Background: The purpose of this study was to evaluate the accuracy and reproducibility of three-dimensional volume measurements by high-resolution endoluminal ultrasound in an esophageal varix model. Methods: An esophageal varix model was made by filling three esophageal dilatation catheters with various volumes of water. A 20 MHz ultrasonography transducer was then pulled along the length of the catheters at a constant rate (1.25 mm/sec) while videotaping the procedure. Cross-sectional surface area measurements of each catheter were taken every second and the cross-sectional surface area was multiplied by the length of each catheter, as determined by high-resolution endoluminal ultrasound, to determine the volume in each catheter. Interobserver variability was calculated, and three-dimensional reconstruction was performed. Results: The measured volumes corresponded closely with the actual volumes with an error ranging from 0% to 15.4%. The correlation between actual and measured volumes was r = 0.988. The interobserver variability ranged from r = 0.951 to r = 0.994. Actual esophageal varices were then imaged in a similar fashion to determine the feasibility of this method in patients with esophageal varices. Conclusions: High-resolution endoluminal ultrasound is an accurate and reproducible method of measuring volumes in an esophageal varix model and can be used in a clinical setting to determine variceal volume. Volume studies are now underway in human subjects.