Assessment of 3D automated software to predict postoperative impingement free range of motion after reverse shoulder arthroplasty

Background: In the past few years 3-dimensional (3D) automated software for preoperative planning in shoulder arthroplasty has increased in practice. It has been used to improve accuracy on implant placement, size, and selection, especially on the glenoid side. New additions have been added however, to look at prediction of range of motion (ROM) for after surgery. Therefore, the purpose of this study was to assess the accuracy of the software feature of predicting impingement-free range of motion compared to postoperative functional outcomes. Methods: A shoulder arthroplasty database including 192 patients was reviewed and 55 patients who underwent a primary RSA without evidence of preoperative fracture, trauma, or neurological associated diseases which could impact postoperative rehabilitation were included in the study. Preoperative plans for RSA were created for all patients based on preoperative shoulder CT scan and each patient's preoperative predictive ROM (PROM) suggested by the software was recorded. These PROM were then compared to postoperative final follow-up clinical range of motion for accuracy (CE ROM). Demographic information and Walch classification were recorded. Results: Our cohort was composed of 30 males and 25 females with an average age of 71.2 years old and mean BMI 30.0 kg/m². The predominant diagnosis was rotator cuff arthropathy (RCA: 32 patients) followed by severe osteoarthritis (SOA: 19 patients). The mean difference between clinical and predicted ROM for active FF was 42° (SD = 46°), passive FF was 61° (SD = 39°), passive ABD was 81° (SD = 33°), active ER was 22° (SD = 22°), and passive ER was 24° (SD = 25°). Our results also showed that the CE ROM was significantly higher than PROM for each clinical measurement (P< .001). We also found weak correlations between the groups for active FF (r = 0.106), active ER (r = 0.246), passive ER (r = 0.209) and no correlation for remaining groups. Conclusion: Not all features available in preoperative planning for RSA using 3D automated software can be translated into practice. Our results demonstrated PROM weakly correlated with CE ROM for patients who underwent RSA. Surgeons need to be aware that surgical decisions should not be altered based on this feature. More research needs to be done to validate this new tool incorporating both scapulothoracic motion and clinical correlation with predictive ROM. As we move into the era with 3D preoperative simulation and optimizing outcomes for shoulder arthroplasty these features incorporating clinical outcomes and function will be critical to consider. Level of Evidence: Level IV, retrospective case series.