In clinical and experimental studies we assessed images of digital subtraction coronary angiography (DSA) for evaluating regional myocardial perfusion. Myocardial perfusion was assessed by injecting contrast medium into the coronary artery, and by imaging the regional myocardium using DSA. On the time-density curve obtained from the myocardial region of interest, we calculated the time to peak concentration (TPC) and the exponential washout rate (T). TPC and T were measured in five patients with stable effort angina pectoris (AP) and left anterior descending (LAD) lesions before and after percutaneous transluminal coronary angioplasty (PTCA). The values of 1/T increased significantly from 0.09 +/- 0.02 l/sec to 0.21 +/- 0.04 l/sec (p less than 0.01) after PTCA, but l/TPC did not change. No significant difference in ejection fractions was observed between the patients with AP and the normal subjects (n = 7), while the regional percent area shrinkage in the anterolateral and apical regions supplied by the LAD was significantly decreased in the patients with AP compared with those of normal subjects (anterolateral: 39.8 +/- 8.8% vs 51.3 +/- 6.8%, apical: 36.6 +/- 8.4% vs 52.4 +/- 13.4%, both p less than 0.01). In 10 anesthetized dogs with varying degrees of reduction in the left circumflex coronary artery (LCX) blood flow (CBF: categories of stenosis (S1-S5), we compared 1/TPC and 1/T with regional myocardial function (systolic wall thickening: %WTh). With varying LCX stenosis, there were no significant changes in heart rate and mean aortic pressure and significant linear correlations were observed between %WTh and 1/TPC (r = 0.51), between %WTh and 1/T (r = 0.55). At S1 (CBF: 100-90% of the control), neither %WTh nor 1/TPC differed from that of the controls, but 1/T was significantly decreased (80% of the controls, p less than 0.01). From S3 (CBF: 79-60%) to S5 (CBF: 39-0%), %WTh, 1/TPC and 1/T were significantly decreased from those of the control levels (all p less than 0.01). However, at S5 (CBF: 39-0%) the values of 1/TPC (71% of controls) and 1/T (33%) did not differ from those at S4; whereas, %WTh was markedly reduced and the systolic thinning of the ventricular wall occurred at S5. Therefore, in critical coronary stenosis, 1/T was more sensitive than 1/TPC or wall dynamics for assessing myocardial ischemia. Both 1/TPC and 1/T, as well as %WTh, were useful for assessing moderate myocardial ischemia; however, these DSA indices had considerable limitations for evaluating the severity of myocardial ischemia when CBF was markedly reduced.