Related Experiment Video
Updated: Jun 13, 2026
06:04
Upper-extremity Approach for Secondary Access in Transfemoral Transcatheter Aortic Valve Implantation
Published on: August 8, 2025
Machine Learning-Driven Probability of Permanent Pacemaker Implantation After Transcatheter Aortic Valve Replacement
Marcel Abras1,2, Daniela Bursacovschi3, Ecaterina Pasat2
1Department of Cardiology, Nicolae Testemitanu State University of Medicine and Pharmacy, MD-2004 Chisinau, Moldova.
Diagnostics (Basel, Switzerland)
|June 12, 2026
Summary
Machine learning accurately predicts permanent pacemaker implantation (PPI) risk after transcatheter aortic valve replacement (TAVR). Anatomical factors, not clinical history, are key predictors, aiding procedural planning.
Area of Science:
- Cardiology
- Medical Informatics
- Artificial Intelligence
Background:
- Permanent pacemaker implantation (PPI) is a frequent complication after transcatheter aortic valve replacement (TAVR).
- Predicting PPI risk is crucial for optimizing procedural planning and patient management.
- Accurate risk stratification can improve outcomes for patients undergoing TAVR.
Purpose of the Study:
- To develop and validate a machine learning (ML) model for predicting PPI risk post-TAVR.
- To identify key predictors of PPI using explainable AI (XAI) methods.
- To enhance clinical decision-making for TAVR procedures.
Main Methods:
- Prospective study of 179 patients undergoing TAVR (2019-2025).
- Development of an ML model using clinical, echocardiographic, and imaging data.
- Evaluation of model performance using accuracy, F1-score, and confusion matrix.
- Application of SHapley Additive exPlanations (SHAP) for feature importance analysis.
Main Results:
- The ML model achieved high predictive performance (accuracy=0.944, F1-score=0.947).
- Anatomical variables (valve size, sinus diameter, annulus diameter) were primary predictors of PPI.
- Baseline clinical factors (e.g., LVEF, MI history) had a lower predictive impact.
Conclusions:
- Machine learning models can reliably predict PPI risk following TAVR.
- Aortic root anatomy and prosthesis characteristics are critical determinants of PPI.
- Explainable AI enhances risk assessment and supports procedural planning for TAVR.
