Three-dimensional fat-saturated T1-weighted Cartesian volumetric interpolated breath-hold examination (VIBE) for the diagnosis of aortitis in patients with suspected large vessel vasculitis: a comparative study with 18F-FDG PET applying fully integrated PET/MRI

AIM: To evaluate the feasibility of T1-weighted (T1W) three-dimensional (3D) fat saturated Cartesian volumetric interpolated breath-hold examination (VIBE) magnetic resonance imaging (MRI) sequence for the diagnosis of aortitis in patients with suspected large vessel vasculitis (LVV) applying fully integrated 2-[F-18]-fluoro-2-deoxy-D-glucose (F-18-FDG) positronemission tomography (PET)/MRI. MATERIAL AND METHODS: Fourteen patients with aortitis and 14 patients with a negative study for aortitis using F-18-FDG PET as the standard of reference for the evaluation of inflammatory aortic involvement were included retrospectively. All patients were imaged at 3 T using T1W VIBE pre- and post-contrast. Four aortic segments were evaluated for image quality (IQ), diagnostic confidence (DC), and the degree of inflammatory activity (IA) using a Likert scale. Binomial and generalised estimating equation model tests were used to assess the diagnostic performance of T1W VIBE. Cohen's k was applied to test for interobserver reproducibility with respect to IA. Spearman's rank correlation coefficient was calculated to examine correlations between IQ, DC, IA, and PET results. RESULTS: On a patient- and segment-based analysis, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 85.7% and 59.8%, 100% and 100%, 100% and 100%, 87.5% and 68%, and 92.9% and 82.1%, respectively. IQ and DC were acceptable to good in all examinations and substantial interobserver agreement was observed for IA (Cohen's k = 0.69). IQ and DC as well as IA and F-18-FDG vessel wall uptake were significantly correlated (r=0.763 and 0.679, respectively; p<0.0001). CONCLUSION: T1W 3D fat saturated VIBE MRI allows diagnosis of aortitis and may aid in the management of patients with suspected LVV. (C) 2019 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.