Lupus & antiphospholipid syndrome
Validity and applicability of machine learning models for systemic lupus erythematosus diagnosis
Authors' conclusion
We developed an open-access, clinically user-friendly tool-AI-PSLE-based on 35 routine features, aimed at facilitating the early identification of SLE patients from healthy populations.
Study context
- Population
- Systemic lupus erythematosus (SLE) patients and healthy controls from the Luzhou and Enshi cohortstwo cohorts
- Comparison
- Luzhou vs Enshi
- Outcome
- Performance of the LGB model on 35 reproducible features
Methodological detailsmachine learning, discovery and external validation cohorts, 13 binary classification metrics, Model integrated into an open-access tool, AI-PSLE
- Machine learning
- Discovery and external validation cohorts
- 13 binary classification metrics
- Model integrated into an open-access tool, AI-PSLE
Performance
As reported in the abstract
- 50 candidate features selected in Luzhou; Light gradient boosting (LGB) model performed best after preprocessing
- 35 reproducible features retained after scaling in Enshi
- LGB model on 35 features maintained superior performance in Luzhou and was successfully validated in Enshi and combined Luzhou+Enshi cohorts
Original publication
Validity and applicability of machine learning models for systemic lupus erythematosus diagnosis. Lupus Sci Med. Published 27 April 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →