OSTEOARTHRITIS: NEW TREATMENT, STRATIFICATION, AND MACHINE LEARNING

Osteoarthritis affects over eight million people in the UK alone, causing significant pain and limited activity and function. Joint replacement is an excellent end stage treatment for advanced osteoarthritis to relieve such pain and restore function. However, there is a clear unmet need at the earlier stages of disease, often in younger patients, before joint replacement. Early stages of osteoarthritis and cartilage injury form a heterogeneous grouping, which can include the rate of disease progression, causation, and clinical phenotype. In order to identify early triggers and modify them appropriately, a stratified approach is required so that the right patient gets the right treatment at the right time. A variety of treatment options exist that would benefit from a stratified approach, ranging from physical therapy through to advanced and regenerative therapeutics, for example those based on cell therapy or a molecule. There are many possible types of measurements that could be useful to facilitate stratification, including information derived from patient reported outcome and symptoms, imaging, biomarkers, and those related to genetics and the genome. Collected over time, such measurements can lead to a large amount of information for analysis and interpretation. Machine learning falls within the field of artificial intelligence and seeks to use algorithms to learn from data sources and in such a way that can be used to make predictions. This presentation will include consideration of how machine learning can be used to stratify patients with osteoarthritis in general terms and more specifically consider how this approach could be used to predict progression in patients with osteoarthritis. The value of such approaches will be considered in the sense of clinical trial design and selection for optimal treatment, particularly at the early stages of the disease.