New AI method to pick up arthritis, lupus early in high-risk patients

Artificial intelligence (AI) can significantly aid in early detection of autoimmune diseases, like rheumatoid arthritis and lupus especially in high-risk patients, leading to better outcomes, finds a study.

In people with autoimmune diseases, the immune system mistakenly attacks their body’s healthy cells and tissues. Some well-known diseases include type 1 diabetes, multiple sclerosis, lupus, and rheumatoid arthritis.

Early diagnosis is critical and may improve treatment and better disease management, said the team led by researchers from the Penn State College of Medicine.

Using machine learning, a type of AI, the team developed a new method that could predict the progression of autoimmune disease among people with preclinical symptoms.

These diseases often include a preclinical stage before diagnosis that’s characterised by mild symptoms or certain antibodies in the blood.

The method dubbed Genetic Progression Score or GPS, could predict the progression from preclinical to disease stages.

In the study, the team used GPS to analyse real-world data to predict the progression of rheumatoid arthritis and lupus.

Compared to existing models, this methodology was found between 25 and 1,000 per cent more accurate in determining mild symptoms that would move to advanced disease stage, the findings showed.

“By targeting a more relevant population—people with a family history or who are experiencing early symptoms—we can use machine learning to identify patients with the highest risk for disease,” said Dajiang Liu, Professor, Penn State College of Medicine. Liu noted that this can also help “identify suitable therapeutics that may be able to slow down the progression of the disease”.

Accurate prediction of disease progression using GPS can enable early interventions, targeted monitoring, and personalised treatment decisions, leading to improved patient outcomes, Liu said.