Visual analysis of the severity of coronary stenosis is limited by observer variability. However, more complex techniques of proved accuracy are tedious and costly. Therefore, a new digital electronic caliper (DEC) was evaluated as a potentially more accurate, rapid and less costly alternative for measuring stenosis severity. Stenosis minimum diameter (Dmin) and percent diameter reduction (% S) were measured from the screen of the cine projector using a DEC. These measurements were compared with visual estimates (VIS) by 4 experienced angiographers and with measurements made by a computer-assisted method (QCA) of proved accuracy. In routine cineangiograms from 7 patients, 10 lesions were significant (greater than 50% S) and 8 were mild (less than 50% S). Variability, the standard deviation of multiple estimates of Dmin and % S, averaged 0.09 mm and 3.1% for QCA; 0.18 mm and 5.9% for DEC; and 0.26 mm and 7.4% for VIS. Compared with QCA, the visual determination of % S significantly underestimates (-5%; p less than 0.02) mild and overestimates (+11%; p less than 0.002) significant stenosis. VIS underestimates Dmin in significant lesions by 20% (p less than 0.04). In contrast, the mean error for DEC measurement of Dmin and % S was not significantly different from 0 in either lesion group. For the entire group of lesions, and particularly in significant lesions, the mean error for measurement of these 2 indexes of disease was significantly less with DEC than with VIS. Thus, variability and error with DEC are acceptably low for clinical use.(ABSTRACT TRUNCATED AT 250 WORDS)