Quantitative measurements of atherosclerotic plaque volumes were performed in 13 patients with a mechanical system, consisting of a frame in which the ultrasound transducer was moved triggered by electrocardiography. The transducer had a variable angle of insonation to prevent the ''ping-pong effect'' of repeated echos and a lift adaptation to reduce ultrasound beam resistance between different tissues. Sixty axial sections at 0.9 mm distance along the carotid artery were collected. Three different measurement methods were used for measuring plaque volume (manual tracing, threshold procedure and watershed algorithm). The intraexaminer variability (examiner I or II) was 3.2 % or 3.4 % (manual tracing), 3.5 % or 4.4 % (threshold procedure) and 3.0 % or 3.3 % (watershed algorithm). The interexaminer variability ranged from 4.0% (manual tracing) to 7.3 % (threshold procedure). The combination of system and examiner variability (method variability) in follow-up measurements ranged from 5.8 % (watershed algorithm) to 10.6 % (threshold procedure). This new method is capable of monitoring the possible progression or regression of atherosclerosis in the carotid arteries.