Research papersAccuracy of MR neurography in the diagnosis of brachial plexopathy

•Established MR neurography criteria for diagnosis of brachial plexopathy had a sensitivity of -- 41% to 71% compared to electrodiagnostic studies.•Interobserver variability in diagnosis was good, with kappa value of 0.658.•Interobserver consistency was improved at 3 T compared to 1.5 T.•Administration of gadolinium contrast did not change the diagnosis.

ObjectiveTo assess the accuracy of MR neurography (MRN) for the diagnosis for brachial plexopathy.Materials and methodsForty-three 43 consecutive cases or MRN with electrodiagnostic study correlation were reviewed independently by 2 readers. Established diagnostic criteria were employed: increased signal intensity, enlargement, loss of fascicular pattern, perineural edema and enhancement after administration of gadolinium contrast. An additional 37 cases without electrodiagnostic correlation were included in a second portion of the study, which evaluated interobserver variability.ResultsThe sensitivity of MRN ranged from 41.2% to 70.6%. Specificity was 97.7% to 100%. There was good interobserver agreement, with kappa value of 0.658. Odds ratio for agreement between observers at 3 T vs. 1.5 T was 1.30.ConclusionsMRN of the brachial plexus showed only moderate sensitivity to brachial neuritis. Interobserver agreement was slightly higher at 3 T than at 1.5 T. MRN can be helpful in confirming an abnormal brachial plexus, but a normal MRN appearance of the plexus should not negate a clinical diagnosis.