Purpose: Pathoanatomic studies suggest that plaque surface disruption, particularly ulceration, plays a key role in the destabilization of internal carotid artery stenosis. Until now, the validity of color-flow Doppler-assisted duplex imaging in detecting such pathoanatomically defined plaque surface abnormalities is unclear.
Methods: We prospectively determined the interobserver reliability and validity of detecting plaque ulceration by means of preoperative color-flow Doppler-assisted duplex imaging in 43 consecutive patients with high-grade (> or = 70%) internal carotid artery stenosis, comparing these ultrasonographic findings with pathoanatomic evaluations of the corresponding endarterectomy specimens.
Results: Interobserver reliabilities for detecting carotid plaque ulceration were kappa= 0.57 for ultrasonography and kappa = 0.82 for the pathologic reference method. Color-flow Doppler-assisted duplex imaging (observer consensus) failed to detect pathoanatomically defined ulceration (chi square = 0.43; p = 0.51). Likewise, sensitivity, specificity, overall accuracy, and positive predictive value were poor (33%, 67%, 56%, and 46%, respectively).
Conclusions: We conclude from our data that color-flow Doppler-assisted duplex imaging is not a reliable or valid means to identify plaque ulceration in high-grade carotid artery lesions.