Although use of videotape for the recording of coronary angiograms continues to grow, the validity of quantitative coronary angiographic analysis of video images remains unknown. To estimate the reliability of angiographic images recorded on videotape, experimental and clinical angiograms were recorded simultaneously on both 35 mm cinefilm and super-VHS videotape with normal images and with spatial filtering of the images (edge enhancement) on a digital cardiac imaging system. The experimental angiographic studies were performed with plexiglass blocks and stenosis phantom of 0.5 to 3.0 mm in diameter. The clinical angiograms were recorded in 20 patients undergoing percutaneous transluminal coronary angioplasty (31 frames before and 20 frames after percutaneous transluminal coronary angioplasty). The cinefilm and corresponding videotapes were analyzed off-line with the new version of the coronary angiography analysis system. For the experimental study, measurements of minimal luminal diameter obtained from cinefilm, normal-image videotape, and edge-enhanced videotape were compared with the true phantom diameter. In the clinical study the agreement between measurements obtained from cinefilm and measurements from normal-image videotape and edge-enhanced videotape was examined. In the phantom series the accuracy and precision of quantitative coronary angiography measurement for cinefilm were -0.10 +/- 0.08 mm, for normal-image videotape -0.11 +/- 0.18 mm, and for edge-enhanced videotape -0.10 +/- 0.11 mm (mean +/- SD). In the clinical series, the differences between measurements from cinefilm and normal-image videotape were 0.14 +/- 0.20 mm and from cinefilm and edge-enhanced videotape 0.04 +/- 0.13 mm.(ABSTRACT TRUNCATED AT 250 WORDS)