LAUSR

BACKGROUND Commercially available preoperative planning software is now widely available for shoulder arthroplasty. However, without the use of patient-specific guides or intraoperative visual guidance, surgeons have little in vivo feedback to ensure proper execution of the preoperative plan. The purpose of this study was to assess surgeons' ability to implement a preoperative plan in vivo during shoulder arthroplasty. METHODS Fifty primary shoulder arthroplasties from a single institution were retrospectively reviewed. All surgical procedures were planned using a commercially available software package with both multiplanar 2-dimensional computed tomography and a 3-dimensional implant overlay. Following registration of intraoperative visual navigation trackers, the surgeons (1 attending and 1 fellow) were blinded to the computer navigation screen and attempted to implement the plan by simulating placement of a central-axis guide pin. Malposition was assessed (>4 mm of displacement or >10° error in version or inclination). Data were then blinded, measured, and evaluated. RESULTS Mean displacement from the planned starting point was 3.2 ± 2.0 mm. The mean error in version was 6.4° ± 5.6°, and the mean error in inclination was 6.6° ± 4.9°. Malposition was observed in 48% of cases after preoperative planning. Malposition errors were more commonly made by fellow trainees vs. attending surgeons (58% vs. 38%, P = .047). CONCLUSIONS Despite preoperative planning, surgeons of various training levels were unable to reproducibly replicate the planned component position consistently. Following completion of fellowship training, significantly less malposition resulted. Even in expert hands, the orientation of the glenoid component would have been malpositioned in 38% of cases. This study further supports the benefit of guided surgery for accurate placement of glenoid components, regardless of fellowship training.