Assessment of carotid plaque composition in hypertensive patients by ultrasonic tissue characterization: a validation...

Ultrasonic tissue characterization of epi-aortic vessels may be useful to define the composition of atherosclerotic plaques. Videodensitometry provides a histogram representing the frequency distribution of gray levels corresponding to different compositions of the carotid wall. However, lack of standardization limits the clinical application of this technique. In the present study, the echoreflectivity (ER) pattern of atherosclerotic plaques in vivo was compared with their histological pattern after surgical removal, and the reproducibility of measurement was tested.

We studied 19 hypertensive patients with a carotid artery stenosis >or= 70%, eligible for carotid thromboendarterectomy (TEA). Before TEA, all patients underwent standard high-resolution B-mode carotid ultrasound. ER parameters (mean gray level, broad band, skewness, and kurtosis) were obtained in a region of interest selected along the whole plaque, between the intima-blood and the media-adventitia interfaces. The plaques removed during TEA were examined by a histologist and classified into three groups on the basis of fibrous tissue (FT) content: lipidic (FT < 20%), fibrolipidic (20 50%). Discriminant function analysis was used to evaluate classification efficacy of different histological groups based on ER parameters.

Histologically, five lesions were classified as lipidic, six as fibrolipidic and eight as fibrous. Analysis of variance showed significant between group differences in all ER parameters. The combined use of all ER parameters provided correct classification of plaques in 94.73% of cases (P < 0.0001), improving the classification made using single parameters. Intra-observer and inter-observer variabilities (Bland-Altman method) of mean gray level measurements were small.

Videodensitometry can discriminate between tissue composition of carotid lesions and complement the quantitative assessment of intima-media thickness by additionally providing a well-reproducible semiquantitative evaluation of vascular wall constituents.