Multicenter development and validation of machine-learning risk models to predict procedural complete revascularization and in-hospital heart failure in STEMI patients treated with primary PCI - Summary - MDSpire
Advertisement
Multicenter development and validation of machine-learning risk models to predict procedural complete revascularization and in-hospital heart failure in STEMI patients treated with primary PCI
To create and externally validate machine-learning models for predicting in-hospital heart failure and the feasibility of achieving complete revascularization during primary PCI in STEMI patients.
Key Findings:
CatBoost model showed the highest performance for predicting in-hospital heart failure (AUC: 0.973; accuracy: 88.6%).
CatBoost also excelled in predicting procedural complete revascularization (AUC: 0.970; accuracy: 92.0%).
Key predictors included LAD involvement, age, symptom-to-guidewire crossing duration, and various clinical indicators.
Interpretation:
The developed models provide strong discrimination and calibration, enhancing decision-making in catheterization laboratories for STEMI patients undergoing PPCI.
Limitations:
The study's findings may not be generalizable beyond the specific multicenter cohort.
External validation may still be limited in diverse clinical settings.
Conclusion:
The study successfully developed and validated two machine learning models that can aid in predicting in-hospital heart failure and the likelihood of achieving complete revascularization during primary PCI, thus supporting clinical decision-making.
