Efficacy and safety of embolization for arteriovenous malformations of the basal ganglia and thalamus via the transarterial approach.

Background: To evaluate the effectiveness and safety profile of transarterial embolization in the treatment of brain arteriovenous malformations (bAVMs) within the basal ganglia and thalamus.
Methods: A retrospective clinical study was performed on 22 patients with bAVMs localized within the basal ganglia and thalamus who were treated with transarterial embolization (December 2012 and January 2019) in our center. The bAVMs were embolized via the transarterial approach with Onyx or Glubran according to the anatomical structure. A detachable or undetachable microcatheter was used in the procedure according to the length of the feeding artery. The data of these patients were retrospectively analyzed.
Results: Among the 22 patients, 9 bAVMs were located in the basal ganglia and 13 were located in the thalamus. Twenty patients presented with hemorrhage (90.9%), leaving 2 patients (9.1%) who had no symptoms. According to the Spetzler-Martin grading classification, 13 bAVMs (59.1%) were grade 3, 7 (31.8%) were grade 4, and 2 (9.1%) were grade 5. Procedure-related complications occurred in only 1 patient (4.5%). No deaths related to the operation occurred. All patients achieved anatomic stabilization and no bleeding was observed in the follow-up. Conclusions: Selective embolization via the transarterial approach is safe and effective for bAVMs originating within the basal ganglia and thalamus. Our results demonstrate a low rate of complications and an elevated degree of anatomical disruption in the endovascular treatment of bAVMs stemming from the basal ganglia and thalamus. 2022 Annals of Translational Medicine. All rights reserved.

Introduction

Brain arteriovenous malformations (bAVMs) consist of aberrant fistulas presenting across arteries and veins which lack any interconnecting capillary beds, leading to various symptoms including intracranial hemorrhage and seizure (1). Deep bAVMs often involve important structures such as the thalamus and basal ganglia and have a higher annual bleeding rate, disability rate, and mortality rate than bAVMs in superficial parts (2). The treatment strategies for bAVMs involving important structures are challenging and variable due to their unknown pathogenesis and complicated internal structure (3).

For neurosurgeons, the treatment of deep bAVMs is difficult and risky. Arteriovenous malformations carry a significant cumulative risk of bleeding. Treatment options include microsurgery, stereotaxic radiosurgery (SRS), embolization. Treatment selection is based on bAVMs characteristics, including size, location, venous drainage, surgical access, rupture status, and previous treatment. The Spetzler-Martin classification is the most commonly used classification system in selecting treatment. Microsurgery can achieve high cure rates with low morbidity and mortality, especially for I–II grade bAVMs (4). SRS is suitable for small or medium-sized bAVMs, unruptured bAVMs and residual bAVMs after operation or endovascular treatment. However, the effect is slow, and it takes 1 to 3 years for the brain AVM to completely disappear (5). Embolization was considered an adjunct to the treatment of bAVMs in the past and was used to assist microsurgery and SRS. With the development of embolic agents, catheter guide wires, etc, embolization as an independent treatment method is increasingly used in bAVMs. For deep bAVMs, microsurgery is difficult to locate accurately and may cause greater trauma. Radiation-related complications and rebleeding rates are relatively high with radiation therapy for deep bAVMs, while occlusion rates are lower. In such patients, embolization is suitable if there is a feeding artery or a draining vein that can reach the nidus (6). Compared to microsurgical resection, endovascular embolization is less invasive and results in less sequelae, and is becoming the first treatment option for more and more intracranial cerebrovascular diseases (7). However, at present, the efficacy and safety of transarterial and intravenous, or radiotherapy in the treatment of this kind of bAVM are still controversial (8,9).

There have been many studies on bAVMs, but they are not applicable to deep bAVMs such as those located in the basal ganglia and thalamus (10,11). Regardless of the choice of microsurgery, stereotactic radiation therapy or embolization, the treatment of deep bAVMs faces more risks. This study aimed to evaluate the efficacy and safety of single-center transarterial endovascular treatment of bAVM deep within the basal ganglia and thalamus. This investigation aimed at evaluating the effectiveness and safety profile of transarterial endovascular treatment of deep bAVMs within the basal ganglia and thalamus at a single center. We present the following article in accordance with the STROBE reporting checklist (available at https://atm.amegroups.com/article/view/10.21037/atm-22-384/rc).

Methods

This investigation retrospectively collected the data of patients with basal ganglia and thalamus bAVMs treated with embolization via the transarterial approach from August 2013 to January 2019. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the Ethics Committee of Renmin Hospital of Wuhan University (No. WDRY2021-K070) and informed consent was taken from all the patients or their guardians. Patient demographic data, medical history, medical manifestation/s, analytical scans, surgical procedure outcomes, post-surgery analyses, and medical monitoring data were collected. The bAVMs were categorized depending on dimensions, localization (basal ganglia or thalamus), and type of drainage (superficial-/deep-/deep- and superficial-venous), and were consequently classified according to the Spetzler-Martin grading classification. Clinical cases with pial arteriovenous fistulas, as well as those a previous history of treatment for this condition, were all excluded from this investigation. All study participants were preoperatively evaluated through head computed tomography (CT) and digital subtraction angiography (DSA) and were examined by the in-house multidisciplinary cerebrovascular team prior to and post intervention.

Clinical data and follow-up data were documented for each patient. Follow-up and evaluation of postoperative neurological function were performed at 12 months after embolization. All patients were followed up with head DSA after embolization. Major adverse events after embolization were assessed, such as novel ischemic stroke, intracranial bleeding, and death. Novel ischemic stroke was defined as a novel, symptomatic, thromboembolic incidence confirmed through magnetic resonance imaging (MRI). Intracranial bleeding was linked to embolization, encompassing intrasurgical vascular perforation, venous occlusion, and bleeding. The modified Rankin scale (mRS) was used to assess neurological recovery.

The purpose of the treatment was to correct the abnormal fistulas between arteries and veins that resulted in the obliteration of venous endothelial walls and consequently led to profuse internal bleeding of bAVMs. During the procedure, we prioritized the undetachable microcatheter (Marathon, EV3, Irvine, CA, USA) advanced with microwire guidance (MirageTM, EV3, USA) via the transarterial approach within the nidus for cases without a sufficiently long reflex feeding artery. For bAVMs with long feeders, a removable-tip microcatheter (ApolloTM, EV3, USA) was used from the transarterial approach to achieve curative results. Onyx was used to penetrate adequately into the nidus if the tip of the microcatheter could be advanced into the nidus, otherwise Glubran with a concentration between 15% to 20% would be used to embolize the nidus in cases of feeders with excessive tortuosity. The treatment was stopped when the arterial approach was no longer feasible or bAVMs could not be reached by total vascular area (TVA).

Statistical analysis

SPSS20.0 software was used was used for statistical analysis, and the mean ± standard deviation expresses the measurement data ().

Results

Patient profiles and outcomes are listed in . A total of 22 bAVM clinical cases, with symptoms stemming from the basal ganglia (n=9) and thalamus (n=13), underwent embolization. The clinical case group was 45% female (mean age 35±18 years). The majority of clinical cases manifested bleeding (91%; n=20), while 7 (32%) patients had worsening neurological condition (mRS >2) at admission. The mean size of bAVMs was 3.77±0.69 cm. Eight cases were grade 3, 11 cases were grade 4, and 2 cases were grade 5 according to the Spetzler-Martin grading classification. The total embolization rate was 40–100%, with an average rate of 86.19% in this series. Onyx was used in 18 (82%) cases and Glubran was used in 4 (18%) cases. are individual clinical cases.

Table 1

Investigational bAVM features and results

Patient	Age (years), sex	Initial mRS	Location	bAVM size (cm)	Presentation	Feeding artery	Spetzler-Martin grade	Venous drainage	Final mRS	Angiographic results	Embolic agents
1	16, M	1	Right basal ganglia	4	Ruptured	2	4	Deep	1	65%	Glubran
2	22, F	2	Left basal ganglia	4	Ruptured	3	4	Deep	1	100%	Onyx
3	24, F	2	Right thalamus	3	Ruptured	4	4	Deep	1	80%	Onyx
4	16, M	3	Right basal ganglia	1.5	Ruptured	1	3	Deep	1	100%	Onyx
5	40, M	1	Left basal ganglia	3	Ruptured	1	4	Deep	1	100%	Onyx
6	26, F	4	Right basal ganglia	2	Ruptured	2	3	Deep	1	100%	Onyx
7	64, M	3	Left thalamus	3.5	Ruptured	1	4	Deep	1	60%	Onyx
8	19, F	3	Left thalamus	2	Ruptured	2	3	Deep	1	100%	Onyx
9	38, M	1	Left thalamus	2.5	Ruptured	2	3	Deep	1	100%	Onyx
10	51, F	1	Left thalamus	2	Ruptured	1	3	Deep	0	100%	Glubran
11	35, F	1	Right thalamus	3.5	Ruptured	2	4	Deep	0	100%	Onyx
12	30, M	1	Right thalamus	3	Ruptured	2	4	Deep	0	90%	Glubran
13	42, F	3	Right thalamus	4	Ruptured	2	4	Deep	1	50%	Onyx
14	38, F	1	Right thalamus	4	Ruptured	2	4	Deep	1	40%	Onyx
15	29, F	1	Right thalamus	3.5	Ruptured	4	4	Deep	0	90%	Onyx
16	21, M	4	Right thalamus	2	Ruptured	2	3	Deep	1	100%	Glubran
17	38, F	1	Right basal ganglia	6	Ruptured	2	5	Deep	1	80%	Onyx
18	66, M	1	Left thalamus	2.5	Ruptured	1	3	Deep	0	100%	Onyx
19	16, M	3	Right basal ganglia	2	Ruptured	1	3	Deep	0	100%	Onyx
20	62, M	1	Right basal ganglia	6	Unruptured	1	5	Deep	0	40%	Onyx
21	67, M	1	Left basal ganglia	3	Unruptured	1	4	Deep	1	100%	Onyx
22	9, M	1	Right thalamus	6	Ruptured	2	5	Deep	1	80%	Onyx

bAVM, brain arteriovenous malformation; mRS, modified Rankin scale.

Figure 1

A 30-year-old clinical case manifested severe headache and left hemiparesis. (A) Nonenhanced CT imaging (axial view) demonstrated high density within the right caudate-capsular area with ventricular extravasation. Selective right ICA DSA from the AP (B) and lateral (C) views highlighted the AVM supply through the posteromedial choroidal arteries, with a nidus of 3.5 cm. Deep drainage was conducted through the Galen vein. (D,E) Three-dimensional DSA better delineated the angioarchitecture. (F) Intraoperative single-shot imaging from the lateral view exhibited the cast of Glubran (GEM Srl, Viareggio, Italy). (G,H) Twelve-month control DSA in the AP (G) and lateral (H) views demonstrated anatomical exclusion of nidus. CT, computed tomography; ICA, internal carotid artery; DSA, digital subtraction angiography; AP, anteroposterior; AVM, arteriovenous malformation.

Figure 2

A 35-year-old clinical case manifested severe headache. (A) Nonenhanced CT imaging (axial) demonstrated ventricular hemorrhage. Selective right vertebral DSA in the AP (B) and lateral (C) views highlighted the AVM supply through the posterior choroidal artery, with a nidus of 3.5 cm. Deep drainage was conducted through the Galen vein. (D) Three-dimensional DSA better delineated the angioarchitecture. Intraoperative single-shot imaging from the lateral (E) view highlights arterial methodology via a microcatheter (Marathon, EV3, Irvine, CA, USA) within the posterior choroidal arteries. Non-subtracted single-shot imaging from the lateral (F) view revealing the cast for Onyx (EV3, Neurovascular, Irvine, CA, USA). Twelve-month control DSA within the AP (G) and lateral (H) views highlighted anatomical exclusion of nidus. CT, computed tomography; DSA, digital subtraction angiography; AP, anteroposterior; AVM, arteriovenous malformation.

All patients achieved a positive clinical outcome (mRS 0–2) after 12 months follow-up. Only 1 (5%) patient presented visual field defects related to embolization. No bleeding or recurrence occurred during the follow-up.

Discussion

It is well known that bAVMs can occur in any part of the brain, with affected regions accounting for more than 60% of the cerebral hemisphere. The first symptoms include headache, epilepsy, and cerebral hemorrhage. bAVMs that occur in deep structures of the brain, such as the basal ganglia or thalamus, account for about 1% to 3% of the total incidence (12). The incidence of hemorrhage in basal ganglia and thalamus bAVMs was shown to be significantly higher compared to superficial bAVMs (13). Within this investigation, 20 basal ganglia and thalamus bAVM cases were admitted with intraventricular hemorrhage, and 2 patients had asthenia as the first symptom.

Rebleeding basal ganglia and thalamus bAVMs are the greatest threat in such patients. The strategies for bAVM treatment in the basal ganglia and thalamus include endovascular embolization, radiosurgery, and microsurgery, as well as the combination of each of them. Microsurgery can clear hematomas and relieve the space-occupying effect, remove the lesions completely, and eliminate the risk of bleeding. In this way, this approach may still be beneficial for bAVMs of the basal ganglia and thalamus (4). Several approaches such as transsylvian fissure and the transcallosal and transcortical approach are usually used to achieve total resection of bAVMs in the basal ganglia and thalamus, especially for patients with hemorrhagic presentation and young age (14). For compact and small size bAVMs of the basal ganglia, microsurgical total resection can be achieved because the bAVMs are supplied by the middle cerebral artery (MCA) branch, which provides early accessibility to the neurosurgeon to seal supplying artery mass (15). However, bAVMs of the thalamus are also supplied by the anterior choroidal artery, thalamoperforating arteries (from P1 posterior-cerebral/posterior-communicating arteries), and posterior choroidal arteries (lateral > medial), which block accessibility to supplying arteries until late dissection phase. As a result, when dealing with bAVMs of the basal ganglia and thalamus, it is hard to avoid damaging surrounding tissues and nerves, and may even lead to massive hemorrhage. Potts reported that although 71% to 94% of insular cortex and basal ganglia bAVMs could achieve total resection, total resection could only be achieved in 46% of thalamus bAVMs. The incidence of complications was 29% (14/48), and the complications were mainly epilepsy, hydrocephalus, and infection (16). Thus, basal ganglia and thalamus bAVMs require multi-strategy treatment compared to other bAVMs.

Radiosurgery comprises part of this treatment, and is recommended for patients with deep or functional area bAVMs or bAVMs that cannot be resected completely. A total of 23% to 44% of bAVMs in the basal ganglia and thalamus were treated with stereotactic radiosurgery (SRS) because of its non-invasiveness. To date, it is well recognized that SRS is suitable for small or medium-sized bAVMs, unruptured bAVMs, and residual bAVMs after surgery or endovascular treatment. It was reported that the disruption rate was 43% to 81% for deep bAVMs which were treated by SRS. However, the disruption is achieved 6 months later, which has a hemorrhagic rate of 8% to 15% during the latency period and is not recommended for ruptured bAVMs (17).

Endovascular embolization is suitable for basal ganglia and thalamus bAVMs as this treatment prevents trauma to the parenchyma caused by microsurgery and SRS. Previously, embolic agents including solid particles such as polyvinyl alcohol (PVA) and sutures, as well as liquid agents such as N-butyl-2-cyanoacrylate (NBCA), had low rates of curative outcome due to their high level of complications. Therefore, endovascular embolization can resolve the condition in a minority of patients and is also a corollary to surgical procedures/SRS (18). With advancements in microcatheter technologies and new embolism-related drugs over the past decade, interventional therapy has become safer and more efficient for bAVMs and is becoming the first choice for nonbleeding cerebrovascular diseases (19). NBCA was the most commonly used agent before Onyx was introduced in clinical practice. The most serious disadvantage of NBCA is the risk of adhesion to the microcatheter after injection, which requires rapid removal of the microcatheter as soon as the glue retrograde, even if the bAVM has not been embolized completely. Onyx, a nonadhesive liquid agent, has a lower risk of catheter entrapment, allowing it to be injected slowly. Moreover, its good perfusion characteristics mean it can penetrate into the nidus adequately, which improves the rate of total embolization and decreases recurrence (20).

Transarterial bAVM embolization is a standalone treatment, with the advantages of reaching the nidus easily, less risk of adhesion to the microcatheter, and staging of embolization (21). The transvenous route is deemed as a rescue treatment. Relative indications for transvenous embolization include small nidus (<3 cm maximum diameter), single draining vein, absence of a navigable arterial feeder, and nil access to the nidal remnant. However, the transvenous approach may cause premature occlusion of the drainage vein and perforation (22). In this series, the microcatheter could reach all of the niduses by the feeding artery, so the transarterial approach was chosen in all cases for safety.

For basal ganglia and thalamus bAVMs, the main feeding arteries usually include perforations within the anterior cerebral artery (ACA), MCA, and posterior cerebral artery (PCA). Thus, embolic agents without good penetration characteristics cannot penetrate into the nidus well. Combined with a detachable microcatheter such as the Apollo or Sonic, Onyx can be injected very slowly and diffuse into the nidus fully without fear of Onyx reflux and microcatheter entrapment. In our series, Onyx was used when the microcatheter could be introduced into the nidus, otherwise different concentrations of NBCA were used to embolize the feeding artery and nidus. In one bAVM with 3 feeding arteries, total embolization was achieved with the combination of Onyx and the Apollo detachable microcatheter. For 2 bAVMs with 4 feeding arteries, to avoid occluding the drainage vein and the perforating arteries, we partially embolized the nidus. The follow-up results showed that there was no rebleeding. To avoid complications, the microcatheter tip should be introduced within the nearest proximity to the nidus and without letting the microwire come out when the microcatheter is near the nidus, otherwise there is a risk of puncturing the bAVM.

Because most of the feeding arteries are from the anterior choroidal artery perpendicular to the internal carotid artery (ICA), it is difficult to introduce the microcatheter because of the acute angle between the parent artery and feeding artery. Thus, a balloon can be used to assist the microcatheter as it goes into the nidus. In this way, embolization from the venous approach is an alternative method for patients with bAVMs in the basal ganglia and thalamus which cannot be occluded from arterial approach, although there was no such case in our series (23).

Conclusions

This is a retrospective clinical study of endovascular treatment for bAVMs of the basal ganglia and thalamus. Embolization treatment appears to have a good safety profile for basal ganglia and thalamus bAVMs. Our results demonstrate a low rate of complications. Under the appropriate treatment strategy, intravascular treatment is safe and reliable as a single treatment approach for basal ganglia and thalamus bAVMs.

The article’s supplementary files as