Machine learning algorithms to accelerate etiological diagnosis of congenital disorders of adrenal steroidogenesis
Authors' conclusion
Machine learning-assisted steroid profiling provides an accurate and highly interpretable diagnostic approach for CDAS, with potential for integration into pediatric endocrine diagnostics and decision-support systems.
Study context
- Population
- Genetically confirmed congenital disorders of adrenal steroidogenesis (CDAS) patients representing 8 subtypes and controlsdevelopment cohort
- Participants
- 1,027 participants
- Outcome
- Diagnostic performance of ML-assisted decision-tree model for CDAS classification (development cohort)
Methodological detailsmachine learning-assisted decision tree model, 5-fold cross-validation, independent validation cohort
- Machine learning-assisted decision tree model
- 5-fold cross-validation
- Independent validation cohort
Performance
As reported in the abstract
Mean overall accuracy
97.1%
Sensitivity
99.5%
Specificity
93.7%
Macro-AUC
0.97
Further findings
Diagnostic performance (validation cohort)
- Overall accuracy
- 98.9%
- Sensitivity
- 93.6%
- Specificity
- 99.8%
Subtype-level accuracy (development cohort)
- Subtype-level accuracy exceeded 98% for most major CDAS subtypes
- >98%
Strongest discriminators
- Feature importance and SHAP identified 11-deoxycortisol, 17-hydroxyprogesterone, 21-deoxycortisol and corticosterone as strongest discriminators
Principal component analysis (PCA) and Uniform Manifold Approximation and Projection (UMAP) clustering
- Distinct clustering of CDAS subtypes
Original publication
Machine learning algorithms to accelerate etiological diagnosis of congenital disorders of adrenal steroidogenesis. J Clin Endocrinol Metab. Published 26 June 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →