Harnessing Artificial Intelligence to Predict RA

By Vandana Suresh | Feb. 21, 2025

While studying rheumatic diseases in children in the 1950s, physician and mathematician Alvan Feinstein observed a significant gap in distinguishing between benign and pathological heart murmurs. At that time, diagnoses were largely reliant on the physician's experience. To address this issue, Feinstein classified heart murmurs based on scientific principles, providing foundational work for evidence-based medicine. This evolution eventually led to the development of precision medicine, which tailors treatment to meet individual patient needs.

Precision medicine for rheumatoid arthritis (RA) is still in its early stages. Recognizing its benefit for patients, the Arthritis Foundation has awarded a Rheumatoid Arthritis Research Program award to Fan Zhang, PhD, assistant professor at the University of Colorado Anschutz Medical Campus. This funding will support her research into developing artificial intelligence tools for predicting RA onset and potentially therapeutic targets for precision medicine.

“Currently, there are no robust predictors for when a patient will start showing noticeable clinical signs of rheumatoid arthritis,” said Zhang. “Artificial intelligence has the potential to transform the field of disease prediction. My team will develop machine learning tools that combine a wide range of patient data to produce the best prediction strategy for disease onset, and novel clinical targets for intervention.”

Before the characteristic clinical symptoms of rheumatoid arthritis appear, such as joint pain and swelling, patients often exhibit immunological abnormalities. These include the production of autoantibodies like anti-citrullinated protein antibodies (ACPA) and rheumatoid factor (RF), which can be detected through blood tests. However, the duration of the preclinical phase of RA varies widely among patients, and not all individuals progress to full disease. This complexity of rheumatoid arthritis’ natural history presents a significant hurdle for early interventional strategies for preventing disease progression.

Therefore, there is a pressing need to identify accurate biomarkers that can serve as indicators of future disease. To address this challenge, Zhang and her team will use their expertise in computational artificial intelligence to develop algorithms that can identify predictive biomarkers within large and complex data sets, including single-cell transcriptomics and proteomics. Specifically, the Zhang lab will employ artificial intelligence tools to analyze RNA profiles and protein expression patterns, among other data, in individual immune cells obtained from participants in the NIH StopRA clinical trial.

By comparing individuals who progress to RA with those who remain disease-free, this longitudinal, data-driven systematic approach can potentially uncover key molecular signatures, forecasting changes in the immune system that are crucial for the transition from preclinical to clinical RA. Furthermore, this research can also open doors to identifying novel targets for treatment.

“The funding from the Arthritis Foundation is a tremendous encouragement,” said Zhang. “We are very excited to push the boundaries of artificial intelligence-driven research and develop computational tools that will provide deeper insights into the evolution of rheumatoid arthritis and other autoimmune diseases.”

Zhang added that the funding also strengthens her collaboration with leading rheumatologist Kevin Deane, MD, PhD, also at Anschutz Medical Campus, who is the principal investigator of the StopRA trial and oversees the preclinical rheumatoid arthritis cohorts in the clinical study.