Mixed-Reality Improves Execution of Templated Glenoid Component Positioning in Shoulder Arthroplasty: A CT Imaging Analysis.

Introduction Glenoid placement is critical for successful outcomes in total shoulder arthroplasty (TSA). Preoperative templating with three-dimensional (3D) imaging has improved implant positioning, but deviations from the planned inclination and version still occur. Mixed-Reality (MR) is a novel technology that allows surgeons intra-operative access to 3D imaging and templates, capable of overlaying the surgical field to help guide component positioning. The purpose of this study was to compare the execution of preoperative templates using MR versus standard instruments (SI). Methods Retrospective review of 97 TSAs (18 anatomic, 79 reverse) from a single high-volume shoulder surgeon between January 2021 and February 2023, including only primary diagnoses of osteoarthritis, rotator cuff arthropathy, or a massive irreparable rotator cuff tear. To be included, patients needed a templated preoperative plan and then a postoperative computed tomography scan. Allocation to MR vs. SI was based on availability of the MR headset, industry technical personnel, and the templated preoperative plan loaded into the software, but preoperative or intraoperative patient factors did not contribute to the allocation decision. Postoperative inclination and version were measured by two independent, blinded physicians and compared to the preoperative template. From these measurements, we calculated the mean difference, standard deviation (SD), and variance to compare MR and SI. Results Comparing 25 MR to 72 SI cases, MR significantly improved both inclination (p<0.001) and version (p<0.001). Specifically, MR improved the mean difference from preoperative templates (by 1.9º inclination, 2.4º version), narrowed the SD (by 1.7o inclination, 1.8o version), and decreased the variance (11.7 to 3.0 inclination, 14.9 to 4.3 version). A scatterplot of the data demonstrates a concentration of MR cases within 5° of plan relative to SI cases typically within 10° of plan. There was no difference in operative time. Conclusion MR improved the accuracy and precision of glenoid positioning. Although it is unlikely that 2o makes a detectable clinical difference, our results demonstrate the potential ability for technology like MR to narrow the bell curve and decrease the outliers in glenoid placement. This will be particularly relevant as MR and other similar technologies continue to evolve into more effective methods in guiding surgical execution.