Predicting Recurrent Instability of the Shoulder (PRIS): A Valid Tool to Predict Which Patients Will Not Have Repeat Shoulder Instability After First-Time Traumatic Anterior Dislocation

Shoulder MRI parameters in anticipating anterior shoulder dislocation: Are they a reliable and reproducible tool in clinical practice?

BACKGROUND

Anterior shoulder dislocation (ASD) is a frequently observed musculoskeletal injury that is often encountered in the context of sports activities or as a result of trauma. Several magnetic resonance imaging (MRI) parameters have been previously investigated for the purpose of characterizing the anatomical features, which could potentially be responsible for the episodes of instability. These measurements have the potential to identify patients who are susceptible to dislocation. Consequently, ensuring the reliability and consistency of these measurements is crucial in the diagnosis and the management of athletic or traumatic shoulder injuries.

METHODS

A group of four students, who had no previous experience in reading MRI series, were selected to perform radiographic measurements on specific parameters of MRI scans. These parameters were glenoid version, glenoid depth, glenoid width, humeral head diameter, humeral containing angle, and the ratio of humeral head diameter to glenoid diameter. The four participants conducted two distinct readings on a total of 28 sets of shoulder MRI scans. Simultaneously, the aforementioned measures were assessed by a consultant shoulder surgeon.

RESULTS

A total of 1512 measurements were categorized into nine sets: eight from students' measurements (two per student) and one from the consultant. Intra-rater reliability assessed by the intra-class correlation (ICC) test indicated excellent or good reliability for all parameters (p < 0.05), with glenoid depth showing the highest (0.925) and humeral-containing angles the lowest (0.675) ICC value. Inter-rater correlation, also evaluated using ICC, demonstrated strong correlation (p < 0.05), with glenoid diameter having the highest ICC score (0.935) and glenoid depth the lowest (0.849). Agreement analysis, expressed by Cohen's Kappa test, revealed substantial agreement (p < 0.05) for all parameters, with humeral head diameter having the highest agreement (0.90) and humeral-containing angle the lowest (0.73).

CONCLUSION

In this study, intra- and inter-rater MRI parameters are substantially concordant. Credibility comes from these reliability and agreement analyses' statistical significance. Glenoid diameter and depth are the most reliable intrarater and interrater, respectively. Best agreement was with the humeral-containing angle. These data demonstrate repeatability and clinical relevance.

LEVEL OF EVIDENCE

Level IV.

Patients with recurrent anterior shoulder instability exhibit altered glenohumeral and scapulothoracic joint kinematics during upper limb movement: A prospective comparative study

BACKGROUND

Altered shoulder kinematics in patients with recurrent anterior shoulder instability remains poorly understood. This prospective study aimed to quantify in vivo glenohumeral and scapulothoracic joint kinematics and joint-contact positions in patients with shoulder instability and healthy controls.

METHODS

Twenty patients with recurrent anterior shoulder instability (mean 28 years) and five patients without shoulder pathology (mean 39 years) were scanned using open CT in six static upper limb positions including 90° of abduction, combined abduction and external rotation, 90° of flexion, lift-off position (i.e. reaching behind the back) and the neutral shoulder with external rotation. Image datasets were digitally reconstructed to quantify shoulder joint kinematics and glenohumeral translation.

FINDINGS

At 90° of abduction, instability patients demonstrated significantly less glenohumeral abduction and a reciprocal increase in upward scapulothoracic rotation compared to controls (mean difference: 13.3°, p = 0.038). With the shoulder in combined abduction and external rotation, instability patients showed a significant increase in glenohumeral rotation and a reciprocal decrease in scapulothoracic rotation compared to controls (mean difference: 5.0°, p = 0.042). There were no significant differences in humeral head translation in the sagittal plane (anterior-posterior axis) for all motions tested (p > 0.05).

INTERPRETATION

Scapulothoracic and glenohumeral kinematics are significantly different between patients with recurrent anterior shoulder instability and those with a healthy shoulder. Instability patients compensate for reduced glenohumeral function during abduction by increasing scapular rotation. With the shoulder in combined abduction and external rotation position, greater glenohumeral joint angles without significantly increased humeral head translation suggest altered neuromuscular control in the unstable shoulder.

Understanding Anterior Shoulder Instability Through Machine Learning: New Models That Predict Recurrence, Progression to Surgery, and Development of Arthritis

Background

Management of anterior shoulder instability (ASI) aims to reduce risk of future recurrence and prevent complications via nonoperative and surgical management. Machine learning may be able to reliably provide predictions to improve decision making for this condition.

Purpose

To develop and internally validate a machine-learning model to predict the following outcomes after ASI: (1) recurrent instability, (2) progression to surgery, and (3) the development of symptomatic osteoarthritis (OA) over long-term follow-up.

Study Design

Cohort study (prognosis); Level of evidence, 2.

Methods

An established geographic database of >500,000 patients was used to identify 654 patients aged <40 years with an initial diagnosis of ASI between 1994 and 2016; the mean follow-up was 11.1 years. Medical records were reviewed to obtain patient information, and models were generated to predict the outcomes of interest. Five candidate algorithms were trained in the development of each of the models, as well as an additional ensemble of the algorithms. Performance of the algorithms was assessed using discrimination, calibration, and decision curve analysis.

Results

Of the 654 included patients, 443 (67.7%) experienced multiple instability events, 228 (34.9%) underwent surgery, and 39 (5.9%) developed symptomatic OA. The ensemble gradient-boosted machines achieved the best performances based on discrimination (via area under the receiver operating characteristic curve [AUC]: AUC_recurrence = 0.86), AUC_surgery = 0.76, AUC_OA = 0.78), calibration, decision curve analysis, and Brier score (Brier_recurrence = 0.138, Brier_surgery = 0.185, Brier_OA = 0.05). For demonstration purposes, models were integrated into a single web-based open-access application able to provide predictions and explanations for practitioners and researchers.

Conclusion

After identification of key features, including time from initial instability, age at initial instability, sports involvement, and radiographic findings, machine-learning models were developed that effectively and reliably predicted recurrent instability, progression to surgery, and the development of OA in patients with ASI. After careful external validation, these models can be incorporated into open-access digital applications to inform patients, clinicians, and researchers regarding quantifiable risks of relevant outcomes in the clinic.

Editorial Commentary: Management of First-Time Anterior Shoulder Instability Requires Risk Stratification and Surgery for Many, But Not All

The management of a patient with a first-time anterior shoulder dislocation has been the subject of longstanding debate among shoulder surgeons. A number of prognostic factors for recurrent instability have been proposed, including younger age, male sex, contact sports, and glenoid bone loss. Predictive tools and scores have been developed to assist in risk stratifying this patient population; however, no universally agreed upon, clinically validated algorithm exists. More recently, there has been emerging evidence favoring early surgical stabilization, as it has been shown to result in better overall outcomes compared with patients undergoing surgery following episodes of recurrent instability. With each subsequent dislocation or subluxation event, there is increased glenoid bone loss (and development of inverted-pear glenoid), a greater prevalence of engaging (i.e., off-track) Hill-Sachs lesions, more extensive labral tears, a greater risk of rotator cuff involvement (in the older patient), and increased plastic and/or permanent deformation, elongation, and compromise of the antero-inferior glenohumeral joint capsule and associated inferior glenohumeral ligament complex. Moreover, there is now sufficient evidence to suggest that recurrence comes at a cost, as it is a major risk factor for poor outcomes following arthroscopic stabilization. However, one risk is overtreatment, potentially exposing those individuals who would not have had another instability event due to an unnecessary procedure. We should continue to use the available evidence within the literature to help risk-stratify patients and develop an individualized treatment plan through a shared decision-making process with the patient.