Artificial Intelligence: Magic Boon for Detecting Osteoarthritis Years Before Symptoms Appear

Researchers at the University of Pittsburgh School of Medicine and Carnegie Mellon University College of Engineering have designed a machine learning algorithm that can detect subtle signs of osteoarthritis years before the appearance of the symptoms.

These signs are difficult to detect even for an experienced radiologist on a MRI scan taken years before the symptoms appear. As of today arthritis is diagnosed after the bone damage has been done. The above research states that Artificial Intelligence can diagnose arthritis atleast 3 years prior to any bone damage/ appearance of symptoms.

‘Artificial intelligence helps in diagnosis of Osteoarthritis 3 years before the appearance of symptoms or joint damage.’

There are a number of diseases that do not show any visible signs and symptoms in the initial stages, osteoarthritis (OA) is one of them. Osteoarthritis is the most common forms of arthritis affecting millions of people worldwide.

It is a common, progressive disorder primarily affecting the weight bearing joints. It is characterized by progressive destruction of the articular cartilage(Malleable tissue at the ends of the bones) that are bone cell formation units that leads to limitation of motion, deformity & disability.

According to the current scenario, the gold standard for diagnosing arthritis is through radiologic imaging - mainly X rays. By the time the physician can see changes in the X ray that is indicative of arthritis, the irreversible bone damage has already been done.

This study was conducted for approximately 7 years. All the patients enrolled in this study got their knee MRI done regularly. This helped in continuous radiologic monitoring of the joints. The study mainly focused on the subset of patients showing very minimal cartilage damage. These patients were followed up regularly.

The patterns of the radiologic images were being continuously saved in a system. These changes were not visible to the naked eyes of the physician by the conventional techniques.A classifier was trained to differentiate progressors and non-progressors on baseline cartilage texture maps, which achieved a robust test accuracy of 78% in detecting future symptomatic OA progression 3 years prior to the appearance of symptoms.

This strongly indicates that detection of osteoarthritis is possible at a reversible stage. One of the key contribution of the study is direct visualization of the cartilage phenotype (physical appearance) defining predictive ability of the cartilage regeneration. Early biochemical patterns of fissuring or cracks in cartilage marked the future onset of osteoarthritis.

Currently the primary treatment of osteoarthritis is the knee replacement surgery. There are no drugs that can be used to prevent the progression of the pre symptomatic stage to the irreversible stage.

In future, coupling presymptomatic osteoarthritis detection techniques with emergent clinical therapies could modify the outcome of the disease and may help in reversing the joint damage.

Source: Medindia