A machine learning model identified clinical features that can predict the development of "difficult-to-treat” rheumatoid arthritis (D2T RA) at baseline and up to 1 year before patients meet the definition of D2T RA, according to a poster presented at the American College of Rheumatology Convergence in Washington, DC.

Jiri Baloun, PhD, presented the findings on behalf of the team at the Institute of Rheumatology in Prague, Czech Republic.

Even with advances in biologic (b-) and targeted synthetic (ts-) disease-modifying anti-rheumatic drugs (DMARDs), a significant number of patients with rheumatoid arthritis (RA) remain symptomatic, meeting the EULAR definition of D2T RA, the researchers reported.  They aimed to develop machine learning models to predict D2T-RA based on previous clinical profiles at enrollment into the registry or even 1 to 2 years before patients met the definition of D2T RA.

They conducted a retrospective analysis of 8543 patients with RA from the Czech Republic biologics registry ATTRA who began treatment with b/tsDMARDs during the period 2002 - 2023. Patients identified as in sustained clinical remission were those with a Simple Disease Activity Index (SDAI) of < 3.3 and Swollen Joint Counts (SJC) ≤ 1 over 2 consecutive follow-up visits 12 weeks apart. All patients began treatment with b/tsDMARDs.

Of 641 patients with D2T RA  84% were women with a mean age of 50.6 years at baseline. They were compared with 641 RA patients in sustained remission, matched by age, gender, disease duration, and b/tsDMARD treatment at each time point.

The machine learning model demonstrated accuracy and area under the receiver operating characteristics curve (AUC) ranges of 0.606–0.747 and 0.656–0.832, respectively, for predicting D2T RA. Key predictors of D2T RA were identified as clinical disease activity measures, C-reactive protein, and duration of b/tsDMARD treatment. These factors demonstrated the best predictive performance one year before patients met the D2T RA.

These findings provide valuable insights that could enhance the early identification and management of patients at risk for D2T RA, potentially improving treatment strategies and outcomes.

Baloun J, Andrés Cerezo L, Kropáčková T, et al. 2211: Identifying Clinical Predictors of Difficult-to-Treat Rheumatoid Arthritis Using Machine Learning-Based Techniques. Presented at: American College of Rheumatology Convergence. November 18, 2024. Washington, DC.