A groundbreaking artificial intelligence system developed by researchers at the University of Surrey can now predict what a patient’s knee X-ray might look like one year from now, offering a powerful new tool for managing osteoarthritis. The technology, presented at the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2025), generates realistic "future" X-rays alongside personalized risk scores that estimate how the disease is likely to progress. Osteoarthritis, a degenerative joint condition affecting over 500 million people worldwide, is the leading cause of disability in older adults. The new AI model was trained on nearly 50,000 knee X-rays from around 5,000 patients, making it one of the most extensive datasets of its kind. It can forecast disease progression about nine times faster than existing AI systems, while maintaining high accuracy and efficiency—key factors for real-world clinical use. David Butler, lead author of the study and researcher at the University of Surrey’s Centre for Vision, Speech and Signal Processing (CVSSP) and the Institute for People-Centred AI, said the system goes beyond simple numerical predictions. “We’re used to medical AI tools that give a number or a risk score, but not much context. Our system doesn’t just tell you your knee might get worse—it shows you what that future knee could look like. Seeing the current X-ray next to a realistic image of what it might look like in a year is incredibly powerful. It helps doctors intervene earlier and gives patients a tangible reason to stick to treatment plans or make lifestyle changes. We believe this could be a turning point in how we communicate risk and improve care for osteoarthritis and related conditions.” At the heart of the system is a diffusion model, a type of generative AI capable of creating highly detailed and realistic images. The model produces a future X-ray and highlights 16 key anatomical points in the knee to track potential changes. This level of transparency allows clinicians to see exactly which areas of the joint the AI is focusing on, increasing trust and clarity in its predictions. The researchers believe the approach could be adapted for other chronic diseases. Similar AI tools might one day visualize lung deterioration in smokers, track heart disease progression, or forecast changes in other joints—offering patients and doctors the same kind of visual foresight currently available for osteoarthritis. Gustavo Carneiro, Professor of AI and Machine Learning at Surrey’s CVSSP, emphasized the system’s advantages over earlier models. “Previous AI tools could estimate the risk of progression, but they were often slow, lacked transparency, and relied only on numbers. Our system delivers fast, visual predictions with clear insights into the specific joint regions at risk. This enables earlier identification of high-risk patients and supports more personalized, proactive care—something that wasn’t practical before.” The team is now working to bring the technology into clinical settings, seeking partnerships with hospitals and healthcare providers to test and integrate the system into routine patient care. With its ability to combine speed, accuracy, and visual clarity, the AI tool represents a significant leap forward in preventive and personalized medicine for osteoarthritis.