Fast virtual deployment of self-expandable stents: Method and in vitro evaluation for intracranial aneurysmal stenting

Abstract

Introduction

Minimally invasive treatment approaches, like the implantation of percutaneous stents, are becoming more popular every day for the treatment of intracranial aneurysms. The outcome of such treatments is related to factors like vessel and aneurysm geometry, hemodynamic conditions and device design. For this reason, having a tool for assessing stenting alternatives beforehand is crucial.

Methodology

The Fast Virtual Stenting (FVS) method, which provides an estimation of the configuration of intracranial stents when released in realistic geometries, is proposed in this paper. This method is based on constrained simplex deformable models. The constraints are used to account for the stent design. An algorithm for its computational implementation is also proposed. The performance of the proposed methodology was contrasted with real stents released in a silicone phantom.

Results

In vitro experiments were performed on the phantom where a contrast injection was performed. Subsequently, corresponding Computational Fluid Dynamics (CFD) analyzes were carried out on a digital replica of the phantom with the virtually released stent. Virtual angiographies are used to compare in vitro experiments and CFD analysis. Contrast time–density curves for in vitro and CFD data were generated and used to compare them.

Conclusions

Results of both experiments resemble very well, especially when comparing the contrast density curves. The use of FVS methodology in the clinical environment could provide additional information to clinicians before the treatment to choose the therapy that best fits the patient.

Introduction

Intracranial aneurysms are pathological dilatations of cerebral vessels whose rupture leads to catastrophic complications such as diffuse bleeding in the brain cavities. The main treatment options for intracranial aneurysms include clipping of the aneurysm or endovascular treatment. Clipping consists in the placement of a metal clip around the neck of the aneurysm so as to isolate the aneurysm from the parent vasculature and to re-establish physiological blood flow. This technique involves craniotomy and a long post-interventional hospitalization demanding expensive health care costs. Regarding endovascular treatments, we can mention coiling and stent implantation. Aneurysm coiling involves the filling of the aneurysm with coils, reducing blood flow into the aneurysmal sac, promoting clotting and isolating the aneurysm from the main blood stream. Intracranial stenting involves the placement of a flexible self-expanding porous tubular mesh made of nitinol or other alloys on the parent vessel across the aneurysm neck. Coiling and stenting are frequently combined to support coil packing inside the aneurysm (Wanke et al., 2003, Vanninen et al., 2003, Felber et al., 2004, Alfke et al., 2004). In the case of aneurysms, stents are widely used to alter the flow reaching the aneurysm redirecting it from the aneurysm neck or dome towards the original blood stream. Additionally, it has been suggested that the use of multiple stents (Crowley et al., 2009) or flow diverters (Wanke and Forsting, 2008, Sadasivan et al., 2009) (a different family of stents consisting of a finer mesh) might be enough to provide sufficient hemodynamic resistance to blood flow trough the aneurysm neck, restituting the original vessel shape and the physiological blood flow along the parent vessel. Bearing this in mind, a tool providing an estimation of the stent configuration after release in a patient-specific anatomy would provide useful information to the clinician. The need and practicality of such a tool has been previously outlined by Karmonik et al. (2005). The authors state that the ability to visualize a virtually released stent within the parent artery provides information that is not visible in medical images (e.g., stent attachment to vessel wall and covering of the aneurysm neck). Such tool would allow evaluating in silico different alternatives for the intervention assessing its potential outcome prior to treatment. Supplementary to the purely qualitative aspects of knowing the stent configuration after its release, obtaining the released stent geometry beforehand is a first step towards introducing more complex tools for early therapy assessment. In that way, the combination of anatomically accurate vascular geometries with Computational Fluid Dynamics (CFD) has been extensively used to simulate flow in cerebral vasculature (Cebral et al., 2005, Steinman et al., 2003). Furthermore, CFD has been satisfactory used to assess the possible outcome of different stenting procedures in patient-specific geometries (Cebral and Löhner, 2005, Appanaboyina et al., 2007, Radaelli et al., 2008). On this regard, Sadasivan et al., 2002, Sadasivan et al., 2009 proposed a methodology for the evaluation of stenting therapy by the evaluation of contrast residence time in the aneurysm which, in combination with CFD, could well be used for early therapy assessment (Ford et al., 2005, Cebral and Löhner, 2005).

A considerable amount of work has been devoted over the past years to develop computational models of vascular stents, their physical behavior and efficiency for treatment. Most of this effort has been dedicated to coronary stents, which are used in the treatment of coronary artery disease (CAD) to maintain the vessel open after angioplasty. Previous work has focused on structural analysis of stent cells (McGarry et al., 2004, Gu et al., 2005, Theríault et al., 2006, Xia et al., 2007), constitutive modelling of stent materials (Petrini et al., 2004, Migliavacca et al., 2005), structural analysis of the interaction between the stent and the vascular wall (Holzapfel et al., 2002, Migliavacca et al., 2004, Timmins et al., 2007), and modelling of the hemodynamic alteration due to the presence of the stent (Deplano et al., 2004, LaDisa et al., 2004, Seo et al., 2005). In such cases, Finite Element Analysis (FEA) is an appropriate methodology for representing the detailed mechanical behavior of the stent material, its design and effect on the vascular wall. In the case of cerebral aneurysm stenting, more detailed description of the interaction between the stent and the vessel wall has been considered in (Bludszuweit-Philipp et al., 2008), where the stent-vessel interaction is approached and integrated with CFD simulation for the prediction of thrombus formation.

All these approaches are designed to model the stent behavior, its interaction with the vessel wall and the local blood flow from a strictly a mechanical point of view. Still, they are not particularly thought for its application on clinical problems where computationally fast techniques that can be applied as part of the daily clinical practice are required. To the best of our knowledge, so far only two approaches have been proposed to model the self-expandable stent release in patient-specific vessels. A first approach has been introduced by Cebral and Löhner (2005) and further studied by Appanaboyina et al. (2007). The stent geometry is mapped as a texture over a cylinder expanded inside the target vessel and the 3D representation of the stent is recovered. However, this method is purely based on the deformation of a cylindrical mesh and its results may be affected by its non-uniform deformation inside the vessel. Another approach has been proposed by Valencia et al. (2007), where the authors suggest to use a simplex deformable model that is initialized as a cylinder and deformed under the effect of internal and external forces to conform to the vessel centerline. Still, this method is based on deforming an already expanded cylinder, not properly taking into account neither the stent geometry nor its compliance to the vessel morphology.