Machine Learning Prediction of Recurrence in Pediatric Thyroid Cancer: Malignant Endocrine Tumors Cohort Analysis Using XGBoost and SHAP
Authors' conclusion
This interpretable machine learning model reliably predicts early recurrence or nonremission in pediatric DTC and may complement current risk stratification systems to support personalized, risk-adapted treatment decisions.
Study context
- Population
- Pediatric patients (aged <18 years) with differentiated thyroid carcinoma (DTC) in the German Pediatric Oncology Hematology-Malignant Endocrine Tumors Registry (1997–2023), with known age at diagnosis and ≥24 months of follow-up
- Participants
- 250 patients
- Outcome
- Composite:
- Structural recurrence
- Failure to achieve remission
- Follow-up
- Within 24 months of initial therapy
Methodological detailsregistry-based, machine learning prediction model, 80/20 train–test split, 50 randomized stratified train-validation splits
- Registry-based
- Machine learning prediction model
- 80/20 train–test split
- 50 randomized stratified train-validation splits
Main finding
As reported in the abstract
Prediction of structural recurrence or nonremission within 24 months of initial therapy
Final model AUROC, independent test set
0.86
Mean AUROC across 50 validation splits
0.82
SD ± 0.05
Sensitivity across 50 validation splits
0.81
SD ± 0.09
Specificity across 50 validation splits
0.64
SD ± 0.06
Observational data: associations, not effects.
Further findings
Most influential predictors (SHAP)
- Younger age at diagnosis (<10 years), elevated postoperative Tg and distant metastases were the most influential predictors
Original publication
Machine Learning Prediction of Recurrence in Pediatric Thyroid Cancer: Malignant Endocrine Tumors Cohort Analysis Using XGBoost and SHAP. J Clin Endocrinol Metab. Published 1 September 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →