Artificial intelligence empowering the diagnosis and treatment of rotator cuff tears: A systematic review from image optimization to prognosis prediction.
Haitao Wang, Deming Guo, Yingqi Gao et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review examined original studies applying artificial intelligence across the rotator cuff tear care pathway, including imaging, diagnosis, anatomical measurement, and prognosis. It included 66 reports spanning multiple imaging modalities and clinical data. Although promising internal performance was reported, methodological concerns limited clinical inference: risk of bias was high in 27/36 diagnostic studies, and all 8/8 completed prediction-model studies were rated high risk of bias. The authors conclude that current evidence demonstrates proof-of-concept value but does not establish routine clinical readiness or support replacing clinician judgment.
Key findings
- Of 36 diagnostic studies, 27 were classified as having high overall risk of bias, 6 unclear risk, and 3 low risk.
- All 8 completed prediction-model studies were rated high risk of bias; one protocol without model results was not rated.
- The 66 reports covered diagnosis or classification (35), segmentation or quantitative imaging (20), prognosis or outcome prediction (9), and two other applications.
Why it matters
AI applications span several stages of rotator cuff care, but promising internal performance alone does not establish reliable clinical benefit. The review highlights limitations in validation, reference standards, and study methods that can affect how confidently clinicians interpret reported performance. Stronger evidence is needed before these tools can be considered ready for routine decision support.
What to keep in mind
The review found heterogeneous tasks, modalities, models, reference standards, and metrics, preventing statistical pooling. Many included studies had high risk of bias, limited external validation, or small or selected datasets.