A multicenter trial was performed on 140 patients from four centers to determine the accuracy of quantitative analysis of stress/delayed thallium-201 myocardial tomograms using normal limits to assess the relative amount of reversibility of stress-induced defects. The patients were found to have 85 fixed and 124 reversible defects, as determined by visual interpretation. Reversibility bull's-eye polar maps were compared to gender-matched normal limits from 36 normals. Regions were identified as reversible if their normalized difference between stress and 4 hr greater than 1.5 s.d.s. from the mean normal limits. Overall agreement between experts at multicenter sites and reversibility maps was 73% for reversible defects and 80% of fixed defects. Sensitivity in detecting reversibility was highest for the left circumflex (88%) and lowest for the right coronary (60%). These results indicate that reversibility polar maps and normal limits offer an objective, accurate technique for determining the reversibility of stress-induced perfusion defects.