Parametric analysis of carotid plaque using a clinical ultrasound imaging system

• Published in: Ultrasound in medicine & biology Vol. 29; no. 11; p. 1521
• Main Authors: Waters, Kendall R, Bridal, S Lori, Cohen-Bacrie, Claude, Levrier, Claire, Fornès, Paul, Laugier, Pascal
• Format: Journal Article
• Language: English
• Published: England 01-11-2003
• Subjects: Carotid Arteries - diagnostic imaging; Carotid Arteries - pathology; Carotid Artery Diseases - diagnostic imaging; Carotid Artery Diseases - pathology; Humans; Image Interpretation, Computer-Assisted - methods; Statistics, Nonparametric; Ultrasonography
• ISSN: 0301-5629
• DOI: 10.1016/j.ultrasmedbio.2003.08.007

We evaluated quantitative ultrasonic methods for assessment of carotid plaque content. In vitro measurements of fixed, carotid plaque specimens obtained by surgical endarterectomy were performed using a clinical Philips HDI 5000 imaging system connected to a radiofrequency (RF) signal-acquisition system. We acquired RF signals and grey-scale images from carotid specimens (n = 17) and a tissue-mimicking reference phantom. Imaged plaque sections were then classified according to histology. Parametric images were constructed from the integrated backscatter (IBS), and the midband, slope and intercept values of a straight-line fit to the apparent backscatter transfer function. Analysis was performed on 82 regions-of-interest (ROIs). The IBS values for collagen, lipid and hemorrhage plaques were 5.8 +/- 5.4, 3.9 +/- 3.7, 2.8 +/- 2.2 dB, respectively. Midband and IBS parameter images exhibited good agreement in morphology with histology, whereas the slope and intercept parameter images were noisy. Mean IBS, midband, and grey-scale values of complex plaques were found to be statistically different (p < 0.05) from lipid, hemorrhage and fibrolipid plaques. The bias and limits of agreement (1.3 +/- 4.9 dB) between the grey-scale and IBS methods, however, indicated that the two methods were not interchangeable. Results indicate necessary improvements, such as reduction of large measurement variances and identification of robust parameters, that will permit multiparametric characterization of carotid plaque under in vivo conditions.