Abstract WP163: 3D-versus 2D-Ultrasound for Plaque Analysis of Internal Carotid Arteries with Gray-Scale Median
Background: Gray-Scale Median (GSM) analysis of carotid artery stenosis allows quantification of plaque echogenicity, correlates with histology and showed an association with symptoms and cerebral embolic findings. The aim of the study is to demonstrate whether several systematically aquired sections of 3-dimensional Power-Doppler ultrasound (3D-PDUS) datasets comprising whole plaque stenoses provide different GSM than single sections acquired by 2D-PDUS as commonly used for this purpose.
Methods: After informed consent, 27 patients were scanned with a 3D-ultrasound (US) probe at 10 Mhz. GSM-analysis was performed offline with a specially designed software (Matlab) in 2D longitudinal scans adjusted to the vessel center line, as well as 3 parasagital - longitudinal sections parallel to each other and 3 views rotated around the center of the plaque, that were aquired off-line from the 3D-datasets by multiplanar rendering. For comparison of the GSM, the median differences between 3D- and the 2D-images as well as standard deviations (SD) were calculated. Mixed model regression analysis was used for further descriptive statistics.
Results: The GSM differences between the 3D-parallel and the central longitudinal 2D-reference section showed a markedly higher proportion of the variance due to image-to-image (63%) versus inherent technical-methodological scattering (37%), that was not found for rotated 3D sections (54 versus 46%). The GSM of the longitudinal 3D-sections correlates significantly with the GSM of the 2D-reference section (r = 0.60) as do the rotated 3D views (r = 0.52). Correlation coefficients of the central 3D- with the 2D-longitudinal section differed significantly, when rotation was performed (r = 0.63) instead of sagital rendering (r = 0.85, p < 0.05). Significant over- versus underestimations were noted for high versus low GSM values in the 3D parasagital and rotated views compared to the 2D sections.
Conlusions: Variations of the GSM values in between the 3D-sections were largely due to real differences of the echogenicity. The GSM of a single 2D-section may not be representative for the whole 3D plaque structure. Parasagital section rendering should be prefered to analyse whole plaque GSM over rotating sections.
- © 2012 by American Heart Association, Inc.