Lumen reduction measurements of the internal carotid artery before and after Levovist enhancement: reproducibility and agreement with angiography.

Our aim was to assess reproducibility of three different lumen reduction measuring methods--North American Symptomatic Carotid Endarterectomy Trial, European Carotid Surgery Trial, and common carotid--using power Doppler and color Doppler sonography before and after Levovist enhancement. We included 20 symptomatic patients with mild or severe carotid disease. North American Symptomatic Carotid Endarterectomy Trial, European Carotid Surgery Trial, and common carotid measurements on longitudinal views and European Carotid Surgery Trial measurements on transverse views were performed. Examinations were repeated and the results compared to assess reproducibility of measurements. Correlation with angiography was obtained by calculating Pearson correlation coefficients. Reproducibility was significantly better (P < 0.05) for European Carotid Surgery Trial and common carotid measurements (95% limits of agreement between -10% to 10% and -19% to 17%) as compared to North American Symptomatic Carotid Endarterectomy Trial measurements (95% limits of agreement between -11% to 21% and -21% to 23%). Variability of measurements after enhancement increased slightly (not significant) for both power and color Doppler sonography. Additionally, European Carotid Surgery Trial measurements, using nonenhanced power Doppler or color Doppler sonography, did not correlate significantly with angiography, whereas North American Symptomatic Carotid Endarterectomy Trial and common carotid measurements correlated well with angiography, particularly in power Doppler mode after enhancement (r = 0.88 and r = 0.82, respectively). We conclude that for lumen reduction measurements of the internal carotid artery with power and color Doppler sonography, the common carotid method is the only method that is reproducible and has good correlation with angiography, which slightly improves after Levovist enhancement.