ARTIFICIAL INTELLIGENCE IN MUSCULOSKELETAL ULTRASOUND FOR SPORTS INJURIES: DIAGNOSTIC ACCURACY, CLINICAL UTILITY, AND IMPLEMENTATION CHALLENGES (2022-2026): A NARRATIVE REVIEW

Abstract

Background: Musculoskeletal Ultrasound (MSK-US) is a popular tool among Sports Medicine professionals due to its portability, repeatability, no radiation exposure, and ability to provide dynamic, point-of-care assessments. However, MSK-US also has an operator dependent component of accuracy. Objective: This Review will summarize the current literature from 2022-2026 regarding Artificial Intelligence (AI) in MSK-US specifically with respect to its use in assessing sports injuries, including its diagnostic accuracy, utility and readiness. Methods: The search engines PubMed, Google Scholar, and Reference Chaining were utilized to identify peer-reviewed publications that contained information related to AI; MSK Ultrasound; Sports Medicine; Tendons; Muscle Injuries; Rotator Cuff Injuries; Point-of-Care Imaging. Results: Potential uses of AI in MSK-US are based on pattern recognition, automated measurements, segmentation, and as an aid to second readers, particularly when assessing injuries to tendons, muscles, or the rotator cuff. AI can potentially help reduce variability and enable consistent monitoring. The quality of the available evidence was limited by studies using retrospectively collected data; curated databases; heterogeneity of study design; lack of external validity; and limited examination of outcomes relevant to patients. Conclusion: AI-assisted MSK US appears to be a promising technology; however, at present it is not ready for unqualified clinical application. The most likely near term use of AI-assisted MSK US would be as a form of supervised decision support for trained health care providers.