Article Text


Original research
A single-center experience in the endovascular treatment of carotid siphon aneurysms using the Willis covered stent: a retrospective analysis
  1. Lin Ma,
  2. Ji-Chong Xu,
  3. Shuo Yan,
  4. Hao Feng,
  5. Hong-Jie Han,
  6. Hua-Qiao Tan,
  7. Chun Fang,
  8. Hong-En Huang
  1. Department of Interventional Radiology, Shanghai Tongji Hospital, Tongji University School of Medicine, Shanghai, China
  1. Correspondence to Dr Chun Fang, Department of Interventional Radiology, Shanghai Tongji Hospital, Tongji University School of Medicine, Shanghai 200065, China; 694053271{at}


Objective To report the clinical results and initial clinical experience of endovascular isolation with the Willis covered stent for carotid siphon aneurysms.

Methods Between November 2013 and December 2016, a total of 57 patients who presented with carotid siphon aneurysms were treated with the Willis covered stent. Results of the procedures, technical events, and complications were recorded. Clinical and imaging follow-ups were performed at 3 months following the endovascular procedures.

Results Placement of the Willis covered stent was successful in all patients. Immediate angiography revealed complete exclusion of aneurysms in 48 patients (84%), while endoleak occurred in nine patients (16%). Procedure-related complications occurred in three cases, including displacement of the covered stent in one patient, acute in-stent thrombosis in one patient, and microwire-related intracranial hemorrhage in one patient. Angiographic follow-ups were done in 49 patients, with complete exclusion of aneurysms in 47 patients. Endoleak was present in two patients. No aneurysm recurrence occurred. Forty-four patients showed good parent artery patency, while the other five patients showed mild to moderate asymptomatic in-stent stenosis. During the follow-up period, no ischemic or hemorrhagic event occurred. The modified Rankin Scale scores at follow-up were 0–2 in 56 patients and >2 in one patient.

Conclusions The treatment of siphon aneurysms with Willis covered stent implantation resulted in satisfactory clinical outcomes. The Willis covered stent seems safe and feasible for the treatment of siphon aneurysms, which still needs to be confirmed by longer follow-up periods and controlled studies with larger samples.

  • Willis covered stent
  • angioplasty
  • carotid siphon aneurysm
  • angiography

Statistics from


The term ‘carotid siphon’ in this study refers to the penetrating segment of the internal carotid artery (ICA), which includes the C3 (lacerum), C4 (cavernous), C5 (clinoid), and C6 (ophthalmic) segments of the ICA, as described by Bouthiller et al.1 Carotid siphon aneurysms are considered to be more suitable for endovascular therapy because of the anatomic structures adjacent to the segments of the ICA (such as the anterior clinoid process, cavernous sinus, and optical apparatus), which could be more challenging for microsurgical treatment.2 After endovascular isolation with the Willis covered stent had been approved for intracranial aneurysms in China (MicroPort, Shanghai, China), preliminary results with this stent were reported, indicating good flexibility and efficacy in the treatment of cranial ICA aneurysms including distal ICA aneurysms, recurrent intracranial aneurysms after coiling, large fusiform carotid aneurysms, and large or giant intracranial aneurysms.3–7 However, data on the safety and efficacy profiles regarding carotid siphon aneurysms treated with the Willis covered stent remain scarce. In this study we report our experience and the efficacy of the Willis covered stent in a retrospective analysis of 57 carotid siphon aneurysms treated with the stent in our institution.

Materials and methods

Between November 2013 and December 2016, 57 endovascular isolation treatments for carotid siphon aneurysms were performed in our institution, including unruptured aneurysms in 27 patients and ruptured aneurysms in 30 patients. Of the 57 patients, 26 were male and 31 were female and their mean age was 67 years (range 19–73). Aneurysm morphology included saccular aneurysms, pseudoaneurysms, and dissection aneurysms. The size of the aneurysms was >10 mm in 24 cases, 3–10 mm in 16 cases, and <3 mm in 17 cases. The demographic and clinical data of the 57 patients are summarized in table 1.

Table 1

Baseline characteristics of the 57 patients with carotid siphon aneurysms

The clinical manifestations were subarachnoid hemorrhage (SAH) in 22 cases (ruptured aneurysm in 19, pituitary operation-related hemorrhage in 3), including Hunt and Hess grade I in eight cases, grade II in 13, and grade III in one; epistaxis in eight cases (nasopharyngeal carcinoma radiotherapy in 4, traumatic injury in 2, paranasal sinus operation-related hemorrhage in 2); headache in 12 cases; visual defect and/or diplopia in nine cases, and pituitary endocrine abnormalities in one case.

Endovascular procedures

According to the condition of the aneurysms, several different methods of endovascular isolation treatment were performed including single covered stent implantation, double covered stent implantation, and covered stent plus coiling, which are summarized in table 2. In principle, a single covered stent would be used first, and other techniques would be used if problems arose. For example, if the width of the aneurysm neck was >12 mm or the aneurysms were at the curved section of the ICA, we would consider using a double covered stent or a covered stent plus coils in order to decrease the incidence of endoleak and increase the rate of occlusion.

Table 2

Summary of treatment with the Willis covered stent in 57 patients

Endovascular treatment was performed under general anesthesia. A 6 Fr long sheath (Cook, Bloomington, USA) or 8 Fr guiding catheter (Envoy 90 cm, Codman Neurovascular, USA) was initially positioned into the C1 segment of the ICA via the right femoral approach. A 6 Fr Neuron (Penumbra, Alameda, California, USA) or Navien (Ev3/Covidien, USA) support catheter was advanced approximately to the C4 segment of the ICA. A 0.014 inch diameter microguidewire (Transcend, Boston Scientific, USA) was navigated into the distal segment of the parent artery. The selected Willis covered stent was advanced over the microguidewire and bridged the aneurysm orifice. In certain cases the support catheter (Neuron or Navien) was navigated across the lesion segment using the coaxial technique. When the Willis covered stent was placed in the correct position, the support catheter was pulled back to unsheath the covered stent.

The Willis covered stent is available in various diameters (3.5–4.5 mm) and lengths (7–16 mm). The stent must extend at least 2 mm beyond the neck of the aneurysm on both sides. The diameter of the stent needs to be approximately 0.5 mm wider than the diameter of the parent artery.

Multiple control angiograms were obtained during the procedure to confirm the position of the stent and to avoid covering any important side branch. The stent was then deployed with a pressure of 5–6 atmos. If the aneurysm was not completely covered and rapid filling of the aneurysm cavity was observed, the proximal and distal part of the stent was re-inflated to eliminate the endoleak. If the endoleak persisted, another covered stent could be employed. For several large and giant aneurysms, before the covered stent was deployed, coils were placed in the aneurysm cavity through another microcatheter.

Unenhanced brain CT and neurological examinations were performed routinely after the procedure to exclude any intracranial hemorrhage or ischemic event. Brain CT or MRI was also performed when new neurological symptoms occurred.

Antithrombotic regime

Patients with unruptured aneurysms were pretreated with daily doses of 100 mg aspirin and 75 mg clopidogrel for >3 days in preparation for the covered stent technique. For ruptured aneurysms, before the procedure a loading dose of 300 mg clopidogrel and 300 mg aspirin was administered through a nasogastric tube. All patients received systemic intravenous heparin during the procedure to maintain an activated clotting time of 250–300 s. Tirofiban was available for intra-arterial application in case of acute thrombosis or thromboembolism.

After the procedure, low molecular weight heparin was subcutaneously injected at a dosage of 4000 U/12 hours for 3 days, and thereafter all patients continued with dual antiplatelet therapy (clopidogrel 75 mg/day and aspirin 100 mg/day) for a minimum of 6 months. Aspirin (100 mg/day) was often required for at least 2 years.

Clinical and imaging follow-up

After discharge from hospital, patients with any deterioration in neurological symptoms were required to come back to the hospital and a brain CT or MRI was performed if necessary. Angiographic and clinical follow-ups were performed approximately 3 months after the endovascular procedure. The modified Rankin Scale (mRS) was used to evaluate the clinical state of the patients. The status of the aneurysms was assessed by digital subtraction angiography (DSA) to exclude the possibility of residual endoleak as well as aneurysm regrowth and in-stent stenosis.


Primary procedural results

Demographic data and clinical and angiographic findings of the study are summarized in Table 1. The navigation and deployment of the stent was technically successful in all cases, despite extremely sharp curves (figures 1 and 2). During the navigation process, stent graft damage and stent abscission occurred in two cases. As another technique, a support catheter was advanced across the lesion and the new covered stent was then safely navigated in the parent artery (figure 3).

Figure 1

Carotid siphon aneurysm treated with the Willis covered stent. (A) Pre-procedural digital subtraction angiography image shows a wide-necked giant aneurysm on the anterior wall of the left C6 segment. (B) The Navien support catheter was positioned in the C4 segment of the internal carotid artery by the coaxial technique. The Willis covered stent (4.0×16 mm) was successfully deployed with proximal and distal ends of the stent covering the neck of the aneurysm (4 mm) on both sides (arrows). (C) Cerebral angiogram immediately after stent placement demonstrates complete occlusion of the aneurysm with patency of the parent artery.

Figure 2

Carotid siphon aneurysm treated with the Willis covered stent plus coils. (A) Pre-procedural three-dimensional rotational view shows a wide-necked large aneurysm on the C4 segment of the internal carotid artery. (B) Cerebral angiogram immediately after stent-assisted coil embolization shows near total occlusion of the aneurysm with mild endoleak. (C) The Willis covered stent can be clearly seen in the plain film after stent placement (arrows). (D) Follow-up angiogram at 12 months shows complete disappearance of the aneurysm with patency of the parent artery.

Figure 3

Navigation of the Willis covered stent using a support catheter. (A) The pre-procedural digital subtraction angiography image shows a dissecting aneurysm on the C6 segment of the internal carotid artery (ICA). (B) The Navien support catheter was positioned in the C4 segment of the ICA. (C) The damage of stent graft during the navigation process (arrow). (D, E) The support catheter can reach as far as the C7 segment of the ICA with coronary artery balloon assisted technology. (F) The Willis covered stent was successfully passed through the support catheter and precisely bridged the location of the aneurysm (arrows). (G) Balloon dilation and stent implantation (arrows). (H) Cerebral angiogram immediately after stent placement showing good patency of the parent artery.

Immediate angiography revealed complete exclusion of the aneurysms in 48 patients (84%) and endoleak in nine patients (16%), including mild residual filling of the aneurysms in seven patients (intraluminal retention of contrast media). Demographic data, as well as clinical and angiographic findings of the study, are summarized in Tables 1,2.

Procedure-related complications occurred in three cases, including displacement of the covered stent in one patient (after deployment, balloon retraction leading to stent movement resulting in endoleak), acute in-stent thrombosis in one patient (intra-arterial application of 8 mL tirofiban, then thrombolysis, no symptoms transpired), and microwire-related intracranial hemorrhage (wire perforation during navigation across the aneurysm) in one patient (hemorrhage leading to death).

Immediate angiography also showed occlusion of the ophthalmic artery in six patients. Fortunately, they were uneventful after the procedure with no evidence of abnormal vision.

Angiographic and clinical follow-up results

The follow-up results are presented in table 2. The mean follow-up period was 9 months (range 6–24 months). Follow-up angiography was performed in 49 patients. Complete exclusion of aneurysms occurred in 47 patients. Endoleak only occurred in two patients. Aneurysm recurrence was not observed. Forty-four patients showed good parent artery patency, while the other five patients showed mild to moderate asymptomatic in-stent stenosis. During the follow-up period, no ischemic or hemorrhagic event was reported by any of the patients. The mRS score at follow-up was 0–2 in 56 patients and >2 in one patient. The patients with occlusion of the ophthalmic artery showed no related clinical symptoms.


Flow diverter devices such as SILK and Pipeline are intended to reconstruct the parent artery, while blood flow is diverted from the aneurysm resulting in disruption of flow and stasis of blood, followed by thrombosis.8 9 The Willis covered stent is a different flow diverter device which consists of three parts: a bare stent, an expandable polytetrafluoroethylene membrane, and a balloon catheter. Reconstruction is supposed to occur in such a way that the aneurysm is isolated from the flow and intra-aneurysmal thrombosis forms progressively. Our results proved that the Willis covered stent is safe and efficient for the treatment of ICA siphon aneurysms, including both ruptured and unruptured aneurysms.

Isolation of the aneurysm with the covered stent can lead to immediate exclusion of the aneurysm from the circulation without procedural manipulations in the aneurysm sac, which reduces the risk of procedure-related rupture or rebleeding, especially in small aneurysms and blood blister-like aneurysms.7 In addition, no or few embolization materials are deployed into the aneurysm sac, so the procedure causes no mass effects and does not impact aneurysm shrinkage after complete exclusion.10 Despite these advantages, the performance of the Willis covered stent requires further investigation for aneurysms of the cavernous, clinoid, and superclinoid segments.

Navigation of the stent

The Willis covered stent was improved with respect to the stent structure, membrane thickness, and the stent delivery system for intracranial application.3 4 However, the navigation still needs attention, especially in the deep curves of the intracranial vasculature. Although the Neuron and Navien support catheters showed excellent compliance performance and thus could reach as far as the C4 segment of the ICA. Only at this high level, that could provide adequate support for the delivery of the stent.11

The complex tortuous anatomy of the siphon segment of the ICA often makes it difficult to navigate the covered stent and may damage the parent artery. Lee et al reported three patients with blood blister-like aneurysms treated with Jostent, with aneurysm rupture occurring in one patient.12 In addition, it also tends to easily damage the graft of the covered stent during the delivery process. In our study, this incident happened in two cases.

Therefore, in certain cases, the support catheter (Neuron or Navien) was navigated across the lesion segment with the coaxial technique; the delivery channel established with the support catheter more effectively guarantees the safe navigation of the covered stent. This is the first report of the use of this technique for intracranial delivery of a covered stent.

Post-procedure endoleak

Although isolation of the covered stent had a high rate of immediate aneurysm exclusion, endoleak was still a non-negligible issue. The potential causes of endoleak after deployment of the Willis covered stent include the non-homogeneous lumen of the blood vessel, a tear in the graft, incomplete coverage of the aneurysm orifice, retrograde flow from a collateral vessel, and aneurysms located at an acutely angled vessel segment.10 13 According to the source of blood flow causing the leak, endoleaks of intracranial aneurysms in these cases are basically classified into types I to IV.14 15 Type I endoleak, resulting from blood flow that originates from the proximal or distal endograft attachment sites, is the most common type.

In general, the proximal and distal parts of the stent can be re-inflated to eliminate the endoleak. However, due to the severe tortuosity of the paraclinoid segment, balloon post-dilation could cause displacement of the deployed stent. In this study, covered stent displacement occurred in one case because of re-inflation for exclusion of an endoleak, which still existed during follow-up. We should therefore be very careful to re-inflate the stent, particularly in a curved parent artery.

After deployment of the covered stent, if immediate angiography demonstrates only slow and slight filling of the aneurysm, balloon post-dilation should be avoided and further observation could be a better choice. In our study, mild endoleak in seven patients gradually disappeared during follow-up.

Another strategy that can be used to improve the prognosis of patients with endoleak is the combined technique of covered stent and coiling which may promote thrombosis in the aneurysm sac. For large aneurysms with a wide neck and involving a curved segment, we could partially embolize the aneurysm sac with coils before deployment of the covered stent, especially in patients with ultra-wide necked aneurysms who need double covered stents. This method has never been previously reported. During follow-up, satisfactory results were achieved in all patients using this combined technique in our group summarized in Table 3.

Table 3

Endovascular treatment, outcome, and follow-up data for nine patients with endoleak

The incidence and outcome of endoleak in our study was superior to previous studies. The incidence and residual rates were 16% and 4% versus 30.8% and 12.8%.16 The discrepancy could be related to device and technique improvement.

In-stent stenosis

In-stent stenosis is almost always caused by neointimal hyperplasia. Antiplatelet therapy has been thought to play a vital role in the inhibition of in-stent neointimal hyperplasia. Long-term oral antiplatelet drugs may inhibit platelet aggregation and activate smooth muscle hyperplasia.17 It has been proven that poor adherence to dual antiplatelet therapy after the procedure is an independent predictor of in-stent stenosis.18

In this study only five patients showed asymptomatic in-stent stenosis. One possible reason could be that all patients followed stringent dual antiplatelet therapy. Besides, the patients have not yet undergone long-term follow-up. Despite this limitation of our study, long-term parent artery patency has been reported in previous studies. Tan et al reported that, at an average follow-up of 17.5 months, only 10% of patients treated with the Willis covered stent showed an asymptomatic in-stent stenosis.19 In another long-term follow-up study with the Willis covered stent, at the end of 2- and 6-year follow-ups the mean rates of in-stent stenosis were 18.0% and 29.0%, respectively.13

Antiplatelet administration

The use of antiplatelet therapy after stent implantation is controversial for patients with SAH. In the present study, patients with SAH showed immediate complete aneurysm exclusion by implantation of the covered stent, which guaranteed the safety of post-procedural dual antiplatelet therapy. Others reported that dual antiplatelet therapy should be suspended and single antiplatelet therapy should be chosen if severe endoleak occurs.20 Fortunately, our patients with mild endoleak did not show any hemorrhagic events with dual antiplatelet therapy during follow-up.


Complications occurred in three patients, including technical complications in two cases and an acute complication following the endovascular procedure in one case. The technical complications included one case of microwire intracranial arterial perforation and one case of covered stent displacement. The patient in whom guidewire intra-procedure perforation with progressive intracranial hemorrhage occurred died, so it is very important that microwire tension is well controlled during the procedure. Once perforation occurs, it would be disastrous under an anticoagulant and antithrombotic state. Covered stent displacement occurred in one case when re-inflating the stent for exclusion of endoleak. The endoleak still existed during follow-up, but no clinical symptoms appeared. Acute in-stent thrombosis following endovascular isolation occurred in one patient. The thrombus was dissolved spontaneously with systemic and intra-arterial administration of glycoprotein IIb/IIIa inhibitors.

Limitations of this study

This study had three limitations. First, the patients did not undergo long-term follow-up and thus some complications, such as endoleak regrowth and in-stent stenosis, cannot be adequately evaluated. Second, the number of patients is relatively small,so larger clinical trials are required to evaluate the outcomes. Third, this study lacked a control group receiving other treatment such as a flow diverter device or coil insertion, which limits the ability to provide robust conclusions.


The treatment of patients with siphon aneurysms by Willis covered stent implantation resulted in satisfactory clinical outcomes. Thus, use of the Willis covered stent seems safe and feasible for siphon aneurysms treatment, which needs to be confirmed by longer follow-ups and controlled studies with larger samples.


View Abstract


  • LM and J-CX contributed equally.

  • Contributors All authors were involved in planning, conducting and reporting the work. Conception and design: LM, J-CX and CF. Operation procedure: H-QT, H-JH, HF, J-CX and SY. Acquisition, analysis and interpretation of data: LM J-CX and CF. Manuscript preparation: LM and J-CX. Manuscript review and critique: H-QT and CF. HEH operation procedure and acquisition, analysis and interpretation of the data.

  • Funding This work was supported by the Project from Shanghai Science and Technology Commission (14DZ1941205), the Project from Shanghai Municipal Commission of Health and Family Planning (20174Y0095), Priority of Shanghai Key Discipline of Medicine (2017ZZ02020).

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval This study was approved by the ethics committee of Shanghai Tongji Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement For any questions regarding data sharing please contact the corresponding author.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.