The Relationship Between Two Different Measures of Osteoarthritis Bone Pathology, Bone Marrow Lesions and 3-dimensional (3D) Bone Shape: Data from the Osteoarthritis Initiative and Effect of nonsteroidal anti-inflammatory drugs (NSAID) on Cartilage 

Overview:

Changes to the bone structure are important in assessing the development of knee osteoarthritis (OA). Bone marrow lesions (BMLs), sometimes known as bone oedema, are a type of bone structural abnormality thought to contribute to the OA disease process. New research has shown that changes in the shape of the knee bones also plays a role in OA disease progression. However, no research has assessed the relationship between bone marrow lesions and bone shape. We are able to measure bone shape using a novel technology called 3-dimensional (3D) statistical shape modelling to analyse magnetic resonance images (MRIs) of the knee.  
  
Cartilage is another important knee structure, and helps to maintain the healthy functioning of the knee joint. The loss of cartilage is a feature of OA. The volume and thickness of cartilage can be measured automatically on MRIs using machine-learning techniques to produce 3D cartilage models. These have been shown to be as accurate and reliable as manual human measurements. Evidence for this was shown using a large dataset of 4796 knees from people with osteoarthritis. However, no research using these 3D cartilage models has assessed how these cartilage measurements relate to other features of osteoarthritis, such as pain, or characteristics such as age and gender.  

Aims and Objectives:

1. To study the relationship between bone marrow lesions (BMLs) and 3D bone shape at a single point in time and to describe how this relationship changes over time in people with osteoarthritis. 
2. To describe how cartilage thickness varies across different areas of the knee joint. To study whether variation in thickness is determined by certain features, for example taking medications such as NSAIDs, and characteristics such as age, height and gender.
3. To determine the variation in cartilage thickness in healthy and osteoarthritic (OA) knees. To study the effects of clinical factors such as pain and function on cartilage loss and how cartilage loss relates to important OA measures such as total knee replacement (TKR). 

Key Findings:

We found that a weak positive correlation between bone shape and total BML score in cross-sectional analysis and femoral bone shape was associated with prevalent BMLs. We found modest associations between change in bone shape and change in BMLs over time. 

Outputs:

• Bowes, MA, Alabas, OA, Hensor, EM, Brett, A, Kingsbury, SR and Conaghan, PG.  Traditional mechanical and patient factors have negligible effects on cartilage thickness outside of OA disease status: a study using accurate 3D MRI data from 9592 normal and osteoarthritic knees.  Osteoarthritis and Cartilage 26 (2018) S438-S439. doi.org/10.1016/j.joca.2018.02.842
• Dube B, Bowes MA, Hensor EMA, Barr A, Kingsbury SR, Conaghan, PG. The relationship between two different measures of osteoarthritis bone pathology, bone marrow lesions and 3D bone shape: data from the Osteoarthritis Initiative. Osteoarthritis and Cartilage 26 (2018) 1333-1337. doi 10.1016/j.joca.2018.06.011
• Bowes, MA, Alabas, O, Brett, AD and Conaghan, PG. Cartilage thickness is surprisingly consistent in healthy knees and risk of symptoms increases only outside the healthy range - an MRI study of 9,433 knees from the osteoarthritis initiative. Osteoarthritis and Cartilage 28 (2020) S256-S257 doi.org/10.1016/j.joca.2020.02.415

Work Package	Early Disease and Risk Prediction: Prevent
Objective	2.5i and 3.2
Lead	Bright Dube and Oras Alabas
Investigators	Phil Conaghan, Sarah Kingsbury, Elizabeth Hensor, Andrew Barr, Alan Brett Michael Bowes, (Imorphics Ltd)
Institution	University of Leeds