Carotid duplex ultrasound velocity measurements versus intravascular ultrasound in detecting carotid in-stent restenosis.

BACKGROUND Duplex ultrasonography criteria for assessing the severity of carotid artery (CA) in-stent restenosis are not well established. METHODS AND RESULTS We analyzed 39 patients (40 CAs) who underwent CA stenting with baseline and 6-month follow-up carotid duplex ultrasonography and intravascular ultrasound. Intravascular ultrasound measurements included minimum luminal diameter, percent diameter, and lumen area stenosis. Duplex ultrasonography measurements included peak systolic velocity (PSV), percentage change in PSV, end-diastolic velocity (EDV), and internal-to-common CA PSV ratio (ICA/CCA). Receiver operating characteristic curves assessed each duplex measurement to detect >or=50% diameter, >or=75% lumen area stenosis, and minimum luminal diameter <3 mm at follow-up. At 6-month intravascular ultrasound follow-up, >or=50% diameter and >or=75% lumen area CA in-stent restenosis occurred in 20% and 25%, respectively; minimum luminal diameter <3 cm occurred in 48%. Area under receiver operating characteristic curves for PSV, EDV, and ICA/CCA were 0.85, 0.96, and 0.89 for >or=50% diameter stenosis and 0.89, 0.93, and 0.88 for >or=75% lumen area stenosis, respectively. Optimal PSV, EDV, and ICA/CCA criteria to detect >or=50% diameter and >or=75% lumen area CA in-stent restenosis were greater compared with those for native CA. A >98% increase in PSV had the highest specificity, whereas the combination of EDV >41 cm/s and ICA/CCA >2 had the highest sensitivity in detecting >or=75% lumen area CA in-stent restenosis. CONCLUSIONS PSV, EDV, and ICA/CCA PSV ratio were good discriminators for detecting significant diameter and lumen area greater compared with those for native CA. The combination of duplex velocity criteria increases diagnostic accuracy.