INTRODUCTION:
Pedicle screw placement is a widely accepted surgical procedure for spinal fixation. Despite increases in knowledge about and expertise in pedicle screw insertion techniques, overall reported screw misplacement rates are still high. Spinal neuronavigation and intraoperative computed tomography (CT) imaging improves the accuracy and safety of pedicle screw placement through the continuous monitoring of screw trajectory. The purpose of this study is to compare pedicle screw placement under an O-arm intraoperative imaging system assisted by the StealthStation navigation system with screw placement under conventional fluoroscopy (C-arm).

METHODS:
For 222 patients, 1288 implanted pedicle screws in total were evaluated between 2018 and 2020. All patients underwent pedicle screw placement in the thoracic and lumbosacral regions through a posterior approach. Moreover, 107 patients (48.2%), 48 men and 59 women, underwent freehand screw placement under conventional fluoroscopy (C-arm group), whereas 115 patients (51.8%), 53 men and 62 women, underwent pedicle screw insertion under O-arm guidance with the help of the StealthStation neuronavigation system (Medtronic Navigation, Louisville, CO, USA) (O-arm group). Data were recorded and retrospectively analyzed. The accuracy of pedicle screw placement was postoperatively examined by using CT imaging and analyzed according to the Gertzbein-Robbins classification.

RESULTS:
Of the 1288 pedicle screws, 665 (51.6%) were placed with C-arm image-guided assistance with a mean of 6.21 ± 2.1 screws per patient and 643 (48.4%) with O-arm image-guided assistance with a mean of 5.59 ± 1.6 screws. The average time for the screw placement procedure was 3:57 ± 1:07 h in the C-arm group and 4:21 ± 1:41 h in the O-arm group. A correct screw placement was detected in 92.78% of patients in the C-arm group and in 98.13% of patients in the O-arm group. Medial cortical breach was shown in 13 Grade B screws (1.95%), 19 Grade C (2.86%), 14 Grade D (2.11%), and two Grade E (0.3%) in the C-arm group, whereas this was shown in 11 Grade B screws (1.71%) and one Grade C (0.16%) in the O-arm group. Lateral breach occurred in eight screws in both groups. Anterior vertebral body breach was shown in eight screws in the C-arm group, whereas it was shown in four screws in the O-arm group. Reoperation for screw misplacement was mandatory in five patients in the C-arm group and two patients in the O-arm group.

CONCLUSION:
Pedicle screw placement under an O-arm intraoperative imaging system assisted by spinal navigation showed greater accuracy compared with placement under conventional fluoroscopic control, thus avoiding the onset of major postoperative complications. Notably, a reduction in medial and anterior breaches has been demonstrated.