Transvenous embolization of the direct carotid-cavernous fistula via the pterygoid plexus: illustrative case

Abstract

BACKGROUND

Endovascular treatment is the mainstay of treatment for carotid-cavernous fistulas, but endovascular approaches vary widely. The authors report a rare case of a direct carotid-cavernous fistula with cranial nerve symptoms caused by rupture of a giant aneurysm in which selective transvenous embolization via the pterygoid plexus was performed.

OBSERVATIONS

An 81-year-old man presented with headache and various progressive cranial nerve symptoms due to a direct carotid-cavernous fistula caused by a ruptured giant aneurysm. All the draining veins visualized on preoperative examination immediately before the treatment were occluded except for the pterygoid plexus. Therefore, the authors chose the dilated pterygoid plexus to approach the shunted pouch at the cavernous sinus and achieve shunt obliteration by selective embolization with coils and n-butyl cyanoacrylate.

LESSONS

Careful study of the three-dimensional rotational images in the preoperative examination is important when considering the various approaches to surgery. The pterygoid plexus can be an effective venous approach route to reach the cavernous sinus area.

A carotid-cavernous fistula (CCF) is a disease entity that causes venous congestive symptoms, blindness, conjunctival hyperemia and edema, pulsatile exophthalmos, and progressive cranial nerve symptoms, and is composed of an abnormal channel between the internal carotid artery (ICA) or external carotid artery and the cavernous sinus (CS). A direct CCF (dCCF) may occur as a result of trauma, rupture of ICA aneurysms, arterial dissection, Ehlers-Danlos syndrome, fibromuscular dysplasia, pseudoxanthoma elasticum, or iatrogenic trauma from surgery.¹

Transvenous embolization (TVE) and/or transarterial embolization using detachable coils to preserve the ICA has become the mainstay of treatment for dCCF; and a combination of balloon, stent, and liquid embolization materials has also been reported.^2,3

Although many venous approach routes have been reported for accessing the CS,⁴ the inferior petrosal sinus (IPS) approach is usually the first choice for TVE.^5,6 However, if the IPS is occluded, several alternative transvenous approach routes must be considered. Here, we report our experience with a case of TVE via the pterygoid plexus, a rare approach, in a dCCF.

Illustrative Case

An 81-year-old man was determined to have a large aneurysm with a maximum diameter of 22 mm at the CS portion of the right ICA on head magnetic resonance imaging (MRI) 2 years earlier. The aneurysm had enlarged gradually. He had sudden onset of headache and diplopia and was admitted to our hospital for further investigation. At the time of presentation, right abducens nerve palsy was also observed and within 1 week of his visit, right oculomotor and trochlear nerve palsy and trigeminal neuropathy also appeared. Digital subtraction angiography (DSA) after the onset of symptoms revealed a giant aneurysm, superior ophthalmic vein (SOV), inferior ophthalmic vein (IOV), superficial middle cerebral vein (SMCV), and pterygoid plexus in the arterial phase of the right ICA angiogram, suggesting shunt disease (Fig. 1A and B). Three-dimensional (3D) rotational DSA revealed a fusiform aneurysm with a maximum diameter of 28 mm, with the SOV, IOV, SMCV, and pterygoid plexus as the draining veins (Fig. 1C). We performed a balloon occlusion test of the right ICA. There was no dysarthria or paralysis during the blockade of the origin of the right ICA, but the patient was slightly unresponsive and disoriented several times. Single-photon emission computed tomography (SPECT) during occlusion of the right ICA showed decreased blood flow in the right cerebral hemisphere (Fig. 1D). The diagnosis was a right dCCF due to the ruptured right giant ICA aneurysm. The Barrow classification was type A. The slab maximum intensity projection (MIP) image of the 3D DSA revealed that the shunt point of the dCCF was located posterior to the lateral side of the aneurysm.

FIG. 1.

Anteroposterior (A) and lateral (B) views of the right ICA angiogram at the time of symptom onset showing a giant aneurysm in the arterial phase, as well as the SOV (black arrowhead), IOV (white arrowhead), and pterygoid plexus (black arrows), suggesting shunt disease. 3D DSA of the right ICA (C) shows an aneurysm with a maximum diameter of 28 mm, with the SOV (black arrowhead), IOV (white arrowhead), and pterygoid plexus (black arrow) as draining veins. Single-photon emission computed tomography (SPECT) (D) during the balloon occlusion test of the right ICA shows decreased blood flow in the right cerebral hemisphere.

There were several problems to consider in the development of treatment strategy: first, the patient was elderly; second, he had undergone abdominal aortic aneurysm stent-graft implantation 18 months earlier and had a risk of endoleak of the stent graft, which prevented dual antiplatelet therapy (DAPT); and third, the balloon occlusion test revealed that he did not have ischemic tolerance. For these reasons, we decided to attempt TVE, which is a palliative treatment that occludes the venous side of the shunt without radical treatment of the aneurysm. Three access routes for TVE were considered: the occluded IPS, the SOV/IOV via the facial or superficial temporal veins, and the pterygoid plexus. At the DSA immediately before treatment, the drainage routes of the SOV, IOV, and SMCV were occluded, and only the pterygoid plexus drainage route was patent (Fig. 2). Access to the pterygoid plexus appeared to be possible via the internal jugular vein (IJV) and retromandibular vein. Therefore, we decided to access the CS/aneurysm via the pterygoid plexus.

FIG. 2.

In the anteroposterior (A) and lateral (B) views of the ICA angiogram before treatment, the SOV and IOV are not visualized, and the only drainage route is the pterygoid plexus.

Under local anesthesia, a 4-Fr sheath and a 7-Fr shuttle sheath (Cook Medical) were inserted into the right femoral artery and vein, respectively. A 4-Fr diagnostic catheter was placed in the right ICA for control angiography. The 7-Fr shuttle guiding sheath was placed in the right IJV. A 3.4-Fr TACTICS (Technocrat Corporation) was used as the distal access catheter as a triple coaxial system. An Echelon-10 microcatheter (Medtronic) was delivered through the pterygoid plexus and the CS to the fistula and into the aneurysm, which was navigated by a Transend EX Soft Tip microguidewire (Stryker). Then, a Headway-17 preshaped 45° (Terumo)/Transend EX Soft Tip with the same triple coaxial system was inserted using the sheeping technique⁷ near the foramen ovale just before the fistula. The MIP images showed that the Echelon-10 had passed through the foramen ovale and CS and had reached the aneurysm (Fig. 3). For fistula coil embolization, a Target XL 360 4 mm × 12 cm (Stryker) was inserted from the Headway-17 to the CS as the supporting coil, followed by parts of an Axium prime helix 2.5 mm × 6 cm (Medtronic) from the Echelon-10 into the aneurysm. By pulling back the Echelon-10 catheter and the Axium prime helix 2.5 mm × 6 cm, we succeeded in inserting the coil on the venous side outside of the aneurysm. We implanted an additional 5 coils on the venous side. Then, 0.03 mL of 33% n-butyl cyanoacrylate (NBCA) from the Headway-17 was injected to fill the gap at the lower end of the coil mass (Fig. 4). Without any periprocedural complications, post-treatment DSA showed that the arteriovenous shunt had disappeared and the giant aneurysm was residual (Fig. 5).

FIG. 3.

We decided to treat via the pterygoid plexus. A jugular venous approach to the lesion was chosen, and an Echelon-10/Transend EX Soft Tip was delivered through the pterygoid plexus and CS to the fistula and into the aneurysm. Slab MIP images (A–C), sagittal sections, of 3D rotational angiography during the procedure show that the Echelon (black arrows) passes through the drainage route through the foramen ovale (white arrowhead) and reaches the aneurysm. In the anteroposterior (D) and lateral (E) views of the fluoroscopy, a Headway-17 45°/Transend EX Soft Tip was inserted using the sheeping technique to the vicinity of the foramen ovale immediately before the fistula. Black and white arrows indicate the tips of the Echelon-10 and Headway-17 microcatheter, respectively.

FIG. 4.

For the fistula coil embolization (A and B), a coil was inserted from the Headway-17, followed by another coil from the Echelon-10 (aneurysm indicated by dashed line). By pulling back the Echelon-10 catheter and the second coil (C and D), we succeeded in inserting the coil on the venous side outside of the aneurysm. We implanted an additional 5 coils in the same area (E). Then, 0.03 mL of 33% n-butyl cyanoacrylate (NBCA) from the Headway-17 was injected to fill the gap at the lower end of the coil mass (black arrowhead, F).

FIG. 5.

The anteroposterior (A) and lateral (B) views of the posttreatment ICA angiogram show that the arteriovenous shunt has disappeared and the giant aneurysm is residual.

The patient’s oculomotor nerve palsy had been improving for approximately 1 year after the surgery but has remained paralyzed since then. In addition, the patient’s abducens nerve palsy has not improved at all since the surgery. The patient’s headache and trigeminal neuropathy have resolved, and there has been no recurrence of CCF on MRI. We have suggested bypass and ICA ligation as radical treatment for the aneurysm, but the patient has refused.

Discussion

Observations

Treatment Strategies

In the present case, the following 3 treatment strategies were considered, ordered from radical to palliative: (1) sacrifice the ICA by direct trapping or endovascular trapping and embolize the venous side; (2) treat the aneurysm, preserving the ICA and embolizing the venous side; or (3) do not treat the aneurysm and embolize the venous side only.

We did not select treatment strategy “1” because of the patient’s elderly age and lack of ischemic tolerance. With regard to treatment strategy “2”, the use of a stent is essential to treat fusiform aneurysms, but we could not select this option because of the risk of endoleak of the stent-graft by the use of DAPT. Therefore, we had no choice but to select treatment strategy “3” as a palliative treatment.

We then considered a plan to guide the microwire transarterially into the fistula from within the aneurysm as a way to reach the venous side of the shunt. However, we judged that it would be very difficult to guide the micro-guidewire into the fistula from within the giant aneurysm under a high-flow shunt, so we decided to attempt treatment via a transvenous approach.

The remaining drainage route, the pterygoid plexus, was selected for the venous side approach because the drainage to the SOV was occluded immediately before the operation. As a result, the arteriovenous shunt was successfully occluded, but the aneurysm itself remained, and ICA sacrifice must be considered for additional treatment of the aneurysm in the future.

Access Route

Various routes have been reported for the transvenous approach to the CS,⁴ but the IPS is generally the primary route of choice.^5,6 If the IPS is occluded or is anatomically unavailable, the SOV directly,⁸ the facial vein,⁹ or the superficial temporal vein (STV)¹⁰ are the usual approach routes of choice.

In the present case, the veins visualized on preoperative angiography were the SOV, IOV, SMCV, and pterygoid plexus, and the IPS was not visualized. Initially, access from the STV, SOV, and IOV was considered; however, the SOV, IOV, and SMCV were occluded immediately before embolization and could not be selected as routes. An approach from the SMCV would require a craniotomy, which was considered invasive for this elderly patient. Preoperative imaging, especially slab MIP, clearly demonstrated a pathway from the IJV to the shunted point through the pterygoid plexus. It has been reported that the IPS, which does not visualize on angiography, can be accessed by advancing a microwire or a 0.035-inch guidewire.^11,12 After consideration, we selected the approach through the pterygoid plexus as the first choice, and through the IPS as the second. We found that the pterygoid plexus was dilated by the high-flow shunt and the vessel diameter was sufficient for the passage of the catheters. The route to the pterygoid plexus was also relatively straight and easily accessible via the IJV and retromandibular veins. Accurate understanding of the anatomy and a sufficiently supported triple-catheter system¹³ were important in the endovascular treatment of this case.

A literature search revealed only 2 reports of transvenous embolization of dCCF via the pterygoid plexus,^14,15 all due to trauma, and the present case is the first report due to a ruptured aneurysm. In addition, 2 reports approached dural arteriovenous fistulas in the CS via the pterygoid plexus.^16,17 In all reports, good embolization to the CCF was obtained. Access through the pterygoid plexus is technically difficult due to the dense vascular structure and it is generally considered a low-priority option. However, this route is also considered safe because most of the approach route is extracranial. We also considered the possibility that the fistula was located posterior to the lateral side of the aneurysm, resulting in the pterygoid plexus, CS, fistula, and aneurysm being in line with each other, thus improving the maneuverability of the microcatheter. By selectively embolizing only the shunt pouch leading to the fistula, we were able to treat the patient without worsening the cranial nerve symptoms caused by compression of the coil mass. It is important to have no preconceived notions, but to carefully review the preoperative slab MIP and 3D DSA images to identify the shunt point¹⁸ and to consider a variety of access routes.

Lessons

We experienced a case in which TVE was performed for dCCF due to a ruptured cerebral aneurysm to improve neurological symptoms. Using various modalities preoperatively, we were able to confirm that the shunt point was localized and that the point could be reached via the pterygoid plexus. When considering treatment for dCCF, the pterygoid plexus is an important option as an access route to the cavernous sinus.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Hiramatsu, Matsuda, Nishi. Acquisition of data: Matsuda, Nishi, Ebisudani, Kimura. Analysis and interpretation of data: Hiramatsu, Matsuda, Haruma. Drafting of the article: Hiramatsu, Matsuda, Edaki. Critically revising the article: Hiramatsu, Matsuda, Sugiu, Haruma, Nishi, Yamaoka, Ebisudani, Kimura, Date. Reviewed submitted version of the manuscript: Hiramatsu, Matsuda, Sugiu, Haruma, Nishi, Yamaoka, Ebisudani, Kimura, Date. Approved the final version of the manuscript on behalf of all authors: Hiramatsu. Study supervision: Sugiu, Hishikawa, Yamaoka, Date.