Pulmonary vascular disease & thromboembolismDiagnostic or prognostic study
The Pediatric Pulmonary Hypertension International Risk Score: A Prediction Model for Outcomes Using Machine Learning
- Prognostic model development and external validation study
- multicenter
- machine learning (extreme gradient boosting)
Authors' conclusion
This multicenter, validated model provides good 1-year risk prediction in pediatric PH across World Symposium on Pulmonary Hypertension groups 1 and 3, providing a robust tool for clinical risk stratification to guide therapy and addressing a gap in pediatric PH care.
Study context
- Population
- Pediatric patients with pulmonary hypertension (PH) (World Symposium on Pulmonary Hypertension groups 1 and 3), Pediatric Pulmonary Hypertension Network Registry 2014–2020development registry n=345 plus external validation registries
- Comparison
- Test vs Dutch National Registry for Pulmonary Hypertension in Childhood vs Spanish Registry of Pediatric Pulmonary Hypertension
- Dutch National Registry for Pulmonary Hypertension in Childhoodn = 155
- Spanish Registry of Pediatric Pulmonary Hypertensionn = 327
- Outcome
- Time to death, transplant, Potts shunt, or atrial septostomy
- Follow-up
- 1 year
Performance
As reported in the abstract
Model discrimination (1-year composite outcome)
| Test or model | AUC | NPV |
|---|---|---|
| Test | 0.90 (0.79-0.97) | 99% (96%-99%) |
| Dutch National Registry for Pulmonary Hypertension in Childhood | 0.76 (0.70-0.81) | 93% (93%-97%) |
| Spanish Registry of Pediatric Pulmonary Hypertension | 0.77 (0.73-0.82) | 96% (93%-97%) |
Further findings
Kaplan-Meier event-free survival by risk group
- Kaplan-Meier analysis significantly differentiated outcomes by risk group (3 groups)
Original publication
The Pediatric Pulmonary Hypertension International Risk Score: A Prediction Model for Outcomes Using Machine Learning. Circulation. Published 11 August 2026. doi:10.1161/circulationaha.125.077391
Summarised from the published abstract; every number is checked against it