AI-navigated shoulder injection: precision, real-time learning and clinical translation

Abstract

This review focuses on artificial intelligence (AI)-guided ultrasound-guided shoulder joint injections. We systematically retrieved relevant studies from PubMed, Embase, Cochrane Library, IEEE Xplore, and Web of Science (1996 ~ 2025). Literature was screened based on predefined inclusion/exclusion criteria, and evaluated AI technologies using core metrics including anatomical segmentation accuracy (Dice similarity coefficient), first-pass puncture success rate, and clinical outcome indicators (Visual analogue scale scores, VAS; American Shoulder and Elbow Surgeons, ASES scores). It explores the technical principles of AI medical image processing (segmentation, detection, tracking, reconstruction) and deep learning algorithms for shoulder anatomy, addressing limitations of traditional ultrasound guidance through AI-enabled precision targeting and real-time learning. Clinical applications, technological advancements, ethical controversies, and regulatory pathways are summarized. Key findings confirm AI enhances injection accuracy, first-pass success rates, and patient outcomes. This work provides a concise, evidence-based reference for clinicians and researchers, highlighting the paradigm shift of AI in optimizing shoulder injection therapy.

Ähnliche Arbeiten