Enhanced Guidance for Endovascular Repair of Abdominal Aortic Aneurysm

Overview

An increasing incidence of abdominal aortic aneurysms (AAA) is observed in our ageing population. Since it is less invasive, endovascular repair (EVAR) by stent-graft (SG) insertion is frequently indicated in patients with intermediate and high-operative risk. Patient selection and stent planning for EVAR is done on CT-scans based on specific anatomic criteria. SG are inserted in the catheterization laboratory under digital substraction angiography (DSA) and fluoroscopic guidance. With this technology, no soft-tissue differentiation is available during the intervention and important information are lacking such as thrombus and aneurysm extension close to proximal and distal landing zones. Our team has recently patented a software allowing the extraction of the AAA (lumen and thrombus) from pre-operative CT-scanner. In collaboration with Siemens medical, the investigators have integrated this technology to the Siemens workstation in the catheterization laboratory. The investigators can now import the aortic lumen and thrombus meshes segmented from pre-operative CT-scans and perform a rigid registration with fluoroscopy and DSA data to enhance visualization of soft tissue during EVAR. Our preliminary results are encouraging in terms of feasibility and visualization. However, the delivery device of the SG and the guidewire used during the intervention are stiff and induce a deformation of aortic lumen and thrombus. This deformation impairs the accuracy of rigid registration. The investigators propose to improve registration accuracy by implementing an elastic deformation of aortic lumen and wall based on the segmentation of endovascular devices (delivery device, guidewires and catheters) inserted during the procedure and by biomechanical modeling.

This project has 3 objectives: 1. To implement and validate a new optimized workflow enabling rigid registration between AAA meshes extracted from preoperative CT-scan with fluoroscopic images. 2. To develop an elastic registration of the AAA meshes based on 2D or 3D modeling of the endovascular device. 3. To validate the accuracy of elastic registration and compare it with the rigid registration. Experimental protocol: These 3 objectives will be completed in three experimental phases: 1. Improvement and implementation of the current rigid registration software. In this phase, several new features such as an improved correction system by automatic capture of DSA, an automated 3D/3D rigid registration, recognition of the origin of internal iliac arteries will be implemented. This new beta version will be tested in 20 patients at CHUM Research center and Dalhousie University-QEII Health Sciences Centre. 2. Implementation of an elastic registration based on endovascular device segmentation and center line corrections. The luminal path of the segmented lumen on preoperative CT-scan will be aligned with the path of the endovascular devices, then an elastic deformation of the aortic lumen and thrombus meshes generated from CT-scanner will be applied and registered to fluoroscopic and DSA images. A workflow allowing a fast recognition and segmentation of endovascular devices from one to three stereotaxic fluoroscopic views will be implemented in the Leonardo workstation and tested off-line on the previous 20 patient database and also in-vitro in realistic phantoms. 3. Online validation of elastic registration: The best strategy as defined above will be validated clinically on-line. The beta version will be tested in real time in 20 patients requiring SG procedures and compared to the prior cohort having EVAR/FEVAR procedures with rigid registration. The accuracy of rigid and elastic registrations will be compared in the two experimental groups.