Deep Learning-based Multiview Facial Identification as a Screening Tool for Cushing Syndrome
Authors' conclusion
These findings indicate that multiview facial identification using deep learning could provide an accurate and accessible diagnostic method for CS. Given that the Artificial intelligence models outperformed endocrinologists, this approach holds potential as a novel diagnostic tool to assist clinicians in identifying various types of CS.
Study context
- Population
- Patients with cushing syndrome (CS) and age-, sex- and BMI-matched controls; training cohort 42 matched pairs, external validation cohort 13 matched pairs42 + 13 matched pairs
- Comparison
- Deep learning facial identification models vs 18 endocrinologists
- Endocrinologistsn = 18reference
- Outcome
- Diagnostic performance of the 4 models in the external validation cohort
Methodological detailsage-, sex- and BMI-matched controls, fivefold cross-validation, external validation cohort
- Age-, sex- and BMI-matched controls
- Fivefold cross-validation
- External validation cohort
Performance
As reported in the abstract
Accuracy (range across models)
92.3% to 100%
Sensitivity (range across models)
84.6% to 100%
Specificity
100%
AUC (range across models)
94.4% to 100%
Further findings
Diagnostic performance in training cohort
- All 4 models showed stable diagnostic performance across the 5 cross-validation folds
Original publication
Deep Learning-based Multiview Facial Identification as a Screening Tool for Cushing Syndrome. Endocr Pract. Published 28 July 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →