Application of Artificial Intelligence in Musculoskeletal Imaging

Musculoskeletal diseases encompass a wide variety of diseases, affect a large number of patients and are characterized by nonneoplastic lesions caused by trauma and sports injuries. Diagnosing musculoskeletal diseases is highly dependent on imaging examinations. The diagnostic results are affected by the radiologists’ experience, thereby making the diagnostic process time-consuming. With the exponential increase in the number of patients presenting with symptoms indicative of musculoskeletal diseases, improving the diagnostic efficiency and accuracy of imaging examinations is very important for radiologists and clinical decision-makers; thus, because of the advent and rapid development of AI, diagnostic efficiency and accuracy can be improved. Applying AI in assessing the musculoskeletal system has primarily been focused on segmenting anatomical structures, and diagnosing and classifying diseases. As research on the applications of AI in assessing the musculoskeletal system has broadened, the scope has broadened to gradually cover bone age assessment, fracture detection, bone density analysis, sports injury assessment, tumour classification and postoperative prognosis prediction and other disease areas [1, 2]. Current research shows that AI and radiologists have similar diagnostic efficacies [1, 3, 4]. However, because of its speed in determining a reliable diagnosis, AI is beneficial to radiologists. The National Medical Products Administration (NMPA) has approved multiple AI-assisted diagnostic systems for fracture detection and bone age analysis.