aceArtificial intelligence empowering the diagnosis and treatment of rotator cuff tears
Background: Artificial intelligence (AI) is being developed for image acquisition, quantitative tissue assessment, diagnosis, and clinical prediction in rotator cuff tears. The breadth of this literature and its methodological limitations have not been integrated across the complete care pathway. Methods: We searched PubMed, Embase, Web of Science, Scopus, and the Cochrane Library from inception to July 5, 2026 for English-language original studies applying AI to rotator cuff-related reconstruction, segmentation…
In brief
A systematic review searched five databases through July 5, 2026 and included 66 reports on AI applied to rotator cuff tears. The studies covered diagnosis, segmentation and quantitative imaging, prognosis, and image reconstruction across several imaging and sensing modalities.
The review matters because AI tools for shoulder care could shape imaging and treatment decisions, yet the underlying research is methodologically weak. Internal performance often looked promising, but whether the tools work reliably in real clinical settings remains unproven.
Main points
- Sixty-six reports were included: 35 diagnostic or classification studies, 20 segmentation or quantitative imaging studies, and 9 prognosis or outcome prediction studies.
- Two other studies covered MRI reconstruction or image optimization and exploratory unsupervised phenotyping.
- Risk of bias was high in 27 of 36 diagnostic studies, unclear in 6, and low in only 3.
- All 8 completed prediction-model studies were rated high risk of bias; one protocol without results was not rated.
- The authors call for prospective multicenter studies, transparent reference standards, and independent external validation.
The problem
Most diagnostic and all completed prediction-model studies of AI for rotator cuff tears were rated high risk of bias, so the evidence cannot support routine clinical use.
The rundown
The authors searched PubMed, Embase, Web of Science, Scopus and the Cochrane Library for English-language original studies. Two reviewers screened reports and extracted data on models, validation, reference standards and performance.
Diagnostic accuracy studies were assessed with QUADAS-2 and prediction-model studies with PROBAST. Tasks, modalities, models and metrics were too different to pool, so the evidence was synthesized narratively.
The AI applications spanned MRI, radiography, ultrasound, CT, clinical or multimodal data, motion video, and emerging sensing data. The authors concluded that AI shows proof-of-concept value in imaging and decision support but does not replace clinician judgment.
Sources
- Peer-reviewedMedicine2026-10-16
The debate