Virtual histology-intravascular ultrasound in assessment of carotid plaques: Ex vivo study

OBJECTIVE: Virtual histology-intravascular ultrasound (VH-IVUS) has been reported to be useful in detecting the components of coronary plaques in vivo. Recently, the application of VH-IVUS to peripheral interventions has been evaluated. The aim of this study was to examine the extent to which the necrotic core of carotid plaques could be assessed accurately by VH-IVUS compared with histopathology. METHODS: A total of 37 carotid plaques underwent ex vivo VH-IVUS within 24 hours after endarterectomy. Ninety-five segments of virtual histological images were matched to histological sections. The area of the necrotic core on histological sections was compared with that on virtual histological images. Intraplaque hemorrhage (IPH) was histopathologically graded by its severity using immunohistochemical staining for glycophorin A as a marker. The relationship of the severity of the IPH to the necrotic core was histopathologically evaluated. The correlation between the necrotic core or IPH with symptomatology was also evaluated. RESULTS: The area of the necrotic core on virtual histological images (median, 8.0%; interquartile range, 5.0%-13%) was significantly smaller compared with that of the histological sections (median, 50%; interquartile range, 40%-63%) (P < 0.0001). The Bland-Altman analysis showed poor agreement in the necrotic core measurement between virtual histological images and histological sections (mean difference, 39.8%; 95% confidence interval, 35.8%-43.8%). Severe IPH was significantly associated with a larger necrotic core and symptomatology (P < 0.0001 and P = 0.0039, respectively). The area of necrotic core on the virtual histological analysis did not correlate with symptomatology (P = 0.70), but that on pathological analysis tended to correlate with symptomatology (P = 0.059). CONCLUSION: In the present virtual histological algorithm, the underestimation of the necrotic core was revealed. The lack of a hemorrhage component in the virtual histological algorithm is a leading cause of its underestimation.