Magnetic Resonance Imaging (MRI) Cartilage Thickness and its Relationship with Established Imaging Markers of Osteoarthritis (OA) Severity
Overview:
Osteoarthritis (OA) of the knee is a massive problem and a leading cause of disability worldwide. Unfortunately, there are limited therapies available and demand for joint replacement is starting to exceed availability. To meet this challenge drugs to halt, or even reverse, cartilage deterioration are being evaluated. However, to measure the efficacy of these drugs, we need highly accurate tools that readily detect tiny changes in cartilage structure. Currently, x-rays are the only approved imaging method used in osteoarthritis trials for assessing the structural benefit of a therapy. However, they only provide a projectional picture of bones, and from that picture we have to estimate where cartilage is lost from measuring the distance between the knee bones. Not surprisingly this is not a very sensitive measure of true cartilage change. MRI directly visualises cartilage as well as providing a more comprehensive joint assessment. The aim of this work is to assess the difference between x-ray measures of cartilage loss and direct MRI assessment.
Aims and Objectives:
By using a novel computer machine learning approach, this research aims to compare the performance of accurately measured MRI cartilage thickness against existing commonly used MRI measurement techniques as well as the current gold-standard of x-rays. This will validate more accurate, reliable and sensitive markers of osteoarthritis severity and provide evidence on the best imaging tools for OA clinical trials.
Key Findings:
Aim 1 Findings
Objective 1:
In healthy knees, articular cartilage contributed nearly two-thirds of the radiographic JSW in the medial and lateral tibiofemoral compartments. The remaining contribution is from non-articular cartilage structures (the ‘proxy meniscus’). This contribution alters with advancing grades of radiographic severity. In the most advanced stages, there is minimal contribution from the proxy-meniscus indicating that most of the inter-bone gap visualised is due to cartilage. However, it is apparent that radiographs are subject to a ceiling effect; where MRI is able to detect residual cartilage thickness, radiographs under estimate its presence likely due to a combination of the planar nature of this imaging modality, radiographic projection errors and bone shape changes.
Objective 2:
In healthy knees, cartilage thickness has excellent agreement with radiographic joint space width (when measured at optimal locations), indicating that they are measuring the same construct. This supports using carefully acquired knee radiographs as a cartilage surrogate in non-OA knees. Agreement between cartilage thickness and radiographic joint space width progressively breaks down in the presence of OA and advancing joint space narrowing, indicating that their relationship becomes unreliable and that radiographs inaccurately reflect regional average cartilage thickness.
Aim 2 Findings:
Objective 1:
A quantified version of the parameters used in a semi-quantitative score (MRI Osteoarthritis Knee Score; MOAKS) was more responsive than the standard semi-quantitative MOAKS overall, achieving moderate responsiveness when measured with the Standardised Response Mean (SRM) in the central medial femur and tibia for total area affected by cartilage lesions (tAB) and approximating the threshold for large responsiveness in the anterior medial tibia and posterior medial tibia at 2-years. Area affected by denudation was less responsive than tAB both quantitatively and semi-quantitatively, though the quantitative version was slightly more responsive.
Cartilage thickness measured in standard quantitative subregions, was the most responsive measure in all regions at both 1 and 2-years and achieved large responsiveness in the cMF. This is consistent with prior studies which have shown this region to be the most responsive and supports assertations that cartilage thickness provides an efficient summary measure of cartilage thinning and denudation.