Management of scalp arteriovenous malformations: a rising trend towards percutaneous direct puncture embolization technique—our experience

Introduction

Arteriovenous malformations (AVM) are a group of vascular malformations with direct communication between arteries and veins without a normal intervening capillary network. The intervening area where abnormal arterial and venous communication occurs is called the nidus (1). Scalp AVM has also been termed as cirsoid aneurysm since 1833 by Brescht due to its resemblance to venous varix (2,3). These are high-flow vascular malformations with enlarged tortuous feeding arteries and aneurysmally dilated draining veins (4). They present with various clinical complaints ranging from local pain, pulsating mass, bruit, tinnitus, and thrill to other less common presentations such as hemorrhage, skin ulceration, and rarely cardiac failure (5–8). They appear as a visible deformity and hence have a significant cosmetic and psychosocial impact on the patient.

Scalp AVMs are relatively rare clinical entities and require multimodality treatment methods for optimum management and satisfactory outcome. Traditionally, these AVMs were largely managed by surgical excision or ligation of the feeding arteries; however, this approach involved the risk of excessive blood loss, recurrence of the lesion, and less desirable cosmetic results (7). With growing awareness and technical advancements, endovascular or percutaneous embolization alone or as an adjunct to surgical excision has evolved as the preferred treatment with a safe and effective outcome (8–12).

We present a study of 50 patients who underwent embolization of scalp AVM by percutaneous/endovascular approach at our institute. We wish to discuss changes in trends of our practice from initially being a predominantly endovascular approach to gradually inclining towards percutaneous management of these lesions. The aim of the present study was to document the technical and clinical outcomes of patients undergoing endovascular/percutaneous embolization of scalp AVMs.

Material and Methods

We retrospectively evaluated 50 patients who underwent embolization (percutaneous/endovascular) of scalp AVM during 2010–2019 at a tertiary care center. The study was carried out with the approval of the institutional ethics committee. The pre-procedure evaluation included reviewing files for demographic data and other clinical details, localization of AVM, documentation of size of AVM, reviewing computed tomographic angiography (CTA), and pre-procedure digital subtraction angiography (DSA) images, where available, and assigning clinical stage according to Schobinger clinical classification system. The patients, who had already undergone any previous surgical or endovascular treatment elsewhere, were not included in the study. CT angiogram or magnetic resonance angiogram (MRA) was used as the initial diagnostic method in all the patients (Fig. 1). For the purpose of analysis, the lesions were divided into two groups based on the maximum diameter of the nidus: group 1 = small lesions (<3 cm); and group 2 = large lesions (>3 cm). In patients, where the lesions were deep-seated or where there was a suspicion of deep/dural venous drainage, diagnostic cranial DSA was performed to map the arterial feeders and venous drainage for planning the embolization procedure accordingly.

OPEN IN VIEWER

Fig. 1.

(a) Non-contrast CT scan image showing soft tissue scalp thickening in the occipital region on the left side, (b) contrast CT image shows enhancement of the lesion with the presence of multiple dilated vessels within, (c) volume-rendered CECT image shows large scalp AVM, predominantly supplied by superficial temporal artery with nidus in the occipital region on the left side. AVM, arteriovenous malformation; CECT, contrast-enhanced computed tomography; CT, computed tomography.

Before the embolization procedure, all the patients in our series received seven days course of broad-spectrum antibiotics coverage, starting two days before the procedure and continuing until five days after the procedure to prevent infective complications. All patients underwent AVM embolization by endovascular or percutaneous direct puncture technique or both and the procedures were performed in the DSA suite (Allura Xper FD20 Single plane; Philips health care, Amsterdam, Netherlands).

The post-procedure follow-up was done in the form of clinical examination at three and six months along with a Doppler evaluation to look for recurrence. Reappearance of the mass or pulsation or both was considered a clinical sign of recurrence. An increase in the size of the mass and/or reappearance of vascular channels on Doppler were considered imaging signs of recurrence.

Percutaneous method of embolization

The “palpate and puncture method” was used to localize the feeding artery and the draining vein. The point of the maximum thrill was first identified, which corresponded to the AVM nidus site. All linear vascular structures leading to and away from the lesion were identified. The compression of the in-flow artery at a point closest to the nidus resulted in a near-total decrease in the thrill of the lesion. Pre-procedure ultrasound (USG) and fluoroscopy images were also used to localize the nidus. Circumferential manual compression was applied with the help of a compression ring. The compression ring was a sterilized plastic ring that was applied in a way that the antegrade draining veins were compressed by the ring manually (Fig. 2). With all aseptic precautions, the feeding artery closest to the nidus or the nidus itself was directly punctured with a 20-G butterfly needle and an angiogram was taken to look for adequacy of the compression applied. Once the compression applied was adequate and outflow veins were blocked, embolization was done with an injection of n-butyl cyanoacrylate (n-BCA) and lipiodol solution. The total volume of embolizing agent to be injected was decided on the basis of the percutaneous angiogram taken after applying manual compression. The ratio of n-BCA:lipiodol was in the range of 1:1–1:3 during the percutaneous procedure depending upon the flow rate within the AVM and the adequacy of the compression applied. The embolization was stopped as soon as the embolizing agent showed any reflux beyond the boundaries of the compression ring. In case of a large lesion, more than one percutaneous needle puncture was done to adequately cover the malformation in the same sitting. The absence of thrill after embolization was considered a technical success of the procedure.

OPEN IN VIEWER

Fig. 2.

Percutaneous embolization of scalp AVM: (a) clinical image showing scalp swelling in the occipital region; (b) USG Doppler image shows dilated vascular channels within the lesion; (c) plastic compression ring sterilized and wrapped in povidone-iodine soaked gauze; (D) placement of compression ring and percutaneous direct puncture of AVM; (e) DSA image via scalp vein needle showing multiple vascular channels within the lesion; (f) post-embolization angiogram showing glue cast within the lesion and non-opacification of the nidus. AVM, arteriovenous malformation; DSA, digital subtraction angiography; USG, ultrasonography.

Endovascular method of embolization

A 5-F vascular sheath was inserted through the right common femoral artery to maintain access. With the help of a 5-F diagnostic catheter, the supplying external carotid artery (ECA) was cannulated for diagnostic angiography (Fig. 3). Selective cannulation of the arterial feeders (branches of the ECA) supplying the lesion was done with a microcatheter (2–3 F size). The volume of embolizing agent required to occlude the nidus was calculated with the help of pre-embolization diagnostic angiograms. A prior 5% dextrose flush was done to free the catheter lumen completely from the ionizing blood and contrast and this was immediately followed by an injection of n-BCA:lipiodol solution. The ratio of n-BCA:lipiodol was in the range of 1:3–1:4 during the endovascular procedure that depended upon the distance of the microcatheter tip from the nidus and flow rate through the AVM. A post-procedure check angiogram was taken via the diagnostic catheter in ECA to assess for any residual filling. If ECA feeders and the nidus were not filling post-endovascular embolization, then it was considered a technical success of the procedure.

OPEN IN VIEWER

Fig. 3.

Endovascular TAE of scalp AVM: (a) clinical image showing a large swelling over the left temporoparietal region; (b) contrast-enhanced axial CT image shows an exophytic soft tissue mass with multiple dilated vascular channels within; (c) DSA images of left ECA shows dilated tortuous STA supplying the lesion with the presence of a feeding vessel large aneurysm close to the nidus; (d) post-embolization angiogram shows occluded superficial temporal artery, aneurysm, and the nidus with the presence of n-BCA cast within the lesion. AVM, arteriovenous malformation; CT, computed tomography; DSA, digital subtraction angiography; ECA, external carotid artery; STA, superficial temporal artery; TAE, trans-arterial embolization.

The endovascular approach was preferred over the percutaneous method to embolize certain lesions, especially:

If fine arborization of feeding arteries was seen in the AVM, then precise percutaneous needle placement was difficult;

When the lesion was deep-seated and located in areas where manual compression cannot be applied efficiently, such as lesions extending to the pinna or eyelid;

As an adjunct to direct puncture embolization when complete to near total embolization could not be achieved by direct puncture technique.

Results

A total of 50 patients (27 men, 23 women) were included in the study. Three patients had a history of previous trauma to the scalp, 45 patients had complaints of significant swelling, 38 had cosmetic and psychosocial concerns, nine had episodes of bleeding, seven had headaches, five had vision difficulties, and four patients had complaints of an irritating bruit (Table 1). The majority of the lesions were seen in the occipital region (13/50, 26%). Other common locations were the parietal, temporal, and frontal regions. More than one area of involvement was seen in 15 (30%) patients. Of the patients, 42 (82%) had class II lesions and 8 (18%) had class III lesions as per Schobinger's clinical classification system. Thirteen patients had small AVMs (size <3 cm) and 37 patients had large AVMs (size >3 cm).

OPEN IN VIEWER

Table 1.

Clinical presentation of scalp arteriovenous malformation.

Symptoms	No. of patients
Scalp swelling	45
Cosmetic and psychosocial concerns	38
Episodes of bleeding	9
Headache	7
Vision difficulties	5
Irritating bruit	4

The pre-procedure CTA/DSA images showed that the superficial temporal artery (STA) was supplying the AVM in a majority of cases (38/50, 76%) followed by the occipital artery (22/50, 44%). Other arterial feeders included branches of the posterior auricular artery (12/50, 24%) and infraorbital branches of the internal maxillary artery (5/50, 10%). In one patient with frontal AVM, anterior ethmoidal branches of the ophthalmic artery were seen supplying the lesion apart from STA. Drainage was most commonly into external jugular veins through multiple tortuous dilated venous collaterals.

All 50 patients underwent embolization and 36 patients underwent post-embolization surgery with excision of the embolized nidus and placement of thick split-thickness skin graft harvested from the thigh (Fig. 4). A total of 20 (40%) patients underwent endovascular trans-arterial embolization (TAE) and 28 (56%) patients underwent percutaneous embolization of the AVM. Two patients needed additional TAE in addition to the percutaneous procedure to achieve complete obliteration of the lesion in the same sitting. Complete obliteration of the scalp AVM was seen in 28 patients in the percutaneous group and 14 patients in the endovascular group in a single session; however, six patients in the endovascular group required a second session of embolization. A total of 36 patients (21 in the percutaneous group and 15 in the endovascular group) underwent post-embolization surgical excision of the lesion and the rest 14 opted for follow-up without surgery. In the non-surgical percutaneous embolization group, eight patients had small lesions and one had a large lesion. In the non-surgical endovascular embolization group, all five patients had small lesions (Table 2).

OPEN IN VIEWER

Fig. 4.

(a) Clinical image showing post-embolization (percutaneous) left temporoparietal scalp AVM, (b) intraoperative image showing completely excised scalp AVM, (c) postoperative follow-up image shows the presence of postoperative scar with no residual AVM. AVM, arteriovenous malformation.

OPEN IN VIEWER

Table 2.

Management of scalp AVM (n = 50).

Embolization (n = 50)
Percutaneous	28
Percutaneous + endovascular	02
Endovascular	20
Post-embolization surgical excision of AVM (n = 36)	Percutaneous group = 21
	Endovascular group = 15
Post-embolization follow-up (n = 14)	Percutaneous group = 9	AVM size <3 cm = 8
		AVM size >3 cm = 1
	Endovascular group = 5	AVM size <3 cm = 5
		AVM size <3 cm = 0

AVM, arteriovenous malformation.

The mean duration between embolization and surgery for 36 patients was 87 days (range = 4–163 days). One patient in the endovascular group showed recurrence of the lesion postoperatively at the three-month follow-up; hence he underwent a second embolization procedure followed by repeat surgery. The mean duration of follow-up for all patients was approximately 7.5 months.

No major complications were seen in the immediate postoperative period. One patient developed mild scalp cellulitis 10 days after the embolization and was managed conservatively with broad-spectrum antibiotics. One other patient developed a stitch granuloma after surgery. No cases of radiation burns, skin necrosis, or hair loss were seen in our sample after embolization.

Discussion

Scalp AVMs account for nearly 8% of all AVMs in the body (13,14). The exact etiology of these lesions is not known and is obscure in the majority of cases. They may be congenital in origin or rarely may develop after trauma to the scalp (3). Congenital AVMs may initially be quiescent; however, they may show a locally aggressive nature, especially during puberty or adolescence, leading to an expansile mass with cosmetic and/or functional disturbances (9–12,14). Generally, they present as a small lump in the beginning and eventually develop into a large disfiguring lesion. Various hypotheses have been proposed for the cause of these congenital AVMs, which suggest the following: (i) persistence of primitive arterio-venous communications; (ii) development from vascular hamartomas; and (iii) fistula formation at the arteriovenous crossing site (3,7,15). Post-traumatic AVMs have been reported to occur secondary to a variety of head traumas, including trauma during hair transplantation, arthroscopic temporomandibular joint (TMJ) surgery, and acupuncture (16). Absolute indications for treatment of these AVMs include hemorrhage and hemodynamic problems, such as high-output cardiac failure or secondary ischemic complications caused by high-flow arteriovenous (AV) shunting (17). In our study, the most common indication for embolization and surgery was cosmetic concerns. The scalp AVMs are supplied by ECA branches in the majority of cases. The STA is the most commonly seen feeder due to its large course in the scalp (3). Other important arterial feeders include the occipital artery, auricular artery, and other branches of the internal maxillary artery. Periorbital AVMs may sometimes take their supply from the ophthalmic artery (branch of the internal carotid artery).

Complete surgical excision is considered the gold standard treatment. The major complications in surgical excision are intraoperative hemorrhage and postoperative recurrences due to incomplete removal of the nidus (7,14,18,19). The present knowledge of these AVMs states that the only curative treatment is complete removal or ablation of the nidus, and a multimodality approach is considered the most effective way to deal with this disease (20). Traditionally, the role of embolization was limited to preoperative occlusion of the supplying feeders to limit blood loss during surgery. Isolated feeder embolization without complete ablation of the nidus works as similar to surgical ligation of the supplying arteries. It does decrease the vascularity of the lesion; however, relative ischemia thus produced leads to the release of pro-angiogenic factors from the residual nidus resulting in neo-angiogenesis and recruitment of new vessels and ultimately causing recurrence or a paradoxical increase in the size of the AVM (14,15). In addition, further endovascular embolization becomes technically more difficult in these cases, as the already embolized feeder arteries block access to the nidus.

Current difficulties faced in the endovascular embolization of these lesions are their tenacity for high flow, large size, diffuse nature, and numerous feeding arteries. Embolization of the AVM is thus less commonly done as a definitive treatment and more commonly used as an adjunctive method to more definitive surgery. Herein comes the role of percutaneous embolization that represents a potentially more definitive treatment for the scalp AVMs, especially the smaller-sized lesions (21–26). Initially, we performed endovascular embolization in the cases of scalp AVMs. The problem with this approach was incomplete embolization and a high incidence of post-embolization residual disease. This was because of the fact that we were often not able to place our microcatheter tip very close to the nidus due to the difficult anatomy and tortuous arterial feeders. As our experience grew, we learned that better percolation of the embolizing agent into the nidus can be achieved by direct percutaneous injection leading to more complete embolization of the AVM. So, in the latter half of our study, we performed embolization primarily through the percutaneous route unless there were contraindications. In the percutaneous approach, since we puncture the arterial feeder close to the nidus or the nidus itself, it ensures adequate delivery and deposition of the embolizing agent within the nidus resulting in complete embolization in a single session. In addition, innovations, such as the use of compression rings, helped us in preventing non-target embolization and reflux of the embolizing agent into the veins.

We embolized 28 cases of scalp AVMs through percutaneous methods, out of which 21 underwent post-embolization surgery and the rest seven opted for non-surgical management. None of them showed any recurrence on follow-up. In two other cases, where the primary approach was percutaneous, we also had to perform endovascular embolization in the same sitting to achieve complete obliteration of the AVM nidus. The percutaneous technique is preferred in those lesions of the scalp that overly the bone/calvaria, as only then the compression can be given effectively. The percutaneous approach is avoided in lesions that involve the orbit or show direct deep venous drainage. TAE is preferred for these lesions as it gives more control during the embolization procedure.

Various liquid embolizing agents that can be used include n-BCA and ethylene vinyl alcohol copolymer (27–29). These agents induce endothelial damage, inflammation, and eventually thrombosis of the vessels that lead to either total or partial atrophy of the malformation. In our study, we exclusively used n-BCA as an embolizing agent, as it was cost-effective, easily available, and, in appropriate concentrations, led to adequate embolization of the nidus due to its low viscosity, fast action, and high penetrability into the lesion. n-BCA is mixed with lipiodol for radio-opacity and dilution. It polymerizes immediately as soon as it comes in contact with the anions in the blood and occludes the vessel (30). However, proper training and experience are required to master the technique of embolization using n-BCA. The complications include the risk of reflux, non-target embolization, and the possibility of adhesion of the microcatheter to the embolized vessels. Other rare but important complications are skin necrosis, infection, and the development of chronic sinus and ulceration (20,31).

The present study has some limitations. These include the retrospective design and a relatively short postoperative follow-up of the patients.

In conclusion, we observed that endovascular/percutaneous embolization is a less invasive, safe, and more definitive treatment for small-sized scalp AVMs and it can be used as an adjunctive method to surgery for large-sized AVMs to reduce intraoperative blood loss and facilitate surgical excision of the lesion. Moreover, there is a high chance of achieving complete obliteration of the nidus in a single sitting while performing percutaneous embolization using liquid embolic agents. The use of a compression ring during embolization of scalp AVMs effectively prevents non-target embolization as well as inadvertent reflux of the embolizing agent.