Accuracy of Subaxial Cervical Pedicle Screw Placement Using Direct Visualization Versus Computed Tomography-Based Navigation

Study Design: Retrospective analysis of operative data from cadaveric cervical spines. Objective: To evaluate the accuracy of neuronavigation compared with laminotomy with direct visualization (DV) of the pedicle for placement of subaxial pedicle screws. Summary of Background Data: Subaxial pedicle screws provide superior fixation compared with other posterior cervical fixation strategies. However, high accuracy is required for safe placement, given the proximity of critical neurovascular structures. Computed tomography (CT)-based neuronavigation has increased in popularity for placement of spinal implants, including subaxial pedicle screws. However, the accuracy of the technique for this application has not been extensively evaluated. Methods: Six fresh-frozen cadaveric spines (occiput to T2) were prepared. Pedicle screws were placed from C3 to C7 on either side using either the DV or neuronavigation technique (alternating sides between specimens). Pedicles with diameters <4 mm were excluded. For the DV technique, a hemilaminotomy was performed for DV of pedicle borders and to determine appropriate screw medialization and trajectory. Neuronavigation screws were placed using CT-based navigation with a reference frame mounted on the C2 spinous process. Screw position was evaluated using postoperative CT, and breaches were classified using the Neo classification. Results: Fifty pedicle screws were placed at 25 levels in 6 cadaveric spines; 25 screws each were placed using neuronavigation or DV. No significant difference in accuracy was found between the 2 techniques. Three (12%) breaches occurred in the DV group, and 9 (36%) breaches occurred in the neuronavigation group (P=0.10). The breaches were evenly distributed across all levels. There were no high-grade breaches with DV and only 1 (4.0%) with neuronavigation (P>0.99). Average pedicle cortical and medullary bone widths were higher for levels with no breach (P=0.009 and P=0.02, respectively). Conclusions: High accuracy can be achieved with both neuronavigation and DV for placement of subaxial cervical pedicle screws in cadavers.