Cervical nerve root avulsion in brachial plexus injuries: magnetic resonance imaging classification and comparison with myelography and computerized tomography myelography.

Object: The authors describe a new magnetic resonance (MR) imaging technique to demonstrate the status of the cervical nerve roots involved in brachial plexus injury. They discuss the accuracy and reproducibility of a MR imaging-derived classification for diagnosis of nerve root avulsion compared with those of myelography combined with computerized tomography (CT) myelography.

Methods: The overlapping coronal-oblique slice MR imaging procedure was performed in 35 patients with traumatic brachial plexus injury and 10 healthy individuals. The results were retrospectively evaluated and classified into four major categories (normal rootlet, rootlet injuries, avulsion, and meningocele) after confirming the diagnosis by surgical exploration with or without spinal evoked potential (EP) measurements and by referring to myelography and CT myelography findings. The reliability and reproducibility of the MR imaging-based classification was prospectively assessed by eight independent observers, and its diagnostic accuracy was compared with that of traditional myelography/CT myelography classification, correlated with surgical and spinal EP findings in another 50 cervical roots in 10 patients with traumatic brachial plexus injury.

Conclusions: In the retrospective study in which MR imaging and myelography/CT myelography findings involving 175 cervical roots in 35 patients were compared, the sensitivity of detection of the cervical nerve root avulsion was the same (92.9%) with both modalities. In the prospective study, interobserver reliability and intraobserver reproducibility showed that there was no statistically significant difference between MR imaging and myelography/CT myelography and that their accuracy for detecting cervical root avulsion was the same as that in the retrospective study. The overlapping coronal-oblique slice MR imaging technique is a reliable and reproducible method for detecting nerve root avulsion. The information provided by this modality enabled the authors to assess the roots of the brachial plexus and provided valuable data for helping to decide whether to proceed with exploration, nerve repair, primary reconstruction, or other imaging modalities.