Detection and quantitation of coronary artery stenoses from digital subtraction angiograms compared with 35-millimeter film cineangiograms

To assess the ability to detect coronary artery narrowings from computer-acquired angiograms, a panel of 4 observers independently identified and measured focal coronary narrowings from digital subtraction angiograms and compared the results to those obtained from standard 35-mm cine film angiograms. Both cine and digital angiograms were obtained sequentially using selective intracoronary artery injection of standard amounts of iodinated contrast media. Digital images were obtained at 8 frames/s with a 512 × 512 × 8-bit pixel matrix. Modifications in the imaging chain for computer acquisition included a slower pulsed radiographic mode, a progressive scan camera, and initial storage of the images on an 80-megabyte digital hard disk. Postprocessing computer algorithms were used to enhance the unsubtracted digital images; these included single-frame, mask-mode subtraction, vessel boundary edge enhancement, and 4-fold pixel magnification. In 19 patient studies, 32 arteries were reduced more than 25 % in diameter according to at least 1 of 4 observers on either the digital or cine film angiograms. There was no significant difference in the mean percent diameter narrowing for all the narrowings between the digital angiograms (53 ± 31 %) and the cineangiograms (52 ± 31 %). In addition, a 2-way analysis of variance yielded no significant difference between the amount of variability in the measurements between the cine film and the digital technique. This similar variability persisted when subsets of patients based on the degrees of stenosis were considered (e.g., only narrowings from 50 to 90% diameter reduction). Because digital acquisition permits immediate playback with image enhancement and greater ease of coronary artery quantification, digital angiography may have widespread clinical use for the detection and quantitation of coronary artery disease.