Assessment and Reproducibility of Aortic Atherosclerosis Magnetic Resonance Imaging

OBJECTIVES: To investigate image quality and interstudy reproducibility of aortic atherosclerosis imaging at 1.5 T, and to explore the impact of parallel imaging techniques at 3 T. MATERIALS AND METHODS: Institutional review board approval and informed consent were obtained. Thirty-two subjects (20 normal, 12 patients with impaired cardiac function) underwent 4 black-blood T2-weighted imaging studies of the abdominal aorta: 2 conventional studies at 1.5 T, a conventional study at 3 T, and an accelerated 3-T study with parallel imaging (SENSE). Contrast-to-noise ratio and image quality score (1-5 scale, 5 = highest quality) were determined for each study. Studies were analyzed for mean wall thickness and area plaque burden as endpoints for aortic atherosclerosis. Bland-Altman analyses were performed to determine interstudy reproducibility between imaging methods. Wilcoxon signed-rank tests were used to identify significant differences between methods (P < 0.05). RESULTS: Image quality scores were comparable between 1.5 T and 3 T with SENSE (4.0 +/- 0.6 vs. 4.2 +/- 0.6, P = 0.21). Bland-Altman reproducibility for mean wall thickness was -0.03 mm +/- 0.15 (1.5 T vs. 1.5 T), 0.01 mm +/- 0.17 (1.5 T vs. 3 T without SENSE), and -0.01 mm +/- 0.18 (1.5 T vs. 3 T with SENSE), P = 0.83. Detection of the presence or absence of plaque was comparable. Bland-Altman reproducibility for area plaque burden was -0.02% +/- 0.32% (1.5 T vs. 1.5 T), 0.06% +/- 0.41% (1.5 T vs. 3 T without SENSE), and 0.11% +/- 0.33% (1.5 T vs. 3 T with SENSE), P = 0.41. CONCLUSION: Black-blood MR imaging of aortic atherosclerosis is very reproducible. Parallel imaging at 3 T permits shorter scan time compared with conventional 1.5-T imaging with comparable measures of atherosclerosis extent.