We obtained institutional review board approval for the retrospective analysis of this study. Over the last seven years (1999-2006), we treated 256 patients with 271 cerebral aneurysms, by endovascular methods. Of these 256 patients, eight patients (7 female, 1 male; mean age, 53 years; age range, 26-65 years) had saccular PICA aneurysms and were the target of this research. Five of the eight aneurysms were found due to rupturing and three were found incidentally. Of the three incidental discoveries of PICAs, two patients had undergone a magnetic resonance imaging due to chronic headache and revealed an PICA aneurysm. The other incidental case was a patient who had a left posterior communicating artery aneurysm clipped at another hospital three years before, and a follow-up CT-angiography revealed an unruptured aneurysm. The locations of the PICA aneurysms were the VA-PICA junction in three patients, the anteromedullary segment in two, the lateromedullary segment in two, and the telovelotonsillary segment in one. Among the eight patients, four patients (3 ruptured and 1 incidental) underwent intrasaccular coiling with the Guglielmi detachable coils. The remaining four patients (2 unruptured and 2 incidental) had aneurysms which seemed difficult to treat by intrasaccular coiling without compromising the lumen of PICA; therefore, an occlusion of the vertebral artery proximal to the origin of PICA with detachable balloons (2 patients) or with coils (2 patients) was performed. The vertebral artery occlusion site was about 2 cm proximal to the origin of PICA in order to prevent the possible occlusion of the perforator to the brain stem (14).

The patient characteristics are shown in Table 1. Of the four patients who underwent intrasaccular coiling, complete packing of the aneurysm with the coils was achieved in three patients (Fig. 1), whereas intentional subtotal packing was done in one case to preserve the parent artery due to a small wide neck aneurysm at the telovelotonsillary segment. The three patients who underwent VA occlusion with balloons, complete occlusion of the PICA as well as thrombosis of the aneurysm was confirmed on a follow-up digital subtraction angiography (DSA) performed between 3 and 14 months after the VA occlusion (Fig. 2). The distal segment of occluded PICA was reconstituted from the contralateral PICA and the proximal segment of occluded VA. The patients were not found to have any symptoms related to the occlusion of the PICA. In one patient who underwent a VA occlusion with coils, a follow-up CT-angiography performed three months later verified the occlusion of PICA and thrombosis of the aneurysm, however the female patient (patient 6) died of pneumonia soon after. No re-bleed occurred until she died, nor any complication related to the performed endovascular procedure. None of the patients experienced a re-bleed during a follow-up examination (mean, 31 months; range, 3-96 months).

The relatively rare incidence and proximity to the brain stem complicates the surgical management of PICA aneurysms. Additionally, an unfavorable relationship of the aneurysmal sac with the parent artery, which is often the case with PICA aneurysms, cause the surgical clipping or endovascular coiling procedures to be difficult and sometimes even impossible without sacrificing the PICA. Of the eight patients with PICA aneurysms in our series of examined patients, 4 (50%) were considered to be difficult for coiling. A deliberate occlusion of VA proximal to the origin of PICA was performed to reduce blood flow and pressure from the contralateral VA to the PICA and the aneurysm. Furthermore, a follow-up DSA confirmed the shrinkage of the aneurysm and ultimately the thrombosis of both the PICA and the aneurysm. In addition, the collateral reconstitution of the distal cortical branch was identified and all four of the patients with complicated PICA aneurysms had no symptoms related to the occlusion of the PICA. A deliberate VA occlusion has been used to treat vertebrobasilar artery aneurysms in a past study, Steinberg et al. (15), with an 86.5% excellent or good result by surgically ligating the VA to treat the unclippable vertebral artery aneurysms. Although the exact location of the VA aneurysm was not specified, the investigators observed that a unilateral VA occlusion could be an important therapeutic option for the treatment of a VA aneurysm. Groden et al. (16) also treated eight patients with fusiform or dissecting VA aneurysms by balloon occlusion of the VA. Their results indicated that one patient had an ischemic infarct at the region involved in the procedure. Occlusion of the VA for the treatment of a proximal PICA aneurysm arguably has some advantages compared to the occlusion of the PICA or the PICA and the aneurysm together. Namely, it could allow some time for the collaterals to the important perforators and cortical branches to be developed by decreasing the pressure and flow to the PICA. However, one important drawback is that it takes a long time for the aneurysms to be thrombosed, thus a substantial risk of rupture while waiting for the aneurysm to be thrombosed exists. Although three out of four aneurysms in our series showed complete thrombosis between 3 to 14 months after VA occlusion, Peluso at al. (17) reported that four PICA aneurysms treated by proximal VA occlusion did not thrombose during follow-up examination (up to 84 months later). Even so, the four aneurysms did not bleed during follow-up examinations, which is consistent with our findings. The potential risk of rupture should be kept in mind when the ruptured aneurysm is dissecting in nature, since the dissecting aneurysm has been known to have a higher risk for re-bleed (18). Maimon et al. (13) treated six patients with an isolated dissecting aneurysm of the PICA by occluding the aneurysm and PICA. Considering the high re-bleed rate and the associated high mortality rate, it seems reasonable to exclude the dissecting aneurysm as an early treatment of choice, however should remain as a possible subsequent treatment. Of the four patients, who underwent a VA occlusion in our series of patients, the discovery of an aneurysm in three of the patients was incidental. Although we were not completely sure that those aneurysms were not dissecting in nature, the saccular shape of aneurysms and the clinical history reduced the likelihood that the aneurysms were dissecting. Given the benign nature of the aneurysms found incidentally, we chose to treat them by minimally invasive techniques.

In our series, four patients underwent coil embolization of the aneurysmal sac with preservation of the parent artery. In three patients, the complete occlusion of the aneurysm neck was achieved. All three patients had a favorable aneurysm shape, which allowed for the uneventful coiling of the aneurysmal sac. In one patient who had a small aneurysm, at the telovelotonsillary segment, intentional subtotal packing with coils was performed, which preserved the PICA. Although this segment of the PICA is known to have no important perforators and could be sacrificed with little clinical consequences, the patient's neurologic status was so grave that we opted to treat the aneurysm less aggressively. Consistently, several other published studies have reported the successful intrasaccular coiling of a PICA aneurysm (5, 6, 9, 11, 12).

A PICA aneurysm could be missed if the whole segment of the PICA is not thoroughly visualized and examined on DSA. In our series, we failed to identify a small lateral medullary segment aneurysm, although, after a retrospective review, the aneurysm was briefly seen on the contralateral VA injection DSA. A repeat angiography, including both VA injections was performed two weeks later along with an ipsilateral VA injection which was non-dominant, to obtain two orthogonal views to show the aneurysm. A non-contrast enhanced CT showed the dilation of the lateral ventricle and some blood clots in the cisterna magna. The importance of both vertebral artery injections has been emphasized when an intraventricular clot and hydrocephalus is identified (3, 19). Blood in the cerebellopontine angle cistern, cisterna magna, and intracerebellar hemorrhage have been noted to be associated with the rupture of a PICA aneurysm.

In conclusion, our series of eight cases demonstrated the safety, feasibility and efficacy of the endovascular management of saccular PICA aneurysms.