Geometric coronary artery measurements with the Phillips Digital Cardiac Imaging System (DCI) and the Cardiovascular Angiography Analysis System (CAAS) were validated using percutaneous insertion of radiolucent stenosis phantoms in swine coronary arteries. Angiographic visualization of the stenosis lumens (phi 0.5, 0.7, 1.0, 1.4, 1.9 mm) was simultaneously recorded on DCI and cinefilm. The acquisition systems were calibrated by either the diameter of the guiding catheter (catheter CAL) or the isocenter method (isocenter CAL). Minimal luminal diameters (MLD) obtained with CAAS and DCI on 20 corresponding cineframes were compared with the true phantom diameters (PD). The accuracy of MLD measurements with the CAAS using isocenter CAL was -0.07mm, the precision 0.21 mm (r = 0.91; y = 0.30 + 0.79x; SEE = 0.19), with catheter CAL the accuracy was 0.09 mm, the precision 0.23 mm (r = 0.89; y = 0.19 + 0.74x; SEE = 0.19). The accuracy of MLD measurements using the DCI with isocenter CAL was 0.08 mm, the precision 0.15 min (r = 0.96; y = 0.08 + 0.86x; SEE = 0.14), with catheter CAL the accuracy was 0.18 mm, the precision 0.21 mm (r = 0.92; y = 0.09 + 0.76x; SEE = 0.17). DCI underestimated PD with isocenter CAL (p less than 0.05) and with catheter CAL (p less than 0.001). MLD can be measured with high accuracy, both applying on-line digital as well as off-line cineangiographic analysis. The results of digital measurements demonstrate high reliability of the new digital software package.