Machine Learning-Based Prediction of Three-Year Heart Failure and Mortality After Premature Ventricular Contraction Ablation
Machine-Learning-basierte Vorhersage von Herzinsuffizienz und Mortalität über 3 Jahre nach Ablation von vorzeitigen ventrikulären Kontraktionen
Diagnostics · 2025-10-24 · 15 (21) · 2693
DOI: 10.3390/diagnostics15212693 ↗Auf einen Blick
Bei 4.195 Erwachsenen nach PVC-Ablation sagte ein Machine-Learning-Modell 3-jährige Herzinsuffizienz und Mortalität mit guter Diskrimination vorher. Die besten ROC-AUCs lagen bei 0,822 bzw. 0,886/0,882.
Fragestellung
Wie gut lassen sich 3-jährige Herzinsuffizienz und Mortalität nach PVC-Ablation mit Machine-Learning-Modellen vorhersagen?
Methode
Retrospektive Analyse von 4.195 Erwachsenen aus einer nationalen Claims-Datenbank; Vergleich von 5 überwachten Algorithmen mit 5-fold-Cross-Validation und ROC AUC sowie PR-Kurven.
Ergebnisse
Für Herzinsuffizienz erreichte LightGBM mit ROSE die höchste ROC AUC von 0,822. Für Mortalität erzielten logistische Regression mit ROSE und LightGBM mit ROSE ROC-AUCs von 0,886 und 0,882.
Schlussfolgerung der Autor:innen
Die Berücksichtigung der Klassenimbalance und der Vergleich moderner Modelle mit logistischer Regression ergab robuste, klinisch interpretierbare Risikostratifizierung nach PVC-Ablation.
Kernaussagen
- LightGBM mit ROSE: ROC AUC 0,822 für 3-jährige Herzinsuffizienz
- Mortalität: logistische Regression 0,886 und LightGBM 0,882
- Wichtige Prädiktoren: Alter, frühere Herzinsuffizienz, Malignom, ESRD
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
At a glance
Journal article in Diagnostics, published online on 2025-10-24. Written by Chung-Yu Lin and 6 co-authors. Institution: Graduate Institute of Business Administration, Fu Jen Catholic University, New Taipei City 242062, Taiwan. Cited 1 times so far (Crossref); 55 references. The full text is freely available (open access, CC BY 4.0). Funded by: Fu Jen Catholic University, National Science and Technology Council, St. Paul's Hospital.
Topics: Künstliche Intelligenz in der Pflege, Herzinsuffizienz
Abstract (original language)
Introduction
Long-term heart failure and mortality after catheter ablation for premature ventricular contraction (PVC) remain underexplored.
Methods
We retrospectively analyzed 4195 adults who underwent PVC ablation in a nationwide claims database. To address class imbalance, we used synthetic minority over-sampling technique (SMOTE) and random over-sampling examples (ROSE). Five supervised algorithms were compared: logistic regression, decision tree, random forest, XGBoost, and LightGBM. Discrimination was assessed by stratified five-fold cross-validation using the area under the receiver operating characteristic curve (ROC AUC). Because rare events can bias ROC, we also examined precision–recall (PR) curves.
Results
For predicting three-year heart failure, LightGBM with ROSE achieved the highest ROC AUC at 0.822. For three-year mortality, logistic regression with ROSE and LightGBM with ROSE showed balanced performance with ROC AUCs of 0.886 and 0.882. Pairwise DeLong tests indicated that these leading models formed a high-performing cluster without significant differences in ROC AUC. Age, prior heart failure, malignancy, and end-stage renal disease were the most influential predictors by model explainability analysis.
Discussion
Addressing class imbalance and benchmarking modern learners against a transparent logistic baseline yielded robust, clinically interpretable risk stratification after PVC ablation. These models are suitable for integration into electronic health record dashboards, with external validation and local threshold optimization as next steps.
Subjects
Author keywords
Authors and institutions
- Graduate Institute of Business Administration, Fu Jen Catholic University, New Taipei City 242062, Taiwan.
- Department of Cardiology, Fu Jen Catholic University Hospital, New Taipei City 242062, Taiwan.
- Department of Thoracic Medicine, St. Pual's Hospital, Taoyuan City 330058, Taiwan.
- Department of Cardiology, St Paul's Hospital, Taoyuan City 330058, Taiwan.
- Department of Teaching and Research, St. Paul's Hospital, Taoyuan City 330058, Taiwan.
ORCID profiles: Chung-Yu Lin, Chien-Wei Chuang
Funding
- Fu Jen Catholic University
- National Science and Technology Council
- St. Paul's Hospital
- Ben-Chang Shia
- St. Paul’s Hospital
Cite
APA
Lin, C. Y., Lai, Y. T., Chuang, C. W., Yu, C. H., Lo, C. Y., Chen, M., & Shia, B. C. (2025). Machine Learning-Based Prediction of Three-Year Heart Failure and Mortality After Premature Ventricular Contraction Ablation. Diagnostics, 15(21), 2693. https://doi.org/10.3390/diagnostics15212693
Vancouver
Lin CY, Lai YT, Chuang CW, Yu CH, Lo CY, Chen M, et al. Machine Learning-Based Prediction of Three-Year Heart Failure and Mortality After Premature Ventricular Contraction Ablation. Diagnostics. 2025;15(21):2693. doi:10.3390/diagnostics15212693
BibTeX
@article{Lin2025,
author = {Lin, Chung-Yu and Lai, Yu-Te and Chuang, Chien-Wei and Yu, Chih-Hsien and Lo, Chiung-Yun and Chen, Mingchih and Shia, Ben-Chang},
title = {Machine Learning-Based Prediction of Three-Year Heart Failure and Mortality After Premature Ventricular Contraction Ablation},
journal = {Diagnostics},
year = {2025},
volume = {15},
number = {21},
pages = {2693},
doi = {10.3390/diagnostics15212693}
}
EUVIMED