Intraoperative Computed Tomography Versus 3D C-Arm Imaging for Navigated Spinal Instrumentation.

Abstract

A prospective case-series study and a retrospective analysis of historical patients for comparison of data. To compare accuracy and limitations of intraoperative computed tomography (iCT)- versus 3D C-arm-based spinal navigation for posterior pedicle screw implantation. Despite the higher accuracy of navigated compared to non-navigated pedicle screw implantation, it remains a matter of debate whether the use of iCT imaging may further benefit navigated spinal instrumentation compared to more commonly used isocentric 3D C-arm imaging. Between 2013 and 2016, 1527 pedicle screws were implanted in 260 patients with iCT (1219 screws) or 3D C-arm (308 screws)-based spinal navigation. Screw positioning was intraoperatively assessed by a second iCT or 3D C-arm (intraoperative accuracy). If necessary, immediate intraoperative screw revision was performed. Thereafter, a third iCT or 3D C-arm scan was performed to confirm repositioning (final accuracy). Clinical and patient data, intraoperative screw assessability, and accuracy rates were retrospectively reviewed and analyzed by an independent observer. Intraoperative CT permitted immediate intraoperative assessment of each implanted screw. In contrast, 39 of the screws visualized with 3D C-arm imaging were intraoperatively not clearly assessable. Regarding the overall precision, iCT and 3D C-arm navigation yielded a comparable intraoperative accuracy (iCT 94.7% vs 3D C-arm 89.4%) and immediate correction of misplaced screws was feasible with both modalities (final accuracy: iCT 95.4% vs 3D C-arm 91.6%). Regarding the region specific performance, however, iCT-based navigation yielded significantly higher final accuracy rates in the cervical (iCT 99.5% vs 3D C-arm 88.9%, P < 0.01) and thoracic (iCT 97.7% vs 3D C-arm 88.8%, P < 0.001) regions. Both iCT and 3D C-arm-based spinal navigation provides high pedicle screw accuracy rates. Immediate screw assessability and placement accuracy in the cervical-thoracic spine, however, appear to be limited with intraoperative 3D C-arm imaging alone. 3.