โก Quick Summary
This study utilized machine learning techniques to identify clinical features that predict the risk of developing difficult-to-treat rheumatoid arthritis (D2T RA). The models achieved an accuracy range of 0.606-0.747, providing valuable insights for early intervention in RA management.
๐ Key Details
- ๐ Dataset: 8,543 rheumatoid arthritis patients from the ATTRA registry
- ๐งฉ Features used: Clinical measures, patient-reported outcomes, treatment duration
- โ๏ธ Technology: Machine learning models including lasso regression, support vector machines, random forests, and XGBoost
- ๐ Performance: Accuracy range of 0.606-0.747, AUC of 0.656-0.832 for predicting D2T RA
๐ Key Takeaways
- ๐ Machine learning can effectively predict the risk of D2T RA.
- ๐ก Key predictive variables include disease activity measures and patient-reported outcomes.
- ๐ฉโ๐ฌ Early identification of D2T RA can lead to timely therapeutic interventions.
- ๐ฅ The study analyzed data from patients treated with biologic or targeted synthetic DMARDs.
- ๐ Findings support the importance of recognizing early indicators of RA progression.
- ๐ SHAP analysis was used to assess the contribution of individual variables to model predictions.
- ๐ Study period: Data collected from 2002 to 2023.
- ๐ Clinical criteria: D2T RA defined by EULAR criteria.
๐ Background
Rheumatoid arthritis (RA) is a chronic inflammatory disorder that can lead to significant joint damage and disability. A subset of patients, classified as having difficult-to-treat RA, often do not respond adequately to standard therapies. Identifying clinical features that predict this challenging condition is crucial for improving patient outcomes and tailoring treatment strategies.
๐๏ธ Study
This retrospective analysis focused on patients from the ATTRA registry who initiated treatment with biologic or targeted synthetic disease-modifying anti-rheumatic drugs (DMARDs). The study aimed to develop predictive models using various machine learning techniques to identify clinical features associated with D2T RA, assessed at baseline and up to one year prior to meeting the D2T RA criteria.
๐ Results
Among the 8,543 RA patients analyzed, 641 met the criteria for D2T RA, while 1,825 achieved sustained remission. The machine learning models demonstrated an accuracy range of 0.606-0.747, with an area under the receiver operating characteristic curve (AUC) ranging from 0.656 to 0.832. SHAP analysis revealed that disease activity measures, patient-reported outcomes, and treatment duration were significant predictors of D2T RA.
๐ Impact and Implications
The findings from this study underscore the potential of machine learning in enhancing the early detection of difficult-to-treat RA. By identifying key clinical features, healthcare providers can implement timely interventions, ultimately improving long-term patient outcomes. This research paves the way for integrating advanced predictive analytics into routine clinical practice, fostering a proactive approach to RA management.
๐ฎ Conclusion
This study highlights the transformative role of machine learning in predicting difficult-to-treat rheumatoid arthritis. By leveraging clinical data, we can enhance early recognition and intervention strategies, leading to better management of RA. Continued research in this area is essential to refine predictive models and improve patient care in rheumatology.
๐ฌ Your comments
What are your thoughts on the use of machine learning in predicting difficult-to-treat rheumatoid arthritis? We invite you to share your insights and engage in a discussion! ๐ฌ Leave your comments below or connect with us on social media:
Machine learning-assisted screening of clinical features for predicting difficult-to-treat rheumatoid arthritis.
Abstract
To identify clinical features that predict the risk of meeting difficult-to-treat (D2T) rheumatoid arthritis (RA) definition in advance. This retrospective analysis included RA patients from the ATTRA registry who initiated biologic (b-) or targeted synthetic (ts-) disease-modifying anti-rheumatic drugs (DMARDs) between 2002 and 2023. Patients with D2T RA met the EULAR criteria, while controls achieved sustained remission, defined as a Simple Disease Activity Index (SDAI)โ<โ3.3 and a Swollen Joint Count (SJC)โโคโ1, maintained across two consecutive visits 12 weeks apart. Patients were assessed at baseline and at one and two years before fulfilling the D2T RA definition. Predictive models were developed using machine learning techniques (lasso and ridge logistic regression, support vector machines, random forests, and XGBoost). Shapley additive explanation (SHAP) values were used to assess the contribution of individual variables to model predictions. Among 8,543 RA patients, 641 met the criteria for D2T RA, while 1,825 achieved remission. The machine learning models demonstrated an accuracy range of 0.606-0.747, with an area under the receiver operating characteristic curve (AUC) of 0.656-0.832 for predicting D2T RA. SHAP analysis highlighted key predictive variables, including disease activity measures (DAS28-ESR, CDAI, CRP), patient-reported outcomes (HAQ), and the duration of b/tsDMARD treatment. We identified clinical features predictive of D2T RA at baseline and up to one year before meeting the formal criteria. These findings provide valuable insights into early indicators of D2T RA progression and support the importance of earlier recognition and timely therapeutic intervention to improve long-term patient outcomes.
Author: [‘Baloun J’, ‘Cerezo LA’, ‘Kropรกฤkovรก T’, ‘Prokopcovรก A’, ‘Mareลกovรก KB’, ‘Mann H’, ‘Vencovskรฝ J’, ‘Pavelka K’, ‘ล enolt L’]
Journal: Sci Rep
Citation: Baloun J, et al. Machine learning-assisted screening of clinical features for predicting difficult-to-treat rheumatoid arthritis. Machine learning-assisted screening of clinical features for predicting difficult-to-treat rheumatoid arthritis. 2025; 15:34747. doi: 10.1038/s41598-025-18298-y