Introduction

   The association of arteriovenous malformations(AVMs) with intracranial aneurysm has been well documented in numerous reports.^1)2)5)6) Aneurysms related to AVMs can be divided into four groups：dysplastic or remote, unrelated to inflow vessels(Type 1)；proximal, arising at the circle of Willis origin of a vessel supplying the AVM(Type 2)；pedicular, arising from the midcourse of a feeding pedicle(Type 3)；and intranidal, within the AVM nidus itself(Type 4).⁵⁾ In several reports, distal or pedicle feeding vessel aneurysms, as well as intranidal aneurysms, have been linked to an increased risk of hemorrhage.¹⁾⁶⁾
   The posterior cerebral artery(PCA) aneurysm associated AVM is more uncommon lesion. Although Zeal and Rhoton et al.⁸⁾ have indicated that 15% of PCA aneurysms arised on the P1 segment of the PCA and Perata et al.⁵⁾ reported a case of thalamoventricular AVM and a ruptured small pedicle aneurysm on the anterior thalamic perforate artery, there are no reports of a ruptured aneurysm arising from the P1 perforator itself of the PCA that fed into the thalamic AVM. Here we report an operated case with a review of literatures.

Case Report

   A 45-year-old male was admitted in a state of semicoma. A brain computed tomography(CT) scan revealed widespread subarachnoid hemorrhage, mainly in the left ambient and quadrigeminal cistern with intraventricular hemorrhage (Fig. 1). Cerebral angiography revealed about 20×25 mm sized AVM on the thalamus and about 10×15 mm sized fusiform dilatation on the proximal portion of the perforating artery of the P1 segment in the PCA. The AVM was fed by the P1 perforator artery and the branch of the PCA and the superior cerebellar artery, and it was drained to the vein of Galen(Fig. 2). 
   On the following day, a clipping of the perforating artery of the P1 segment was successfully performed through left pterional approach, however, the thalamic AVM could not be removed. In postoperative one month, repeated angiography revealed obliteration of aneurysm but the AVM was remained intact(Fig. 3). A Gamma-knife radiosurgery or an endovascular treatment of the AVM was recommended but patient's relatives refused. In postoperative 6 months, he underwent ventriculo-peritoneal shunt because of communicating hydrocephalus.
   The postoperative course was uneventful, and the patient had been improved to be a simple communication. During the follow-up period of two years, there was no more hemorrhage.

Discussion

   Although little is understood about pathogenesis of AVM-associated aneurysm, it is generally accepted that hemodynamic stress due to focally increased flow is responsible, at least in part, for their occurrence.¹⁾²⁾
   The perforating arteries arising from the P1 segment, which are always present, are divided into two groups^3)4)7)8)：1) the diencephalic branches, also called thalamoperforating branches；2) the mesencephalic branches. All of the perforators should be preserved regardless of their caliber, because it is impossible during neurosurgical procedures to judge their functional significance. In our patients, the proximal part of the perforator suggesting the thalamoperforating artery was sacrificed by making use of a clip because the neck of aneurysm could not be identified. 
   If there is a hemorrhage, the lesion believed responsible should be treated first. Redekep et al.⁶⁾ found that the source of the hemorrhage between an aneurysm and the AVM was almost equal. In contrast, Brown et al.²⁾ reported that a hemorrhage appeared to occur with somewhat greater frequency from the aneurysm than from the AVM, and suggested that if the CT scan showed any non-localized SAH and the source of bleeding in question, the aneurysm was more likely the responsible lesion. There is ongoing debate concerning the need for treatment of flow-related aneurysms that are discovered incidentally in association with AVMs. To avoid hemorrhagic consequence, several reports¹⁾²⁾⁵⁾ recommend control of aneurysms, while the resection of the associated AVM may proceed thereafter.

Conclusion

   This is the first case of an AVM associated with an aneurysm that the P1 perforator of the PCA can serve as the anatomic origin of a ruptured cerebral aneurysm.
   We believe that the most safe approach for this combined lesion is to treat aneurysm before resection of AVM despite of asymptomatic patients. 

REFERENCES

1. Batjer H, Suss RA, Samson D. Intracranial arteriovenous malformation associated with aneurysm. Neurosurgery 18:29-35, 1986

2. Brown RD Jr, Wieber DO, Forbe SG. Unruptured intracranial aneurysms and arteriovenous malformations: frequency of intracranial hemorrhage and relationship of lesions. J Neurosug 73:859-63, 1990

3. Marincovic S, Milisavljevic M, Kovacevic M. Interpeduncular perforating branches of the posterior cerebral artery: Microsurgical anatomy of thier extracerebral and intracerebral segments. Surg Neurol 26:349-59, 1986

4. Pedroza A, Dujovny M, Ausman JI, Diaz FG, Cabezudo AJ, Berman SK, Mirchandani HG, et al. Microvascular anatomy of the interpenduncular fossa. J Neurosurg 64:484-93, 1986

5. Perata HJ, Tomsick TA, Tew JM Jr. Feeding artery pedicle aneurysm: association with parenchymal hemorrhage and arteriovenous malformation in the brain. J Neurosurg 80:631-4, 1994

6. Redekop G, Terbrugge K, Montanera W, Willinsky R. Arterial aneurysms associated with cerebral arteriovenous malformations: classification, incidence, and risk of hemorrhage. J Neurosurg 89:539-46, 1998

7. Sieben G, De Reuck J, Vander EH. Thrombosis of the mesencephalic artery: A clinico-pathological study of two cases and its correlation with the arterial vascularization. Acta Neurol Belg 77:151-62, 1977

8. Zeal AA, Rhoton AL Jr. Microsurgical anatomy of the posterior cerebral artery. J Neurosurg 48:534-59, 1978