Sjögren's disease & overlap diseases
Magnetic resonance imaging-based radiomics for early detection of primary Sjögren's disease using ZOOMit-DWI: a non-invasive diagnostic approach
Authors' conclusion
The RFEP model, leveraging parotid ZOOMit DWI radiomics, shows high diagnostic accuracy and interpretability, offering a non-invasive tool for early parotid gland injury in pSjD.
Study context
- Population
- Early primary Sjögren's disease (pSjD) patients and healthy controls; radiomics features from MRI-derived parotid gland volumes of interest (VOIs)
- Comparison
- Early pSjD vs healthy controls, classified by ZOOMit-DWI MRI radiomics models
- Early pSjDn = 15
- Healthy controlsn = 40reference
- Outcome
- Diagnostic performance (AUC, ROC analysis) for early pSjD
Methodological detailsretrospective, MRI-based radiomics, machine learning
- Retrospective
- MRI-based radiomics
- Machine learning
Performance
As reported in the abstract
Diagnostic performance (AUC)
| Test or model | AUC (training) | AUC (test) | pSjD cases correctly identified | pSjD cases correctly identified (training) | pSjD cases correctly identified (test) |
|---|---|---|---|---|---|
| RFEP | 0.96 | 0.92 | 13 out of 15 | 9/10 | 4/5 |
RFEP
AUC (training)0.96
AUC (test)0.92
pSjD cases correctly identified13 out of 15
pSjD cases correctly identified (training)9/10
pSjD cases correctly identified (test)4/5
- All models, AUC range (training)
- 0.91 to 0.96
- All models, AUC range (test)
- 0.73 to 0.92
Original publication
Magnetic resonance imaging-based radiomics for early detection of primary Sjögren's disease using ZOOMit-DWI: a non-invasive diagnostic approach. Clin Exp Rheumatol. Published 26 October 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →