Background: Two-dimensional image guidance and navigation can help to reduce the number of misplaced pedicle screws, but do not completely prevent misplacement. This experimental, retrospective, non-inferiority study was designed to evaluate and compare the efficacy of a novel 3D imaging technique versus conventional postoperative CT-scan, for intra-operative determination of pedicle screw position accuracy.
Methods: The capacity of C-OnSite® to intraoperatively assess screw placement was evaluated in 28 clinical cases and 23 deliberately misplaced screws in a cadaver model, and compared to placement accuracy determined by standard CT. The position of each implant, as viewed by both modalities, was graded by three neurosurgeons, one orthopaedic-surgeon and one radiologist. The intermodal variance determined the difference between CT- and C-OnSite® results for each observer, while the inter-observer variance measured the difference between ratings of the same modality by different observers.
Results: C-OnSite® successfully assessed 120/138 screws (25/28 cases). Mean procedural fluoroscopy time was 132 ± 51 s, and 40 ± 16s per C-OnSite® scan. The average inter-modality variance was ,15 % with mismatches >1° between C-OnSite® and the gold-standard imaging technique in only 2 % of the comparisons. Average inter-observer variances were about similar (12 % for CT and 18 % for C-OnSite®), with deviations of >1° reaching 1 % for CT and 3 % for C-OnSite®. Individual variances between experienced only observers differed even less.
Conclusions: C-OnSite® is a feasible, reliable and intuitive means of intraoperatively visualizing pedicle screw positions and might render the majority of postoperative CTs superfluous. C-OnSite® might help avoid re-operations for screw re-positioning.