Predicting rheumatoid arthritis progression from seronegative undifferentiated arthritis using machine learning: a deep learning model trained on the KURAMA cohort and externally validated with the ANSWER cohort
Authors' conclusion
The FNN model exhibited robust performance in predicting the progression of RA from seronegative UA and maintained substantial sensitivity in an independent validation cohort.
Study context
- Population
- Patients with seronegative undifferentiated arthritis (UA)
- Participants
- 210 vs 140
- Comparison
- KURAMA vs ANSWER
- KURAMAn = 210
- ANSWERn = 140
- Outcome
- Progression from seronegative UA to rheumatoid arthritis (RA), predicted by ML model (AUC, sensitivity)
Methodological detailsmachine learning, Feedforward Neural Network (FNN), SHAP values
- Machine learning
- Feedforward Neural Network (FNN)
- SHAP values
Performance
As reported in the abstract
FNN model performance for predicting progression to RA
| Test or model | AUC | Sensitivity |
|---|---|---|
| KURAMA | 0.924 | 80.7% |
| ANSWER | 0.777 | 77.8% |
AUC
KURAMA0.924
ANSWER0.777
Sensitivity
KURAMA80.7%
ANSWER77.8%
Further findings
Progression to RA
Variable importance (SHAP)
- MMP-3 had the highest impact on the model
Original publication
Predicting rheumatoid arthritis progression from seronegative undifferentiated arthritis using machine learning: a deep learning model trained on the KURAMA cohort and externally validated with the ANSWER cohort. Arthritis Res Ther. Published 26 March 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →