Lupus & antiphospholipid syndrome
Non-invasive prediction of the chronic degree of lupus nephropathy based on ultrasound radiomics
Authors' conclusion
ML model based on ultrasound radiomics features can accurately predict the chronic degree of LN
Study context
- Population
- Patients with lupus nephritis (LN) who had renal biopsy7:3 training/validation
- Participants
- 136 patients
- Comparison
- Ultrasound radiomics machine learning (ML) models vs clinical model vs US model
- Outcome
- Prediction of the chronicity index of LN (AUC)
Methodological detailsretrospective, machine learning radiomics, random 7:3 training/validation split, LASSO feature selection
- Retrospective
- Machine learning radiomics
- Random 7:3 training/validation split
- LASSO feature selection
Performance
As reported in the abstract
Prediction of LN chronicity index, AUC in test set
| Test or model | AUC |
|---|---|
| Xgboost | 0.826 (95% CI: 0.681-0.936) |
| Clinical model | 0.560 (95% CI: 0.357-0.761) |
| US model | 0.679 (95% CI: 0.489-0.853) |
AUC
Xgboost0.826 (95% CI: 0.681-0.936)
Clinical model0.560 (95% CI: 0.357-0.761)
US model0.679 (95% CI: 0.489-0.853)
- Average ROC of the seven ML models
- 0.683
Original publication
Non-invasive prediction of the chronic degree of lupus nephropathy based on ultrasound radiomics. Lupus. Published 20 December 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →