The accuracy and precision of a fully automated quantitative coronary angiography (QCA) algorithm for use in a cineless environment were determined in phantom studies and in an in vivo canine preparation. Imaging studies of 118 coronary segments in six anesthetized dogs were compared with measurements of the diameters of casts of the canine coronary arteries produced in physiologic conditions. Regression analysis of phantom vessel diameters against QCA measurements revealed slopes of 0.94 to 0.96 and r values > 0.99. The results of the in vivo studies showed good correlation with the coronary cast diameter measurements, with a slope of 0.969 and an r value of 0.987 for the sets of measurements. The high degree of accuracy obtained in a model representative of the clinical situation demonstrates that QCA methods can be applied reliably in the clinical arena with current digital imaging technology and without cinefilm.