The Validity of Radiographic Measurements Compared to 3D-surface Model-Based Measurements in Shoulders with Reverse Total Shoulder Arthroplasty

Background: Postoperative assessment following reverse total shoulder arthroplasty (rTSA) typically involves plain radiographs to evaluate implant positioning parameters, such as humeral distalization and lateralization along with medialization and distalization of the center of rotation (COR). However, the precision of these radiographic measurements remains unclear. This study aimed to validate the accuracy of radiographic two-dimensional (2D) measurements compared to three-dimensional (3D) surface model-based measurements derived from CT data for evaluating postoperative parameters in rTSA. Additionally, we examined how these parameters affect the distances between COR and the acromion (DA), and the COR and the greater tuberosity (DGT).

Methods: The study included 31 shoulders from 30 patients who underwent rTSA at a single institution. Radiographic images and computed tomography (CT) scans were obtained at an average of 16 ± 5 months postoperatively. Implant position parameters measured on both 2D radiographs and 3D surface models included medialization and distalization of the COR, humeral distalization, humeral lateralization offset, DA, and DGT. Measurements were conducted independently on two separate occasions by two certified orthopedic surgeons. Intraclass correlation coefficients (ICCs) were used to assess measurement reliability. Bland-Altman plots analyzed the agreement between 2D and 3D measurements. A linear mixed-effects model was used to analyze the relationship between implant positioning measurements and the DA and DGT.

Results: The ICCs for radiographic and 3D measurements showed excellent reliability (>0.9 for all parameters). Bland-Altman analysis revealed minimal differences between 2D and 3D measurements for most parameters. However, significant fixed biases were noted for humeral distalization and distalization of the COR, indicating potential underestimations in radiographic assessments. The linear mixed-effect model revealed that DA was positively associated with distalization of the COR, humeral lateralization offset, and medialization of the COR (P=0.008, P=0.004, and P=0.005, respectively), but not humeral distalization (P = 0.798). DGT was positively associated with humeral distalization, humeral lateralization offset, and medialization of COR (P=0.002, P<0.001 and P<0.001, respectively), but not distalization of the COR (P=0.468).

Conclusions: Radiographic 2D measurements and 3D-surface model-based measurements are comparable for evaluating postoperative parameters of DA, DGT, humeral lateralization offset, and medialization of COR in shoulders with rTSA. However, fixed biases were present for humeral distalization and distalization of the COR. A larger DA can be achieved by medializing and distalizing the COR, while smaller DGT can be achieved by decreasing humeral lateralization and distalization and by lateralizing the glenoid.