Influence of data acquisition on accuracy in carotid plaque volume measurements with 3D-ultrasound

Abstract

Objective: The aim was to investigate factors influencing reliability of plaque volume measurements by three-dimensional ultrasound in the carotid arteries. Methods: Data acquisition of 100 plaques was performed with longitudinal or transversal sections twice in a 3-day interval. Repeated plaque volume measurements were performed with a three-dimensional ultrasound system. Plaque area in 2D images was estimated by manual outlining of the plaque, or by semi-automatic grey level thresholding. Results: For data acquisition with longitudinal sections, a drop out-rate of 13% occurred; for data acquistion with transversal sections the rate was 3%. In follow-up measurements, intra-sonographer reliability for the manual measurement method was r = 0.97 for data acquisition by transversal and r = 0.92 for data acquisition by longitudinal ultrasound sections. For the threshold procedure, the intra-sonographer reliability was r = 0.95 (transversal sections) or r = 0.93 (longitudinal sections). Intra-reader reliability of plaque volume values for both — manual tracing and threshold procedure — varied between r = 0.98 (longitudinal sections) to r = 0.99 (transversal sections). Conclusion: Three-dimensional ultrasound follow-up measurements of plaque volumes in carotid arteries have a higher reliability with data acquisition by transversal sections than by longitudinal sections. Another advantage of the data acquisition by transversal ultrasound sections is the reduced number of drop-outs.