Validation of automated contour analysis of intravascular ultrasound images after vascular intervention.

摘要

PURPOSE: The purpose of this study was to determine the feasibility of automated contour analysis of intravascular ultrasound images obtained after vascular intervention. STUDY DESIGN: This was a descriptive study. METHODS: Intravascular ultrasound images obtained from patients after balloon angioplasty (n = 10), stent (n = 10), or stent graft placement (n = 10) were analyzed. A comparison was made between lumen area measured with an automated and a manual system. The location showing the smallest lumen area derived from the automated system was compared with the smallest lumen area selected by visual estimation. RESULTS: Images containing a dissection as a result of balloon angioplasty could not be analyzed by the automated system. The coefficient of variation between the lumen area measurements obtained with the automated system and the manual tracing system of images with a stent (n = 76) or stent graft (n = 79) was 2.7% and 2.1%, respectively. Correlation between the two systems was high (r = 1.00, p < 0.01) both for images containing stents or stent grafts. Minimum lumen area measured with the automated analysis system was smaller than minimum lumen area selected by visual estimation (mean difference 0.8 mm2 (4.9%) for stents and 2.4 mm2 (10.9%) for stent grafts). The location of the smallest lumen area determined with both systems was the same (<1 cm) in 16 cases and differed more than 1 cm in 4 other cases. CONCLUSIONS: The automated analysis system shows good agreement with manual contour analysis of lumen area in images with a stent or stent graft and is a reliable tool for determination of the smallest lumen area. The system is not able to analyze an irregular-shaped lumen area caused by a dissection.