This study determined the correlation between the extent of the resting perfusion defect by technetium-99m sestamibi tomographic imaging and the first-pass left ventricular (LV) ejection fraction (EF). A total of 1,955 patients underwent technetium-99m sestamibi tomographic imaging with measurement of first-pass resting LVEF. Twenty-five percent of patients had a prior history of myocardial infarction. First-pass LVEF was measured using a peripheral intravenous injection and a multicrystal gamma camera with standard software. Resting tomographic perfusion defect size (infarct size) was quantitated using previously published methods. Mean LVEF for the study group was 0.60 +/- 0.11. Mean LV infarct size was 5 +/- 11%. For the 1,265 patients (65% of the study group) with no measurable perfusion defect, the prevalence of a normal (> or = 0.50) LVEF was 96% (1,212 of 1,265 patients). For patients with a measurable defect (n = 690, 35%), the inverse linear correlation with LVEF was highly significant (r = -0.60, p <0.0001) but with wide confidence limits (SEE = 10 LVEF points), thereby limiting the predictive value in individual patients. Thus, in the absence of known cardiomyopathy, valvular heart disease, or left bundle branch block, patients without a quantifiable resting perfusion defect are highly likely to have a normal resting LVEF and may not require determination of LV function. For patients with resting perfusion defects, LVEF cannot be predicted with confidence and should therefore be measured.