Despite the diagnostic efficacy of stress myocardial perfusion imaging, the correlation between the actual perfusion tracer activity and diseased state of a coronary artery has not been studied in detail. We estimated exercise-related perfusion augmentation in relation to disease states of a coronary artery in diabetic and non-diabetic patients by a newly developed quantitative technetium (Tc)-99m-tetrofosmin myocardial imaging technique. Tc-99m-tetrofosmin tomographic imaging with an exercise-rest protocol was performed in 26 stable coronary patients and in 8 age-matched controls. Percent increase (%IR) in myocardial count during symptom-limited submaximal exercise-stress was calculated in 16 non-infarcted polar map segments and in each coronary territory by a subtraction technique with corrections for physical decay and injected tracer doses, and the results were compared with those of angiographically quantified coronary diameter stenosis (%DS). Percent IR and peak heart rate during exercise showed a positive linear correlation both in coronary territories with significant stenosis (%DS > or = 75%) and in control or nonstenotic (%DS < 75%) territories. The regression line in stenotic regions was, however. significantly (p < 0.01) shifted downward compared to that in non-stenotic regions. Percent IR in stenotic regions showed a significant inverse correlation with %DS. Coronary stenosis of 75% or more was identified by a %IR cutoff value of 40% with 77% sensitivity, 70% specificity, and an accuracy of 72%. In coronary territories with a %DS of less than 75%, %IR in diabetic patients was significantly lower (46+/-15%) than that in nondiabetic patients (61+/-25%). Thus, blunted exercise-related augmentation of myocardial uptake of Tc-99m-tetrofosmin correlates with the severity of coronary narrowing and diabetic state.