To determine the reproducibility of quantitative indexes of hypoperfused myocardium by exercise thallium-201 single-photon emission computed tomography, duplicate studies were performed in 16 stable patients within 1 month. Twenty-three other patients, with intervals up to 13 months between studies, were retrospectively identified from medical records. Symptoms, weight, heart rate achieved and peak systolic blood pressure during the 2 studies were similar. Maximal counts circumferential profiles' polar maps were generated and divided into 3 vascular territories. The hypoperfused myocardium was defined as the percent stress profile points below the normal level. The observed values were compared between the 2 studies for each patient. Defect size ranged from 0 to 73%. The concordance correlation coefficient (a measure of the closeness of the data points to the line of identity) and mean absolute deviation were 0.94 and 4.5%, respectively, when all patients were considered together, 0.93 and 3.2%, respectively in the reproducibility group, and 0.94 and 5.1%, respectively, in the retrospective group. Inter- and intraobserver reproducibility were also very high. For individual vessels, the concordance correlation coefficients were all > or = 0.89 and the mean absolute deviations between 3.7 and 9%. Thus, in stable patients with repeat thallium studies within a year, the percent hypoperfused myocardium is highly reproducible.