Accuracy of pedicle screw placement using intraoperative neurophysiological monitoring and computed tomography.

Fifty consecutive patients with posterior thoracolumbar spine fusion were included in a prospective study to determine the accuracy of intraoperative neurophysiological monitoring (IONM) for safe pedicle screw placement using postoperative computed tomography (CT). The patients were allocated into two equal groups. Pedicle screw placement was evaluated intraoperatively by using the image intensifier. In group A, the integrity of the pedicle wall was evaluated intraoperatively with monopolar stimulation of each screw head with a hand-held single-tip stimulator; the compound muscle action potentials were recorded. A constant current threshold of 7 mA was considered indicative of pedicle breach; < 7 mA was considered as direct contact with neural elements, and > 7mA was considered normal. In group B, pedicle screw placement was performed without IONM. Overall, 306 pedicle screws were inserted in both groups. Postoperatively, all patients underwent CT scans of the spine to evaluate pedicle screw placement. Intraoperatively, five screws in respective group A patients had to be repositioned after IONM (threshold of < 7 mA); in these patients, postoperative CT scans showed proper screw placement. Postoperative CT scans showed eight misdirected screws; two screws (1.26%) in group A patients and six screws (4%) in group B patients. Two screws were misdirected through the medial pedicle wall and six screws were misdirected through the lateral pedicle wall. Both medially misdirected screws were observed in group B patients (1.35%); these patients developed neurologic symptoms postoperatively and underwent revision surgery, with redirection of the misdirected screws and subsequent resolution of the neurologic symptoms. Two of the six laterally misdirected screws were observed in group A patients (1.26%); the remaining four laterally misdirected screws were observed in group B patients (2.7%). None of these patients had neurologic sequelae; no revision surgery was required. The cut-off value of 7 mA had a 98.73% (> 95%) positive predictive value for accurate pedicle screw placement.