Statistical shape modelling of osteoarthritic knee joint

Abstract

Osteoarthritis (OA) is a degenerative joint disease caused by the deterioration of the cartilage within a joint. It is diagnosed in around 1 in 7 adults, and it is the leading cause of disability among elderly individuals. OA occurs frequently in the knee joint, causing pain, inflammation, swelling and, in its worse cases, reducing functionality. From imaging, it manifests as osteophytes, subchondral sclerosis, bone attrition, and asymmetric joint-space narrowing, leading to a unique shape of the bones. Currently no early biomarkers exist for this pathology, and pharmacological treatments only target symptoms, but not its cause. It has been proven that changes in bone shape can be found even more than one year prior to the onset of radiographic OA, and the joint space narrowing seems a good candidate to be used as biomarker. In order to properly visualize the joint space, however, imaging needs to be performed in a standing position. This is only possible in radiography (which provides 2D images) and not in 3D scan as Magnetic Resonance Imaging or Computed Tomography (which can provide an accurate 3D model of the joint). In this study, different ways to create statistical shape models that can provide accurate information on the anatomy of the knee joint under load are described, and currently registration involving 2D and 3D images seems to be the most suitable solution.