Related Experiment Video
Updated: Jun 18, 2026
06:59
Improved Registration of 3D CT Angiography with X-ray Fluoroscopy for Image Fusion During Transcatheter Aortic Valve Implantation
Published on: June 3, 2018
A novel computational method to predict hypoattenuated leaflet thickening post-transcatheter aortic valve replacement
Aniket Venkatesh1,2, Fateme Esmailie3, Noah Tregobov4,5
1Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, Atlanta, Ga.
JTCVS Structural and Endovascular
|June 17, 2026
Summary
A new AI pipeline predicts hypoattenuated leaflet thickening (HALT), a sign of blood clots after TAVR valve replacement, using pre-procedure CT scans. This tool aids in preventing valve dysfunction and improving patient outcomes.
Area of Science:
- Cardiovascular Imaging and Interventions
- Artificial Intelligence in Medicine
- Biomedical Engineering
Background:
- Hypoattenuated leaflet thickening (HALT) is a computed tomography (CT) finding post-transcatheter aortic valve replacement (TAVR) indicating bioprosthetic valvular thrombosis.
- Current methods for predicting HALT are not standardized or validated, posing risks for bioprosthetic valve dysfunction and adverse patient outcomes.
Purpose of the Study:
- To develop a novel, rapid, artificial intelligence (AI) and machine learning (ML)-driven computational pipeline for predicting HALT using preprocedural CT scans.
Main Methods:
- The pipeline integrated pre-TAVR CT reconstruction and reduced order modeling for predicting postprocedural geometric parameters.
- Automated left ventricle segmentation was employed to predict hemodynamic parameters.
- Statistical and ML analyses were performed to derive HALT predictive metrics.
Main Results:
- The pipeline achieved 84.40% accuracy and an AUC of 0.87 in predicting HALT.
- Statistically significant relationships were found between HALT and peak systolic blood velocity, peak systolic blood flow, left ventricular ejection time, ejection volume, and right coronary height.
- ML-derived metrics of neosinus circulation strongly correlated with HALT occurrence.
Conclusions:
- A computational pipeline utilizing pre-TAVR CT scans can predict post-TAVR geometric and hemodynamic measurements indicative of HALT risk.
- This predictive tool may enhance clinical decision-making and understanding of postprocedural thrombosis prevention.
