Editorial Commentary: Artificial Intelligence Models Using Machine Learning Can Improve Preoperative Identification...

Area of Science:

• Orthopedics
• Radiology
• Artificial Intelligence

Background:

• Subscapularis tendon tears are notoriously difficult to diagnose accurately using traditional magnetic resonance imaging (MRI) compared to other rotator cuff tendons.
• Existing diagnostic methods often lack the sensitivity and precision required for early and accurate detection of subscapularis pathology.

Discussion:

• Machine learning models analyzing physical examination findings and MRI data demonstrate a strong correlation between confirmed subscapularis tears and specific indicators.
• These indicators include abnormal subscapularis tendon length, concurrent tears of the long head of the biceps, and subscapularis fatty atrophy.
• Physical examination signs such as internal rotation weakness and positive lift-off, belly press, and bear hug tests are also highly associated with these tears.

Key Insights:

• Artificial intelligence (AI) evaluation of patient data reveals significant associations between specific clinical signs, MRI findings, and arthroscopically confirmed subscapularis tears.
• Abnormal subscapularis tendon length, long head of the biceps tears, and fatty atrophy are key MRI indicators.
• Weakness in internal rotation and specific physical tests (lift-off, belly press, bear hug) are crucial clinical markers.

Outlook:

• While current machine learning models may not yet revolutionize clinical practice, they represent a significant advancement in diagnostic tools.
• Continued rapid research and development in AI for medical imaging and diagnostics are expected.
• Future AI models hold the potential to significantly improve the timely identification of subscapularis pathology, optimize preoperative planning, and enhance physician training.