Assessment of the Carotid Bodies in Magnetic Resonance—A Head-to-Head Comparison with Computed Tomography

Abstract

Objectives: To evaluate carotid body visibility in contrast-enhanced magnetic resonance (MR) studies and to compare the results to contrast-enhanced computed tomography (CT). Methods: Two observers separately evaluated MR and CT examinations of 58 patients. MR scans were acquired with contrast-enhanced isometric T1-weighted water-only Dixon sequence. CT examinations were performed 90 s after contrast agent administration. Carotid bodies’ dimensions were noted and their volumes calculated. To quantify the agreement between both methods, Bland–Altman plots were computed. Receiver operating characteristic (ROC) and its localization-oriented variant (LROC) curves were plotted. Results: Of the 116 expected carotid bodies, 105 were found on CT and 103 on MR at least by a single observer. Significantly more findings were concordant in CT (92.2%) than in MR (83.6%). The mean carotid body volume was smaller in CT (19.4 mm3) than in MR (20.8 mm3). The inter-observer agreement on volumes was moderately good (ICC (2,k) 0.42, p < 0.001), but with significant systematic error. The diagnostic performance of the MR method added up to 88.4% of the ROC’s area under the curve and 78.0% in the LROC algorithm. Conclusions: Carotid bodies can be visualized on contrast-enhanced MR with good accuracy and inter-observer agreement. Carotid bodies assessed on MR had similar morphology as described in anatomical studies.

Keywords

• carotid body
• magnetic resonance imaging
• computed tomography
• Dixon sequence