Pre-procedural simulation for precision stent-assisted coiling of cerebral aneurysm

Abstract

Artificial vascular models are emerging as a newly-inexpensive and accurate way to simulate a procedure before the treatment. Through utilization of precision three-dimensionally printed, silicone-reconstructed, patient-specific models of aneurysms, we can compare the performance of devices including stents, and accurately predict the behavior of the microcatheter and stent-assisted coiling in the aneurysm to not only reduce procedural time, but also make the procedure safer. Here we report two challenging cases of wide-necked aneurysms, which could be safely treated with stent-assisted coiling as simulated in the patient-specific aneurysm models.

Introduction

Stent-assisted coiling is an effective treatment option for wide-neck aneurysms with lower rates of recanalization and retreatment.^1,2 Lower profile stents such as the LVIS Jr. (Microvention-Terumo, Tustin, CA, USA) and Neuroform ATLAS (Stryker, Kalamazoo, MI, USA) are available and useful to treat aneurysms in distal and tortuous vessels.

Current endovascular treatment utilizes virtual three-dimensional (3D) imaging technology, which relies upon physician perception of a 3D reconstruction on a two-dimensional (2D) screen. However, a challenge exists to predict the degree of expansion of the stent in a specific aneurysm anatomy, the ability to protect a side branch, and behavior of a microcatheter and coils after stent deployment based on the 2D screen images.

One emerging solution to counteract the anatomical discrepancies between the patient’s vessel and reconstructed 3D images is the utilization of 3D printed, patient-specific vascular models. With our advancements in 3D vascular modeling methodologies, a patient-specific, silicone vascular hollow replica can be created within 24 h for as little as $2.³ Here we present two cases of challenging wide-neck aneurysms with an eloquent branch incorporation. In both cases, pre-treatment simulations provided invaluable information to the treating physicians in terms of device selection and device behavior.

Case presentation

Digital imaging and communication in medicine (DICOM) data were exported and loaded to Osirix Lite (Pixmeo SARL, Bernex, Switzerland). Post-processing was performed to add the necessary inlet, outlet, and casing for the vascular model. The 3D vascular image was converted to a stereolithography file (.stl) and transferred to a 3D printer. The 3D solid vascular mold was formed from acrylonitrile butadiene styrene (ABS) plastic using a 3D printer (Mojo; Stratasys, Eden Prairie, MN, USA). The vascular mold was immersed in ABS solvent (eSolve, Kaneko Chemical, Saitama, Japan) for 1 min to smooth the stair-like surface of the printed objects.³ After drying, the vascular mold was filled with polydimethylsiloxanes (PDMS; Dow Corning Corp., Midland, MI, USA), cured at 60℃ overnight, and then immersed in acetone to dissolve away the ABS plastic. To simulate endovascular treatment, the silicone replica vascular model was connected to a silicone tube with a peristaltic pump (WPX1; Welco, Tokyo, Japan).

Case 1

A 63-year-old female presented to our hospital with a left A3 aneurysm. Due to her risk factors of its size and multiple cerebral aneurysms, stent-assisted coiling was recommended. Because of the complex nature of her wide-neck aneurysm which incorporated an eloquent frontal branch (Figure 1(a)), we elected to perform a patient-specific simulation for treatment. In our simulation, the Neuroform ATLAS stent had a significantly larger degree of herniation into the aneurysmal sac relative to the LVIS Jr. stent (Figure 1(b) and (c)). Additionally, the simulation elucidated the simplest route of microcatheter advancement for trans-cell coil embolization, which was consistent with the microcatheter route in the operation. The Neuroform ATLAS stent herniated into the aneurysm sac and the frontal branch was protected as observed in our simulation (Figure 1(d)).

Figure 1.

(a) Left internal carotid artery angiogram showing an A3 wide-neck aneurysm. The neck measures approximately 6.71 mm while the remainder of the aneurysm measures 6.23 mm × 3.79 mm × 2.29 mm (CC × TR × AP). The diameter of the proximal A2 segment measures 2.0 mm. (b) Silicone block model photograph showing the deployment and herniation of the Neuroform ATLAS stent (3.0 mm × 21 mm) into the neck of the aneurysm. (c) Silicone block model angiogram showing the trans-cell coil embolization of the aneurysm through the Neuroform ATLAS stent. (d) Post-operative image showing the successful coil embolization of the wide-neck, A3 aneurysm with a protected, patent frontal branch.

Case 2

A 69-year-old woman presented to our hospital with a left basilar-superior cerebellar artery (SCA) aneurysm. Due to her risk factors of multiple aneurysms, family history of ruptured cerebral aneurysms, and presence of a bleb, she was recommended for treatment. Because her wide-neck aneurysm incorporated the origin of the SCA (Figure 2(a)), stent-assisted coiling was considered. Through use of our patient-specific aneurysm simulator we determined that both the LVIS Jr. and Neuroform ATLAS stents herniated into the aneurysm sac and protected the SCA, but coil-embolization was more stable in the Neuroform ATLAS (Figure 2(b) and (c)). Based on the simulation result, Neuroform ATLAS stent-assisted coil-embolization was performed. The Neuroform ATLAS stent herniated into the aneurysm sac and the SCA was protected as observed in our simulation (Figure 2(d)).

Figure 2.

(a) Right vertebral artery angiogram showing an irregular shaped saccular aneurysm in the distal basilar artery at the origin of the left superior cerebellar artery (SCA). The origin of the SCA is incorporated in the neck of the aneurysm which measures 2.22 mm. Distance from the contralateral wall of the parent artery to the lateral origin of the SCA measures 3.5 mm. The remainder of the aneurysm measures 2.77 mm × 2.44 mm and projects laterally, horizontally, and posteriorly. (b) Silicone block model photograph showing the coil placement with jailing technique after the deployment and herniation of the Neuroform ATLAS stent (4.0 mm × 21 mm) into the aneurysm. (c) Silicone block model angiogram showing the protection of the origin of the SCA with the Neuroform ATLAS stent after the coil placement. (d) Right vertebral artery angiogram showing the coil embolization of the wide-neck, basilar-SCA aneurysm with a protected, patent SCA branch.

Discussion

The pre-procedural simulations using the patient-specific vascular models are useful to test the new device or to predict the behavior of the devices in challenging cases.^4,5 In our cases, the simulations assisted our development of a treatment plan for each complex aneurysm case by allowing an understanding of what kind of stent type to deploy, what stent size to choose, and the behavior of microcatheters and coils after stenting. In both cases, we successfully treated the aneurysms according to the plan optimized by the findings obtained in the simulations, and observed no significant difference in the behavior of the devices between our simulation model and in practice.

In the simulations, we tested different types of stent and evaluated the degree of expansion of the stent in the aneurysm and the ability to protect a side branch. In the first case, the open-cell Neuroform Atlas herniated into the aneurysm more than the closed-cell LVIS Jr., resulting in better protection of the side branch. After the stent deployment, the simulation showed a realistic trans-cell route of the microcatheter, and how the coil spread in the aneurysm after stenting, which saved time during the case intervention and gave the physician firm confidence.

One of the limitations of the simulation was the difference in elasticity of the vascular model compared to the human vessel. In the silicone block model, the shapes of the vessels are fixed and inelastic, even with the passage of the microguidewire and stent placement. Thus, it is difficult to represent the vascular elasticity. Furthermore, silicone possesses a higher coefficient of friction relative to arterial vasculature. However, we observed no differences in stent or coil manipulation related to friction. This could be because of our newly developed technique to smooth the vascular models.³ Finally, the stents are expensive and using them on simulations can increase cost. However, we were able to remove the stents from the simulated models after deployment and return to the manufacturers for repackaging and subsequent testing. Ideally, medical insurance would cover the cost of patient-specific vascular modeling.

This is the first report of the utilization of 3D patient-specific vascular models for preoperative simulations of stent-assisted coiling for treatment of complex cerebral aneurysms. The results indicate the usefulness of patient-specific vascular models for preoperative simulation and are consistent with the principles of precision medicine in stroke and cerebrovascular disease.⁶

Conclusion

Stent-assisted coiling has been efficient but difficult to predict the behavior of the devices based on the 2D screen images. Pre-procedural simulation using patient-specific vascular models can provide invaluable information regarding device selection and device behavior with a goal of reducing intra-procedural complications.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship and/or publication of this article: ST has been a consultant for Cerenovus, Medtronic, and Stryker. GC has been a consultant for Medtronic, Microvention-Terumo, and Stryker. RJ and GRD have been consultants for Medtronic. The other authors have no personal or financial interest in any of the materials or devices described in this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by a UCLA Cardiovascular Theme Discovery Award (grant number CVTDA-0001-2016) to JH.