Radiographic evaluation of the implant configuration after reverse shoulder arthroplasty (RSA) is time-consuming and subject to interobserver disagreement. The final configuration is a combination of implant features and surgical execution. Artificial intelligence (AI) algorithms have been shown to perform accurate and efficient analysis of images. The purpose of this study was to develop an AI algorithm to automatically measure glenosphere inclination, humeral component inclination, and the lateralization and distalization shoulder angles (DSAs) on postoperative anteroposterior radiographs after RSA.
The Digital Imaging and Communications in Medicine files corresponding to postoperative anteroposterior radiographs obtained after implantation of 143 RSAs were retrieved and used in this study. Four angles were analyzed: (1) glenoid inclination angle (GIA, between the central fixation feature of the glenoid and the floor of the supraspinatus fossa), (2) humeral alignment angle (HAA, between the long axis of the humeral shaft and a perpendicular to the metallic bearing of the prosthesis), (3) DSA, and (4) lateralization shoulder angle (LSA). A UNet segmentation model was trained to segment bony and implant elements using manually segmented training (n = 89) and validation (n = 22) images. Then, an image-processing–based pipeline was developed to measure all 4 angles using AI-segmented images. Measures performed by 3 physician observers and the AI algorithm were then completed in 32 additional images. The agreements among human observers and between observers and the AI algorithm were evaluated using intraclass correlation coefficients (ICCs) and absolute differences in degree.
The ICCs (95% confidence interval) for manual measurements of LSA, DSA, GIA, and HAA were 0.79 (0.55, 0.90), 0.90 (0.80, 0.95), 0.96 (0.93, 0.98), and 0.99 (0.97, 0.99), respectively. The AI algorithm measured the 32 images in the test set in less than 2 minutes. The agreement between observers and the AI algorithm was lowest when measuring the LSA for observer 2, with an ICC of 0.77 (0.52, 0.89), and an absolute difference in degrees (median [interquartile range]) of 5 (4). Better agreements were found between the AI measurements and the average manual measurements: absolute differences in degree for LSA, DSA, GIA, and HAA were 3 (5), 2 (3), 2 (2), and 2 (1), respectively; ICCs for LSA, DSA, GIA, and HAA were 0.89 (0.79, 0.95), 0.96 (0.93, 0.98), 0.85 (0.68, 0.93), and 0.98 (0.95, 0.99), respectively.
The AI algorithm developed in this study can automatically measure the GIA, HAA, LSA, and DSA on postoperative anteroposterior radiographs obtained after implantation on RSA.
A significant proportion of revisions after reverse shoulder arthroplasty (RSA) is attributed to the humeral component. The purpose of this study is to evaluate the radiographic and clinical outcomes of the hybrid humerus technique for RSA using a Grammont-style humeral prosthesis in an onlay fashion with metaphyseal bone impaction grafting technique and undersized stem to avoid humeral stress shielding, notching, and loosening.
This is a prospective case series of patients who underwent RSA using the hybrid humerus technique with a minimum 2-year follow-up. Key steps of this technique include the use of undersized Grammont-style stem, impaction bone grafting of the proximal 5 cm of the humerus, and adjusting the height and offset of the stem depending on the patient stature, desired lateralization and distalization, and joint and soft tissue tension. Radiographic assessments were performed immediately after surgery, and at 1 and 2 years after surgery. These included assessment of metaphyseal and diaphyseal filling ratio, cortical narrowing, radiolucent lines, cortical lucencies, spot welding, scapular notching, and stem alignment. Preoperative and 2-year postoperative clinical assessments included American Shoulder and Elbow Surgeons score, Constant-Murley Score, University of California Los Angeles score, visual analog scale for pain, and active range of motion. Correlation between the filling ratios and clinical outcomes were also evaluated.
Sixty-one patients were included in the study. The average metaphyseal and diaphyseal filling ratio on the postoperative X-ray was 0.66 and 0.54, and 0.67 and 0.54 at 2 years, respectively. Stress shielding was graded as none in 24 (40.7%), mild in 33 (55.9%), and moderate in 2 (3.4%). No stem had a change in position of more than 5°. At 2 years of follow-up, no humeral implant loosening was noted, with only 2 (3.4%) of the stems at risk of loosening. Thirty-nine (66.1%) had no notching, 14 (23.7%) were graded as mild, and 6 (10.2%) had moderate signs of notching. All clinical assessments significantly improved at 2 years (P < .001), with a weak negative correlation between visual analog scale and metaphyseal filling ratio (r = −0.268, P = .036) but none between diaphyseal filling ratio and clinical outcomes.
The hybrid humerus technique of metaphyseal bone grafting with a low filling ratio stem presents a promising solution for reducing humeral complications in RSA. This technique demonstrates a low incidence of stress shielding and loosening, with excellent clinical outcomes at 2 years.
Preoperative planning has gained popularity in the management of reverse shoulder arthroplasty (RSA). Commercially available software provides 3-dimensional segmentation of scapula and humerus, as well as providing arc of motion for the implanted articulation and identifying potential areas of bony impingement. However, these software algorithms use a fixed scapula model, disregarding the preoperative clinical range of motion (C-ROM) of the patient, be it glenohumeral or scapulothoracic, as well as any soft tissue parameters. This study aims to compare the ROM based on preoperative planning software by using the implant position from postoperative computed tomography (CT) images (predicted ROM using preoperative planning software [P-ROM]), with the C-ROM assessed at minimum of 2 years of follow-up.
Preoperative and postoperative CT scans of 46 patients who underwent primary RSA between 2017 and 2021 were analyzed. At the postoperative 2-year review, each patient was assessed for active ROM. Implant size and position based on operative notes and postoperative CT scans were used to replicate the performed surgery in the planning software. Abduction, flexion, and external rotation motion were simulated and recorded. The relationship between C-ROM and P-ROM was investigated using linear regression analysis, Pearson correlation coefficient, and paired t-test.
P-ROM was significantly lower than C-ROM at 2 years postoperatively (P < .001), with an average discrepancy of 78° in abduction, 47° in flexion, and 37° in external rotation (C-ROM: abduction 155° ± 21° [80°-180°]; flexion 160° ± 17° [90°-180°]; external rotation 52° ± 14° [10°-80°] vs. P-ROM: abduction 77° ± 13° [53°-107°]; flexion 112° ± 25° [67°-180°]; external rotation 15° ± 21° [0°-79°]). The linear regression analysis indicated weak agreement between C-ROM and P-ROM (abduction R2 = 0.03; flexion R2 = 0.01; external rotation R2 = 0.04). Pearson’s correlation coefficients revealed weak correlations of −0.18, 0.03, and 0.21 for abduction, flexion, and external rotation, respectively.
P-ROM based on preoperative software in its current form does not allow the prediction of the C-ROM at 2 years of follow-up for patients undergoing RSA.