Development and validation of a model to predict full thickness rotator cuff tear.

INTRODUCTION

Full-thickness rotator cuff tear (FTRCT) often requires surgical treatment. Magnetic resonance imaging (MRI) is the diagnostic gold standard; however, it is costly and resource-intensive. We aimed to develop and validate a clinical prediction model for FTRCT in patients with RCI, with the goal of identifying those at high risk at the initial visit.

METHODS

This prospective study consecutively enrolled patients clinically diagnosed with RCI from May 2025 to September 2025. All patients underwent shoulder MRI to confirm the presence or absence of FTRCT. Potential predictors involving demographic characteristics, clinical history, and physical examination findings, were collected. Predictor selection was performed using least absolute shrinkage and selection operator (LASSO) regression, followed by multivariable logistic regression to construct a risk prediction model, and then a nomogram was developed.

RESULTS

357 patients were included, MRI confirmed FTRCT in 166 patients (46.5%). Patients were randomly divided into a training set (n = 249) and a test set (n = 108). LASSO regression identified 10 candidate predictors, and multivariable logistic regression identified physical labor, history of trauma, age, daily pain score, pain duration, and physical examination shoulder strength as independent predictors of FTRCT. The nomogram achieved an AUC of 0.910 (95% CI, 0.88-0.95), with 88% sensitivity, 82% specificity, and a C-index of 0.890 (95% CI, 0.83-0.95). Calibration and decision curve analyses supported its accuracy and clinical utility.

CONCLUSIONS

This study developed a preliminary clinical nomogram for predicting FTRCT in patients with RCI, which showed good performance in internal validation and may help support early risk assessment and MRI triage. Its generalizability should be confirmed in external cohorts.

LEVEL OF EVIDENCE: Level III.