Sclerotherapy of voluminous venous malformation in head and neck with absolute ethanol under digital subtraction angiography guidance

Introduction

Venous malformation (VM) is the most common symptomatic low-flow vascular malformation. ¹ VM may cause pain, induce neuropathy, ulcerate, bleed, compress or invade adjacent structures, and lead to cosmetic disfigurement. In addition, VM may impede speech, impair swallowing and obstruct upper airway. Treatments of VM include irradiation, electrocoagulation, cryotherapy, laser therapy, copper implantation, surgical excision, sclerotherapy or a combination of the above procedures. ^2–5 However, none of the above treatment procedures can provide an acceptable rate of recurrence for the management of VM. Ethanol is the most powerful sclerotherapeutic agent, and direct percutaneous treatment with 99% ethanol has proven to be effective for VM. ^6–13 Nonetheless, it has the highest risk of complications and morbidity. In this article, we describe and discuss a technique in the treatment of voluminous VM in the head and neck using direct ethanol injection under digital subtraction angiography (DSA) guidance, by which the sclerosing agent injection was performed with higher precision.

Methods

From October 2005 to November 2008, 23 patients with voluminous VM (lesion ≥15 cm in maximum diameter or lesion invading more than one anatomical space) in the head and neck region were admitted to our hospital and received treatment. Among 23 patients, 13 were males and 10 were females, with an average age of 19 years (ranging from 13 to 25 years). The primary lesions involved parotid masseter region (3), buccal region (8), neck (3), tongue and floor of mouth (7) and bilateral sides of face, tongue and palate (2) (Figure 1). All patients had symptoms and signs due to malformations, including swelling (23), disfigurement (23), sleeping problems (9) and oral dysfunctions (8). The diagnosis was confirmed by clinical history, thorough physical examination and magnetic resonance imaging (MRI). The medical records of all patients were available and permitted for retrospective review. All patients had received sclerotherapy with other sclerosing agents previously, such as sodium morrhuate and pingyangmycin. Due to the limitation of the total dose of these agents, the VMs did not shrink remarkably after several sessions of sclerotherapy. The patients still had symptoms and signs, and were referred to our department for further treatments. Routine blood examination, liver and renal function tests were performed before and after the treatment. If the results were within normal limits prior to treatment, percutaneous ethanol sclerotherapy was conducted under DSA guidance using intravenous general anaesthesia. Nine patients with lesions involving the tongue, palate and floor of mouth had tracheotomy performed before ethanol injection.

OPEN IN VIEWER

Figure 1

(A, B, C) Images of a young man with voluminous venous malformation (VM) of the face, lip, neck and tongue (A and B are frontal and lateral images, respectively; C is an intraoral image). (D) Illustrates magnetic resonance imaging findings of VM before sclerotherapy (coronal plane). (E) Depicts direct puncture phlebographic finding of compartment of VM. The draining vein is indicated by arrows

Endovascular procedures: All sessions of ethanol sclerotherapy were performed under general anaesthesia, and the cardiopulmonary functions were monitored simultaneously. The VM lesions were punctured percutaneously with a butterfly needle until free flow of blood was obtained from the soft tube. Direct puncture phlebography with the DSA technique was performed until the draining veins were filled with the contrast. The volume of contrast was recorded. The amount of absolute ethanol used was approximately one-half to two-third of the volume of the contrast media needed to fill the lesion. Absolute ethanol injection was performed through the needle's tube slowly. No contrast media was added to the ethanol. If the venous blood flow was fast, the draining vein was compressed with fingers in the neck for five minutes during the injection. Punctures and injections of different compartments were performed at the same session, and the maximum volume of absolute ethanol was 1 mL/kg. Patients were also given dexamethasone (0.1 mg/kg) immediately before the procedure and every eight hours for three days after the treatment. After sclerotherapy, urinary output was monitored for volume and colour. Patients with gross haemoglobinuria had their urine alkalinized by replacing the standard intravenous solution with 75 mEP of sodium bicarbonate per litre of 5% dextrose.

All patients were followed up for three months to three years with a mean follow-up period of 25 months, and MRI was undertaken to detect possible remaining VM every three months after injection. The treatment results were evaluated clinically by reduction in lesion size and by patient or family satisfaction. The changes of lesion size were assessed with pre- and post-treatment MRI divided into four grades: excellent response (≥75% reduction in size), good response (≥50% reduction in size, <75% reduction in size), fair response (≥25% reduction in size, <50% reduction in size) and poor response (<25% reduction in size). Subjective evaluation was based on the improvement in swelling, facial contour and oral function after sclerotherapy. If there was no swelling, and the patients were satisfied with their facial contour and oral function such as chewing, swallowing and breathing during follow-up, then the goal of the treatment was achieved. Otherwise, the treatment was continued until the goal was achieved, or until the patients could tolerate the existing symptoms. The interval between the two sessions was four weeks.

Results

Twenty-three patients received an average of eight sessions of ethanol sclerotherapy (3–22 sessions). The average volume of ethanol used was 73 mL (28–700 mL). Three sessions were required in seven patients, seven sessions in seven patients, nine sessions in six, 10 sessions in one, 17 sessions in one and 22 sessions in one. All patients were satisfied with their improvement in facial contours and oral functions (Figure 2) after treatment. All pretreatment symptoms disappeared. Swelling in the head-down position was reduced in all patients. Seventeen patients (73.9%) achieved excellent response and six patients (26.1%) achieved good response in MRI assessment (Table 1). One patient with VM in the parotid region developed transient facial palsy, but recovered in three months. One patient with VM in the tongue and floor of mouth developed mucosal necrosis, and the ulceration healed in three weeks. Two patients developed haemoglobinnuria, which disappeared one day later. Complications such as acute pulmonary hypertension, cardiovascular collapse and pulmonary embolism were not encountered in this series. There was no liver or renal function impairment after ethanol injection.

OPEN IN VIEWER

Figure 2

(A, B, C) Show that near-complete obliteration of the venous malformation (VM) was achieved after 22 sessions of injection. The total volume of ethanol used was 700 mL (A and B are frontal and lateral images, respectively; C is an intraoral image). (D) Illustrates magnetic resonance imaging findings of VM after 22 sessions of sclerotherapy using absolute ethanol (coronal plane). (E) Depicts immediate direct puncture phlebographic finding of VM after ethanol injection. The draining vein was embolized (arrows shown)

Discussion

Mulliken and Glowacki ¹⁴ classified vascular lesions as either haemangiomas or vascular malformations. Vascular malformations that enlarge proportionately with the growth of the child and do not undergo spontaneous involution can be subdivided, based on blood flow rate, into ‘slow flow’ (venular, venous, lymphatic or mixed) versus ‘fast flow’ (arterious, arteriovenous, fistulae or shunt) subtypes. ¹⁴ The types of VM are various and complicated, ¹⁵ most of which are diffusely located and involve both superficial and deep vessels and structures. The head and neck region makes up about 40% ^2,5,16 of VM in adults.

The diagnosis of VM is based on clinical history and careful physical examination. ¹³ VMs are typically present at birth, but sometimes are not readily apparent. These lesions do not undergo an involutive stage, and the malformation usually enlarges commensurately with the growth of a child. Symptoms vary depending on the location and the size of the lesions. Functional difficulties are most often encountered when these lesions involve the upper aerodigestive tract. Upon physical examination, the lesions in the head and neck are usually soft, compressive and may become more engorged during Valsalva's manoeuvre or dependent positioning. Phleboliths may be in the lesions that could cause pain. Duplex sonography and MRI are the basic combinations used for the initial diagnosis of VM. Among these non-invasive tests, MRI technology, especially its T2-weighted images, provides the most crucial and advanced information. ^17,18 MRI is now well accepted for the contemporary diagnosis of VM. It provides haemodynamic and anatomic information on the lesion, including its relationship with surrounding tissues and organs. MRI is also essential for the follow-up assessment of the results of these multisession therapies. Ultrasonography is a simple non-invasive method to distinguish ‘slow flow’ from ‘fast flow’ vascular malformations. But in this series, ultrasonography was not performed because of its limitations in defining deep-seated and intraoral lesions, and little valuable information was available for the treatment planning of massive and extensive VM.

The choice of treatment of VM depends on the depth, extent and anatomical location of the lesion, and the experience of the clinicians. The methods available are laser therapy, excision, sclerotherapy and often combinations. ^5,19 All these techniques have their particular indications and limitations. Small lesions can be treated by either excision or sclerotherapy or laser therapy. But lasers can only be applied for the treatment of superficial lesions in oral mucosa. For the voluminous and extensive lesions, complete surgical excision is seldom achieved. Because VMs grow among soft tissues, such as muscles, nerves and blood vessels, they are difficult to delineate during surgery. ^6,7 Recurrences and cosmetic deformities are therefore common after excision. For these reasons, there is increasing interest in sclerotherapy in treatment of VM. ¹¹

Sclerotherapy refers to the introduction of a foreign substance into the lumen of a vessel, causing thrombosis and subsequent fibrosis, and is used for vascular anomalies, particularly for venous and lymphatic malformations. A wide variety of different sclerosing agents, such as ethanol, sodium tetradecyl sulphate, pingyangmycin and bleomycin, have been used in the treatment of VM. ²⁰ The most fearful complication of the use of bleomycin and pingyangmycin is pulmonary fibrosis when they are administered in VM sclerotherapy, and the total dose should not exceed 5 mg/kg. ^21,22 When used for massive VM, a large amount of bleomycin or pingyangmycin and many treatment sessions are required, which render patients at a high risk of pulmonary fibrosis. Therefore, intralesional bleomycin or pingyangmycin injection is indicated to small and medium-sized VM rather than large and extensive VM. Sodium tetradecyl sulphate has also been used for the treatment of VM, but it may lead to chronic facial pain when the injection lasts for a long period of time. ²³ In several articles, ethanol sclerotherapy of VM has been reported to be the most effective and have the lowest rate of malformation recurrence of all sclerosing agents. ^{6–13,24–26} Ethanol is the most often used due to its low cost, antiseptic quality, wide availability and easy usage. It denatures blood protein, dehydrates vascular endothelial cells, precipitates their protoplasm, denudes the vascular wall of endothelial cells and segmentally fractures the vessel wall so that there is total tissue devitalization, which demonstrates its curative potential with permanent thrombosis. ²⁶ Because the procedure of ethanol injection is very painful, general anaesthesia is used during ethanol injection for the control of severe pain. Above all, with general anaesthesia, cardiopulmonary function can be monitored closely to prevent pulmonary hypertension. Direct percutaneous contrast injection into the cavity is also required to detect both the lesion volume and the possibility of multiple compartments. We use a direct puncture under DSA guidance ‘direct version’ – the procedure is safe, convenient and easy to perform. This technique reduces the risk of having sclerosing solution misinjecting into veins or penetrating undesirable spaces, which may lead to morbidities. Through this procedure, not only the diagnosis can be confirmed, but also the volume of the ethanol used can be evaluated. The volume of ethanol to be used is determined from the contrast study. ¹³ The amount of absolute ethanol used was usually one-quarter to one-third the volume of contrast media used to define the compartment, ^12,13 and it would require multiple sessions for voluminous VM to achieve good results. However, the amount of ethanol used in our patients was one-half to two-third as described in the methods section. In order to prevent the flow of ethanol draining into veins quickly and prolong the existing time of ethanol in the compartments, the draining vein should be compressed during the injection. The interval between the two sessions was four weeks to let the patients recover from the previous therapy. Furthermore, it allows the local tissue reaction to be subsided.

Conflict exists in evaluation of results of various VM treatments. ^11,12,27 In patients with voluminous head and neck VM, concerns frequently focus on cosmetic considerations. Therefore, treatment goal must be clearly established before the procedure. If patients were satisfied with their facial contours and oral functions, and there is no pain or swelling, the goal of treatment has been achieved. In this study, all patients were satisfied with their facial contours and oral functions after treatment, and all patients achieved good to excellent results in MRI evaluation. More importantly, patients should be followed up for a long time to see whether lesions recurred and to decide whether sclerotherapy should be continued.

Potential complications of sclerotherapy include local skin or mucosal necrosis, transient neural palsy, haemoglobinuria, blood loss, acute pulmonary hypertension, cardiovascular collapse, pulmonary embolism and anaphylaxis, ^11,26 and the incidence rate was about 27.9%. ²⁸ Due to these possible complications, ethanol sclerotherapy was once considered too dangerous to be accepted as a routine treatment method for VM. ²⁸ In our study, complications developed during the procedures in four patients (17.4%) out of a total of 23 patients, and these complications included transient facial paralysis, haemoglobinuria and oral mucosa necrosis. All complications were successfully managed with full recovery during three months of follow-up. In order to prevent the potential complications, the maximum recommended dose of ethanol is 1 mL/kg of body weight. ^11,29 Dexamethasone (0.1 mg/kg) was used to minimize tissue reaction before and after the treatment. After sclerotherapy, urinary output was monitored for volume and colour. Patients with gross haemoglobinuria had their urine alkalinized to prevent renal function impairment. If the lesions involved oral cavity, tissue swelling after ethanol injection could result in airway obstruction. Tracheotomy was performed on nine patients in our series before ethanol injection to prevent airway obstruction. Pulmonary hypertension is potentially fatal and occurs when a large amount of ethanol reaches the pulmonary circulation in a short time. ²⁸ As DSA technique was used in our study to drain the veins and to determine the volume of ethanol necessary to be used, severe complications such as acute pulmonary hypertension, cardiovascular collapse and pulmonary embolism reported in other studies were not encountered. In this study, all patients experienced symptomatic or cosmetic improvement without major complications.

In conclusion, sclerotherapy with absolute ethanol can be an important treatment modality for voluminous and complex VMs in the head and neck region, because the procedure is reasonably safe and render effective results. Sclerotherapy should be performed under general anaesthesia with DSA guidance and should only be used by those with appropriate expertise to minimize the risk of complications.