Digital coronary arteriography has advanced from a curiosity to a powerful clinical tool. This development has been motivated by the new imaging demands of interventional cardiology, by the need for quantifying atherosclerotic disease, and by advances in computer and video technology. Digital imaging has now essentially replaced cinefilm for clinical decision making in some catheterization laboratories, although uncertainty remains regarding the diagnostic comparability of the two modalities. Therefore, we compared simultaneously acquired digital and cine arteriograms from 18 patients with multivessel coronary artery disease. Digital utilized pulse-progressive technique, a 512 x 512 pixel x 8 bit image matrix, and no post processing. Four angiographers interpreted the arteriograms, identifying 131 stenotic lesions for measurement with manual calipers. Measurements of percent diameter stenosis by digital and cine correlated closely (r = 0.88), but digital estimated stenoses to be significantly more severe than cine (p less than 0.0001). This difference was most significant in small (less than 2 mm diameter) arteries, in branch arteries, and with mild stenoses. The differences between digital and cine were not statistically significant for stenoses greater than 50% diameter narrowing. Interobserver variability was similar for digital and cine imaging. Thus, unprocessed digital and film-based coronary arteriography yield similar assessments of atherosclerotic stenoses, but the techniques may not be diagnostically interchangeable. This paper discusses the technical advances which have occurred in digital coronary arteriography, the comparability of digital and cine imaging, and the factors which may be responsible for observed differences between the two modalities.