Artificial intelligence can be used in the identification and classification of shoulder osteoarthritis and avascular necrosis on plain radiographs: a training study of 7,139 radiograph sets.

BACKGROUND AND PURPOSE: Knowledge concerning the use AI models for the classification of glenohumeral osteoarthritis (GHOA) and avascular necrosis (AVN) of the humeral head is lacking. We aimed to analyze how a deep learning (DL) model trained to identify and grade GHOA on plain radiographs performs. Our secondary aim was to train a DL model to identify and grade AVN on plain radiographs. PATIENTS AND METHODS: A modified ResNet-type network was trained on a dataset of radiographic shoulder examinations from a large tertiary hospital. A total of 7,139 radiographs were included. The dataset included various projections of the shoulder, and the network was trained using stochastic gradient descent. Performance evaluation metrics, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were used to assess the network's performance for each outcome. RESULTS: The network demonstrated AUC values ranging from 0.73 to 0.93 for GHOA classification and > 0.90 for all AVN classification classes. The network exhibited lower AUC for mild cases compared with definitive cases of GHOA. When none and mild grades were combined, the AUC increased, suggesting difficulties in distinguishing between these 2 grades. CONCLUSION: We found that a DL model can be trained to identify and grade GHOA on plain radiographs. Furthermore, we show that a DL model can identify and grade AVN on plain radiographs. The network performed well, particularly for definitive cases of GHOA and any level of AVN. However, challenges remain in distinguishing between none and mild GHOA grades.

Deep learning classification of shoulder fractures on plain radiographs of the humerus, scapula and clavicle.

In this study, we present a deep learning model for fracture classification on shoulder radiographs using a convolutional neural network (CNN). The primary aim was to evaluate the classification performance of the CNN for proximal humeral fractures (PHF) based on the AO/OTA classification system. Secondary objectives included evaluating the model's performance for diaphyseal humerus, clavicle, and scapula fractures. The training dataset consisted of 6,172 examinations, including 2-7 radiographs per examination. The overall area under the curve (AUC) for fracture classification was 0.89, indicating good performance. For PHF classification, 12 out of 16 classes achieved an AUC of 0.90 or greater. Additionally, the CNN model had excellent overall AUC for diaphyseal humerus fractures (0.97), clavicle fractures (0.96), and good AUC for scapula fractures (0.87). Despite the limitations of the study, such as the reliance on ground truth labels provided by students with limited radiographic assessment experience, our findings are in concordance with previous studies, further consolidating CNN as potent fracture classifiers in plain radiographs. The inclusion of multiple radiographs with different views from each examination, as well as the generally unselected nature of the sample, contributed to the overall generalizability of the study. This is the fifth study published by our group on AI in orthopaedic radiographs, which has consistently shown promising results. The next challenge for the orthopaedic research community will be to transfer these results from the research setting into clinical practice. External validation of the CNN model should be conducted in the future before it is considered for use in a clinical setting.

Validity, reliability, and responsiveness of the Swedish version of Western Ontario Osteoarthritis of the Shoulder index.

BACKGROUND: The Western Ontario of the Shoulder index (WOOS) is a patient-reported, disease-specific instrument, designed to measure quality of life in patients with osteoarthritis of the shoulder. The Swedish Shoulder Arthroplasty Registry (SSAR) uses WOOS and EuroQoL 5-dimensions 3 levels (EQ-5D-3L) as patient reported outcome measures. The purpose of this study was to test the validity, responsiveness, and reliability of the Swedish translation of WOOS for patients with osteoarthritis of the shoulder. METHODS: Data was collected from three shoulder arthroplasty studies performed during 2005-2013, with 23, 21, and 19 patients respectively. Forms were collected preoperatively, and postoperatively between 12 and 24 months. WOOS and EQ-5D-3L were used in all three studies. Additionally, the Oxford Shoulder Score (OSS) (n = 23) was used in one study, and the Constant-Murley score (CMS) (n = 40) in two of the studies. Validity was analysed by calculating Pearson's correlation coefficient (PCC). Cronbach's alpha (CA) was used to estimate internal consistency and reliability. The responsiveness of WOOS was measured by calculating effect size and standardized response mean. To assess the performance of WOOS over time, we present repeated measures of WOOS in the registry over a 10-year period. RESULTS: The validity analysis showed excellent correlations of WOOS to CMS, OSS and EQ-5D 3L, with Pearson's correlation coefficient of 0.72, 0.83, and 0.62 respectively (P < 0.001). There were adequate floor effects in the sport and lifestyle domains preoperatively, and adequate ceiling effects in all domains postoperatively. There were no floor effects and adequate ceiling effects for total WOOS. Analyzing reliability, Cronbach's alpha was 0.95 for the pre- and postoperative WOOS scores combined. The analysis of responsiveness for WOOS showed an effect size of 2.52 and a standardized response mean of 1.43. The individual results measured by WOOS within the registry shows stable levels from 1 to 10 years. CONCLUSION: The Swedish translation of WOOS is valid, reliable, and responsive for use in a clinical setting for patients with glenohumeral osteoarthritis treated with shoulder arthroplasty, and we regard it as an appropriate instrument for use in the Swedish Shoulder Arthroplasty Registry.

Lower age increases the risk of revision for stemmed and resurfacing shoulder hemi arthroplasty.

Background and purpose - The number of patients where shoulder hemiarthroplasty (SHA) is an option is still substantial. Descriptive analyses performed by the Swedish Shoulder Arthroplasty Registry (SSAR) showed that while patients receiving SHA designs, i.e. resurfacing hemi (RH) and stemmed hemi (SH), reported similar shoulder functionality and quality of life, the revision rate for RH (12%) was larger than for SH (6.7%); this difference was studied. Patients and methods - All primary SHA (n = 1,140) for OA reported to SSAR between 1999 and 2009 were analyzed regarding risk factors for revision and PROM outcome, 950 shoulders with primary OA (POA), and 190 secondary OA (SOA). Mean age was 67.4 years (SD 10.8). PROM including WOOS and EQ-5D were collected at 5 years, until December 31, 2014. Results - 76/950 prostheses because of POA and 16/190 prosthesis because of SOA were revised. Age at primary surgery was the main factor that influenced the risk of revision, lower age increased the risk of revision, and was also the explanation for the difference between SH and RH. We also found that SH and RH had similar outcomes measured by PROM, but the POA group had higher scores than the SOA group with a clinically relevant difference of 10% in WOOS. Interpretation - The risk of revision for SH and RH is similar when adjusted for age and does not depend on primary diagnosis or sex. A lower age increases the risk of revision. Patients suffering from POA experience better shoulder functionality than SOA patients irrespective of implant type.