Evaluating the Accuracy of Emerging Pedicle Screw Placement Technology — A Systemetic Review

Introduction: In spine surgery, technology and technique enjoy a mutually beneficial relationship. Thoracolumbar instrumentation in particular has seen tremendous growth from the mature description of safe pedicle screw placement to widespread adoption of assistive technologies like three-dimensional navigation and intraoperative robotics. Despite significant costs, the marginal benefit of these systems on pedicle screw accuracy remains incompletely described. To evaluate how these systems affect pedicle screw placement accuracy relative to freehand technique, we performed a large-scale systematic review and meta-analysis surveying the outcome of 37,032 screws.

Methods: Our study followed PRISMA guidelines. The PubMed, Medline, and PubMed Central databases were queried for papers from 2020 to 2023 using the terms [pedicle screw accuracy] and [navigation]. We excluded articles not published in English, not peer-reviewed, non-human experiments, or did not report accuracy convertible to standard scales. We identified 91 full-length papers, yielding data on 6138 free-hand, 1821 3D-Navigation, and 12677 robotic screws. Data analysis was performed via custom scripts written in Matlab using standard statistical tests of significance.

Results: Pedicle screws were placed without medial breach in 76.6% of freehand screws, 90.2% of navigated screws, and 89.1% of robot-assisted screws. This reflects a significant improvement between freehand & navigation (p = 0.00005), and freehand & robotic assisted (p = 0.016), but not between technologies (p = 0.70). Screws placed with 2mm breach or less occurred in 91.8% of freehand screws, 96.3% of navigated screws, and 97.5% of robotic-assisted screws. This difference is significant between freehand & navigation (p = 0.0024), and freehand & robotic assisted (p = 0.018), but not between navigation and robotics (0.35). Technological assistance further improved precision: placement standard deviation was reduced from 16.7% in the freehand group to 8.7% in the navigated and 11.8% in the robotic group (F-statistic 0.048, 0.0027, respectively; between technologies 0.09).

Conclusion : Technological assistance during pedicle screw placement is associated with modest improvements in accuracy (5-15%) and substantial improvements in precision (20-50%) via reductions in relative variability. Further effects, such as those on screw size, pseudarthrosis, and adjacent segment disease, remain to be elucidated. Toward this goal, accuracy scales based solely on breach extent may be insufficient.