Non-component clinical feature-based machine learning for disease activity risk stratification in juvenile…

AI / informaticsFrontiers in pediatrics· 2026

Non-component clinical feature-based machine learning for disease activity risk stratification in juvenile idiopathic arthritis: a multi-center retrospective validation study.

Peipei Dong, Fei Song, Bin Wang … Chuansheng Wu

RheumatologyAI / informaticsDiagnosisQ1IF 2.2

Machine learning with non-component features accurately stratifies juvenile idiopathic arthritis disease activity.

Non-component clinical feature-based machine learning for disease activity risk stratification in juvenile idiopathic arthritis: a multi-center retrospective validation study.

Peipei Dong … Chuansheng Wu

Frontiers in pediatrics · 2026

Background

JADAS27 is widely used to assess juvenile idiopathic arthritis (juvenile idiopathic arthritis) disease activity, but complete scoring is impractical in many clinical settings because it requires simultaneous physician global assessment (physician global assessment), patient/parent global assessment (parent global assessment), active joint count (active joint count), and erythrocyte sedimentation rate (erythrocyte sedimentation rate).

Methods

In this retrospective multi-center study, 800 patients with juvenile idiopathic arthritis were enrolled from CARRA ( n = 400), PRCSG ( n = 240), and LHTCM ( n = 160).

n = 800 patients

accuracy 0.731

Results

accuracy 0.731

the computer model correctly identified juvenile arthritis activity levels nearly three-quarters of the time

n = 800 patients

More results

Class-wise recall was highest for inactive (82.4%) and high activity (77.4%), with most errors between adjacent classes.

Removing all three proxy variables reduced accuracy to 0.619 and macro AUC to 0.843, indicating that both proxy and non-proxy features contribute independently.

More results

Sensitivity analysis with subtype-specific cutoffs yielded comparable performance (accuracy 0.713).

“

Conclusion · 1 of 2

Strictly non-component clinical features can stratify JADAS27-defined disease activity with clinically meaningful external performance.

Conclusion · 2 of 2

This approach may support early risk stratification when formal JADAS27 scoring is unavailable, and complements rather than replaces physician assessment.

Non-component clinical feature-based machine learning for disease activity risk stratification in juvenile idiopathic arthritis: a multi-center retrospective validation study.

Non-component clinical feature-based machine learning for disease activity risk stratification in juvenile idiopathic arthritis: a multi-center retrospective validation study.

Related papers