Related Experiment Video
Updated: Jan 3, 2026
09:57
Four-Dimensional Computed Tomography-Guided Valve Sizing for Transcatheter Pulmonary Valve Replacement
Published on: January 20, 2022
3.0K
Enabling Automated Device Size Selection for Transcatheter Aortic Valve Implantation
Patricio Astudillo1,2, Peter Mortier1, Johan Bosmans3
1FEops, Technologiepark-Zwijnaarde 122, Ghent, Belgium.
Journal of Interventional Cardiology
|November 29, 2019
Summary
A new deep learning method accurately and quickly predicts aortic annulus size for transcatheter aortic valve implantation (TAVI) device selection. This AI tool aids both experienced and new operators, improving efficiency and accuracy in TAVI procedures.
Area of Science:
- Cardiovascular Imaging and Intervention
- Artificial Intelligence in Medicine
- Medical Device Technology
Background:
- Transcatheter aortic valve implantation (TAVI) volumes are increasing, necessitating enhanced efficiency for high-volume operators and training support for new practitioners.
- Accurate aortic annulus sizing is critical for optimal TAVI device selection and patient outcomes.
- Current manual measurement methods can be time-consuming and subject to interobserver variability.
Purpose of the Study:
- To develop and validate a fast deep learning method for automatic prediction of aortic annulus perimeter and area from annular plane images.
- To assess the accuracy and reproducibility of the deep learning method compared to manual measurements and interobserver variability.
- To evaluate the utility of the automated measurements for TAVI device size selection.
Main Methods:
- A deep learning approach combining two deep convolutional neural networks followed by a postprocessing step was developed.
- The models were trained on 355 patient cases and evaluated on an independent set of 118 patients.
- Validation included comparison against manual measurements and an interobserver variability study.
Main Results:
- The automated method demonstrated prediction accuracy for aortic annulus area and perimeter comparable to interobserver variability.
- Automated device size selection for Edwards Sapien 3 and Medtronic Evolut devices showed good agreement with operator selections.
- The entire analysis from image to prosthesis size suggestion was completed in under one second.
Conclusions:
- The proposed deep learning method provides a fast, accurate, and reproducible approach for automated TAVI device size selection.
- The method's reliability, validated against interobserver variability, suggests its potential for clinical integration.
- Embedding this AI tool into preoperative planning can significantly enhance TAVI procedure efficiency while maintaining accuracy.
