Reliability and reproducibility of automated contour analysis in intravascular ultrasound images of femoropopliteal arteries

Ultrasound Med Biol. 1998 Jan;24(1):43-50. doi: 10.1016/s0301-5629(97)00238-x.

Abstract

An automated contour analysis system was previously developed to increase reproducibility and facilitate quantitative analyses of intravascular ultrasound (IVUS) images. The aim of this study was to compare measurements by this automated system with those obtained by conventional manual tracing, and to determine the intra- and interobserver variability of the automated system. IVUS images obtained in the femoropopliteal artery (n = 12) were analyzed with both systems. Area measurements by the automated system agreed well with the results obtained by manual tracing, displaying low coefficients of variation (8.5 to 15.7%) and high correlation coefficients (r = 0.92 to 0.98). Intra- and interobserver comparison of lumen area, vessel area, plaque area and percentage area stenosis showed low coefficients of variation (6.0 to 15.3% and 5.7 to 14.0%, respectively) and high correlation coefficients (both: r = 0.93 to 0.99). These data indicate that the automated analysis system is a reliable tool for the quantitative assessment of vessel dimensions in IVUS images obtained during clinical examination of peripheral arteries.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Aged, 80 and over
  • Algorithms
  • Arterial Occlusive Diseases / diagnostic imaging*
  • Female
  • Femoral Artery / diagnostic imaging*
  • Form Perception
  • Humans
  • Image Processing, Computer-Assisted
  • Linear Models
  • Male
  • Middle Aged
  • Observer Variation
  • Popliteal Artery / diagnostic imaging*
  • Reproducibility of Results
  • Ultrasonography, Interventional*