Artificial Intelligence Applications in Musculoskeletal Imaging.

Overview

abstract

PURPOSE OF REVIEW: The demand for AI-driven solutions in musculoskeletal (MSK) imaging has risen alongside the surge in orthopedic imaging studies, reflecting the need for tools that enhance diagnostic accuracy, reduce healthcare costs, and alleviate physician workload. This review explores recent applications of AI-particularly computer vision and deep learning (DL)-in MSK imaging, from trauma and surgery to specialized and point-of-care technologies. The review also highlights existing challenges and limitations hindering the integration of these tools into clinical practice. RECENT FINDINGS: AI applications are abundant in MSK imaging, with DL models showing remarkable versatility and success across multiple use cases. These include but are not limited to fracture detection, segmentation for preoperative planning, surgical navigation and tracking, tumor detection and classification, pediatric bone age estimation, and bone density measurement. Specialized use cases also target injury detection in sports medicine, and AI has been integrated into point-of-care technologies, such as motion-monitoring systems, underscoring AI's broad potential to improve diagnostic accuracy, reduce interpretation times, and increase efficiency. AI has shown promise in transforming MSK imaging, suggesting improvements in diagnostic performance, speed, and cost-efficiency. Despite research advances, challenges remain in deploying AI in real-world clinical settings, where model generalizability, data quality, and high computational demands pose obstacles. However, recent developments in AI, including the rise of adaptable foundation models and advancements in model efficiency, offer promising solutions that may accelerate the integration of AI into clinical workflows, bringing the field closer to realizing the full potential of AI in patient care.