A machine-learning model may help predict how patients with juvenile idiopathic arthritis respond to treatment with methotrexate, according to a recent study published by Shoop-Worrall et al in eBioMedicine. Although methotrexate is the standard therapy for this patient population, only about 50% of patients respond to or tolerate the treatment. Researchers used four UK-based multicenter prospective cohorts within the CLUSTER consortium to recruit 657 pediatric patients with idiopathic arthritis who initiated methotrexate prior to January 2018. They applied the machine-learning model to determine clusters with distinct disease patterns following treatment initiation, predict cluster membership, and compare clusters to existing treatment response measures. The researchers then verified their results in a group of 1,241 young patients. The model was able to identify six clusters, which they called fast improvers (11%), slow improvers (16%), improve-relapse (7%), persistent disease (44%), persistent physician global assessment (8%), and persistent parental global assessment (13%). They noted that the last two clusters were comprised of patients who demonstrated improvement in all disease features except one, and that the factors they used to characterize the clusters included age, ethnicity, International League of Associations for Rheumatology criteria, and parental global assessment and erythrocyte sedimentation rate scores at treatment initiation. The researchers reported that the area under the curve values were 0.65 to 0.71 for the novel model, and that singular American College of Rheumatology Pediatric 30/90 scores were unable to measure improvement speed, relapsing courses, or diverging disease patterns at 6 and 12 months. The researchers hope their novel machine-learning model can lead to the development of a stratified medicine program and help physicians prescribe more effective first-line therapies for patients with juvenile idiopathic arthritis.