Predicting treatment outcome in congenital adrenal hyperplasia using urine steroidomics and machine learning
Authors' conclusion
Using ML on 24 h urine steroid profiling predicted treatment outcome in children with CAH, even in the absence of clinical data, suggesting that routine comprehensive 24 h urine steroid profiling could improve treatment monitoring in CAH.
Study context
- Population
- Children and adolescents with congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency; 24 h urines of 2 consecutive visits, 40 steroids measured by gas chromatography-mass spectrometry
- Participants
- 59 children112 visits
- Exposure
- Sparse partial least squares discriminant analysis (sPLS-DA) model using 24 h urine metabolites only vs model adding clinical variables
- Outcome
- Prediction of treatment outcome (clinically classified as undertreated, optimally treated or overtreated) by machine learning (ML) on 24 h urine steroid profiling (AUC)
Main finding
As reported in the abstract
Prediction of treatment outcome (optimal vs undertreatment) by sPLS-DA
| Group | AUC |
|---|---|
| Urine metabolites | 0.88 |
Urine metabolites
AUC0.88
- Addition of clinical variables marginally improved classification (AUC)
- 0.90
Observational data: associations, not effects.
Original publication
Predicting treatment outcome in congenital adrenal hyperplasia using urine steroidomics and machine learning. Eur J Endocrinol. Published 14 June 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →