Abstract

Introduction: Ultrasound is often used for monitoring of carotid disease. In current clinical practice, degree of stenosis is an important predictor to assess stroke risk. Lumen narrowing from plaque is currently measured via techniques such as computed tomographic scan, magnetic resonance angiogram or conventional angiogram. Duplex ultrasound measures degree of stenosis based on peak systolic velocities and other parameters and only provides a wide range of level of stenosis. Pursuing lumen size measurement in ultrasound via manual quantification of lumen diameter is tedious. Furthermore, non-uniformity in plaque growth makes it more challenging and time-consuming. There has been an increasing interest in the automatic and robust delineation of the lumen boundaries of the carotids and to measure the lumen diameter via ultrasound given its non-invasive and safe approach. Methods: Deidentified carotid ultrasound images were obtained on patients retrospectively who underwent carotid ultrasound at Toho University Ohashi Medical Center, Tokyo, Japan. A higher order derivative Gaussian filter is applied on these images to highlight the edges. Using pixel classification, lumen region is detected and lumen boundaries are estimated. Results: Of the 202 patients with common carotid artery images, 155 were males and 47 were females.Mean age 69 ± 15.9 years. Mean HbA1c, LDL, HDL and Cholesterol of patients were 6.28±1.1 mg/dl, 101.27±31.6 mg/dl, 50.26±14.8 mg/dl and 175.04±38 mg/dl, respectively. Specialist trained in carotid ultrasound manually traced lumen diameter. Automated tracing and lumen measurements were obtained. The coefficient of correlation between automated diameter and manual diameter was: 0.88, 0.91 and 0.93. The mean diameter error between automated and manual tracing were: 0.50±0.37 mm, 0.36±0.34 mm and 0.30±0.28 mm. Precision of merit between automated diameter and manual diameter was: 93.28%, 95.33% and 96.32% corresponding to manual tracers. Conclusions: The automated lumen diameter measurement is near real time, quick, accurate, fully automated and reliable to assess carotid lumen diameter and narrowing.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call