Hand Surg Rehabil. 2026 Jan 31:102587. doi: 10.1016/j.hansur.2026.102587. Online ahead of print.
ABSTRACT
Osteoarthritis (OA) is a leading cause of disability, with diagnosis and management limited by inter-observer variability and the absence of individualized therapeutic strategies. This review critically examines recent applications of artificial intelligence (AI) in OA diagnosis, treatment planning, rehabilitation, and drug discovery, with a particular focus on clinically relevant imaging-based and predictive models. We synthesize evidence from radiographic and MRI-based AI systems used for disease grading, progression prediction, and surgical outcome forecasting, highlighting their performance, limitations, and translational barriers. Particular emphasis is placed on how AI-generated outputs can inform clinical decision-making, including treatment selection and rehabilitation monitoring. Current challenges related to dataset bias, external validation, and workflow integration are discussed using concrete examples from published studies. Finally, we outline future directions aimed at improving clinical utility through explainable AI, multi-modal data integration, and prospective validation. This focused synthesis underscores both the promise and the practical constraints of AI-driven osteoarthritis care.
PMID:41628715 | DOI:10.1016/j.hansur.2026.102587