From traditional to digital: Artificial intelligence, transforming ultrasound in rheumatology

Abstract

The use of artificial intelligence (AI) in musculoskeletal ultrasound has grown significantly in recent years, particularly in rheumatology, where standardization of diagnostic accuracy and image acquisition is essential. This article provides a narrative review of the available scientific literature on AI applications in musculoskeletal ultrasound, focusing on the recent advances in machine learning and deep learning algorithms. Sixteen studies published between 2020 and 2023 were reviewed, applying AI to anatomical segmentation and diagnostic prediction in musculoskeletal conditions, such as rheumatoid arthritis and other inflammatory diseases. Most of the reviewed studies employed deep learning algorithms to enhance the diagnostic accuracy and reduce operator-dependent variability. However, a common limitation across studies is the lack of external validation, indicating that future research should focus on confirming the generalizability of these models to larger populations and diverse clinical settings. Despite this limitation, the results suggest that AI has the potential to optimize clinical decision making and improve patient care in the context of musculoskeletal ultrasound in rheumatology.