Abstract
Background
This study aimed to evaluate the detailed location and the number (single or multiple) of cavernous sinus dural arteriovenous fistula (CSDAVF) shunted pouches as well as the relationship between the characteristics of shunted pouch(es) and the treatment outcome of transvenous embolization for CSDAVF.
Methods
A total of 23 consecutive patients with CSDAVFs who underwent angiogram and transvenous embolization were retrospectively analyzed. Shunted pouches were assessed using three-dimensional angiogram and multiplanar reformatted image obtained from the rotational angiogram data.
Results
Of the 23 patients with CSDAVFs, 40 shunted pouches were identified. Twelve CSDAVFs had a single shunted pouch, and 11 had multiple shunted pouches. The mean CSDAVF with multiple shunted pouches was 2.5. The shunted pouches were more often found in the posterior compartment of the CS, which was connected with the intercavernous sinus (23/40; 57.5%). In 12 CSDAVFs with a single shunted pouch, 10 were treated with selective embolization and complete occlusion was achieved during the follow-up. Two CSDAVFs with single shunted pouch were just observed without intervention, and DAVFs disappeared spontaneously during the follow-up period. In 11 CSDAVFs with multiple shunted pouches, eight were treated with selective embolization and three with sinus embolization. In six of eight (75%), complete occlusion was achieved following selective embolization, but two of eight (25%) recurred and required retreatment.
Conclusions
Rotational angiography data suggested that the shunted pouches of CSDAVFs were mostly located in the posterior compartment of the CS connected with the intercavernous sinus. Selective embolization for CSDAVFs with a single shunted pouch is the first-line treatment alternative to sinus packing, and selective embolization with multiple shunted pouches will be a considerable treatment option.
Keywords: Cavernous sinus dural arteriovenous fistula, shunted pouches, transvenous embolization, selective embolization, rotational angiography
Introduction
Transvenous coil embolization of affected sinus (sinus packing) is one of the treatment options for cavernous sinus dural arteriovenous fistula (CSDAVF). Packing the cavernous sinus (CS) with detachable coils is an effective treatment, but complications occur related to dense coil packing, which impairs the cranial nerve along the CS. 1 In addition, sinus coil packing usually requires a large number of detachable coils to embolize the CS densely. Since recent advances in endovascular devices and imaging modalities, selective embolization of shunted pouches has been used for dural arteriovenous fistulas (DAVFs) including CS lesions. 2 To selectively embolize the shunt of a DAVF, precise identification of the location of shunted pouches by using multimodal imaging is indispensable. In recent years, several papers have investigated the location of shunted pouches and reported that shunted pouches in CSDAVFs are often located in the posterior part of the CS by analyzing with multimodal imaging including rotational angiography.2,3
In the present study, we analyzed the detailed location of single and multiple shunted pouches, as well as the treatment outcome of selective embolization in CSDAVFs.
Materials and methods
We retrospectively analyzed 23 patients with CSDAVFs (six males and 17 females, mean age 67.2 years) who underwent cerebral angiogram at our institute from January 2010 to December 2015. The 23 patients were divided into two groups according to the number of shunted pouches, and the difference between the two groups was compared. All cases had undergone conventional angiography and rotational angiography. To evaluate the shunted pouch of a CSDAVF, especially one with a high-flow shunt, we often performed selective angiography from microcatheterization of the ascending pharyngeal artery, internal maxillary artery, and/or feeding arteries from the contralateral side. Using the reconstructed images from rotational angiography data, the locations of the shunted pouches were identified in detail. Angiography was performed using a biplane angiomachine (Artis Q BA Twin; Siemens, Munich, Germany). The injection speeds for rotational angiography were 3.5 ml/s for the common carotid angiogram, 1.0–2.0 ml/s for the external carotid angiogram, and 0.5 ml/s for the internal maxillary angiogram, using nonionic contrast material (iopamidol 300 mg/ml). The X-ray delay time from injection was set as 1.0–3.0 s. With this additional 1.0–3.0 s and the 5-second acquisition time, the total injection time was 6.0–8.0 s. The total injection volumes were 20–22 ml, 9–13 ml, and ∼3 ml in the common carotid angiogram, external carotid angiogram, and internal maxillary angiogram, respectively. The data matrix was 512 × 512, and the size of the flat panel detector was 38.2 cm × 29.6 cm. The data acquired from rotational angiography were transferred to the workstation (Syngo X workplace; Siemens), which produced a three-dimensional (3D) angiogram and multiplanar reconstruction (MPR) images. The MPR slice thickness was set as 1.0–2.0 mm, and the MPR images included axial, coronal, and sagittal sections. The 3D angiogram and MPR images of the three directions were simultaneously manipulated in the workstation display, and the shunted pouches of the CSDAVF were determined in detail. The angiographic and MPR images were reviewed by two experienced neurointerventionists, and the locations of the shunted pouch were plotted into a model of the CS. For better comprehension, a complicated CS structure was changed to a simple model. The outline of the CS and intercavernous sinus was detected using 3D-angiogram and MPR images, and the CS was adapted in the model.
In this model, the CS was bilaterally divided into three compartments by two horizontal lines. The authors defined these two lines depending on the position of the anterior clinoid process (AC) and the posterior clinoid process (PC). One of them (line A in Figure 1(a) and (b)) was described horizontally according to the position of the PC. The other line was described horizontally according to the position at the posterior one-third of the AC-PC. In addition, these three compartments were separately divided into 12 equal small compartments. The intercavernous sinus was also compartmented and vertically trisected.
Figure 1.
Model of the cavernous sinus (CS) including the intercavernous sinus (a). The CS is bilaterally divided into three compartments by two horizontal lines. The authors define these two lines depending on the position of the anterior clinoid process (AC) and the posterior clinoid process (PC). Line A is described horizontally according to the position of the PC. Line B is described horizontally according to the position at the posterior one-third of the AC-PC. In addition, these three compartments were separately divided into 12 equal small compartments. The intercavernous sinus is compartmented and vertically trisected. The model of the CS (a) is projected onto the lateral X-ray-like image for a better understanding of the location (b). Lines A and B have the same relationship in (a) and (b).
In actual transvenous embolization for CSDAVFs, a microcatheter is navigated based on a plain X-ray image. Therefore, the model of the CS was projected onto the lateral X-ray-like image to better understand the location (Figure 1(b)).
Transvenous embolization was performed for all patients except for two who were observed without intervention. Selective embolization was usually attempted first, and sinus packing was performed when selective embolization was regarded to be difficult. All interventions were performed by experienced neurointerventionists. The indications of endovascular intervention for CSDAVF were treatment for cases with symptoms including extraocular signs and cranial nerve palsy, and for asymptomatic cases with cortical venous reflux. In all cases, transvenous embolization was performed under local anesthesia and a transfemoral venous approach. The endovascular procedure was finished when complete occlusion or marked decrease of shunt flow with disappearance of cortical venous reflux was achieved. When the procedure was long, intervention was separated into two or three sessions. Magnetic resonance imaging (MRI) follow-up was usually scheduled three to five days and three, six, and 12 months after the intervention. After 12 months, symptoms and MRI were assessed at the regular annual checkup. In addition, unscheduled MRI was obtained when the patient presented with probable symptoms related to CSDAVF and then angiography was considered. The recurrence or reworsening of CSDAVF was evaluated by the expansion of the high-intensity area in the CS detected by time-of-flight MRA or neurological deterioration/extraocular sign, which was considered attributable to DAVF.
Statistical analysis was performed using a statistical software (SPSS Statistics Version 22; IBM, Armonk, NY, USA). However, no significant difference was found because of the small sample size (p < 0.05 was considered significant). Therefore, statistical results were omitted in this article.
Results
Location of the shunted pouches
Of the 23 CSDAVF patients, 12 had a single shunted pouch and 11 had multiple shunted pouches. They had a total of 40 shunted pouches, and the mean number of shunted pouches was 1.7 (range 1–4). CSDAVFs with multiple shunted pouches had a mean of 2.5 shunted pouches. The distribution of the shunted pouches in the CS was not statistically different including the intercavernous sinus, but they were often located in the posterior compartment of the CS.
In 11 CSDAVFs with a single shunted pouch, the locations of the shunt points were not significantly different although they were mainly located at the posterosuperior and posteromedial wall of the CS (Figure 2(a)). On the contrary, no shunt was detected in the anterior part of the CS.
Figure 2.
Locations of shunted pouches in cavernous sinus dural arteriovenous fistula (CSDAVF) plotted on the model of cavernous sinus (CS) (Figure 1(a)). Figure 2(a) and (b) show the shunt distribution of CSDAVFs with single or multiple shunted pouches, respectively. Figure 2(c) demonstrates the number of shunted pouches in each compartment. The shunted pouches of CSDAVFs are often located posteromedially in the CS both with single and multiple shunts.
A total of 28 shunted pouches were found in 11 CSDAVFs with multiple shunted pouches. Multiple shunted pouches tended to be located more medially than single shunted pouches. The medial wall of the left and right CS, which is located just posterior to the PC and connects to the intercavernous sinus, had six and five pouches (21.4% and 17.9%, respectively) (Figure 2(b)). Six shunted pouches were located in the intercavernous sinus without significant distribution. In total, 23/40 (57.5%) shunted pouches were located in the superomedial compartment of the CS, which connected to the intercavernous sinus (Figure 2(c)).
In the lateral X-ray-like image, almost all the shunts were located in the posterior compartment. Moreover, shunt points were aggregated just above the vertical line of PC and the horizontal line of AC (Figure 3).
Figure 3.
Locations of shunted pouches are projected to the lateral X-ray-like image. Almost all the shunts are located in the posterior compartment (posterior one-third of the anterior clinoid process (AC)-posterior clinoid process (PC)). Moreover, the shunt points are aggregated just above the vertical line of the PC and the horizontal line of the AC both in single and multiple shunts.
Relation between shunted pouches and draining vein
The main drainage route of CSDAVFs with a single shunted pouch were the ipsilateral superior orbital vein (9/12: 75.0%), bilateral superior orbital vein (3/12: 25.0%), intracranial drainages including superficial middle cerebral vein and superior petrosal sinus (5/12: 41.7%) and/or other drainages including inferior petrosal sinus and pterygoid venous plexus (5/12: 41.7%). Main drainers of CSDAVFs with multiple shunted pouches were the ipsilateral superior orbital vein (2/11: 18.1%), bilateral superior orbital vein (5/11: 45.4%), intracranial drainages (8/11: 72.3%) and/or other drainages (7/11: 63.6%).
Treatment outcome
Among the 23 patients with CSDAVFs, two with single shunts were observed without intervention because the shunt flow was extremely slow. Among the 21 patients treated, 18 (10 with a single shunt and eight with multiple shunts) were treated with selective embolization. The shunts disappeared immediately in five patients (four with single and one with multiple) after embolization. In the remaining 13 patients (six with single and seven with multiple), a marked decrease of shunt flow was achieved after selective embolization and 11 had become completely occluded during the follow-up period. However, two patients with multiple shunts experienced recurrence several months after selective embolization and required retreatment (Figure 4). The median follow-up period was 25 months (5–65 months).
Figure 4.
Treatment results for cavernous sinus dural arteriovenous fistulas (CSDAVFs) with single and multiple shunted pouches.
Treatment outcome of CSDAVFs with a single shunted pouch
In 12 CSDAVFs with a single shunted pouch, 10 were treated with selective embolization and complete occlusion was achieved during the follow-up. Two CSDAVFs with a single shunted pouch were just observed without intervention, and DAVFs disappeared spontaneously during the follow-up period.
Treatment outcome of CSDAVFs with multiple shunted pouches
In 11 patients with multiple shunts, eight were initially treated with selective embolization and two of eight patients presented with a recurrence. Three of 11 patients were initially treated with sinus packing, and no recurrence was noted.
CSDAVFs with multiple shunted pouches initially treated with sinus packing
In three CSDAVFs treated with sinus packing for the initial treatment, the numbers of shunted pouches were two, three, and four, and several shunted pouches were located in the same compartment of the CS. In two of the three CSDAVFs, shunt flow was markedly decreased after sinus packing, and complete occlusion was subsequently achieved. One of them required three sessions of treatment for sinus packing. This case had four shunted pouches with much shunted flow and both of the CS were affected. The first session took time to attempt selective embolization, which finally failed. The strategy was changed to sinus packing, but it resulted in partial packing because of the long operation time. In second and third sessions, total sinus packing was performed, and no recurrence was found.
CSDAVF with multiple shunted pouches initially treated with selective embolization
In 11 patients with multiple shunts, eight were initially treated with selective embolization. Two of these eight patients presented with a recurrence and required retreatment. They initially underwent selective embolization, and sinus packing was then performed to treat recurrent shunt. The numbers of shunted pouches were two and three, and at least two shunted pouches were located in the same compartment of the CS. The shunted pouches were located in the junction of the CS and intercavernous sinus. In one case, the recurrent shunted pouch was close to the shunted pouch at the first treatment, but the correct location of the recurrent shunt was unclear because of coil artifact. Main feeders were changed from the external carotid artery at the initial treatment to those from the ipsilateral internal carotid artery in the retreatment. In the other case, another shunted pouch, which was not recognized at the first treatment, and a residual pouch appeared. Angiograms acquired immediately after retreatment of sinus packing showed a marked decrease in shunt flow in both cases and had no evidence of worsening at the follow-up.
Case presentation
In this case, a CSDAVF with a single shunted pouch was successfully treated with selective embolization. The patient was a 50-year-old female who presented with the following symptoms: right chemosis, proptosis, and diplopia. Selective right external carotid angiogram demonstrated a CSDAVF with only one shunted pouch located in the posteroinferior area of the clinoid process, mainly feeding from the right middle meningeal artery and the ascending pharyngeal artery (Figure 5(a) and (b)). The orifice of the shunted pouch was clearly identified using 3D external carotid angiogram (Figure 5(c)) and MPR images from rotational external carotid angiogram (Figure 5(d)). The angiogram injected from the microcatheter navigated to the CS demonstrated that the tip of the microcatheter was located in the shunted pouch (Figure 5(e)). Selective embolization was performed (Figure 5(f)), and the shunt was completely occluded. Hereafter, her ocular symptoms improved.
Figure 5.
A case of cavernous sinus dural arteriovenous fistula (CSDAVF) with a single shunted pouch treated with transvenous selective embolization. (a) and (b) Right external carotid angiogram shows a CSDAVF with reflux to the superior orbital vein (anterioposterior and lateral view, respectively). (c) and (d) Three-dimensional angiogram and multiplanar reconstruction sagittal image from the right rotational external carotid angiogram clearly reveal the single shunted pouch of the DAVF. The line in (d) was vertically described according to the posterior clinoid process. (e) and (f) Angiogram injected from the microcatheter transvenously navigated to the cavernous sinus indicates a shunted pouch (arrowhead) and CS (asterisk). (g) Postoperative X-ray image lateral view shows the small coil mass with selective embolization. The coil mass is located just above the posterior clinoid process (arrow).
Discussion
Location of the shunted pouches
Recently, the efficacy and usefulness of selective or targeted embolization for DAVFs have been reported. 4 For selective/targeted embolization of shunted pouches of DAVFs, identification of the shunt location and precise navigation of the microcatheter to the target shunted pouch are essential. 5 In this series, shunted pouches were evaluated by carefully analyzing image data from rotational angiography, and selective angiogram was useful to identify the shunted pouch in the CSDAVF.
In the present study, the CS was divided into three major compartments and 68 small aspects in detail. In this study, we demonstrated that 97.5% of the shunted pouches were located in the posterior compartment of the CS. In particular, 70% of them were located in the posteromedial and posterosuperior compartments of the CS, which were connected with the intercavernous sinus (Figure 2(c)). This tendency (shunted pouches were located in the posterior compartment of the CS) is remarkable in CSDAVFs with multiple shunted pouches. Kiyosue et al. 3 reported that 16 of 19 CSDAVFs (84.2%) had shunted pouches posteromedially in the CS. Satow et al. 2 reported that eight of 20 cases (40%) had shunted pouches posterosuperiorly in the CS. The results of these studies are comparable with the present findings. The intercavernous sinus could also be the site affected by shunted pouches. 6 This study demonstrated that the sites of shunted pouches in the intercavernous sinus did not show laterality and distribution (Figure 4).
Almost all the shunted pouches were located around the vertical line of the PC. Projecting the identified shunted pouches onto the lateral skull X-ray-like image, they were aggregated just above the vertical line of the PC and the horizontal line of the AC (Figure 3). In actual clinical settings, when the MPR images of rotational angiogram identified the shunt(s), the X-ray projected image was instructive because plain X-ray or conventional angiography was used for selectively navigating the microcatheter to the affected lesion.
The posteromedial part of the CS receives various arterial supplies from several arterial systems including meningeal, clival (branches from the internal carotid artery), and ascending pharyngeal arteries, sometimes bilaterally. On the contrary, the lateral and anterior part of the CS receives less and usually a unilateral arterial supply compared with the posteromedial part of the CS. This difference in arterial supply could induce more localization of shunted pouches in the posteromedial part of the CS. In addition, the lateral wall of the CS is made of two dural layers (meningeal layer and periostic layer), whereas the medial wall of the CS is made of a single meningeal layer.7,8 Therefore, the medial wall of the CS has simpler membranous structure. This difference in wall structure could lead to more frequent arteriovenous shunts in the medial wall of the CS.
Treatment for CSDAVF
Recently, transarterial embolization with Onyx for DAVF has been reported, 9 but Onyx use for DAVFs in the CS is uncommon. Therefore, transvenous embolization for CSDAVFs is still an indispensable procedure. Understanding the location of shunted pouches in the CS is helpful for transvenous embolization, particularly in selective embolization. Moreover, even if liquid embolic materials are used for feeder embolization, correctly identifying the orifice of the shunted pouch is important.
The factors of recurrence are multiple shunted pouches and multilocalization of shunted pouches. 3 In this study, no recurrences were observed in 10 CSDAVFs with single shunted pouches receiving selective embolization as an initial treatment. Thus, selective embolization for CSDAVFs with a single shunted pouch is the first treatment alternative to sinus packing. On the contrary, two of eight (25%) CSDAVFs with multiple shunts receiving selective embolization as an initial treatment presented with a recurrence and required retreatment of sinus packing. In both cases, two shunted pouches were located in the same compartment and located nearby. When multiple shunted pouches were localized closely, shunt flow seemed to be underlooked or underestimated by complicated angioarchitecture around the shunted pouches. However, even in recurrent cases, DVAFs had only a low shunt flow and they seemed to be completely occluded with fewer coils required for sinus packing compared with selective embolization used as an initial treatment. Intracranial drainage with cortical reflux was observed more often in CSDAVFs with multiple shunted pouches than in those with single shunted pouches. Previously coiling for intracranial drainage might have been considered before selective embolization for a shunted pouch because one-quarter of cases that multiple shunted pouches treated by selective embolization had presented with a recurrence and required retreatment. In addition, there are cases of difficult-to-identify shunted pouches owing to a diffuse shunted area. These cases seem to be suitable for sinus packing. By contrast, three-quarters of cases of CSDAVFs with multiple shunts could be occluded by selective embolization. This result may be acceptable for the first treatment option for CSDAVF. The treatment strategy for CSDAVFs with multiple shunts (selective embolization or sinus packing) should be selected depending on the angioarchitecture in individual cases.
Conclusions
Based on the data from rotational angiography and the detailed definition of the compartment of the CS, the orifices of the shunted pouches of CSDAVFs were located in the posteromedial and posterosuperior compartment of the CS, which was connected with the intercavernous sinus. Selective embolization for CSDAVFs with single shunted pouches is the first-line treatment alternative to sinus packing. Furthermore, selective embolization for CSDAVFs with multiple shunted pouches should be considered a viable treatment option.
Table 1.
Characteristics of the 23 patients with cavernous sinus dural arteriovenous fistulas.
| Number of patients | Mean number of shunted pouches | Gender (male/ female) | Initial treatment |
Retreatment | |||
|---|---|---|---|---|---|---|---|
| Selective embolization | Sinus packing | Observation | |||||
| Total | 23 | 1.7 | 6/17 | 18 | 3 | 2 | 2 |
| Single shunted pouch | 12 | 1 | 3/9 | 10 | 0 | 2 | 0 |
| Multiple shunted pouches | 11 | 2.5 | 3/8 | 8 | 3 | 0 | 2 a |
Two of eight patients initially treated with selective embolization presented with a recurrence and required retreatment with sinus packing.
Acknowledgments
Author contributions include the following: M Sato: Data collection and manuscript writing, T Izumi: Project development and operation, N Matsubara: Operation and manuscript writing, M Nishihori: Data collection, S Miyachi: Data collection and operation, T Wakabayashi: Data collection.
We declare that this study has been approved by the institutional review board of our university hospital and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. We declare that all patients gave informed consent prior to angiography and intervention.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
References
- 1.Nishino K, Ito Y, Hasegawa H, et al. Cranial nerve palsy following transvenous embolization for a cavernous sinus dural arteriovenous fistula: Association with the volume and location of detachable coils. J Neurosurg 2008; 109: 208–214. [DOI] [PubMed] [Google Scholar]
- 2.Satow T, Murao K, Matsushige T, et al. Superselective shunt occlusion for the treatment of cavernous sinus dural arteriovenous fistulae. Neurosurgery 2013; 73(ONS Suppl 1): ons100–ons105. [DOI] [PubMed] [Google Scholar]
- 3.Kiyosue H, Tanoue S, Hori Y, et al. Shunted pouches of cavernous sinus dural AVFs: Evaluation by 3D rotational angiography. Neuroradiology 2015; 57: 283–290. [DOI] [PubMed] [Google Scholar]
- 4.Nakamura M, Tamaki N, Kawaguchi T, et al. Selective transvenous embolization of dural carotid-cavernous sinus fistulas with preservation of sylvian venous outflow. Report of three cases. J Neurosurg 1998; 89: 825–829. [DOI] [PubMed] [Google Scholar]
- 5.Kannath SK, Rajan JE, Sarma SP, et al. Anatomical localization of the cavernous sinus dural fistula by 3D rotational angiography with emphasis on clinical and therapeutic implications. J Neuroradiol 2017; 44: 326–332. [DOI] [PubMed] [Google Scholar]
- 6.Takahashi S, Sakuma I, Tomura N, et al. Transvenous embolization of dural arteriovenous fistula of the cavernous sinus. Fistulous points and route of catheterization. Interv Neuroradiol 2004; 30(Suppl 1): 85–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Campero A, Campero AA, Martins C, et al. Surgical anatomy of the dural walls of the cavernous sinus. J Clin Neurosci 2010; 17: 746–750. [DOI] [PubMed] [Google Scholar]
- 8.Kural C, Simsek GG, Guresci S, et al. Histological structure of the medial and lateral walls of cavernous sinus in human fetuses. Childs Nerv Syst 2015; 31: 699–703. [DOI] [PubMed] [Google Scholar]
- 9.Ambekar S, Gaynor BG, Peterson EC, et al. Long-term angiographic results of endovascularly “cured” intracranial dural arteriovenous fistulas. J Neurosurg 2015; 124: 1123–1127. [DOI] [PubMed] [Google Scholar]