Magnetic resonance imaging (MRI) offers direct three dimensional (3D) visualization of the articular cartilage in the knee joint. New high-field three-Tesla (3T) MRI scanners and specialized pulse sequences can provide images with excellent spatial resolution and high contrast, which show the cartilage and other structures of the knee in great detail. These new techniques are ideal for the application of image processing software to segment (identify and outline) cartilage plates on MRI slices for the quantitative assessment of osteoarthritis (OA) progression. Several academic and industry affiliated research groups have developed software tools to segment the cartilage plates on knee MRI scans. However measures derived from the full cartilage plate or even specified subregions can be relatively insensitive to change. We propose to test the hypothesis that an analysis based on comparing local regions of the knee cartilage will be more responsive than global measures of cartilage morphometry for detecting knee OA changes. We will use a software tool developed in our laboratory to segment the articular cartilage on knee MRI image sets. 3D registration software will be used to align baseline and follow-up segmented scans. Once the baseline and follow-up segmented images are registered (aligned) the technique will measure volume change in a region in close proximity to a known cartilage area of thinning. We will also develop and validate new quantitative metrics based on measurements of bone marrow edema (BME), bone cysts, and osteophyte volume. Our project will use data and images from the osteoarthritis Initiative (OAI), a large National Institutes of Health (NIH) study with approximately 4,800 enrolled subjects subdivided into three groups: the Incidence Cohort, the Progression Cohort, and the Control Cohort. A subset of 400 subjects from the Progression Cohort will be used to test our hypotheses. Our existing analysis software will be further developed to facilitate the aims of the study.
This project will address public health concerns by validating an improved technique for evaluating osteoarthritis of the knee. The work will contribute to the public health indirectly by assisting others in developing and evaluating OA therapies.
