ScholarWorks@중앙대학교

초록

Background Pedicle screw fixation is the gold standard for spinal stabilization, but accurate placement remains challenging. Conventional navigation systems have limitations including workflow disruption and divided attention between the surgical field and external monitors. This study evaluates the feasibility and accuracy of a novel augmented reality navigation for pedicle screw fixation (ARNF) system in human cadaveric specimens. Methods Two human cadaveric specimens were used for placement of 18 pedicle screws (10 in cadaver 1, 8 in cadaver 2) across thoracolumbar vertebrae (T12–L4). A custom-developed head-mounted display with optical tracking was employed to project 3-dimensional reconstructed spine models and planned screw trajectories into the surgeon's field of view. Registration was performed using 15 anatomical landmarks with a point-pair matching technique. Accuracy was assessed by postoperative computed tomography scan, measuring entry point deviation, trajectory deviation, and angular difference, and using the Gertzbein-Robbins classification. Results Of the 18 pedicle screws placed, 16 (88.89%) achieved grade A accuracy (complete intrapedicular placement) and 2 (11.11%) were grade B (minor breach < 2 mm), yielding 100% clinically acceptable placement. Technical accuracy measurements revealed a mean entry point deviation of 0.17 mm (standard deviation [SD], 0.11 mm), mean tip deviation of 0.80 mm (SD, 0.33 mm), and mean angular deviation of 2.14° (SD, 0.69°). The average placement time was 3.1 minutes per screw. Conclusions The ARNF system demonstrated exceptional accuracy for pedicle screw placement in a human cadaveric model with high precision at critical anatomical landmarks. This technology allows surgeons to maintain visual focus on the surgical field while providing precise navigation guidance, potentially addressing key limitations of conventional navigation systems.

키워드