Diagnostic accuracy of imaging studies for diagnosing root avulsions in post-traumatic upper brachial plexus traction injuries in adults

Background There is no consensus about which type of imaging study, computed tomography myelography (CTM) or magnetic resonance imaging (MRI), provides better information concerning root avulsion in adult brachial plexus injuries. Methods Patients with upper brachial plexus traumatic injuries underwent both CTM and MRI and surgical exploration. The imaging studies were analyzed by two independent radiologists and the data were compared with the intraoperative findings. The statistical analysis was based on dichotomous classification of the nerve roots (normal or altered). The interobserver agreement was assessed using Cohen’s Kappa. The accuracy of CTM and MRI in comparison with the intraoperative findings was evaluated using the same methodology. Results Fifty-two adult patients were included. CTM tended to yield slightly higher percentages of alterations than MRI The interobserver agreement was better on CTM than on MRI for all nerve roots: C5, 0.9960 (strong) vs. 0.145 (poor); C6, 0.970 (strong) vs. 0.788 (substantial); C7, 0.969 (strong) vs. 0.848 (strong). The accuracy regarding the intraoperative findings was also higher on CTM (moderate, kappa 0.40–0.59) than on MRI (minimal, kappa 0.20–0.39) for all nerve roots. Accordingly, the overall percentage concordance (both normal or both altered) was superior in the CTM evaluation (approx. 70–75% vs. 60–65%). CTM was superior for both sensitivity and specificity at all nerve roots. Conclusion CTM had greater interobserver agreement and higher diagnostic accuracy than MRI in adult patients with root avulsions due to brachial plexus injury.