A Fast Method of Virtual Stent Graft Deployment for Computer Assisted EVAR

TL;DR

This paper presents a rapid, reliable simulation method for virtual stent graft deployment in EVAR, combining intraoperative images and mechanical models to improve accuracy and assist surgeons.

Contribution

A novel multi-step algorithm for simulating stent graft deployment that enhances accuracy and speed using intraoperative data and mechanical modeling.

Findings

Average deployment error of 1.73 mm, within clinical threshold.

Maximum fenestration reconstruction error under 2.5 mm.

Method enables fast, accurate EVAR planning with low-cost framework.

Abstract

In this paper we introduce a new method simulating stent graft deployment for assisting endovascular repair of abdominal aortic aneurysms. The method relies on intraoperative images coupled with mechanical models. A multi-step algorithm has been developed to increase the reliability of simulations. The first step predicts the position of the stent graft within the aorta. The second step is an axisymmetric geometric reconstruction of each individual stent. The third step minimizes the rotation of each stent around its main axis. Finally, the last step virtually deploys each stent within a deployment box extracted from the preoperative CT scan. A proof of concept is performed on a patient. The accuracy is compatible with the clinical threshold of 3 mm: the average distance between target and simulated stents is 1.73 $\pm$ 0.37 mm. Fenestrations of the stent-graft are reconstructed with a maximum error of less than 2.5 mm, which is enables a secure catheterization of secondary arteries. In summary, the method is able to assist EVAR practitioners by providing all necessary information for a fast and accurate stent graft positioning, combining intraoperative data and a mechanical model in a very low cost framework.