Treating rheumatoid arthritis can be like throwing darts—and doctors don’t always hit the bullseye. Trial and error approaches have typically been used to administer methotrexate (a common arthritis drug) to patients, where some patients' symptoms continue to progress even with the treatment.
Work by researchers at Mayo Clinic is helping to create more targeted, effective strategies for helping patients with rheumatoid arthritis using artificial intelligence. Their findings were presented in the journal Arthritis Care & Research.
Rheumatoid arthritis is a chronic autoimmune disorder, meaning that the patient’s immune system mistakenly launches an unyielding attack against a variety of tissues, including those lining the joints, the skin, eyes, and even the heart. One of the classic symptoms of rheumatoid arthritis is pain, swelling, and deformity of the joints, which can take a massive toll on the patient’s quality of life.
In the study, the researchers used a variety of data inputs—patient genomes, clinical reports, and demographic information—in an attempt to filter out the key factors that influence responses to methotrexate in patients at the beginning stages of rheumatoid arthritis. Machine learning algorithms helped to sift through vast volumes of data to reveal trends hidden within them.
The team, led by researchers Liewei Wang, Arjun Athreya, and Richard Weinshilboum, says that the concept of using computers to predict drug treatment outcomes was initially used for individuals with depression.
In their latest work, the team highlighted pharmacogenomic biomarkers that can be used to predict how well patients would respond to methotrexate, particularly in the first three months of treatment.
Future research will focus on integrating AI-powered tools into clinical practice such that more rheumatoid patients can benefit from these groundbreaking developments.