Fractal- and Entropy-based Textural Analyses of Subchondral Bone Offer Better Osteoarthritis Prediction in Older, At-Risk Patients: Presented at WCO

By Chris Berrie

KRAKOW, Poland -- April 25, 2018 -- In older patients deemed to have a pre-existing increased risk, a more effective prediction of incident osteoarthritis can be made by using a combination of fractal- and entropy-based textural analyses of plain subchondral bone radiographs as well as the usual joint-space width and area (JSW/A) and clinical features, according to study reported at the World Congress on Osteoporosis, Osteoarthritis and Musculoskeletal Diseases (WCO).

Current assessment parameters and indicators for osteoarthritis remain insufficient for the prediction of osteoarthritis, explained lead author Richard Ljuhar, PhD, Image Biopsy Lab GmbH, Vienna, Austria, speaking here on April 21. “Changes in trabecular bone are present prior to any visible anatomical changes in the knee joints,” he noted.

Dr. Ljuhar and colleagues defined bone-microstructure measures for use in their predictive models of osteoarthritis, with data derived from conventional posterior-anterior knee radiographs of men and women in the Multicentre Osteoarthritis Study database (MOST1).

Following screening for eligibility of 1,092 knee radiographs from the Birmingham, Alabama area, a total of 574 subjects (344 males) met the inclusion criteria of Kellgren and Lawrence (KL) score 0 at baseline. At the follow-up of 84 months, 41 women (17.8%) and 79 men (23.0%) had developed a KL score of 1.

A texture analysis included oriented fractal- and entropy-based texture algorithms from 4 regions of interest for the proximal tibia and 1 for each condyle of the distal femur. The team assessed JSW/A in parallel.

The initial models described by Dr. Ljuhar included clinical features (age, body mass index) and JSW/A, which provided areas under the curve (AUCs) for incident osteoarthritis of 0.67 for the women and 0.61 for the men. The texture analysis defined for this study (adjusted for body mass index), provided AUCs of 0.78 for the women and 0.66 for the men.

Combination of the clinical features, JSW/A, and the defined texture analysis further improved AUCs of 0.80 for the women and 0.69 for the men.

“Based on these results, using textural algorithms in combination with clinical risk factors, it is possible to improve the prediction of incidence of osteoarthritis, and this is certainly superior to using clinical features alone,” Dr. Ljuhar concluded.

The online Multicentre Osteoarthritis Study provides access to a community-based sample of men and women aged 50 to 79 years of age from the general United States population, as those likely to either have pre-existing osteoarthritis or to be at high risk, as indicated by weight, knee symptoms, or a history of knee injuries or operations.

As a leading cause of pain and disabilities in our aging population, osteoarthritis prevention and treatment is of increasing importance. There is a need not just for treatment of disease progression, but also early detection and quantification of osteoarthritis onset and severity, the authors noted.

The WCO Congress is sponsored by the International Osteoporosis Foundation (IOF) and the European Society for Clinical and Economic Aspects of Osteoporosis, Osteoarthritis and Musculoskeletal Diseases (ESCEO).

[Presentation title: Combining Fractal- and Entropy-Based Bone Texture Analysis for the Prediction of Osteoarthritis: Data from the Multicenter Osteoarthritis Study (MOST). Abstract P378]

To read more Conference Dispatch articles, click here.