This study is a retrospective multicenter analysis retrieving data of endovascular treatment of six different tertiary referral centers for cerebrovascular diseases in a timespan between January 2015 through December 2020.

We included patients $>$18 year-old with complete radiological and clinical reports and close follow-up. We excluded patients secondarily treated with endovascular devices after the failure of a previous clipping. We report on the pure radiological technical aspects of the procedures, not taking into account relevant clinical data. Patients gave written informed consent to use their medical and radiological data. The Institutional Review Board (AOSPFe EC) approved this study [EC 1254/M/2017].

We collected for each patient: (1) demographics: age, gender; (2) neurological status at admission; (3) complications; (4) radiological outcome measured as presence or absence of subarachnoid hemorrhage (at CT-scan); aneurysm size and morphology; neck width; fetal-type PComA; treatment type; occlusion rate at post-operative imaging and follow-up (6–12–24 months); need for second re-treatments. In particular, we considered fetal-type a PComA/P1 ratio $>$1 with hemodynamic prevalence in the posterior cerebral artery from the anterior circulation (meaning the posterior cerebral artery territory filled from internal carotid artery injections) [18]. Angiographic features were evaluated by cerebral and rotational angiography with three-dimensional (3D) image reconstruction.

The treatment strategy was based on a multidisciplinary consensus between vascular neurosurgeons and interventional neuroradiologists in all included patients. All endovascular procedures were performed under general anesthesia and via a femoral artery approach. Simple coiling, stent-assisted coiling, flow-diversion, flow-diversion plus coiling and intravascular devices deployment procedures were performed according to single-center devices availability and protocols. Antiplatelets/anticoagulant drugs administered during the procedure and pre-and post-operatively were managed according to single-center protocols.

MR angiography (MRa) or digital subtraction angiography (DSA) were performed at single or multiple follow-up times: 6, 12 and 24 months. The aneurysm occlusion rate was defined according to the Raymond-Roy (RR) classification incomplete occlusion (class I), residual neck (class II), and incomplete occlusion or residual aneurysm (class III) [19].

Kruskal–Wallis one-way analysis of variance and Chi-Squared Test were used to assess differences among different endovascular treatments with respect to aneurysms’ characteristics and treatment outcomes. To detect all pairs of a group that show significant differences, significant associations at the above omnibus tests were further analyzed through Dunn’s posthoc comparisons and multinomial Generalized Linear Models with posthoc Bonferroni’s correction, respectively. Contingency tables were also used to evaluate the association between radiological outcomes and aneurysms morphology and treatment. Independent samples Student’s T-Test and Chi-Squared Test were used to analyze the differences in continuous and categorical variables among fetal and non-fetal PComA type. The significance level was set at 0.05. Statistical analysis was performed using JASP 0.13.1.0 - Chocolatey.org.

Demographics and aneurysms’ characteristics are summarized in Table 1. After institutional databases screening, 250 patients treated for PComA aneurysms with different endovascular techniques were included. Mean age was 60 $\pm$ 14 years (range: 22–92 years), 217 were female (87%), and 140 (56%) of the included patients presented with SAH due to PComA aneurysm rupture at the time of hospitalization.

The great majority of aneurysms were saccular (235/250; 94%) with a mean dome and neck size of 9 mm and 4 mm, respectively. However, ruptured aneurysms had significantly smaller maximum size. On average, ruptured aneurysms had a significantly smaller maximum diameter and a narrower neck than unruptured ones.

Simple coiling was performed in 171 patients (68%), whereas flow-diverter stenting was performed in 32 (13%). Other treatments, such as aneurysm remodeling, intravascular devices placement, stent-assisted coiling, and combination of coiling and flow-diverter stenting, were overall performed in 47 patients (19%) (Tables 2,3).

The choice of the treatment modality significantly differed according to aneurysm presentation (ruptured/unruptured) and aneurysm angioarchitecture features (Table 2). In particular, among the 110 patients with unruptured aneurysms, coiling was performed in 54 (49%), flow diversion in 28 (25%), balloon remodeling in 10 (10%), and other treatments in 18 (16%) (Tables 2,3). Among the 140 patients with ruptured aneurysms, coiling was performed in 117 (83%), remodeling in 14 (10%), flow diversion in 4 (3%), and other treatments in 5 (3%) (Table 2).

Moreover, coiling was preferred in the majority of ruptured aneurysms (84%) and, regardless of the ruptured status, in the case of aneurysms of small size (mean 8.4 mm) and narrow neck (mean 3.4 mm). Contrariwise, the use of flow-diverters was mainly preferred in patients harboring unruptured aneurysms or, in general, aneurysms of larger size (mean 12.9 mm) and wider neck (mean 5 mm) (Tables 2,3,4). Tables 4 and 5 show significant differences between the two treatment types at the posthoc analysis of those quantitative or qualitative variables, which showed a significant difference at omnibus tests.

At immediate post-procedural imaging (MRa and/or DSA), a complete occlusion was reported in 163 patients (65%), a near-complete occlusion in 43 (17%), and an incomplete occlusion in 44 (18%) (Table 2). A 6-months radiological follow-up was available in 245 (98%) patients (5 patients lost at follow up) showing complete occlusion in 148 (61%), near-complete occlusion in 74 (30%), and incomplete in 23 (9%) (Table 2).

A delayed improvement of occlusion rate in those patients with non-complete occlusion at post-procedural imaging (87 cases) was observed in 40 cases. In contrast, aneurysm recanalization during follow-up was observed in 51 patients (Tables 2,3).

A long-term radiological follow-up (24 months) was available in 105 patients (145 patients lost at follow-up). It showed complete occlusion in 64 (61%) cases, near-complete occlusion in 32 (30%) and incomplete in 9 (9%) (Table 2). However, at the last follow-up, the occlusion rate was stable in 149 (60%) of cases, improved in 49 (19%) and worsened in 51 (21%) (Table 3). The mean time of follow-up was 11 months, while the median follow-up time was 12 months.

Results according to aneurysm presentation are summarized in Table 2.

Complete post-procedural occlusion was obtained in 103 (74%) ruptured and 60 (55%) unruptured aneurysms. At early post-procedural control, near-complete exclusion was seen in 30 (21%) ruptured and 13 (12%) unruptured aneurysms, while an incomplete exclusion in 7 (5%) ruptured and 37 (34%) unruptured, respectively. This difference appeared significant (p 0.001).

However, 33 (30%) unruptured aneurysms showed an occlusion rate improvement over time, showing a complete occlusion at last follow-up in 69 (63%) cases, and a near-complete exclusion in 31 (28%) patients. On the other hand, a delayed improvement was less frequent in ruptured aneurysms, which showed a complete occlusion rate of 71% and a near-complete exclusion rate of 24% at the last follow-up. Therefore, no significant difference in exclusion rate was seen between ruptured and unruptured aneurysms at the last follow-up (p = 0.4). Recanalization during follow-up was documented in 34 (24%) ruptured and 17 (16%) unruptured aneurysms, respectively. However, this difference was not significant (p = 0.1).

A subgroup analysis comparing ruptured and unruptured aneurysms showed that 47/54 patients (87%) with unruptured PComA aneurysms undergoing coiling showed a complete occlusion at post-procedural imaging, compared with 88/117 (75%) in the ruptured group. A near-complete occlusion after coiling was obtained in unruptured 7 (13%) and 27 (23%) ruptured aneurysms, respectively. An incomplete occlusion was observed only in 2 cases treated by simple coiling of the ruptured group. Nonetheless, these differences in occlusion rate were not statistically significant (p = 0.1). The frequency of recanalization after simple coiling was similar between the unruptured and ruptured groups, with 14 (26%) unruptured and 32 (27%) ruptured aneurysms recanalized during follow-up.

Overall, 3 patients with ruptured aneurysms died in the early post-procedural period for complications not related to endovascular treatment (2 sudden cardiac deaths and 1 respiratory failure). Intraprocedural PComA thrombosis was the most frequent complication reported in 23 patients (9%). All of them were immediately pharmacologically treated, obtaining recanalization. Transient ischemic attacks (TIA) or strokes with no evidence of intraoperative PComA thrombosis were instead reported in 6 cases (2%). Two cases of intraprocedural aneurysm rupture during coiling occurred, although no major neurological sequelae were registered.

Thirty-eight aneurysms (15%) with incomplete occlusion or recanalization underwent 41 secondary procedures as 3 patients needed 2 re-treatments. In particular, a flow-diverter stent was positioned in 17 (41%) cases, a stent-assisted recoiling in 12 (29%), and a simple recoiling in 6 (14%). Microsurgical clipping and aneurysm remodeling were preferred in 4 (10%) and 2 (5%) of recanalization, respectively (Tables 2,3).

A complete occlusion at first treatment was significantly more common in small (p 0.001) and ruptured (p 0.001) aneurysms. A 6-months complete occlusion was mostly observed in patients who underwent coiling (135/163; 79%) compared with other endovascular treatments (28 cases, 80%) (p 0.001) (Tables 2,5). At long-time radiological follow-up instead, on the one hand, we observed a significant improvement of the occlusion rate (20/49 out of cases) in patients treated with flow-diverter stents (p 0.001), while on the other a significant rate of recanalization (considered as an increase in contrast filling of the aneurysm once it was completely occluded) (46/51 out of cases) in those treated with coiling (p 0.001) (Tables 3,5).

Finally, no significant difference in occlusion rate was observed at the last follow-up according to the different endovascular treatments (Tables 3,5).

Comparison between fetal and non-fetal type PComA is reported in Table 6.

A Fetal PComA type was found in 59 patients (24%). Simple coiling was performed in 43 cases (73%), flow-diversion in 6 (10%), remodeling in 5 (9%) and other treatments in 5 (9%). This subgroup of patients showed 7 blister-like aneurysms, differently from the non-fetal variant (2 cases) (Table 6). This difference was statistically significant (p 0.001). However, no statistically significant differences were seen in demographics, mean aneurysms size, clinical presentation, type of endovascular treatments, early and delayed occlusion rate, and frequency of recanalization between the two PComA type groups.

Our paper mainly focuses on the pure radiological technical aspect of the procedures, lacking relevant clinical data. Our results seem to suggest that endovascular Treatment for PComA aneurysms is safe and effective. Proximal location, clear angiographic anatomy and comfortable catheter access represent favorable elements when considering embolization. Conversely, whenever the PComA has to be sacrificed, the fetal variant has been reported a higher risk for ischemic complications [20, 21, 22].

To the best of our knowledge, the present work represents the largest multicenter case series investigating radiological outcomes after endovascular treatments for PComA aneurysms at long follow-up ($>$24 months).

In this series, endovascular treatments showed an occlusion rate (complete and near-complete occlusion) as high as 65% on post-procedural imaging and improved to 90–93% in long-term follow-up evaluations (Tables 1,2,3). When comparing the different techniques, there were no significant differences in aneurysm occlusion nor a higher rate of hemorrhagic and thromboembolic complications at the last imaging follow-up. However, we should consider that the different groups of treatment are heterogeneous in terms of aneurysms size or neck width (larger for FD or coiling + FD) and aneurysm presentation (more frequently ruptured in the coiling group) Table 3.

Raymond et al. [23], in their series of 43 patients suffering from PComA aneurysms and treated endovascularly, reported a recurrence rate of 37.2%. In addition, in the ISAT trial, coiled aneurysms in the PComA location reported a significantly higher risk of recurrence requiring re-treatment [24].

The difference between our results showing a lower recurrence rate (15%) and the studies above could be explained by considering we included aneurysms treated with different techniques. In contrast, patients were mainly treated by coiling in Raymond [23] and the ISAT trial. Indeed, a possible explanation could be that a better patient-based selection of the devices was performed. More than 1/3 of patients were treated differently than coiling alone.

In our series, coiling was the most common procedure (68%), associated with an 18% recurrence rate. Indeed, recanalization (increase in contrast filling of the aneurysm once it was completely occluded) had a significant association with simple coiling (p = 0.01) and never occurred with flow diversion. This is not unexpected since aneurysms treated with flow diverters present a minimal risk of recanalization once occlusion is achieved [25]. Higher rates of recanalization with coiling are also reported by Choi et al. [18] in their series of 480 coiled PComA aneurysms, 33.1% showed a recanalization during the follow-up.

In our series, recanalization (increase in contrast filling of the aneurysm once it was completely occluded) during follow-up was overall reported in 20% of patients. Re-treatment was performed in 15%. Flow-diversion was chosen for re-treatment in most of these cases (41%). Even if most of the patients with a recanalization were retreated, in the choice of proposing re-treatment, we should consider the relatively low risk of rebleeding (varying from 0.6% to 2.6% [26]) and the immediate risks of morbidity and mortality associated with each procedure in the re-treatment session. Moreover, our data showed a tendency towards delayed (at 12 and 24 months) improvement of occlusion rates of patients with an incomplete early occlusion at 6 months follow-up. This could be explained considering patients undergoing flow-diversion usually present a “delayed” occlusion if compared to coiling (more frequently completely occluding the aneurysm at early controls (p 0.001)). Indeed, the occlusion rate significantly improved at final follow-up in patients treated using a flow diverter (p 0.001). The complications rate was 12%, most of which were PComA thrombosis. Indeed, all thrombosis was treated during the endovascular procedure and did not cause any post-procedural neurological deficit.

Simple coiling was the procedure of choice in nearly 70% of patients, preferred for small and ruptured aneurysms. The early occlusion rate was 79%, and the complication rate was 13%, being thrombosis the most frequent. Coiled aneurysms tend to recanalize over time (Fig. 1), with a statistically significant difference (p = 0.01) compared to other treatments.

Fig. 1.

DSA of coiling of a PComA aneurysm. 68 years old woman with an acute onset of III cranial nerve palsy and evidence of PComA aneurysm at DSA (a,b). Coiling procedure with complete occlusion of the aneurysm (c). Six months follow-up DSA showing recanalization (d). Re-treatment with coiling (e) and 6 months follow-up DSA showing a small residual (f).

Stent-assisted coiling was not common as the first treatment (2%) but was the treatment of choice in 29% of re-treated cases. The use of this technique as the first line is probably limited, especially in wide-neck PComA aneurysms with a fetal PCA because these tend to incorporate the origin of the fetal PCA and the ICA [27].

Flow-diversion was performed in 13% of patients, and it was associated with coiling in 8 (3%) patients (Fig. 2). Although it is usually preferred in unruptured aneurysms (because of the need for dual antiplatelet therapy), in our series 4 ruptured aneurysms were treated with flow-diverters. Indeed, the safety of dual antiplatelet administration in acute SAH is still controversial for the possible increased risk of bleeding. However, evidence exists supporting the safety of this therapy even in hemorrhagic settings [28, 29]. Moreover, flow-diversion was the chosen procedure in more than 40% of re-treatments. The association of coiling and subsequently flow diversion could represent a good option to benefit from the immediate effect of occlusion of coils and the long-term effect of vessel reconstruction of flow diverters (Fig. 2). However, the drawback of this strategy is the need for long-term antiplatelet therapy compared to the coiling alone group.

Fig. 2.

DSA of stent assisted coiling of a ruptured PComA aneurysm. 70 years old woman with a history of subarachnoid hemorrhage and evidence of a giant PComA aneurysm at DSA (a). Endovascular treatment with balloon-assisted coiling (b) and complete aneurysm (c) occlusion. Because of the size of the aneurysm, 3 months later, a flow diverter was placed (d,e). Last follow-up (5 years) DSA shows the aneurysm’s complete occlusion with no recanalization (f).

Intravascular devices are increasingly emerging as endovascular treatment, especially for bifurcation aneurysms [30]. We reported a relatively large number (10 patients) of intravascular devices Woven EndoBridge (WEB)© placement for PComA aneurysms if compared to the present knowledge [31, 32, 33]. In 3 cases, the aneurysms were ruptured. The radiological outcome for all the treated aneurysms (ruptured and unruptured) at long-term was complete/near-complete occlusion in 100% of patients, with no complications (Fig. 3).

Fig. 3.

DSA of intravascular device placement for the treatment of a ruptured PComA aneurysm. 45 years old woman with unruptured PComA aneurysms (a). Intravascular device deployment (b,c) with complete aneurysm (d) occlusion. Six-months follow-up DSA shows the aneurysm’s complete occlusion (e,f).

This procedure could be a good alternative to simple coiling in ruptured aneurysms and patients with contraindications for antiplatelet therapy. However, since this technique is relatively new, the learning curve, the device’s prompt availability, and its costs should be carefully considered.

Lastly, aneurysm remodeling, namely coiling with the assistance of a removable device (either a balloon or a Comaneci device in our series), was used in 24 (10%) patients, with 7 (29%) cases requiring a second treatment during follow-up. No significant differences in aneurysm characteristics, early and late occlusion compared to simple coiling were found.

Fetal-type PComA was present in 59 patients, and aneurysms more frequently presented a blister morphology (12% vs. 1% of non-fetal PComA). These kinds of aneurysms can be considered a challenge for treatment because the origin of the fetal PComA arises from the aneurysm sidewall [34].

Our results did not show differences between fetal and non-fetal type PComA aneurysms in terms of the type of treatment, complications (higher rate of thrombosis, 12% vs. 8%, but not statistically significant (p = 0.7)), early and delayed occlusion and recanalization.

These data stand out when compared to the previous studies. Indeed, a higher rate of incomplete occlusion was reported when a fetal-type PComA was present, particularly with flow-diversion treatment [12, 13, 15, 35]. Possible explanations for the failure of flow diversion in fetal type PComA aneurysms are based on the fact that flow through the fetal PComA and the aneurysm sac remains plentiful for the high physiologic demand and does not allow the stasis needed for the thrombotic process and occlusion to occur [36].

Discrepancies in our results could be explained considering the previous series reported a smaller number of patients (between 4 and 7), and mean follow-up was between 5 and 9 months. Indeed, as we demonstrated, flow-diversion procedures increase the rate of occlusion at long-term follow-up. Moreover, Enriquez-Marulanda et al. [37], in their series of 35 patients, did not report the presence of fetal PComA associated with increased risk of incomplete occlusion; and Choi et al. [18] in their series of 480 coiled aneurysms reported recanalization rates unrelated to PCA variation. Furthermore, fetal-type PCA failed to emerge as a risk factor for recanalization in multivariate analysis.

Coiling is the most common procedure which guarantees a good rate of immediate occlusion but a higher rate of long-term recanalization. It should be mainly proposed for ruptured aneurysms [38]. Flow-diversion is mainly effective in the long term, with no recanalization once that aneurysm occlusion is achieved. Long-term antiplatelet administration should be considered in general (not only cerebral) hemorrhagic risks.

This is a retrospective study, flawed by all the inherent related bias. We did not include the clinical outcome of patients; as our primary aim was to investigate the radiological occlusion rate for the different types of treatments. While it is the largest study about endovascular Treatment of PComA aneurysms, the sample size could still reduce the power to detect differences between groups. Longer follow-up was available for a reduced number of patients. Results of intravascular devices are promising and should be better investigated with further specific prospective studies.

The multi-centric nature of this study could cause discrepancies in the type of deployed devices and drug administration protocols, with consequently outcome evaluation biases. Comorbidities would be very useful for a complete evaluation. Unfortunately, we only have data about hypertension, which was present in 45.3% of patients and was not shown to be a statistically significant factor (p = 0.6) for occlusion rate, recanalization, or complications.