Identification of early atherosclerotic lesions in carotid arteries with quantitative characteristics measured by 3D MRI

Purpose To evaluate the usefulness of quantitative characteristics of morphology and signal intensity of arterial wall measured by 3D multicontrast magnetic resonance vessel wall imaging (MRVWI) in identification of carotid early atherosclerosis (CEAS). Materials and Methods In all, 61 older subjects (mean age 71.8 ± 5.6 years old; 25 males) without cardiovascular symptoms in the last 6 months were recruited. The carotid arteries without advanced plaque features on 3.0T MRI were included for analysis. Ultrasound imaging was used as a reference to identify CEAS. The morphological parameters including lumen area (LA), wall area (WA), wall thickness (WT), and normalized wall index (NWI = WA/[WA+LA] × 100%) and the signal intensity on 3.0T MR T2‐weighted images (T2SI) of the carotid arterial wall were measured. Three regression models were built to identify CEAS with the following parameters: Model 1 with both morphological and T2SI parameters; Model 2 with T2SI parameters; and Model 3 with morphological parameters. All models were adjusted for age and sex. Area under the curve (AUC) was calculated to validate models. Results Of the 86 carotid arteries without advanced plaques, 47 (54.7%) were found to have early plaques determined by ultrasound. Among three regression models, Model 1 showed the highest AUC values in identifying CEAS (left: AUC = 0.856, P < 0.001; right: AUC = 0.867, P < 0.001), followed by Model 2 (left: AUC = 0.843, P < 0.001; right: AUC = 0.798, P = 0.001), and Model 3 (left: AUC = 0.790, P = 0.002; right: AUC = 0.806, P < 0.001). Conclusion The combination of morphology and normalized T2SI of arterial wall measured by MRVWI is more effective than each characteristic alone in identification of CEAS. J. Magn. Reson. Imaging 2016;44:1270–1276.