Galenic dural arteriovenous fistula in neurofibromatosis type 1 treated with Onyx

Abstract

Galenic dural arteriovenous fistula (GDAVF) represents a unique, hard-to-treat subgroup of tentorial dural arteriovenous fistulae. Neurofibromatosis type 1 (NF1) has been infrequently associated with different cerebrovascular conditions that may lead to either ischemic or haemorrhagic stroke. Intracranial GDAVF has not been described in NF1 patients. We present an unusual case of GDAVF in a 37-year-old man with NF1. The fistula drained directly to the vein of Galen through multiple feeders. Complete occlusion of the fistula was achieved through trans-arterial embolisation with Onyx (ethylene vinyl alcohol copolymer) in a single treatment session. Deep venous drainage remained intact, and the patient recovered well. To our knowledge, this is the first report on complete closure of GDAVF with NF1 using trans-arterial embolisation. The preservation of functioning of the straight sinus may have contributed to the success of treatment.

Introduction

Galenic dural arteriovenous fistula (GDAVF) is a rare condition. The treatment of GDAVF is a therapeutic challenge and often requires multimodality treatment. It is well known that neurofibromatosis is associated with various types of arteriovenous fistulae (AVF), but association with a GDAVF has not been reported. We describe a case of spontaneous GDAVF associated with neurofibromatosis type 1 (NF1). He was successfully treated by endovascular embolisation alone and achieved a complete cure with a good clinical outcome.

Case report

This 37-year-old male with a history of NF1 presenting to our hospital complained of progressive dizziness and tinnitus for half a year. He was in good general condition, and his neurological examination was unremarkable. He exhibited typical features of neurofibromatosis, including multiple cafe au lait spots and numerous subcutaneous neurofibromas. Magnetic resonance imaging (MRI) of the brain showed large flow voids of the cisterna venae magnae cerebri, representing DAVF with pial vein reflux (Figure 1(a)–(c)). Cerebral angiography showed an A–V shunting zone at the lateral vein of Galen wall (Figure 1(d)–(g)). The entire scope of the fistula itself was very extensive, with variceal dilatation of the vein of Galen, mainly supplied by petrosal and tentorial branches of both middle meningeal arteries, the maxillary artery, the mastoid branches of the occipital arteries, the meningohypophyseal trunk, and drained caudally through the patent straight sinus (Figure 1(d)–(g)). The inverted direction of flow in the basal vein of Rosenthal (curved arrow) was noted (Figure 1(h)). The straight sinus was patent (Figure 1(h)). Endovascular therapy was undertaken with embolisation of a dural contribution from the left middle meningeal artery through which the Marathon microcatheter (Medtronic, Minneapolis, MN) advanced over the fistula, and the tip was placed in the fistulous variceal dilatation segment (Figure 1(i) and (j)). After unremarkable injections with dimethyl sulfoxide, trans-arterial embolisation of the GDAVF was performed with 5 mL Onyx 18 within 20 minutes. When injection was in progress, close attention was paid to prevent Onyx backflow and avoid crossing the orifice of vein of Galen and the straight sinus. Immediate post-embolisation angiogram showed the complete occlusion of the fistula (Figure 1(k) and (l)). The patient was uneventful postoperatively. The achieved endovascular occlusion of the fistula in a single session proved to be persistent on 10-year follow-up MRI and angiography (Figure 2).

Figure 1.

(a) Axial, T2-weighted image and (b) magnetic resonance angiography show a large area of flow void in the quadrigeminal cistern, indicating a dilated vein of Galen (asterisks) with dilation of feeding arteries of tentorium cerebelli (white arrows). (c) Sagittal magnetic resonance venography reveals dilated venous structures of Rosenthal’s vein (dashed white arrows). (d) and (e) Digital subtraction angiography (DSA) of bilateral internal carotid arteries reveals enlarged tentorial artery, a branch of meningohypophyseal trunk, supplying the Galenic dural arteriovenous fistula (GDAVF) through enlarged tentorial artery into the enlarged tortuous vein of Galen. (f) and (g) DSA of the selective left external carotid artery (arterial phase) reveals the petrosal and tentorial branches of both middle meningeal arteries (black arrow), the maxillary artery (white arrow), the mastoid branches of the occipital arteries (curved arrow) feeding the GDAVF and (h) the venous phase shows the inverted direction of flow in the basal vein of Rosenthal (white arrowheads) and drainage to the straight sinus (black arrowheads). (i) Intra-procedural magnified angiographic view after microcatheterisation of the fistulous venous segment (see asterisk). (j) Super-selective angiography of left-middle meningeal arteries shows the angioarchitecture of the fistula in which the region circled by white dashed line is considered the site of therapeutic target and could be embolised with no risk of the venous drainage obstruction. However, the region circled by red dashed line should be reserved because of its anterograde outflow function (white and black arrowheads). (k) Non-subtraction angiography shows the Onyx cast has embolised the petrosal and tentorial feeding arteries (white arrow), as well as the fistulous venous and lateral drainage vein (black arrow). (l) Immediate post-embolisation angiography shows the complete obliteration of the fistula, with preservation of patency of the straight sinus (white dashed line).

Figure 2.

Ten-year follow-up magnetic resonance images with contrast (a) axial, (b) coronal, (c) sagittal and (d)–(f) angiograms show the absence of the GDAVF. Note the patency of the straight sinus on contrast enhancing imaging (see white arrowheads).

Discussion

DAVFs are classically defined as abnormal connections that are located within the dura mater, usually within the walls of a dural venous sinus or involving an adjacent cortical vein. They represent 10–15% of cerebrovascular malformations, and most commonly develop in the region of the transverse and sigmoid sinuses.¹ The exact mechanism of development of such DAVF is unclear, and the theories are controversial, with the most accepted being the pathological recanalisation of a thrombosed sinus. Galenic DAVFs are rare and are a heterogeneous group defined by drainage into the deep venous system at various sites and represent approximately 23% of all tentorial DAVFs, which is most frequently associated with aggressive clinical behaviour, presenting the clinician with diagnostic and therapeutic challenges.² The term GDAVF is sometimes confused with vein of Galen malformations (VGMs) or vein of Galen aneurysmal dilatation (VGAD).^1,3,4 VGMs are direct arteriovenous shunts in the subarachnoid space of the velum interpositum cistern and quadrigeminal cistern, supplied by the choroidal and quadrigeminal arteries and drained by the dilated median prosencephalic vein of Markowski, the embryonic precursor of the vein of Galen. VGAD indicates the dilatation of the ‘true’ (embryologically matured) vein of Galen by vascular lesions, including pial or dural arteriovenous malformation.

There are a limited number of case reports of a spontaneous DAVF associated with NF1, mostly located in the cervix related to the vertebral artery. The precise mechanisms involved in DAVF associated with NF1 are poorly understood but are likely related to the function of neurofibromin, the protein product of the NF1 gene. It has been hypothesised that the loss of neurofibromin expression in endothelial cells may cause vascular smooth-muscle cells to proliferate.^5,6 Deans et al.³ investigated the developmental mechanism of AVF in patients with neurofibromatosis, and proposed the following two scenarios: (1) an aneurysm develops in an area of the arterial wall weakened by smooth-muscle dysplasia or neurofibroma, and a fistula involving adjacent veins eventually forms; and (2) an AVF develops congenitally owing to mesodermal dysplasia. Because our patient was an adult, we considered the GDAVF was unlikely to be a congenital lesion and was most likely to have been triggered by any one of a number of inciting events, such as dural sinus thrombosis, thrombophlebitis, infection or inflammation, and promoted by a weakened arterial wall related to NF1. Cases with NF1 involving in the vein of Galen are considered to be a further rarity. We should remembered that GDAVF maybe occur in NF1.

Endovascular approaches are generally recommended as the initial treatment in most DAVFs, and open surgical techniques have high risk of bleeding due to the vascular fragility common in NF1.^7,8 Depending on the anatomy, a trans-arterial, transvenous, combined transarterial/transvenous approach can be used to cure a DAVF.⁹ Multiple embolisation agents can be used. However, liquid embolisation agents and coils are primarily used. We believe that transarterial Onyx embolisation is the best way to achieve on-target obliteration at the fistula site, leading to effective shunt reduction. Onyx is a non-adhesive liquid embolic agent and allows a longer injection time with high volume of embolic materials and good penetration, hence producing more stable results with fewer complications. To maximise the chance for long-term closure of an arteriovenous fistula, it is usually recommended to treat either the fistula directly with some form of occlusive device or both supplying arteries and the fistulous venous segment. In any case, a too proximal occlusion of the arterial feeders without reaching and closing the arteriovenous fistulous connections should be absolutely avoided. However, the Galenic system is a complex venous network, and embolisation of the vein of Galen or its tributaries is dreaded by most neurointerventionalists, although the consequences of venous sacrifice have not been consistently reported in the literature. Treatment of GDAVFs is so challenging that endovascular approach usually cannot obliterate it completely in a single session, and in most cases, multiple-staged embolisation is needed.^2,10–12 Steven De Vleeschouwer et al. reported a case of GDAVF in which a session of embolisation of the falcotentorial feeding vessels followed by additional surgical transection of the remaining tentorial arterial feeders failed to exclude the Galenic DAVF.¹² However, in cases where those entire feeders end in the vein of Galen, the presence of obliterated, non-functioning straight sinus allows the take-down of all feeders in one treatment session, simply by filling the vein of Galen through a single artery. In our case, the Galenic venous pouch supplied by multiple feeders is considered the site of therapeutic target base on the angioarchitecture of the patient (Figure 1G). Complete occlusion of the complex GDAVF was achieved by careful endovascular embolisation alone using the key point of placing the microcatheter in the fistulous venous segment, allowing the embolic material to diffuse into the feeding arteries in a controlled fashion with preservation of the lumen of vein of Galen and anterograde flow in the deep venous system. In the present study, endovascular therapy using Onyx 18 was successful in obliterating the GDAVF. However, because of the presence of a NF1-induced congenital mesoderm abnormality, the fistula could still recur, and a new lesion may develop. Hence, it will be important to observe this patient over a long period of time.

Conclusion

Galenic DAVFs are a rare form of DAVF. Their deep location and proximity to crucial vascular and anatomical structures make treatment challenging. To achieve the difficult goal of selective occlusion of the fistula without damaging the deep venous structures, a trans-arterial approach such as through the middle meningeal artery is often a good choice. In the present study, we performed Onyx embolisation in a NF1 adult, successfully obliterating an AVF in a single session. To our knowledge, this is the first case of Galenic DAVF in NF1.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethical Standards and Patient Consent

We declare that all human and animal studies have been approved by the Peking University Shenzhen Hospital Ethics Committee.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: funded by Shenzhen Healthcare Research Project (no. 201601005).

Informed consent

Informed consent was obtained from the participant included in the study.