Machine Learning Can Predict Level of Improvement in Shoulder Arthroplasty

Abstract

Background: The ability to accurately predict postoperative outcomes is of considerable interest in the field of orthopaedic surgery. Machine learning has been used as a form of predictive modeling in multiple health-care settings. The purpose of the current study was to determine whether machine learning algorithms using preoperative data can predict improvement in American Shoulder and Elbow Surgeons (ASES) scores for patients with glenohumeral osteoarthritis (OA) at a minimum of 2 years after shoulder arthroplasty. Methods: This was a retrospective cohort study that included 472 patients (472 shoulders) diagnosed with primary glenohumeral OA (mean age, 68 years; 56% male) treated with shoulder arthroplasty (431 anatomic total shoulder arthroplasty and 41 reverse total shoulder arthroplasty). Preoperative computed tomography (CT) scans were used to classify patients on the basis of glenoid and rotator cuff morphology. Preoperative and final postoperative ASES scores were used to assess the level of improvement. Patients were separated into 3 improvement ranges of approximately equal size. Machine learning methods that related patterns of these variables to outcome ranges were employed. Three modeling approaches were compared: a model with the use of all baseline variables (Model 1), a model omitting morphological variables (Model 2), and a model omitting ASES variables (Model 3). Results: Improvement ranges of ≤28 points (class A), 29 to 55 points (class B), and >55 points (class C) were established. Using all follow-up time intervals, Model 1 gave the most accurate predictions, with probability values of 0.94, 0.95, and 0.94 for classes A, B, and C, respectively. This was followed by Model 2 (0.93, 0.80, and 0.73) and Model 3 (0.77, 0.72, and 0.71). Conclusions: Machine learning can accurately predict the level of improvement after shoulder arthroplasty for glenohumeral OA. This may allow physicians to improve patient satisfaction by better managing expectations. These predictions were most accurate when latent variables were combined with morphological variables, suggesting that both patients’ perceptions and structural pathology are critical to optimizing outcomes in shoulder arthroplasty. Level of Evidence: Therapeutic Level IV . See Instructions for Authors for a complete description of levels of evidence.

Artificial intelligence-based applications in shoulder surgery leaves much to be desired: a systematic review

“…Three studies have applied ML to predict functional clinical outcomes following shoulder arthroplasty. 30 , 31 , 42 Kumar et al used 4782 primary aTSA or rTSA patients from a multicenter database to build and evaluate the performance of three ML models in predicting functional outcomes, such as American Shoulder and Elbow Surgeons (ASES) scores, global shoulder function scores, active abduction, and external rotation at multiple time points postoperatively. 30 Moreover, Kumar et al built upon this study using 5774 aTSA and rTSA patients from a multicenter database by evaluating if an ML model with a minimal feature set of only 19 input parameters is as accurate as a full feature set with hundreds of inputs in predicting functional clinical outcomes postoperatively.…”

Modern Applications of Machine Learning in Shoulder Arthroplasty

“…McLendon et al analyzed the ability of algorithms to predict improvement in American Shoulder and Elbow Surgeons (ASES) scores for 472 TSA patients at a minimum of 2 years of follow-up 44 . Each patient was grouped into 1 of 3 classes based on postoperative ASES score improvement.…”

Current clinical applications of artificial intelligence in shoulder surgery: what the busy shoulder surgeon needs to know and what’s coming next

“…In shoulder surgery, advancements are being made in applying AI for patient-specific risk predictions, including predicting complications, outcomes, and costs. 2 , 6 , 13 , 21 , 31 , 36 , 37 , 46 , 47 , 48 , 53 , 60 Most recently published studies are focused on predicting perioperative complications in shoulder arthroplasty, particularly within the 30-day postoperative period. 2 , 6 , 13 , 14 , 21 , 31 , 46 In 2021, Lopez et al used 21,544 elective primary shoulder arthroplasty cases from a national database to develop and test ML models for predicting nonhome discharge and the occurrence of 1 or more postoperative complications within 30-days.…”

“…AI is also being increasingly explored for predicting functional outcomes, patient satisfaction, and costs following shoulder arthroplasty. 36 , 37 , 53 Kumar et al used 2153 primary anatomic total shoulder arthroplasty (TSA) and 3621 reverse arthroplasty patients to develop and test a ML model for predicting achievement of the minimal clinically important difference and substantial clinical benefit for the American Shoulder and Elbow Surgeons score, Constant Score, Global Shoulder Function score, and other functional outcomes at 2-3 years postoperatively. 36 Similarly, ML models for predicting achievement of the minimal clinically important difference and substantial clinical benefit patient satisfaction-based thresholds for active internal rotation following anatomic TSA and reverse TSA have been tested.…”