Latest insights in ultrasound imaging using artificial intelligence for musculoskeletal disorders: A narrative review

This review examines future directions by focusing on the latest findings, limitations, and challenges in the clinical application of artificial intelligence (AI)-based ultrasound imaging for carpal tunnel syndrome (CTS) and osteochondritis dissecans (OCD). In these conditions, ultrasonography (US) imaging is clinically applied. Over the past few years, AI has significantly enhanced the diagnostic capabilities of US by providing real-time imaging and accurate lesion delineation. The integration of deep learning models has led to significant advancements in automatic segmentation and object detection, enabling high prediction accuracy in diagnosing these conditions. Numerous studies have demonstrated the effectiveness of AI models, such as YOLO and Mask R-CNN, in accurately identifying and classifying musculoskeletal disorders. These advancements have the potential to streamline clinical workflows and improve diagnostic outcomes. The continued development of AI-based US imaging has significant potential for improving diagnostic accuracy and efficiency. However, there are limitations to its clinical application, and further studies, including the use of large-annotated datasets for training AI models and establishing standardized protocols for US imaging diagnostics, are crucial.