The purpose of this study was to assess the potential for rapid acquisition computed axial tomography (Imatron C-100) to quantify regional myocardial perfusion. Myocardial and left ventricular cavity contrast clearance curves were constructed after injecting nonionic contrast (1 ml/kg over 2 to 3 seconds) into the inferior vena cava of six anesthetized, closed chest dogs (n = 14). Independent myocardial perfusion measurements were obtained by coincident injection of radiolabeled microspheres into the left atrium during control, intermediate and maximal myocardial vasodilation with adenosine (0.5 to 1.0 mg/kg per min, intravenously, respectively). At each flow state, 40 serial short-axis scans of the left ventricle were taken near end-diastole at the midpapillary muscle level. Contrast clearance curves were generated and analyzed from the left ventricular cavity and posterior papillary muscle regions after excluding contrast recirculation and minimizing partial volume effects. The area under the curve (gamma variate function) was determined for a region of interest placed within the left ventricular cavity. Characteristics of contrast clearance data from the posterior papillary muscle region that were evaluated included the peak myocardial opacification, area under the contrast clearance curve and a contrast clearance time defined by the full width/half maximal extent of the clearance curve. Myocardial perfusion (microspheres) ranged from 35 to 450 ml/100 g per min (mean 167 +/- 125). Two flow algorithms derived from characteristics of the contrast clearance curves showed a good correlation with regional myocardial flow determined by microspheres: the ratio of the peak myocardial opacification from baseline to the area under the left ventricular cavity curve (r = 0.7, p less than 0.001, SEE = 44.4 ml/min), and the ratio of the left ventricular cavity to posterior papillary muscle curve areas divided by the full width/half maximal contrast transit time in the region of the posterior papillary muscle (r = 0.82, p less than 0.001, SEE = 52.2 ml/100 g per min). The form of these two flow algorithms was derived from classical indicator dilution theory. In conclusion, indices derived from these data correlated well with regional myocardial perfusion in the posterior papillary muscle region of the dog as assessed by microspheres. This approach offers promise for the quantitation of regional myocardial perfusion and myocardial flow reserve in patients.