Transvenous embolization of vein of galen aneurysmal malformations with coils as a final procedure for cure: A single-institution experience of 18 years

Abstract

Introduction

Staged, transarterial embolization (TAE) is currently considered the gold standard for the treatment of vein of Galen aneurysmal malformation (VGAM); however, as transarterial access becomes restricted, further staged TAE may become ineffective or carry an increased risk of hemorrhagic or ischemic stroke when attempting complete obliteration.

Objective

To describe the first consecutive, retrospective series of VGAM treated with transvenous embolization (TVE) with coils alone, as the final treatment in staged endovascular therapy, at a single institution between January 2004 and September 2021.

Results

A total of 10 patients with a median age of 5.5 (IQR: 9.25) years were treated with coiling TVE. Patients were treated with a median number of 5 (IQR: 2.75) TAEs prior to the final TVE treatment. Complete or near-complete immediate angiographic obliteration was achieved in eight patients. Immediate post-procedural (within 48 h) hemorrhagic complications were noted in two patients (20%), one of whom passed away while the second suffered from hemiparesis. Stereotactic radiosurgery was performed in two patients with incomplete obliteration after TVE. The median follow-up time after TVE was 17 (IQR: 9) months. At long-term follow up (17 months) for the remaining nine patients, all VGAMs were completely obliterated. Long-term clinical deterioration compared to pre-TVE was noticed in one case.

Conclusion

Transvenous coil embolization is a technically feasible but risky option, as a final-stage treatment for cure of VGAMs with restricted trans-arterial access. Although TVE with coils remains an effective therapeutic modality, we recommend continuing investigation of safer TVE techniques to achieve cure.

Introduction

Vein of Galen aneurysmal malformation (VGAM) is a rare, congenital vascular malformation (VM) characterized by abnormal arteriovenous communication from choroidal arteries to the median prosencephalic vein of Markowski (the precursor of the mature vein of Galen). ¹ It is found in 1 out of 58 thousand live births ² and still remains the most common, prenatally detected, cerebral VM. Left untreated, VGAM may be associated with neonatal morbidity and mortality, mostly attributed to congestive heart failure (CHF) or neurologic sequelae. ³

Treatment for VGAM is typically performed across several staged embolizations over a period spanning from identification of the malformation (most commonly, neonatal or fetal period) to as far as adolescence. ⁴ Staged transarterial embolization (TAE) is considered the gold standard therapy, achieving complete or near-complete angiographic obliteration in 80% of cases in our practice, with good long-term clinical outcome.^4–7 The success of TAE is strongly dependent on the morphology of the arterial feeding supply and there is a subset of patients for whom VGAM obliteration with TAE alone is ineffective or risky. In instances where the residual arterial supply is through small, trans-mesencephalic feeders, or a fine, plexiform network, total obliteration through TAE is technically difficult. Restriction of transarterial access is the most commonly described indication for transvenous embolization (TVE) in the literature.^6,8 In these cases, the transvenous approach to coil off the venous pouch or the fistulas appears to be the technically simplest choice once the venous component becomes smaller and more amenable to coiling. ⁸

So far, the risks of this practice have not been widely studied. Herein, we describe our experience regarding the transvenous coil embolization approach as the final-stage endovascular treatment for cure of VGAMs previously treated with multiple rounds of TAE.

Methods

Patient selection

A retrospective review was conducted of our endovascular database to identify all VGAM patients who underwent TVE as the final-stage procedure between January 2004 and September 2021. Immediate angiographic outcome, long-term angiographic outcome, and long-term clinical outcomes were the primary outcomes of our cohort. This study was approved by the appropriate institutional review board under IRB-STUDY-21-00749 including a waiver of informed consent due to the retrospective nature of the study.

We collected information regarding the patients’ VGAM classification, age at the time of the TVE procedure, number of previously attempted TAE sessions, pre-procedural clinical exam, intraoperative technique of occlusion, immediate and long-term angiographic outcomes, postoperative hemorrhagic complications, follow-up time since the TVE procedure, additional treatments that were performed in order to achieve complete obliteration of the VGAM, as well as long-term clinical outcome and mortality.

Technique description of TVE coiling

Initially, full diagnostic cerebral angiography was performed. Here, we ensured that none of the internal cerebral veins (ICVs) drained into the malformation. When the ICV was visible in the venous phase, especially in the carotid artery injection, the mask image was selected in the venous phase and subtracted from the arterial phase where the dilated vein of Galen was clearly seen in the lateral image, so that we could delineate the morphological relationship between the malformation and the ICV. Once the coil-amenable venous segment was confirmed, a 6F guiding catheter, with femoral approach, was navigated into either sigmoid or transverse sinus in the dominant side. One or two 0.021-inch microcatheters were used to access the segments using the roadmap from the arterial angiogram. The venous sac was coiled using detachable coils. Fibered coils were preferably added to achieve better occlusion.

Results

In total, 124 VGAM patients with choroidal or mural type were treated at our institution between January 2004 and September 2021, 10 (8.1%) of whom underwent transvenous embolization with coiling once the TAE approach was deemed to carry too high risk or be technically difficult. The median age of the cohort was 5.5 (Interquartile Range-IQR: 9.25) years. Patients were treated with a median number of 5 (IQR: 2.75) TAEs prior to final TVE treatment. Two patients presented with significant complications from previous interventions conducted at outside institutions which resulted in moderate developmental delay prior to our endovascular intervention.

Complete or near-complete immediate angiographic obliteration was achieved in 8/10 (80%) patients. Immediate post-procedural (within 48 h) hemorrhagic complications were noted in two patients treated with coils: one patient expired during hospitalization while the second patient suffered from motor delay and hemiparesis. Stereotactic radiosurgery (SRS) was performed in two patients with incomplete obliteration post-procedurally. The median follow-up time after TVE was 17 (IQR: 20) months. At long-term follow-up of the remaining nine patients, all VGAMs were completely obliterated. Long-term clinical deterioration compared to pre-TVE was noticed in one case. In total, six patients developed normally, three experienced a moderate developmental delay with hemiparesis, and one patient expired (Table 1).

Table 1.

Clinical characteristics and treatment outcomes of our cohort.

Patient	VGAM Classification	Pre-Treatment Examination	Age at TVE (y)	Treatments before TVE	TVE technique	Immediate Angiographic Outcome	Long-Term Angiographic Outcome	Follow-up Interval Time From TVE	Post-Operative Bleeding	SRS	Clinical Outcome
1	Choroidal	Normal	15	2	Coiling	Near-Complete Obliteration	Complete Obliteration	36 Months	No	No	Normal
2	Choroidal	Normal	9	5	Coiling	Incomplete Obliteration	Complete Obliteration	17 Months	No	Yes	Normal
3	Mural	Normal	4	7	Coiling	Complete Obliteration	Complete Obliteration	18 Months	No	No	Normal
4	Choroidal	Normal	5	11	Coiling	Near-Complete Obliteration	Complete Obliteration	9 Months	Yes	No	Moderate Delay and Hemiparesis
5	Choroidal	Normal	5	6	Coiling	Complete Obliteration	Complete Obliteration	16 Months	No	No	Normal
6	Choroidal	Normal	18	5	Coiling	Complete Obliteration	N/A*	N/A*	Yes	No	Expired
7	Choroidal	Normal	20	11	Coiling	Incomplete Obliteration	Complete Obliteration	57 Months	No	Yes	Normal
8	Choroidal	Normal	1	4	Coiling	Near-Complete Obliteration	Complete Obliteration	17 Months	No	No	Normal
9	Choroidal	Treatment-related stroke;  Moderate delay; Hemiparesis	1	4	Coiling	Near-Complete Obliteration	Complete Obliteration	36 Months	No	No	Moderate Delay and Hemiparesis
10	Choroidal	Severe delay; Hemiparesis	6	4	Coiling	Near-Complete Obliteration	Complete Obliteration	6 Months	No	No	Moderate Delay and Hemiparesis

*Patient 6 expired due to hemorrhagic complication 48 h post-procedurally.

Demonstrative cases

Case 8

A choroidal VGAM was identified in a female neonate during workup for congestive heart failure. Initial TAE treatment took place at the age of two months due to failure to thrive (Figure 1(a)). In total, she underwent four sessions of TAE with n-BCA. A small residual venous pouch, mainly fed by thalamoperforators was seen in the angiogram at the age of 19 months (Figure 1(b) and (c)). The fistulous venous pouch was embolized with detachable platinum coils and fibered coils via TVE (Figure 1(d)) and near-complete obliteration was achieved (Figure 1(e)). In the PICU, strict blood pressure control and frequent neurological exams were performed for 48 h post procedurally. Complete occlusion of the VGAM was confirmed in the follow up angiogram (Figure 1(f) and (g)) at the age of three.

Figure 1.

A choroidal type VGAM before the first TAE treatment (a) After 4 sessions of TAE with n-BCA, left CCA (b) and right VA angiograms (c) revealed a small residual shunt mainly fed by thalamoperforators at the age of 19 months (red arrows in b and c indicate the falcine sinus; yellow arrow in c indicates the small residual venous pouch of VGAM). The fistulous venous pouch was embolized with detachable platinum coils and detachable fibered coils (d; green arrow demonstrates the coil mass). Near complete occlusion was achieved (e) Complete occlusion of the VGAM was confirmed in the follow up left CCA (f) and right VA angiograms (g) at the age of 3. VGAM: Vein of Galen aneurysmal malformation, TAE: Transarterial embolization, CCA: Common carotid artery, VA: Vertebral artery.

Case 5

A choroidal VGAM (Figure 2(a)) was diagnosed prenatally in female fetus the day before delivery. In total, she underwent 11 sessions of TAE with n-BCA, after which no safe arterial access could be achieved in order to embolize the small residual VGAM. This residual was fed by thalamoperforators and a choroidal, arterial network (Figure 2(b)). As a result, the patient underwent transvenous coil embolization at the age of five. Two fistulous venous pouches were embolized with detachable and fibered coils (Figure 2(c)). Near complete occlusion was achieved (Figure 2(d)). In PICU, strict blood pressure control and frequent neurological exams were performed after the procedure until she was found unresponsive 36 h post-embolization. Immediate CT scan showed a left thalamic hemorrhage extending into the midbrain with intraventricular extension and acute hydrocephalus (Figure 2(e)). External ventricular drain was placed emergently and she underwent VP shunt placement later. The patient underwent aggressive rehabilitation and her language/memory improved significantly, but had persistent right hemiparesis. Follow-up MRI T2-weighted images showed post-hemorrhagic changes in the left thalamus (Figure 2(f)). Follow up angiogram at age of eight showed total obliteration of the VGAM (Figure 2(g)).

Figure 2.

Initial vertebral artery angiogram revealed a choroidal type VGAM (a). After 11 sessions of TAE with n-BCA, no more safe arterial access for TAE was seen (b) Two fistulous venous pouches were embolized with detachable fibered coils (c; yellow arrows indicate microcatheters in the falcine sinus and green arrow shows the coil mass). Near complete occlusion was achieved, with stagnation in the late venous phase (d) A post-operative CT scan showed left thalamic hemorrhage with intraventricular extension (e) Follow-up MRI T2-weighted images showed post-hemorrhagic changes (red arrow) in the left thalamus (f) and a follow-up angiogram at the age of eight revealed complete obliteration of the VGAM (2). VGAM: Vein of Galen aneurysmal malformation, TAE: Transarterial embolization.

Discussion

Summary of our findings

To our knowledge, this is the first consecutive series of VGAM patients treated with TVE coiling after multiple rounds of TAE. This technique was utilized in an attempt to cure partially obliterated VGAMs with restricted transarterial access, previously treated with the TAE approach. According to the authors, TVE coiling appears to carry a significant technical benefit under these circumstances. Once the fistula has shrank enough, coil embolization becomes technically easy and coils can be precisely placed at the desired anatomic location within the VGAM, in order to obliterate the fistula, leaving normal vascular territory unaffected.

From this study, a few important points can be made: Firstly, there is a high chance of incompletely obliterating the VGAM if only coils are used. Secondly, the postoperative risk of bleeding can be high. Within our cohort, the prevalence of post-procedural hemorrhage was 20% (2/10). In order to minimize the risks emerging from TVE coiling, the following aspects may be considered: In instances where normal deep venous tributaries drain in the VGAM, TVE coiling is deemed unsafe because of the risk of hemorrhage or venous infarction. Meticulous preoperative assessment of the venous anatomy is essential in order to identify and preserve these veins. Notably, a severely dilated VGAM is considered unfavorable, so we shrank the VGAM as much as possible with multiple TAEs before TVE with coils. Lastly, the fine vascular network formation around the VGAM may be the risk for hemorrhagic complication.

In this report, the authors want to emphasize the potential postoperative hemorrhage risk of TVE coiling, related to all the aforementioned factors, which may decrease the future adoption of this technique. Especially after the fine vascular network develops around the fistula, closing the vein simply with coils without addressing the small vessels seems to be associated with a higher risk of delayed hemorrhage. On the contrary, when the fistula or the arterial pedicle is the target (TAE), this technique is simple, effective, and safe. In our practice, closing a VGAM encompasses a dual goal: Preventing/reversing delay in developmental milestones and preventing the risk of future intracranial hemorrhage and as such, completely shutting the fistula flow is considered clinically desirable. The goal of the interventionists for all cases presented in this cohort was to close the fistula, in order to eliminate the risk of delayed future bleeding or VGAM enlargement due to fine vascular network formation. In addition, since VGAM is a fistulous arteriovenous shunt disease without “nidus” (unlike AVM), and since the fistulous point is easily accessible, TVE theoretically seemed a technically feasible option for the cases with restricted transarterial access. The endpoint of TVE is to densely pack the fistula using detachable and/or fiber coils within a safe segment, in a way that does not compromise the flow within the vein of Galen.

Some of the patients of our cohort had a certain degree of developmental delay and some did not. One of the fatal cases was a young adult who returned to our practice in a decade after multiple transarterial embolizations but with residual developmental delay. In one of our previous case series of VGAM, mortality was 4.4%, 66.6% of patients were neurologically and developmentally intact, while the rest had mild-to-moderate impairment of function. ⁵ As discussed earlier, the interventionalists’ primary goal was to obliterate the fistula, in order to eliminate the risk of any future hemorrhage or neurological deterioration. The benefits and risks of TVE were always discussed with the family before the procedure. Emphasis was given to the distinct nature of TVE as well as the risks inherent to the procedure and informed consent was obtained prior to the procedure in all cases, after the family agreed that the benefits of surgery supersede the risks.

Prior literature experience on TVE

In a meta-analysis by Brinjikji et al., it was shown that endovascular embolization of VGAMs is associated with a good clinical outcome in more than 60% of cases. Neonatal (vs. infantile) age and lack of CHF were important predictors of good clinical outcome. ⁴ Historically, the first attempt of a transvenous approach for a VGAM traces back to Mickle and Ronald. who performed coiling after achieving transtorcular venous access through a small craniotomy.^9,10 In this study, even though the symptoms of the subjects’ CHF were controlled, the ultimate neurological outcome was poor (consistent with unpublished data from our institution), for which reason we don't take this approach nowadays. Dowd et al. achieved transfemoral vein access with subsequent retrograde catheterization of the feeding artery and closure as close as possible to the arteriovenous shunt using coils. ¹¹ Notably, Casasco et al. reported primary transvenous coiling (Gianturco Coils) of previously untreated VGAMs, achieving partial or total obliteration without any reported postoperative hemorrhage and with resolution of clinical symptoms on follow-up. ¹² The “kissing technique” was described by Melia et al., encompassing a combination of transarterial and transvenous coiling in order to ensure obliteration of the fistula. ¹³ Unfortunatley, this technique is difficult to apply to cases of VGAM wherein arterial feeders remain small.

Transarterial embolization and transvenous embolization carry unique advantages and drawbacks dependent on context. For example, a notable adverse event encountered with TAE is thalamic stroke. Fine, vascular network formation has been observed in cases of incomplete VGAM obliteration resulting in secondary recruitment of thalamoperforators which complicates subsequent TAE sessions. This is likely due to the fragility of the newly identified feeder vessels. Embolizing through these feeders may be ineffective and carries considerable risk as it can cause mechanical rupture and intracerebral hemorrhage, thalamic stroke, or proximal reflux of the nBCA. ¹³ Prominent risks of TVE include postoperative hemorrhage, cerebral edema and venous infarct.^6,8,11 The most common cause of these complications is occlusion of normal deep venous tributaries draining into the wall of VGAM. Ultimately, TVE was reported to be inherently associated with higher mortality and morbidity risk compared to TAE.^8,11 Nevertheless, when most of the arterial flow has been reduced and the remaining supply is derived from trans-mesencephalic perforators, TVE becomes an attractive option. In these cases, transarterial catheterization becomes more challenging and carries a higher risk compared to transvenous navigation to the fistulous point.

The adjunct use of liquid embolic agents

Transvenous embolization of VGAMs with liquid embolic agents (LEAs) was limited to individual cases and case reports until recently.^14–16 Orlov et al. applied this practice in VGAM cases with restricted transarterial access with or without the adjunct use of coiling. ⁸ In addition, creating a plug allows for slower and safer injection of a higher volume of Onyx, giving the ability to embolize the small arterial feeders in a retrograde fashion. ¹⁵ Contrastingly, TVE performed purely with coils does not address these smaller feeders and the subsequent increase in pressure within the malformation can lead to rupture of these fragile feeders. Our practice has also reported the Chapot Pressure Cooker Technique, which serves as an additional TVE alternative to coiling alone. ¹⁷

Limitations

The sample size of this study is notably small and there is a lack of a formal control group. While comparison to the efficacy of other well established techniques may be sufficient, the collective literature on this subject would benefit from a systematic review. This sample stems from a single institution and it would be prudent to perform a future analysis from a wide array of hospitals and interventionalists. This is an inherently challenging and risky procedure and it is important to recognize that other neurointerventionists may encounter other complications and technical challenges. While the eligibility criteria are well defined in this study, the threshold for this operation and outcomes may differ across different practices. All the subjects within our cohort received the same treatment and under the same indications, thus achieving homogeneity, but this may be difficult to apply to a wider analysis. Overall, this small-sample cohort study serves as a benchmark for the development of new endovascular techniques for the cure of challenging VGAM cases.

Conclusion

To the best of our knowledge, this is the first retrospective cohort study of TVE purely with coiling as a final treatment for VGAMs after multiple staged TAE sessions. We propose that TVE is a technically feasible option as a final-stage attempt for cure, particularly in cases where the dilated VGAM constricts over time to become more amenable to coiling. The risk of hemorrhage and greater potential for incomplete obliteration of the malformation suggest that TVE coiling should be performed in conjunction with LEAs rather than on its own, since it may increase postoperative hemorrhage rates. Further research is currently taking place to identify techniques that may ensure lower post-operative intraparenchymal hemorrhage risk.