Introduction

The complex region of the anterior communicating artery (ACoA) is one of the most common sites of cerebral aneurysms.1 ACoA aneurysms have a higher risk of rupture at smaller sizes than aneurysms at other sites,2 3 and in patients with multiple aneurysms, the ACoA aneurysms have higher rupture rates than aneurysms at other sites.4 Because of this, treatment of unruptured ACoA aneurysms is usually indicated. Due to the anatomical variability of this region, a microsurgical approach is challenging,5 and so endovascular treatment becomes the first option.6–9 However, coiling has a higher risk of recurrence compared with clipping,7 and so flow diversion represents a feasible type of treatment for ACoA aneurysms10–15 and for recurrence after an initial coiling approach.

We present a large single center series of flow diversion treatments of ACoA aneurysms with several types of stents (pipeline embolization device (PED; Covidien/ev3, Irvine, California, USA); flow redirection endoluminal device (FRED Jr; MicroVention, Tustin, California, USA); Silk (Balt Extrusion, Montmorency, France); Surpass flow diverter (FD) (Stryker Neurovascular)),16 and propose a new classification of this anatomical region to improve treatment of these aneurysms.

Methods

Our institutional review board approved this retrospective study. Individual patient consent was not required as data were collected in an anonymized manner and there was no risk to patients. Two investigators independently reviewed and identified all ACoA aneurysms treated by any FD stents between April 2014 and November 2018 and characterized the type of anatomy of the region. A third investigator reviewed cases where there was inconsistency. Data collection included the following: demographics, aneurysm characteristics, clinical presentation, follow-up, and clinical outcome.

To analyze the best strategy for the initial treatment of ACoA aneurysms with FD stents, we propose a new classification of this region based on comparing the diameters of both A1 segments. This may help to make strategic decisions for stent deployment, to predict rates of thromboembolic complications, and to improve long term results. Here we describe four types of A1 configuration (figure 1): H1, both A1 segments have the same diameter; H2, A1 segments have different diameters, with <50% difference between the sides; H3, A1 segments have different diameters, with ≥50% difference between the sides; and Y type, no A1 segment on one side. All patients treated with an FD stent in our service received daily dual antiplatelet medication with aspirin (160 mg daily) and ticagrelor 90 mg twice daily; all patients treated before 2015 received aspirin (160 mg daily) and clopidogrel (75 mg daily). The treatment occurred under general anesthesia and systemic heparinization was via a transfemoral approach; using a long sheath (Neuron Max 088; Penumbra, Alameda, California, USA), a 6 F guiding catheter (FargoMax; Balt, Montmorency, France or Navien, Covidien) was advanced into the intracranial internal carotid artery. Vessel and aneurysm features were analyzed via biplane and three-dimensional rotational acquisition angiography. Depending on the type of FD stent used, a microcatheter measuring 0.027 or 0.021 inches was used according to techniques previously described.15 17–19

For the stent placement strategy, there are three techniques: the FD stent is placed on the same side (A1 segment of one side to ipsilateral A2), on the side with increased flow relative to the aneurysm neck (the ‘ipsilateral technique’ or ‘I technique’); the ‘H technique’,13 when two FD stents are placed in each side; and the ’Chicane technique’, where the FD stent is crossed from the A1 segment of one side to the contralateral A2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Anatomical classification of A1 segments for flow diversion treatment. H1, both A1 segments have the same diameter; H2, A1 segments have different diameters but there is <50% difference between the sides; H3, A1 segments have different diameters but there is ≥50% difference between the sides; Y, no A1 segment on one side. AcoA, anterior communicating artery.

Results

We reviewed data from 30 patients who received FD stents for ACoA aneurysms. Characteristics included: average age 52 years (range 32–69 years); 17 women and 13 men; and average aneurysm size 5.5 mm20 (range 2.1–15 mm). Twenty patients had previously been treated: 16 for ruptured aneurysms treated with coil embolization in the acute phase and 4 for unruptured aneurysms treated with a Medina embolization device (Covidien/eV3) (2 patients) or a Woven EndoBridge device (WEB; Sequent Medical, Aliso Viejo, California) (2 patients). Procedures were successful in all 30 cases (100%).

Twenty-two patients (73.3%) experienced no procedure related complications; eight patients experienced thromboembolic complications. Of these, one patient experienced clinical repercussions due to an ischemic lesion with a modified Rankin Scale score of 1 (patient No 19). Of the seven other thromboembolic complications without clinical repercussions, one was caused by a twist of the Silk stent, which was resolved by catheterization with the Echelon 10 microcatheter and after angioplasty with a 4×10 mm Scepter balloon (patient No 7). In another patient, the Silk stent did not open appropriately and a 4×20 mm Solitaire AB was deployed inside the FD stent after in situ injection of abciximab and nimodipine (patient No 22). In one patient, the FRED Jr was difficult to deploy at the left A1 segment to the left A2 segment. Stent migration occurred, with clot formation that was resolved by in situ injection of abciximab. At the 6 month DSA follow-up, the left A1 segment was occluded and both A2 segments were opacified by the right A1 segment. The four other procedure related complications were caused by clot formation, resolved by the in situ injection of abciximab. The global rate of thromboembolic complications in our case series was 26.7%, with 3.3% of patients experiencing symptomatic ischemic stroke. No hemorrhagic complications were observed. All complications were thromboembolic, three caused by difficulties with stent deployment and five by clot formation in conjunction with adequate stent apposition.

Follow-up angiography (table 1) was available for 23/30 patients (76.6%) (range 6–42; mean 19 months). Total aneurysm occlusion occurred in 17/23 patients (73.9%), neck remnant in 3/23 patients (13%), and aneurysm remnant in 3/23 patients (13%). The distribution of each type of FD stent in our group, angiographic follow-up, total aneurysm occlusion, neck remnants, and aneurysm remnants rates are shown in table 2.

Classifying aneurysms using this new proposed anatomical classification, we found that 17/30 aneurysms were H1 (56.6%), 8/30 were H2 (26.6%), 5/30 were H3 (16.6%), and none were Y type.

For all 23 patients with angiographic control, the number of patients with total occlusion by each anatomical type was as follows: H1, 10/12 (83.3%); H2, 4/6 (66.7%); and H3, 3/5 (60%). Analyzing all thromboembolic complications according to each anatomical classification revealed complications in 3/17 procedures (17.6%) in the H1 group, 2/8 procedures (25%) in the H2 group, and 3/5 procedures (60%) in the H3 group. When we excluded all technical issues of stent deployment and analyzed only the thromboembolic complications not caused by issues with stent deployment, we found rates of 12.5%, 14.3%, and 50% in the H1, H2, and H3 groups, respectively. Although not statistically significant, the H3 group had four times higher rates of thromboembolic complications (P=0.287), and 1.38 times lower rates of total occlusion than the H1 group (P=0.258).

Discussion

This is a case series of FD treatment of ACoA aneurysms (30 patients) with different types of stents. We found high rates of total occlusion (73.9%) and adequate occlusion (86.9%), with low rates of symptomatic complications (3.3%). Because of the complex anatomy of this region, there are many challenges in treating these aneurysms, not only for microsurgical clipping treatment but also for standard endovascular treatment. In our case series, the average dome to neck ratio was 1.46.21 Because of this, techniques for treating wide neck aneurysms were required. New technology for treating wide neck aneurysms (such as the WEB device)22 and new techniques (such as the balloon remodeling assisted WEB technique)23 help, but do not completely solve the problem of treating aneurysms with this type of topography.

Although it is known that ACoA aneurysms rupture more frequently and at smaller sizes than other aneurysms, the reasons for this are not well understood.12 The configuration of the A1 segment that may induce aneurysm development by hemodynamic stress has been demonstrated,24 but how this relates to flow diversion treatment is not yet fully understood. It has been reported that FD stents may not be suitable for middle cerebral artery bifurcation aneurysms due to high rates of ischemic complications (43%).25 However, some variations are seen in the current literature.26 The experience of the surgeons within our institution leads to a belief that Y type aneurysms are likely to function like middle cerebral artery bifurcation aneurysms and that the rate of thromboembolic complications would be similar. Because FD treatment is avoided as a consequence, other endovascular techniques are chosen for treatment of patients with the Y configuration27–30 ; this may explain why no patients in our series had the Y configuration.

In our study, one patient (patient No 22) had a H3 configuration before treatment, but this functioned as a Y configuration because we observed no opacification in the ACA area. Twelve month DSA follow-up after FD treatment revealed that the ACA area had become a H3 configuration, anatomically and functionally (figure 2), due to an enlargement of the hypoplastic A1 segment. This type of change between the anatomical classification before and after treatment was also seen in patient No 4 (H3 became H1) (figure 3) and has previously been described by Colby et al.13 In patient No 4, the first follow-up was carried out only in the carotid artery with the larger A1 segment, and revealed total occlusion. The second follow-up included both arteries and revealed an aneurysm recanalization. The patient was retreated with the coiling technique, through the hypoplastic A1 that had become enlarged after FD treatment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Pipeline embolization device (PED) treatment of a recanalization of a previously coiled ruptured anterior communicating artery (ACoA) aneurysm (patient No 22). (A) Right internal carotid artery (ICA) DSA before flow diverter (FD) treatment. Left corner shows the position of the PED from the right A1 segment to the left A2 (the ’Chicane technique'). (B) Left ICA DSA before FD treatment showing a H3 type (>50% difference between A1 segment diameters) that functioned as a Y type (no filling of the ACA territory). (C) Late phase left ICA DSA before treatment showing no perfusion of the anterior cerebral artery (ACA) territory.  (D) Twelve month follow-up. Right ICA DSA showing filling of the left ACA territory by the right A1 segment (PED in this topography). (E) Twelve month follow-up. Left ICA DSA showing the left A1 segment (H3 type anatomically and functionally) filling the right ACA territory crossing the FD. (F) Twelve month follow-up, late arterial phase left ICA DSA, showing all the right ACA territory being supplied by the left A1 segment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Silk treatment of a recanalization of a ruptured anterior communicating artery (ACoA) aneurysm previously coiled (patient No 4), with incomplete wall apposition and a Solitaire AB inside the flow diverter (FD). (A) Right and left internal carotid artery (ICA) DSA before FD and coiling treatment showing the H3 configuration of the right A1 segment. (B) Three month follow-up. Left ICA three-dimensional DSA showing recanalization. (C) Left ICA DSA, showing FD position (left A1 segment to left A2) with distal occlusion of the FD. Right corner: VasoCT acquisition showing Silk stent with the Solitaire AB inside the FD. (D) Left ICA DSA showing FD treatment. (E) Twenty-four month follow-up after FD treatment (recanalization) and the flow change of the right A1 segment (H1 type). Left corner: coiling treatment through the right A1 segment. (F) Thirty month follow-up. Right ICA DSA. H1 configuration of the right A1 segment with total occlusion of the aneurysm. (G) Thirty month follow-up. Left ICA DSA.

The ‘I technique’ was used in all patients except one in each group. In the H1 group, the ‘H technique’13 (patient No 18) was used and the ‘Chicane technique’ was used in the other two patients (H2 group (patient No 23) and H3 group (patient No 22) (figure 2).

These numbers raise some questions. Why does the H3 configuration have the lowest rates of total occlusion (60% vs 83% in the H1 group) and the highest rates of thromboembolic complications (50% vs 12.5% in the H1 group)? The thromboembolic complications could be explained by the similarities between the flow competition observed in the H3 and Y anatomical configurations and middle cerebral artery bifurcation aneurysms; it is possible that FD treatment should also be avoided in cases of H3 anatomical configuration. Another important point is that in the H1 group, the ‘Chicane technique’ was not used, but a high rate of total occlusion (83.3%) occurred with the ‘I technique’. It may be that the change in flow alone is sufficient to induce aneurysm thrombosis. This effect is not seen in the H3 group, in which the flow change would be higher because of the difference in A1 segment diameters. The high total occlusion rates achieved with the ‘I technique’ in the H1 group suggest we can avoid the use of more complex techniques, such as the ‘Chicane technique’, which can cover small branches of the ACoA, and the ‘H technique’, which requires two stage treatment,13 although Colby et al reported low complication rates with this technique. These techniques can therefore be reserved for patients with failure during follow-up of the ‘I technique’. More studies are needed to confirm this hypothesis.

The main limitations of this study are its retrospective, non-randomized nature and the small number of patients. However, it appears there is a high chance of total occlusion of H1 type ACoA aneurysms using the ‘I technique’ for FD stent deployment. Another important point is the need to assess the two carotid arteries, even if the contralateral A1 segment is hypoplastic before the treatment. This is because there is a high possibility of flow change following flow diversion treatment and, subsequently, of enlargement of the A1 segment.

Conclusion

The treatment of ACoA aneurysms with FD stents is feasible, safe, and effective as firstline treatment or in recanalization cases (coils, WEB devices, or Medina embolization device). The system we propose, classifying the ACoA complex into four anatomical types, may facilitate decision making in FD stent deployment, prediction of thromboembolic complications, and may lead to improved long term outcomes. The most suitable anatomical configuration for FD treatment seems to be the H1 configuration with the ‘I technique’. It is possible that the H3 configuration is not a good indication for FD treatment. However, further studies are needed to evaluate the feasibility of this anatomical classification and the reproducibility of our findings regarding complication rates and long term follow-up in the FD treatment of these aneurysms.