stop, 64x64 matrix (pixel size 5.8 mm), 8 time bins (75% forward/backward framing). Projection data were reconstructed by filtered back-projection using a Butterworth filter. LV volumes calculated from gBPT correlated well with LV volumes measured on gMPT (r=0.93 for end-diastolic volume and 0.95 for end-systolic volume). Volumes above 200 ml, however, were substantially higher with gMPT than with gBPT. These discrepancies were related to the severity, but not the extent, of the perfusion defects. There was also good agreement between gBPT and gMPT for the LVEF (r=0.91). On the Bland-Altman plot, no trend but a systematic error of 5.57% and a random error of 6.85% could be found. For the LVEF, the differences between the gated tomographic techniques were related neither to the extent nor to the severity of the perfusion defects. In conclusion, LV volumes and EF computed on gMPT correlated well with those measured on gBPT. Discrepancies were observed for large volumes presumably because of inaccuracies of gMPT in patients with severe perfusion defects.