Image registration-based method to reconstruct transcatheter heart valve geometry from patient-specific CT scans

0

This article was originally published here

Ann Biomed Eng. 2022 Apr 15. doi: 10.1007/s10439-022-02962-9. Online ahead of print.

ABSTRACT

Accurate reconstruction of transcatheter aortic valve (TAV) geometries and other stented heart devices from computed tomography (CT) images is challenging, primarily associated with bloom artifacts caused by metal stents. Additionally, the bioprosthetic leaflets of VATs are difficult to segment due to low tissue signal intensity. This article describes a method that exploits known device geometry and uses an image registration-based reconstruction method to accurately recover in vivo stent and leaflet geometries from patient-specific CT images. Error analyzes showed that the geometric error of the stent reconstruction is approximately 0.1 mm, less than 1/3 of the stent width or at most CT scan resolutions. Moreover, the method requires only a few human inputs and is robust to input biases. The geometry and residual stress of the sheets can then be calculated using finite element analysis (FEA) with displacement boundary conditions derived from the recording. Finally, the stress distribution in self-expanding stents can be reasonably estimated by FEA-based simulation. This method can be used in pre-surgical planning for TAV-in-TAV procedures or for in vivo evaluation of surgical outcomes of post-procedural CT scans. It can also be used to reconstruct other medical devices such as coronary stents.

PMID:35428905 | DOI:10.1007/s10439-022-02962-9

Share.

About Author

Comments are closed.