Artikel
Quality assurance in CT-navigated spinal instrumentation – accuracy of postoperative screw positioning in relation to 3D-navigated trajectory planning
Qualitätssicherung in der CT-navigierten Wirbelsäuleninstrumentierung – Genauigkeit der postoperativen Schraubenpositionierung in Bezug auf die 3D-navigierte Trajektorienplanung
Suche in Medline nach
Autoren
Veröffentlicht: | 4. Juni 2021 |
---|
Gliederung
Text
Objective: Screw accuracy in spinal instrumentation has so far mainly been evaluated from a clinical point of view according to positioning within cortical bone compartments or in relation to nervous structures. In the light of the anticipated advancements of robotic assistance devices for spinal instrumentation however, it is important to also evaluate the screw accuracy in relation to a preoperative 3D-trajectory plan for quality assurance in navigated screw implants. We sought to evaluate how accurate a 3D-trajectory plan for lumbar pedicle screws can be transferred to the surgical site using CT-Navigation.
Methods: This retrospective analysis was performed on 15 consecutive cases of CT-navigated spinal instrumentations for lumbar spondylosis between February 2014 and June 2016. Intraoperatively, a dedicated 3D-trajectory was planned for all screws using a commercial navigation system. On postoperative CT, implanted screws were manually segmented to derive achieved screw positions. Comparison of 3D-trajecotry plans and implanted screws was performed on the Medical Imaging Interaction Toolkit (MITK) by computing minimal absolute differences (MAD) of screw head and tip points (in mm) and screw axis (in degree), respectively. The percentual overlap of planned and implanted screws was calculated for each screw (Dice-Coefficient). Clinical acceptability of each screw was evaluated by the Gertzbein-Robins classification. Results were evaluated considering the registration error obtained during navigation.
Results: We obtained data for 104 screws covering levels L2-S1. A median axis deviation of 5° (range 1-22°) was shown between planned and implanted screws. Screw head and tip points showed median MAD of 2mm (range 0-11mm) and 4mm (range 0-16mm), respectively. Median Dice Coefficient between planned and implanted screws was 65% (range 34-99%). Screw positioning was clinically acceptable in all cases using the Gertzbein-Robins classification, showing no or only minor (<2mm) cortical breach in 83 (80%) and 21 (20%) cases, respectively. Mean registration error was 7mm (range 5-9).
Conclusion: Despite the clinically acceptable screw placements in all the reviewed screws, considerable deviation between planned and actual trajectory was observed, which exceeded the mean registration error. Hence, factors contributing to screw variations should be evaluated in the future, helping to improve CT-navigated as well as robot-assisted placement of spinal screw implants