Neuromonitoring with pulse-train stimulation for implantation of thoracic pedicle screws: a blinded and randomized clinical study. Part 1. Methods and alarm criteria

Object. Reports of the accuracy of existing neuromonitoring methods for detecting or preventing medial malpositioning of thoracic pedicle screws have varied widely in their claimed effectiveness. The object of this study was to develop, test, and validate a novel neuromonitoring method for preventing medial malpositioning of pedicle screws in the thoracic spine during surgery. Methods. This is a prospective, blinded and randomized study using a novel combination of input (4-pulse stimulus trains delivered within the pedicle track) and output (evoked electromyography from leg muscles) to detect pedicle track trajectories that once implanted with a screw would cause that screw to breach the pedicle's medial wall and encroach upon the spinal canal. For comparison, the authors also used screw stimulation as an input and evoked electromyogram from intercostal and abdominal muscles as output measures. Intraoperative electrophysiological findings were compared with postoperative CT scans by multiple reviewers blinded to patient identity or intraoperative findings. Results. Data were collected from 71 patients, in whom 802 screws were implanted between the T-1 and L-1 vertebral levels. A total of 32 screws ended up with screw threads encroaching on the spinal canal by at least 2 mm. Pulse-train stimulation within the pedicle track using a ball-tipped probe and electromyography from lower limb muscles correctly predicted all 32 (100%) of these medially malpositioned screws. The combination of pedicle track stimulation and electromyogram response from leg muscles proved to be far more effective in predicting these medially malpositioned screws than was direct screw stimulation and any of the target muscles (intercostal, abdominal, or lower limb muscles) we monitored. Based on receiver operating characteristic analysis, the combination of 10-mA (lower alarm) and 15-mA stimulation intensities proved most effective for detection of pedicle tracks that ultimately gave rise to medially malpositioned screws. Additional results pertaining to the impact of feedback of these test results on surgical decision making are provided in the companion report. Conclusions. This novel neuromonitoring approach accurately predicts medially malpositioned thoracic screws. The approach could be readily implemented within any surgical program that is already using contemporary neuromonitoring methods that include transcranial stimulation for monitoring motor evoked potentials.