Machine learnIng and Multimodal imaging for knee OSteoArthritis prediction – MIMOSA

MIMOSA focuses on a public health priority medical application: Kkee OsteoArthritis (OA). In knee OA, the changes are currently assessed primarily by visual grading of joint space width and osteophytes on plain radiographs. However, visual assessment on radiographs is often inaccurate as the changes of cartilage volume are small and the grading requires experienced observers in order to be reproducible and also, it is not sensitive to early osteoarthritis changes. There is an urgent need to develop methods that would allow a reliable assessment of knee OA on radiographs for the prognosis and diagnosis of early knee OA. The question is how to predict the onset and the progression of knee OA using plain radiographs, to improve and make reproducible its early diagnosis? To address this issue different image modalities and machine learning models will be performed. The aim is to introduce a new transparent Computer-Aided Diagnosis system based on machine learning to automatically score knee OA severity.