Augmented reality-assisted intraoperative navigation increases precision of glenoid inclination in reverse shoulder arthroplasty.

Background: Reverse total shoulder arthroplasty (RTSA) is the standard of care for patients with glenohumeral osteoarthritis and rotator cuff deficiency. Preoperative RTSA planning based on medical images and patient-specific instruments has been established over the last decade. This study aims to determine the effects of using augmented reality-assisted intraoperative navigation (ARIN) for baseplate positioning in RTSA compared to preoperative planning. It is hypothesized that ARIN will decrease deviation between preoperative planning and postoperative baseplate positioning. Moreover, ARIN will decrease deviation between the (senior) more (>50 RTSAs/yr) and less experienced (junior) surgeon (5-10 RTSAs/yr).

Methods: Preoperative CT scans of 16 fresh-frozen cadaveric shoulders were obtained. Baseplate placements were planned using a validated software. The data were then converted and uploaded to the augmented reality system (NextAR; Medacta International). Each of the 8 RTSAs were implanted by a senior and a junior surgeon, with 4 RTSAs using ARIN and 4 without. A postoperative CT scan was performed in all cases. The scanned scapulae were segmented, and the preoperative scan was laid over the postoperative scapula by the nearest iterative point cloud analysis. The deviation from the planned entry point and trajectory was calculated regarding the inclination, retroversion, medialization (reaming depth) and lateralization, anteroposterior position, and superoinferior position of the baseplate. Data are reported as mean ± standard deviation (SD) or mean and 95% confidence interval (CI). P values < .05 were considered statistically significant.

Results: The use of ARIN yielded a reduction in the absolute difference between planned and obtained inclination from 9° (SD: 4°) to 3° (SD: 2°) (P = .011). Mean difference in planned-obtained inclination between surgeons was 3° in free-hand surgeries (95% CI: -4 to 10; P = .578), whereas this difference reduced to 1° (95% CI: -6 to 7, P = .996) using ARIN. Retroversion, medialization (reaming depth) and lateralization, anteroposterior position, and superoinferior position of the baseplate were not affected by using ARIN. Surgical duration was increased using ARIN for both the senior (10 minutes) and junior (18 minutes) surgeon.

Conclusions: The implementation of ARIN leads to greater accuracy of glenoid component placement, specifically with respect to inclination. Further studies have to verify if this increased accuracy is clinically important. Furthermore, ARIN allows less experienced surgeons to achieve a similar level of accuracy in component placement comparable to more experienced surgeons. However, the potential advantages of ARIN in RTSA are counterbalanced by an increase in operative time.