Augmented reality in minimally invasive spine surgery: early efficiency and complications of percutaneous pedicle screw instrumentation.

Abstract

Augmented reality (AR) employs an optical projection directly onto the user's retina, allowing complex image overlay on the natural visual field. In general, pedicle screw accuracy rates have improved with increasingly use of technology, with navigation-based instrumentation described as accurate in 89%-100% of cases. Emerging AR technology in spine surgery builds upon current spinal navigation to provide 3-dimensional imaging of the spine and powerfully reduce the impact of inherent ergonomic and efficiency difficulties. This publication describes the first known series of in vivo pedicle screws placed percutaneously using AR technology for MIS applications. After IRB approval, 3 senior surgeons at 2 institutions contributed cases from June, 2020 - March, 2022. 164 total MIS cases in which AR used for placement of percutaneous pedicle screw instrumentation with spinal navigation were identified prospectively. 155 (94.5%) were performed for degenerative pathology, 6 (3.6%) for tumor and 3 (1.8%) for spinal deformity.  These cases amounted to a total of 606 pedicle screws; 590 (97.3%) were placed in the lumbar spine, with 16 (2.7%) thoracic screws placed. Patient demographics and surgical metrics including total posterior construct time (defined as time elapsed from preincision instrument registration to final screw placement), clinical complications and instrumentation revision rates were recorded in a secure and de-identified database. The AR system used features a wireless headset with transparent near-eye display which projects intra-operative 3D imaging directly onto the surgeon's retina. After patient positioning, 1 percuntaneous and 1 superficial reference marker are placed. Intra-operative CT data is processed to the headset and integrates into the surgeon's visual field creating a "see-through" 3D effect in addition to 2D standard navigation images. MIS pedicle screw placement is then carried out percutaneously through single line of sight using navigated instruments. Time elapsed from registration and percutaneous approach to final screw placement averaged 3 minutes and 54 seconds per screw.  Analysis of the learning curve revealed similar surgical times in the early cases compared to the cases performed with more experience with the system.  No instrumentation was revised for clinical or radiographic complication at final available follow-up ranging from 6-24 months. A total of 3 screws (0.49%) were replaced intra-operatively. No clinical effects via radiculopathy or neurologic deficit postoperatively were noted. This is the first report of the use of AR for placement of spinal pedicle screws using minimally invasive techniques.   This series of 164 cases confirmed efficiency and safety of screw placement with the inherent advantages of AR technologies over legacy enabling technologies.