9 March 2010 Three-dimensional ultrasound-based texture analysis of the effect of atorvastatin on carotid atherosclerosis
Author Affiliations +
Abstract
Carotid atherosclerosis is the major cause of ischemic stroke, a leading cause of death and disability. This is driving the development of image analysis methods to quantitatively evaluate local arterial effects of potential treatments of carotid disease. Here we investigate the use of novel texture analysis tools to detect potential changes in the carotid arteries after statin therapy. Three-dimensional (3D) carotid ultrasound images were acquired from the left and right carotid arteries of 35 subjects (16 treated with 80 mg atorvastatin and 19 treated with placebo) at baseline and after 3 months of treatment. Two-hundred and seventy texture features were extracted from 3D ultrasound carotid artery images. These images previously had their vessel walls (VW) manually segmented. Highly ranked individual texture features were selected and compared to the VW volume (VWV) change using 3 measures: distance between classes, Wilcoxon rank sum test, and accuracy of the classifiers. Six classifiers were used. Using texture feature (L7R7) increases the average accuracy and area under the ROC curve to 74.4% and 0.72 respectively compared to 57.2% and 0.61 using VWV change. Thus, the results demonstrate that texture features are more sensitive in detecting drug effects on the carotid vessel wall than VWV change.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Joseph Awad, Adam Krasinski, David Spence, Grace Parraga, Aaron Fenster, "Three-dimensional ultrasound-based texture analysis of the effect of atorvastatin on carotid atherosclerosis", Proc. SPIE 7626, Medical Imaging 2010: Biomedical Applications in Molecular, Structural, and Functional Imaging, 762621 (9 March 2010); doi: 10.1117/12.844371; https://doi.org/10.1117/12.844371
PROCEEDINGS
10 PAGES


SHARE
Back to Top