Carotid IMT measurements: a comparison between s-VHS or cd-ROM recorded ultrasonic scans

Objective: To assess whether the use of new B-mode imaging technologies, which allow the direct storage of series of single digitalized images in electronic supports (CD-rom) in place of movies continuously recorded on analogical supports (s-VHS videotapes), may compromise the accuracy, precision and repeatability of carotid artery intima media thickness (IMT) measurements. Methods: The agreement between 121 replicated-scans stored as movies on s-VHS videotapes was compared with the agreement obtained with 121 replicated-scans stored as series of single frozen images on CD-rom. Results: The intra-method agreement was highest with movies recorded on s-VHS videotapes, with a bias between log-transformed readings of 0.0017\ub10.0305, mean absolute difference of 0.0210\ub10.0222 and repeatability coefficient of 0.061. The digital system provided a slightly lower repeatability with a bias between readings of 120.0040\ub10.0343, mean absolute difference of 0.0266\ub10.0219 and repeatability coefficient of 0.069. When the two methods were compared, the log-transformed bias between readings was 0.0169\ub10.06 with higher values obtained with s-VHS, the repeatability coefficient being 0.124, with almost all points within \ub12SD of the mean difference. The mean absolute difference between the two measurements was 0.042\ub10.046. Conclusions: Compared with movies recorded on videotapes, the direct storage of series of single digitalized images in CD-rom for IMT evaluation associates with a lower agreement (-28%) between replicate scans. This implies the need of sample size larger of about 28% when this technology is used in IMT-progression clinical trials. Funding: European Project, IMPROVE QLG1-CT-2002-00896