A fully automated artificial intelligence-driven software for planning of transcatheter aortic valve replacement

Area of Science:

• Cardiovascular Imaging
• Medical Artificial Intelligence
• Surgical Planning Software

Background:

• Transcatheter aortic valve replacement (TAVR) is a growing treatment for aortic stenosis.
• Computed tomography (CT) is crucial for TAVR pre-procedural planning.
• Current TAVR planning software demands significant manual input.

Purpose of the Study:

• To develop and validate an artificial intelligence (AI) powered software for automated anatomical measurements in TAVR planning.
• To reduce human interaction in the TAVR planning process.

Main Methods:

• Automated measurements from 100 CT scans were compared against expert clinician measurements.
• Precision and accuracy were assessed using correlation coefficients and mean differences.
• AI software performance was evaluated against commercially available planning tools.

Main Results:

• AI-generated annular measurements showed excellent agreement with expert manual measurements (correlation coefficients of 0.97 for perimeter and area).
• Minimal bias was observed in AI measurements for annular perimeter (-0.07 mm) and area (-1.4 mm²).
• AI-driven sizing demonstrated high agreement with effective implant sizes for both self-expanding (87-88%) and balloon-expandable (88%) valves.

Conclusions:

• A fully automated AI software provides accurate and precise anatomical segmentation and measurements for TAVR planning.
• The software operates with high reliability, eliminating the need for human interaction.
• This AI solution streamlines TAVR planning and implantation processes.