Using Blood Tests and AI to Predict Treatment Response in Rheumatoid Arthritis
Research Background
Rheumatoid arthritis (RA) is a chronic autoimmune disease that causes pain, swelling, and stiffness in the joints. If left untreated, it can lead to joint damage, reduced mobility, and a lower quality of life. Biological medicines such as adalimumab have transformed the treatment of RA and helped many people manage their symptoms and slow disease progression.
However, not everyone responds well to these treatments. Currently, there are no reliable tests that can predict which patients will benefit from a specific biologic therapy before treatment begins. As a result, many people experience a trial-and-error approach to treatment, which can delay symptom relief, expose patients to unnecessary side effects, and increase healthcare costs.
This study aimed to address this challenge by investigating whether patterns in how active different genes are in blood samples could help predict how patients with RA will respond to adalimumab treatment.
Key Findings
Researchers analysed blood samples from people with rheumatoid arthritis before and during treatment with adalimumab. Using machine learning techniques, a form of artificial intelligence (AI) that can identify patterns in large amounts of data, they examined thousands of genes to look for differences linked to treatment response.
The study identified a group of genes that were more or less active depending on how well patients responded to treatment. One gene, called MZB1, stood out as a particularly promising biomarker – a measurable biological sign that could provide information about how someone may respond to treatment. Researchers found that MZB1 may provide clues about why adalimumab does not work as well for some people. The gene is active in certain immune cells, called B cells, which produce antibodies. In some patients, B cells can produce antibodies against biologic medicines such as adalimumab. These are known as ‘anti-drug antibodies’ and can reduce how well the medicine works. The findings suggest that MZB1 may therefore be linked to biological processes that influence treatment response.
By combining information about how active different genes are with machine learning, the researchers identified patterns that could help predict how patients would respond to treatment more accurately than using clinical information alone.
Research Impact
These findings represent an important step towards more personalised treatment for rheumatoid arthritis. In the future, blood-based biomarkers such as MZB1 could potentially help doctors identify which patients are most likely to benefit from adalimumab before treatment begins.
If validated in larger studies, this approach could reduce the need for trial-and-error prescribing, helping patients receive effective treatments sooner while avoiding therapies that are less likely to work. Earlier access to the right treatment could improve symptom control, reduce disease progression, and enhance quality of life. Before these biomarkers can become part of routine clinical care, they will also need to undergo further validation and regulatory review to ensure they are safe, reliable, and effective for use in patients.
The study also highlights promising biological pathways for further research and may support the development of future diagnostic tests that help guide treatment decisions in rheumatoid arthritis. More broadly, the research demonstrates how advanced data analysis and molecular profiling can support the development of precision medicine approaches. By helping to identify which treatments are most likely to work for individual patients, this research contributes to the long-term goal of delivering the right treatment to the right patient at the right time, improving outcomes for patients while making healthcare more efficient and effective.
Original Publication