Max Wintermark MD, Presenter: Consultant, PAION AG

Joseph Rapp, Abstract Co-Author: Nothing to Disclose

Elizabeth Tong, Abstract Co-Author: Nothing to Disclose

Gabriel Acevedo-Bolton, Abstract Co-Author: Nothing to Disclose

X.M. Pan, Abstract Co-Author: Nothing to Disclose

David Saloner PhD, Abstract Co-Author: Research grant, Koninklijke Philips Electronics NV

To evaluate the ability of multidetector-row CT-angiography (CTA) studies to assess the histological composition and characteristics of carotid artery atherosclerotic plaques, using histology as the gold standard.

Eight patients with transient ischemic attacks who underwent both a CTA and carotid endarterectomy were enrolled in a prospective study. Two analyses, a quantitative one and a qualitative one, were performed, comparing in-vivo CTA to histology, used as a gold standard. The quantitative analysis consisted of a systematic comparison of CTA images with histological sections to determine what CT density is associated with each component of the atherosclerotic plaques. The qualitative analysis used these results to develop a computer algorithm that automatically interpretes the components of the carotid atherosclerotic plaques. The reading of this computer analysis by a neuroradiologist was compared to the interpretation of histological samples by a pathologist with respect to the types and characteristics of the carotid atherosclerotic plaques based on the American Heart Association classification.

There was a 72.6% agreement between CTA and histology in terms of carotid plaque characterization. CTA showed perfect concordance when classifying calcifications. There was a significant overlap between densities associated with classifying lipid-rich necrotic core, connective tissue and hemorrhage, which limited the reliability of individual pixel Hounsfield readings to indicate either of these components. However, CTA showed good correlation with histology when only large lipid cores (kappa = 0.796, p<0.001) and wide hemorrhages (kappa = 0.712, p=0.102) were considered. CTA performed well in detecting ulcerations (kappa = 0.855) and in measuring the fibrous cap thickness (R2 = 0.77, p<0.001).

The composition of carotid atherosclerotic plaques determined by CTA accurately reflects plaque composition defined by histology.

The ability to analyze carotid plaque components on CTA using an automated classification algorithm could provide a convenient, repeatable, noninvasive method of studying carotid atherosclerotic disease.

Wintermark, M, Rapp, J, Tong, E, Acevedo-Bolton, G, Pan, X, Saloner, D, High-Resolution CT Imaging of Carotid Artery Atherosclerotic Plaques.  Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL. http://archive.rsna.org/2007/5005634.html