Image Registration-Based Method for Reconstructing Transcatheter Heart Valve Geometry from Patient-Specific CT Scans

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Computational Mechanics

Background:

• Accurate reconstruction of transcatheter aortic valve (TAV) geometries from CT images is challenging due to metallic stent blooming artifacts and low tissue signal strength for leaflet segmentation.
• Existing methods struggle with precise in vivo geometry recovery of cardiac devices.

Purpose of the Study:

• To develop and validate an image registration-based method for accurate in vivo reconstruction of TAV stent and leaflet geometries from patient-specific CT images.
• To enable subsequent computation of leaflet stress and self-expandable stent stress distribution.

Main Methods:

• Exploitation of known device geometry combined with an image registration-based reconstruction approach.
• Utilizing displacement boundary conditions derived from registration for finite element analysis (FEA) of leaflet geometry and residual stress.
• FEA-based simulation for estimating stress distribution in self-expandable stents.

Main Results:

• Achieved geometric error of approximately 0.1mm for stent reconstruction, significantly lower than CT resolution.
• Demonstrated robustness to input biases with minimal human input required.
• Enabled accurate computation of leaflet geometry, residual stress, and stent stress distribution.

Conclusions:

• The developed method accurately reconstructs in vivo TAV stent and leaflet geometries from CT scans, overcoming common imaging artifacts.
• This technique facilitates pre-surgical planning (e.g., TAV-in-TAV) and in vivo assessment of post-procedural outcomes.
• The method is applicable to other medical devices like coronary stents and aids in understanding device biomechanics.