Musculoskeletal Segmentation

In 2001, 6.1 million Australians (32%) had long-term arthritis or a musculoskeletal disorder with almost 1.2 million reporting an associated disability (Bhatia 2005). Osteoarthritis (OA) is the most common form of arthritis affecting nearly 1.4 million Australians and a major cause of chronic pain and disability (Bhatia 2005). This chronic joint disease is characterized by progressive degenerative changes in the anatomical structure of articular cartilage with late stage X-rays showing joint space narrowing, bone sclerosis and osteophytes which eventually result in compromised joint integrity and function leading to significant pain and disability.

Clinically, X-ray examinations do not directly demonstrate cartilage and lack the sensitivity to perform either early diagnosis, as by the time radiographic findings are observed 13% of the cartilage tissue is lost (Jones 2003), or to assess changes in cartilage during short-term studies or drug trials into OA (Graichen 2004, Cicuttini 2005). In contrast, magnetic resonance imaging (MRI) provides excellent visualization of articular cartilage (Figure 1) along with other joint structures and this has generated extensive clinical interest in the development of MR technologies to provide quantitative analyses of joint structures to facilitate early stage diagnostic and management options for OA and other significant joint pathologies.

For quantitative MRI-based joint analyses to be clinically viable, image processing approaches which generate fast, accurate and reproducible data on the morphometric and biochemical characteristics of joint cartilage, bone and surrounding structures need to be developed for medical specialities such as orthopaedic surgery, sports medicine, and rheumatology. By far the best clinical solution will be a fully automated joint segmentation and quantitative analysis package rather than the expertise- and time-consuming manual or operator dependant approaches used in past research studies for morphological (Ding 2007, Eckstein 2006b, Koo 2005) and biochemical (Li 2007, Glaser 2005) analyses to demonstrate the capacity of MRI for earlier detection of changes and more accurate monitoring of OA disease progression (Glaser 2005, Eckstein 2006ab, Eckstein 2007, Hunter 2009).

Automatic MR Cartilage Analysis

State-of-the art algorithms for segmentation and analysis (Fripp 2007, 2009) are employed to help objectively measure the health and condition of cartilage tissue.


Figure 2: Saggital slices of 3T weDESS MRI Knee scan with cartilage segmentation colour coded using cartilage thickness.

Saggital slices of 3T weDESS MRI Knee scan with cartilage segmentation colour coded using cartilage thickness.