Magnetic resonance imaging assessment of knee osteoarthritis: current and developing new concepts and techniques.

Magnetic resonance imaging (MRI) is a well-established imaging technique for structural assessment of knee osteoarthritis (OA) particularly in a research context. Conventional MRI allows evaluation of morphological changes in osteoarthritis, and advanced compositional MRI techniques enable assessment of 'premorphologic' biochemical compositional changes of articular and periarticular tissues. Limitations of conventional radiography are well known, although radiography remains the primary imaging modality applied in osteoarthritis clinical trials to date. Hybrid techniques such as PET/MRI have been introduced, which may potentially supplement conventional imaging techniques. Artificial Intelligence (AI) such as deep learning with convolutional neural networks is becoming increasingly recognised as a supportive instrument to deepen our understanding of morphologic OA development and progression. In this narrative review article, we will first give summary of current concepts and widely used MRI assessment techniques of knee osteoarthritis. We will then describe more recent and novel MRI techniques focusing primarily on publications from the last 4 years (2016-2019).