Abstract 2789: Evaluation of Risk of Cerebral Infarction by Tissue Characterization of Carotid Artery Plaques using Integrated Backscatter Ultrasound

Abstract

Background: Although carotid plaques with large lipid pools may be more likely to cause cerebral infarction (CI), there is little clinical evidence supporting this hypothesis. The purpose of this study was to elucidate whether three dimensional (3D) IB color-coded mapping methods could stratify the risk of CI. Methods: IB values of plaques were measured in 63 carotid plaques (>50% stenosis) in 63 patients [acute CI (<2 days after onset): n=14, old CI (>60 days after onset): n=15, asymptomatic: n=34]. IB images were acquired using an ultrasonic system (Philips Medical, Sonos7500). We set regions of interest (0.6 x 0.6mm) on the echo tomography. Our definition of IB values for each tissue type was determined by comparing the histology reported in our previous study (7<lipid pool ≤ 13; 13<fibrosis ≤ 18; 18<dense fibrosis ≤ 27; 27<calcification ≤ 33dB). The % lipid volume (%LV: lipid volume/plaque volume) was automatically calculated in each plaque. We also characterized lipid-rich plaques using a magnetic resonance imaging (MRI) at the same sites. Results: % lipid volume of acute CI, old CI and asymptomatic were 80±14, 22±15 and 37±24%, respectively. Receiver-operating-characteristic curve analysis showed that a cutoff point of 51% was the most reliable predictor of the CI risk [sensitivity: 93%, specificity: 93%, positive predictive value (PPV): 93%, negative predictive value (NPV): 93%]. These diagnostic accuracies were higher than those of MRI (sensitivity: 85%, specificity: 71%, PPV: 75%, NPV: 83%). Conclusion: Carotid plaques with greater volume of lipid pool are more likely to cause CI. The 3D IB method is more useful to predict CI rather than depending on MRI.