Article Text

other Versions


Case series
DSA-Dynavision in pretreatment planning for coil embolization of indirect carotid–cavernous fistula
  1. Alex Botsford,
  2. Jai Jai Shiva Shankar
  1. Department of Diagnostic Imaging, QE II Health Sciences Centre, Halifax, Nova Scotia, Canada
  1. Correspondence to Dr J J S Shankar, Department of Diagnostic Imaging, QEII Health Sciences Centre, 1796 Summer St, Room 3305A, Halifax, Nova Scotia, Canada B3H 3A7; shivajai1{at}


Introduction Indirect carotid cavernous fistulas are treated with coil embolization when they present with orbital/visual symptoms or if there is cortical venous reflux. Most of the time, the treatment is done by non-specifically packing the whole cavernous sinus with coils. The purpose of this case series was to examine whether DSA-Dynavision before embolization would improve treatment by shortening the procedure time, requiring fewer coils, or reducing the complication rate.

Materials and method 8 patients with 9 fistula sites were retrospectively identified. DSA-Dynavision and non-DSA-Dynavision patients were compared in a retrospective cohort study.

Results Mean total coil length was significantly shorter for the group who had DSA-Dynavision than for those who had non-DSA-Dynavision (130.5 cm vs 190 cm, p=0.034) and mean procedural time was significantly shorter for the DSA-Dynavision group (171.1 min vs 280.3 min, p=0.025). A transient neurological complication was seen in only one patient.

Conclusions The use of DSA-Dynavision in pre-procedural planning facilitates selective coil embolization of the foot of the vein.

Statistics from


An abnormal vascular shunt between the internal carotid artery (ICA), external carotid artery (ECA), or their branches and the cavernous sinus is called a carotid cavernous fistula (CCF).1 Complex anastomoses between both ICA/ECA branches and the trabeculated meshwork of the cavernous sinus itself leave a large number of possibilities with respect to which vessels are involved in the formation of a CCF.2 ,3 The most commonly used classification system was devised by Barrow et al4 and divides fistulae according to arterial supply. Indirect CCFs (Barrow types B, C, D) are shunts of branches of the ICA, ECA, or both. They generally present with a much more insidious onset, which varies largely according to rate of flow and subsequent pressure build-up within the cavernous sinus.5 Indirect CCFs frequently present with chemosis, proptosis, and headache, with visual disturbance and other more serious clinical presentations often occurring later.

Although indirect CCFs are rare, these represent a sight threatening condition with a variety of occasionally insidious clinical presentations. The standard at most centers is to treat these fistulas with coil embolization, particularly in the presence of cortical venous reflux. Transarterial and transvenous approaches have been documented for the treatment of indirect CCFs,6 but transvenous is generally preferred. Small diameter of the feeder artery and inherent rich anastomosis between the ECA and cavernous segment of the ICA makes the fistulae difficult to treat with a transarterial approach, resulting in a higher rate of complications, including ischemic complications.7 Halbach et al8 demonstrated substantially higher numbers of complete occlusion of indirect CCFs with fewer complications with the transvenous approach compared with the transarterial approach. A further advantage of the transvenous approach is the many venous outflows that can be used to access the fistula. Unfortunately, this can also present challenges. One of the most frequently cited issues with the transvenous approach is procedure time.9 Most of this time is spent deploying a large number of coils for dense packing of the cavernous sinus. In the absence of a clear understanding of the exact fistula point in the complex trabecular anatomy of the cavernous sinus, a large number of coils are used to occlude all possible avenues of venous drainage, resulting in a long procedural time.

DSA-Dynavision from rotational angiography has not been used in pretreatment planning of indirect CCFs.5 ,10 ,11 The purpose of this study was to examine whether DSA-Dynavision done for pretreatment planning of indirect CCFs helps identify the exact fistula point in the cavernous sinus and whether this helps make the treatment more efficient. Our hypothesis is that localization of the exact fistula point on DSA-Dynavision for indirect CCFs would result in reducing the overall procedure time, overall coil length, and avoid packing most of the cavernous sinus, with similar treatment outcome.


This study was approved by our institutional research ethics board. A retrospective search (using the keywords ‘CCF’, ‘CC fistula’, ‘carotico-cavernous fistula’, ‘carotid cavernous fistula’, and ‘dural AV fistula’) of our PACS database was done to find all cases of indirect CCFs at our institution between 2005 and 2015. These results were then matched with those available in the interventional neuroradiology database to ensure inclusion of all patients with indirect CCFs. The inclusion criteria were patients with a diagnosis of indirect CCFs who were treated with endovascular embolization and had imaging on our PACS. Patients with arteriovenous malformations, dural arteriovenous fistulas, or direct CCFs were excluded. Patients in whom DSA-Dynavision was used for pretreatment planning were compared with patients in whom DSA-Dynavision was not used for pretreatment planning. Both groups of patients were classified according to location and type, based on their diagnostic CT, MR, and DSA.

In rotational angiography, the vascular region of interest is acquired from several projection directions during an acquisition series by rotation of the C arm. DSA-Dynavision is the angle triggered acquisition technique with digital online subtraction. Mask and contrast images (mask and fill frames) are acquired in the same angle position of the C arm. Because the mask and fill phases occur in the same direction, the images are acquired under the same conditions. The images were acquired using 5 s rotational angiogram acquisition (70 KV and radiation dose 12.5 mS) on a biplane angiogram unit (Artis Zee, Siemens) with selective injection into the ICA, ECA, or both, depending on where the main feeders came from. For the ICA, the contrast was injected at a rate of 4 mL/s with a total dose of 24 mL and an X-ray delay of 1 s. For the ECA, the contrast was injected at a rate of 2 mL/s with a total dose of 12 mL. This yielded 2×133 frames, which were then reconstructed using multiplanar reconstruction on a Leonardo workstation unit (Siemens). A parallel range was used to reconstruct images in different planes to demonstrate the exact location of the fistula site, and images were saved and sent to PACS. In some patients, the DSA-Dynavision was part of their diagnostic angiogram workup and in others it was done as part of their treatment planning at the time of coil embolization. In the latter situation, the time for acquisition and analysis of DSA-Dynavision added up to the total procedure time.

All patients were treated using transvenous coil embolization. The choice of access microcatheter and coils was dependent on the availability and preference of the operator. All patients underwent general anesthesia for their embolization. Patients were heparinized throughout the procedure. In patients where the exact fistula site was localized on DSA-Dynavision, the foot of the draining vein was approached in the same way as coiling of an aneurysm using a three-dimensional coil for initial framing of the foot of the vein. The initial frame was then subsequently packed densely using shorter and softer coils until flow through the fistula was completely obliterated on check angiograms. This method possibly resulted in focal occlusion and shorter coil length requirement. In patients where the exact fistula site was not localized, the affected cavernous sinus was packed with coils until the fistula was obliterated.

The cases and controls were compared for demographic characteristics, presenting symptoms, obliteration rates, procedural complications, procedure time, and total coil length used. In patients where DSA-Dynavision was done at the time of transvenous coiling, the time used for acquisition and analysis of DSA-Dynavision was subtracted from the total procedure time.

Frequency measures were used for descriptive statistics. For comparison of different parameters between cases and controls, the Student's t test and χ2 test were used; 95% CIs were calculated. A p value of 0.05 was considered significant.


We found 23 patients with CCFs, 13 of which were indirect. We excluded three patients who had no treatment for their fistulas and two other patients who were treated with Onyx or glue. The final eight patients with nine fistulas (one patient with two fistulas) included in the study are shown in table 1.

Table 1

Demographics of the patient population, clinical presentation, and carotid cavernous fistula characterization

Pre-procedural information for each patient is summarized in table 1. We had equal sex distribution with a mean age of 52.3 years. Patients generally presented with proptosis (75%), headaches (25%), chemosis (25%), or diplopia (25%), and one patient presented with new abducens nerve palsy. The majority of the fistulas were Barrow type D (75%) and 62.5% of cases also had cortical venous reflux.

Procedural and follow-up information is summarized in table 2. DSA-Dynavision was used in five (62.5%) patients. Transfemoral venous access to the fistula was possible in seven (87.5%) patients, with one patient requiring direct surgical access to the ophthalmic vein for venous access. All patients reported resolution of their presenting symptoms on clinical follow-up. One patient continued to experience headaches on follow-up but his presenting symptom (proptosis) resolved.

Table 2

Summary of treatments, outcomes, and complications

Mean total coil length for patients where DSA-Dynavision was used for pretreatment planning was significantly shorter (130.5 cm vs 190 cm, p=0.034) compared with patients where DSA-Dynavision was not used for pretreatment planning. Patient No 5 had a complex tortuous access and took almost 5 hours (307 min out of a total procedural time of 412 min) to get to the inferior petrosal sinus. Controlling for the time for access in this patient, the mean procedural time for patients where DSA-Dynavision was used for pretreatment planning was significantly shorter (171.1 min vs 280.3 min, p=0.025) compared with patients where DSA-Dynavision was not used for pretreatment planning.


Our study examined the role of DSA-Dynavision in the pretreatment planning of endovascular treatment of indirect CCFs. This is the first time this has been reported in the literature. Use of DSA-Dynavision in pre-procedural planning for the treatment of indirect CCFs resulted in a shorter procedural time (171.1 min vs 280.3 min, p=0.025) and the use of shorter total coil length for treatment (130.5 cm vs 190 cm, p=0.034). This was achieved with long term treatment outcome similar to what has been reported in the literature. The rate of cure in our series was 8/9 fistulas (88.89%) (mean follow-up of 33 months), similar to the 81–90% reported rate in the literature.5 ,10 ,11 One patient who did not demonstrate clinical cure had persistent headache at the clinic visit 3 months later, but reported resolution of proptosis. Immediate angiographic occlusion was seen in five (62.5%) patients. Three patients had a small residual fistula at the end of treatment but none had any residual cortical venous reflux or residual ophthalmic venous drainage. Follow-up imaging revealed stable fistula occlusion in six (75%) patients and no patient demonstrated ophthalmic venous drainage at follow-up (mean follow-up time of 12.4 months). Of note, no patient in our series needed repeat endovascular intervention. Retreatment rates of 9–29% have been reported in the literature.5 ,10 ,11 Overall, our patients needed fewer coils than most patients reported in the literature to achieve similar, if not better, outcome.10

One (12.5%) patient in our study developed complications following the procedure: headaches, which began in the recovery room and required 24 hour admission to hospital and IV steroids. This rate of symptomatic complications was similar to previous studies.5 ,10 ,11 We had no patient with procedure related permanent morbidity, and no patient developed cranial nerve palsy or diplopia following the procedure, which had previously been documented in 6–14% of patients. Although this was not evaluated in our study, we believe that a lower rate of permanent morbidity and lack of new post-treatment cranial nerve palsy or diplopia may be explained by the shorter total coil length (overall mean length of 173.2 cm) used. Overpacking of the cavernous sinus has been a proposed mechanism of complications in previous studies.10 The large volume of coils inserted into the cavernous sinus often exceed 2 m in total length,10 causing high pressure, resulting in symptoms of cavernous sinus syndrome, similar to those from the fistula itself. The abducens nerve is at particular risk of compression as it is located just lateral to the ICA as opposed to within the lateral wall of the sinus, like the other cranial nerves.10 Use of DSA-Dynavision helped reduce the total coil length deployed and thereby allowed only partial coiling of the cavernous sinus, reducing complications related to compression of the cranial nerves.

Identifying the exact fistula site also helped us change our approach to treatment of indirect CCFs in our patients. In patient No 7 (figure 1), the diagnostic angiogram revealed a large fistula with arterial supply from both ECAs and ICAs, and predominantly left-sided venous drainage. Without information from DSA-Dynavision, this fistula would have been treated with dense transvenous coiling either on the left side or bilaterally. However, DSA-Dynavision images showed that the exact fistula site was in the right cavernous sinus. The foot of the vein in the right cavernous sinus was selectively catheterized and coiled. We believe the information from DSA-Dynavision helped in selectively treating the correctly identified fistula site, thereby reducing procedure time and coil length. This also helped prevent re-routing of the venous drainage and potential complication thereof if we would have coiled what appeared to be the left cavernous sinus fistula on routine angiogram.

Figure 1

(A) Right and (B) left internal carotid artery (ICA) angiograms demonstrate a large C-C fistula with predominantly left-sided drainage via the ophthalmic vein (solid arrow) and the inferior petrosal sinus (broken arrow). Analysis of DSA-Dynavision localized the fistula only in the left cavernous sinus. (C) Selective catheterization with microcatheter tip in the right cavernous sinus, confirming the location of the fistula (arrow) with left-sided venous drainage. (D) Immediate post-coiling angiogram demonstrating no residual fistula following coiling of the foot of the vein in the right cavernous sinus only (arrow). (E) Six month follow-up right and (F) left ICA angiogram demonstrating stable occlusion of the carotid cavernous fistula.

In patient No 6, the diagnostic angiogram showed bilateral ICA/ECA supply and bilateral ophthalmic venous drainage (figure 2). Without the information from DSA-Dynavision, the fistula site was not very clear. DSA-Dynavision revealed that this was not one fistula but in fact two separate fistulas, with a clearly identifiable fistula site in each cavernous sinus. Each fistula site was independently catheterized and selectively embolized. Although we considered the total coil length used to be long (243 cm), in reality coil length was for embolization of two fistulas and not just one.

Figure 2

(A) Right and (B) left internal carotid artery (ICA) angiogram demonstrates bilateral ophthalmic venous drainage (arrows). Analysis of DSA-Dynavision images confirmed two fistula sites—one on each side. (C) Selective microcatheter angiogram confirming a left-sided fistula (arrow) that was successfully coiled (arrow in D). (E) Confirmation of a separate right-sided fistula (arrow) on selective microcatheter angiogram. (F) Right ICA angiogram at the end of the procedure showing two separate coil masses (arrows) for two separate fistulas and no residual fistula. Three month follow-up left (G) and right (H) left ICA angiogram showing stable closure of the fistula.

Our study is significantly limited in the small sample size. However, despite this limitation, we have shown a very promising technique in the planning of treatment of rare but complex indirect CCFs. Our study was also limited by the suboptimal follow-up in four of our patients with CT angiography and MR angiography.

In conclusion, DSA-Dynavision in pre-procedural planning helped identification of the exact fistula site in patients with indirect CCFs. This facilitated selective coil embolization of the foot of the vein, resulting in shorter procedure time and use of shorter total coil length with similar outcome.


View Abstract


  • Contributors AB: substantial contributions to the acquisition, analysis, and interpretation of the data for this work. JJSS: substantial contributions to the conception and design of the work, acquisition, analysis, and interpretation of data for this work. AB and JJSS: major role in drafting the work and revising it critically for important intellectual content;approved the version to be published; agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Competing interests None declared.

  • Ethics approval The study was approved by Nova Scotia Health Authority Research Ethics Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.