Dose-saving isolation procedure in percutaneous ethanol sclerotherapy for venous malformations

Abstract

Objectives:

To evaluate the feasibility and effectiveness of an isolation technique during ethanol injection sclerotherapy for venous malformations (VMs) in the head and neck region.

Methods:

The subjects were 23 patients with 35 VM lesions in the head and neck, treated between 1999 and 2012. The mean lesion area was 3.75 ± 3.09 cm² (± standard deviation). We confirmed the contour of the lesions to be treated on a fully filled image on direct injection cisternography, and observed patterns of communicating drainage to systemic veins. The cisterns were evacuated by squeezing and were isolated by manual compression of the communicators. Ethanol (94.5%) with a contrast agent was then injected into both isolable and unisolable lesions, up to a total volume of 1 mL/cm², avoiding complications. We investigated the relationship between lesion size and injected ethanol dose, and also dose per unit area.

Results:

Both manual evacuation by compression and isolation were performed in 20 (57.1%) isolable lesions, but not in 15 unisolable lesions. The mean injected ethanol dose was 0.65 ± 0.31 mL/cm² overall, 0.70 ± 0.32 in isolable and 0.59 ± 0.30 in unisolable lesions (NS). However, the injected ethanol dose was significantly lower for lesions sized >6 cm. Complete to near-complete shrinkage was observed in all isolable lesions, and in 60% of unisolable lesions (P < 0.05). Clinical outcome seemed unrelated to the injected ethanol dose or the dose per unit area. There was one case of recurrence and one complication in the unisolable lesions. No further relapses or complications were observed during the follow-up period of 38.6 ± 12.3 months.

Conclusions:

Clinical outcome was related to the isolability not to the injected dose. The isolation appears useful for improving the safety and effectiveness of ethanol sclerotherapy for VM.

Keywords

venous malformations draining vein ethanol sclerotherapy communicating vein complication

Introduction

Venous malformation (VM) occurs most commonly in the head and neck area.^1,2 The clinical symptoms include pain, disfigurement and occasional bleeding.^1,3,4 For palliative and cosmetic reasons, sclerotherapy has gained favour over surgery, which can cause tissue loss and further disfigurement. Excellent treatment results are reported for percutaneous sclerotherapy.^5–9 Among various sclerosants in use,^5,9,10 ethanol is the most commonly available^4,7–9,11 but has an injurious nature, causing acute blistering, local tissue ulceration, nerve injury when extravasated^8,9,11 or arrhythmia if overdosed.¹²

The essential principles of safe and effective treatment are control of local ethanol distribution and avoidance of overdosing. It is important to avoid local hazardous events such as ulceration and drainage into the systemic vein. Systemic toxicity can generally be avoided by limiting the total dose^13,14 and possibly by decreasing the draining flow to the systemic veins. Preparatory evacuation of blood from the lesion before instillation has been considered and utilized to increase the success of treatment. Viscous mixtures and foamy preparations containing both ethanol and fibrosing alkyl ethers have been introduced [Kuhne, 1982 #520] [Riche, 1983 #516] [Bergan, 2007 #591]. An anticancer agent, Bleomycin, has also been clinically investigated.¹⁵ Recently reported methods for improving ethanol monotherapy include immediate suction of pooled ethanol in the lesion after instillation¹³ and subcutaneous injection of saline.¹⁴

In a study of predictors of response in percutaneous ethanol sclerotherapy of 158 VM patients, Yun et al.¹⁶ reported significant predictors for good response as no or delayed visualization of drainage, non-extremity location and well defined margin on magnetic resonance imaging (MRI) (P = 0.02, 0.01 and < 0.01, respectively), while the volume of ethanol was not significant (P = 0.9). If the solitariness of a lesion has a positive relation with good response, then the additional promotion of solitariness would enhance the benefits of sclerotherapy. In their representative angiograms of immediate visualization shown in their articles were of only single or a few drainage veins.¹⁶ Some type of isolation procedure may have been beneficial in these cases. To the best of our knowledge, control of these communication veins has been attempted with foams or viscous sclerosants but not by manual obstruction. In our recent study of VM lesions located mainly in the head and neck region, all lesions showed visible communication with systemic veins, either small or large, on digital subtraction angiography after continuous injection.¹⁷ We postulated that obstruction of these visualized communication routes would benefit safety and reduce the required dose of ethanol for sclerotherapy.

In the present study, we attempted to evacuate blood from and isolate the cisterns by blocking the outflow channels with manual compression under fluoroscopic monitoring, with the aim of controlling the distribution of ethanol and reducing the dose required for sclerotherapy. We then evaluated the feasibility and effectiveness of the manual isolation procedure in ethanol sclerotherapy for VM between isolable and unisolable lesions.

Methods

Study design

We analysed serial treatment results using a prospectively designed protocol to evaluate the feasibility, rationale and clinical effect of isolating cisterns during ethanol injection (Figure 1). Isolability was defined as isolation readily achievable by simple digital compression or by ligation of tissue containing communicating veins using a surgical thread. The isolation procedure was performed following evacuation of blood from the cisterns. The lesions were classified as isolable or unisolable according to the success of the isolation procedure (isolable or unisolable groups). No lesions were evacuated without isolation or isolated without evacuation.

Figure 1

Process flow chart

Patients

Enrolled in the study were 23 patients (9 men, 14 women) with 35 head and neck lesions that were clinically and radiologically diagnosed as VMs between December 2000 and March 2012. The median patient age was 32 years (age range, 5–79 years) (Table 1).

Table 1

Patient characteristics and lesion sites

Patients	No. of patients	Isolable	Unisolable
Gender
Male	9	3	6
Female	14	10	4^*
Total	23	13	10
Age
Median (years)	31	31	34.5
Range (years)	5–79	5–79	5–73
Lesions	No. of lesions	Isolable	Unisolable
Site
Lip and perioral	15	11	4
Tongue	8	4	4
Lalate and oropharynx	4	3	1
Cheek	2	2	0
Neck	4	0	4
Shoulder and chest	2	0	2^**
Total	35	20	15

p = 0.18, Fisher's exact test

p = 0.027, Fisher's exact test

The diagnostic criteria were as follows: history of a slow-growing mass that was soft, non-palpitating, collapsible by digital compression, and that spontaneously refilled after release of compression; no high-flow signal in the mass on power Doppler ultrasonography; no evidence of arterial supply; blood flow on direct puncture; and having the ‘bunch of grapes’ appearance¹⁸ on direct-injection contrast study.

VM was defined according to the 2003 ISSVA (International Society for the Study of Vascular Anomalies) classification system.¹⁹

The mean lengths of the long and short axes were 21.7 mm (range, 7–40 mm) and 18.3 mm (5–35 mm), respectively. X-ray computed tomography and MRI appearances were compatible with those reported previously.^20,21

No patient received biopsy or histological study because of the risk of haemorrhage and for cosmetic reasons.

The purpose of treatment was cosmetic improvement in all, pain relief in six and hemostasis in one. All of the patients had good performance status, and none had any other local or systemic disease, including haematological abnormalities. Informed consent was obtained before treatment in all of the patients according to our institutional ethical guidelines. The study was performed in accordance with the principles of the Helsinki Declaration.

Direct puncture cisternography

Under local anaesthesia with subcutaneous injection of lidocaine hydrochloride, the VM lesion was directly punctured with a 24-gauge plastic needle (set of inner needle and plastic outer cannula). On successful puncture, confirmed by observation of blood flow from the needle, the VM was manually compressed to evacuate the lesion (emptying), and the contrast medium (Iopa-midol 300, 300 mg iodine/mL) was gently injected into the lesion under fluoroscopy. We routinely used digital subtraction angiography (DSA) to observe fine vessels.⁸ When a typical ‘bunch of grapes’ appearance¹⁸ showing inflated multiple cisterns was confirmed, a radiological diagnosis of VM was made. We recorded the injected volume of the contrast medium necessary to obtain the characteristic appearance, termed the ‘fully filled dose’. Injection of the contrast medium was continued until the communicating channels were well visualized. We manually evacuated blood from the lesion immediately before performing the following isolation procedure (Figure 1).

Classification of communicating pattern

The pattern of venous communication between capillary VMs and systemic venous flow was categorized according to Dubois et al.'s classification:²² Type 1, no visible drainage; Type 2, draining into a normal venous system; and Type 3, with abnormally ectatic draining veins. Type 2 was further divided into early- and late-appearing types according to Puig et al.²³

Isolation procedure

Among the draining veins that could be visualized, we attempted manual compression of the major draining veins to temporarily isolate the cisterns during ethanol injection (Figures 2 and 3). We evaluated the effectiveness of the isolation manoeuver by comparison of images obtained with and without the manoeuver.

Figure 2

Isolable lesion: direct injection cisternography reveals the characteristic retention pattern in a perioral venous malformation in the mental region under the lower lip. The draining veins (submental and inferior labial veins) both merge into facial veins. Related muscles include the inferior labial depressor, mental and triangular muscles (a). Direct injection cisternography during compression of the veins shows that the outflow was effectively obstructed (b). The arrows indicate fingers compressing the outflow channels

Figure 3

Unisolable lesion: a venous malformation located on the right side of the tongue. The areas of aggregated long fusiform retention match the vertical lingual muscle fascicle, which suggests the location of these lesions in the perimysium (sack of fascicles). Since the draining veins run toward the profound lingual vein anteriorly and posteriorly, it was impossible to perform the isolation procedure, (a) Lateral view and (b) anteroposterior view

Ethanol preparation

We prepared a mixture of 94.5% ethanol, 0.5% water and 5% contrast medium (Iopamiron 300 mgl/mL; Bayer AG, Leverkusen, Germany).

Injection of absolute ethanol

Under isolation, and while a small amount of the residual contrast medium was visible in the cisterns, ethanol was carefully injected through the plastic cannula into the lumen under fluoroscopic monitoring of the gradual spread of ethanol. As institutional policy, the dose of ethanol was strictly limited to less than 1 mL/cm² of surface area of the lesion to avoid the adverse effects of injected ethanol. (This policy was adopted based on a previous case report²⁴ and plot analysis of estimated data from a published study²⁵ These data are later discussed.) Injection was stopped when the ethanol appeared to have spread over the extent of the target lesion, or when the patient complained of pain. No additional ethanol was injected. Acetaminophen (300–500 mg tablet) was given for pain relief after the ethanol injection.

Follow-up and evaluation of treatment effect

The patients were followed up at one and four weeks after the procedure. The effectiveness of the treatment was evaluated at four weeks after the procedure. The patients were either followed up at our institute or by family dentists, who reported the clinical course to our clinic at the regional university hospital.

Evaluation methods

We calculated the area and volume of the lesions as follows: area = π × 1/2 × long diameter (a) × 1/2 × short diameter (b); and volume (V, as an ellipsoid) = (4/3) × π × 1/8 × a × b × (a + b).

Objective treatment effect was categorized as one of five outcomes: complete shrinkage (Ec), almost complete disappearance with a few residual spots on the skin (Eac), shrinkage of more than 50% (E_>50) and shrinkage not more than 50% (E_<50). The subjective effect was classified into three categories: satisfied, unsatisfied but needing no further treatment and requiring further observation or treatment.

Area of the lesion, injected ethanol dose and injected ethanol dose per unit area were plotted with curve fitting. Difference in size, communication type, injected ethanol volume (mL and mL/cm²) and clinical effect between isolable and unisolable lesions were analysed using Student's t-test or Fisher's exact test. A P value less than 0.05 was considered statistically significant.

Results

There were no technique-related complications. Mean lesion area was 3.75 cm² (range, 0.3–9.6 cm²) and mean ellipsoid volume was 6.82 cm³ (range, 0.2–22.4 cm³). Mean fully filled dose was 4.73 mL (range, 0.5–10.0 mL) (Table 2).

Table 2

Lesion characteristics and injected volumes of ethanol

Patients	Total	With isolable lesions	With unisolable lesions
	23	13	10
Lesions	Total	Isolable lesions	Unisolable lesions	P value
Number of lesions	35	20	15
Area (cm²) (α)	3.75 ± 3.09 (0.3–9.6)	2.90 ± 2.86 (0.3–9.4)	3.33 ± 3.44(1.2–9.6)	0.025	S^*
Calculated volume (cm³)	6.82 ± 7.59 (0.1–22.4)	4.95 ± 6.83(0.1)	9.32 ± 8.06(0.1–22.4)	0.038	S^*
Fully filled dose (mL)	4.81 ± 3.60(0.5–10.0)	2.00 ± 2.59 (0.5–10)	6.10 ± 3.42(1.5–10)	0.007	S^*
Communication type
Type 1	0	0	0
Type 2 (early-appearing)	18	10	8
Type 2 (late-appearing)	15	10	5
Type 3	2	0	2	0.302	NS^†
Injected ethanol dose (mL) (δ)	2.27±2.20(0.2–10.0)	1.46±1.15(0.2–3.8)	3.34 ± 3.34(0.6–10)	0.007	S^*
Ethanol dose per area size (δ/α)	0.65 ±0.31 (0.1–1.5)	0.69 ± 0.32(0.1–1.5)	0.59 ± 0.3(0.3–1.4)	0.421	NS^*
	Mean ± standard deviation (range)

NS, not significant; S, significant difference

Student's t-test

†

Fisher's exact test (extended) for 3 × 2 contingency table including Type 2 early-appearing, Type 2 late appearing and Type 3

Communication type

DSA study revealed visible communication channels between VMs and systemic veins in all cases. There were 18 early-appearing Type 2 lesions, 15 late-appearing Type 2 lesions and two Type 3 lesions with small ectatic drainers. There were no Type 1 lesions.

Feasibility of the isolation procedure

Twenty (57%) of the 35 lesions were isolable (Figure 2); 18 of which were isolated by digital compression of communicating veins identified on DSA. Two lesions were isolated by ligation of a section of the tongue and oral floor, using a surgical thread. Fifteen lesions were unisolable: eight because of the deep course of the draining vein (Figure 3), and seven because the physician lacked an assistant to help apply manual compression during the procedure. Without isolation, the lesion refills soon after the release of compression; therefore, in unisolable lesions in which the state of evacuation is achieved by manual compression, the evacuated status could not be perfectly maintained long enough for ethanol injection to be performed. Area size, calculated volume and fully filled dose were larger in unisolable than in isolable lesions (P < 0.05). No statistically significant difference was detected between isolable and unisolable lesions in terms of communication type (Table 2).

Injected sclerosant dose and size of the lesion

Each diagnostic and ethanol injection procedure was accomplished within 60 minutes. Mean ethanol dose was 1.46 mL (range, 0.2–3.8 mL) for isolable lesions, and 3.34 mL (0.6–10.0 mL) for unisolable lesions (P < 0.05) (Table 2). Mean ethanol dose per area was 0.69 mL/cm² (range, 0.1–1.5 cm²) for isolable lesions, and 0.59 mL/cm² (range, 0.26–1.41 cm²) for unisolable lesions (no significant difference) (Table 2). Approximation curves showed an increase in the injected dose with increasing lesion area (Figure 4a), with the curve being steeper in unisolable lesions. When limited to lesion size greater than 6 cm², injected dose and dose per area were significantly less in the isolable than in the unisolable group: 1.43 ± 0.9 versus 6.25 ± 2.3 mL (P = 0.0015) and 0.16 ± 0.1 versus 0.80 ± 0.39 mL/cm² (P = 0.004), respectively.

Figure 4

Dose of ethanol-containing mixture according to venous malformation (VM area. Area size of VMs and dose of ethanol-containing mixture showed moderate positive correlations both in isolable groups and in unisolable groups (a), and area size and injected volume per unit area showed moderate correlations in isolable group (b)

Inverse correlations were observed between lesion area and dose per unit area (Figure 4b), except when only the unisolable group was considered.

Clinical outcome

The mean follow-up period was 38.6 ± 12.3 months (range, 5–144 months). Shrinkage was observed in all lesions (Figures 5 and 6). At four weeks after treatment, there were 23 Ec, six Eac, six E_>50 and no E_<50 lesions (Table 3). Ec, Eac, E_>50 and E_<50 were observed in 15 (Table 4a), 5 (b), 0 (f) and 0, respectively, of the 20 isolable lesions; and in 8 (c), 1 (d), 6 (e) and 0, respectively, of the 15 unisolable lesions. Clinical effectiveness was significantly higher in isolable lesions than in unisolable lesions (P < 0.01, Fisher's exact test) (Table 3). Ec, Eac and E_>50 were distributed throughout the range of lesion sizes (Figures 7a and b). There was no statistically significant difference in size between any two of the following subsets: Ec, and Eac subsets of the isolable group; and Ec, Eac and E_>50 subsets of the unisolable group (statistical data not shown). There was no apparent uneven distribution of clinical effects in terms of dependency to dose or of dose per area (Figures 7a and b). Though mean ethanol dose per unit area of 15 isolable lesions with Ec clinical effect was higher than those of eight unisolable with E_>50, no statistical difference of injected ethanol dose or dose per unit area was found between any two of the subsets (a) – (f), or among the three subsets of all Ec (23 lesions), all Eac (6 lesions) and all E_>50 (6 lesions) (Table 4). There were no immediate complications that required mediation during or immediately after sclerotherapy. Transient pain or a burning sensation was recorded in all patients. All lesions showed transient inflammatory reactions to various extents and patterns after treatment, which resolved without additional medication. In seven patients with lip or superficial lesions, typical redness and acute swelling were observed during the two days following treatment, which gradually diminished within two weeks. These patients complained of mild pain at the injection site, within a tolerable range, lasting for 3–5 days. A skin ulcer developed at the periphery of the lesion after treatment in one patient with an unisolable lesion (Figure 6) (dotted arrows in Figure 7), which resolved without disfigurement after one month. All of the patients reported satisfaction with the treatment results, even when adverse effects occurred. There was one case of relapse one year after treatment in an unisolable lesion. Two re-treatments were performed: for relapse at the same location at a tooth extraction space, and to a satellite relapse.

Figure 5

An isolable case: venous malformation on the right lateral aspect of the tongue. Photographs taken before treatment (a) and four weeks after treatment (b) show disappearance of the lesion. Continuous direct injection cisternography (lateral view) shows outflow to deepen the lingual vein in two roots (c). Anterior-posterior fluoroscopic view shows the effectiveness of manual compression for obstructing these lateral routes (d)

Figure 6

An unisolable case: venous malformation in the left perioral region. Continuous direct injection cisternography reveals a typical pattern with a draining vein (a). Photographs taken before treatment (b) and showing a skin ulcer that developed after treatment but healed after one month (c). The site of ulcer development at the periphery of the lesion suggests the migration of superfluous ethanol via a draining vein

Figure 7

Injected volume of ethanol-containing mixture and volume per unit area of venous malformation (VM). Dots of each subgroup are indicated on the graph of the area size and the injected ethanol volume (a) and on that of the area size and the dose per unit area (b). Dotted arrows indicate the case that developed a skin ulcer

Table 3

Clinical effectiveness by group and subsets (a)-(f)

	Effectiveness
Group	Ec	Eac	E_>50	Total
Isolable	15 (a)	5 (b)	0 (f)	20
Unisolable	8 (c)	1 (d)	6 (e)	15
Total	23	6	6	35

Statistically significant difference was found between the clinical effect in the isolable group and that of the unisolable group [P = 0.0044) Ec, complete shrinkage; Eac, almost complete disappearance with a few residual spots on the skin; E_>50, shrinkage of more than 50% (E_>50); E_<50, shrinkage not more than 50%; and (a)–(f), subsets. Fisher's exact test to compare isolable and unisolable groups: P = 0.0044. P = 0.0031 for combined Ec and Eac

Table 4

Clinical effect and ethanol dose per unit area

Clinical effect	Isolable	Subset	Unisolable	Subset	Subset	Combined isolable and unisolable
Ec	15^*	(a)	8	(c)	All Ec (a + c)	23
	0.76 ± 0.29^†		0.58 ± 0.20			0.69 ± 0.27
Eac	5	(b)	1	(d)	All Eac (b + d)	6
	0.56 ± 0.41		0.26			0.51 ± 0.39
E_>50	0	(f)	6	(e)	All E_>50 (e)	6
			0.90 ± 0.48			0.90 ± 0.48
P value	(a, b): 0.18		(b, c): 0.44		(Ec, Eac): 0.15
	(a, c): 0.053		(b, e):0.12		(Eac, E_>50): 0.06
	(a, e): 0.23		(c, e): 0.067		(Ec, E_>50):0.15

P values shown are for dose per unit area using Student's t-test. (a)–(f) are subsets. Description as (a, c): 0.18 represents P value of difference of ethanol dose per unit area between subset a and subset c

Number of lesions

†

Injected ethanol dose per unit area

Discussion

We evaluated feasibility and effectiveness of the evacuation-isolation procedure. The procedure was feasible in 57% of lesions (isolable), which enabled us to compare lesion size, injected ethanol dose, size and the dose per area unit and clinical effect between isolable and unisolable lesions.

Lesion size, injected dose and dose per unit area

Injected ethanol dose was positively correlated to lesion size (Figure 4a). Size–dose plots (area size was estimated using published diameter data [r] as π × r² × k, for k = 1 [round] when r < 6; and k = 2/3 [oval] when r ≥ 6, because the larger lesions are most likely oval than round shapes) from a data-set published by Ieraldi et al.,²⁵ containing 80 treatment records before and after ethanol injection, fitted a similar and stronger positive correlation curve (Figure 8a). In the present study the correlation curve was flatter in the isolable group, possibly reflecting a dose-saving effect of the isolation technique.

Figure 8

Plotted data calculated from the original data of leraldi et al.²⁵ (a) Lesion area (cm²) and injected ethanol dose (mL). (b) Lesion area (cm²) and injected ethanol dose per unit area (mL/cm²)

In the clinical situation, the dose of ethanol injected is decided by the operator based on assessment of risk and benefit. In the present study, lesion size and ethanol dose per unit area showed inverse correlations. A similar phenomenon was observed in the data plots of Ieraldi et al.²⁵ (Figures 4b and 8b), in which the calculated injected dose per unit area was 2.0 ± 1.0 mL/cm² for lesions smaller than 6 cm² (n = 11), and only 0.4 ± 0.4 mL/cm² for lesions sized 6–20 cm² (n = 25). In both the study of Ieraldi et al. and the present study, a greater volume per unit area was injected into smaller compared with larger lesions. It might be a possible interpretation that the operators’ attitude tended to inject more sufficient volume when the lesion size was small. An important implication of this inverse correlation is that an unnecessarily high dose per unit area may have been injected into smaller lesions. Although we did not study the reason for this phenomenon, we speculate that contributing factors were poor visibility of the small injection volume under fluoroscopy and a lack of information regarding the inverse correlation curve, besides the operators’ attitude. This is the first study to report the inverse correlation between lesion size and injected dose per unit area, though further studies are required to confirm and address the phenomenon.

Size and isolability

There was a significant difference in lesion size between the isolable and unisolable groups. Lesion size is considered to be physiologically related to un-isolability because larger lesions may have a deep location and more drainage veins. The main reason for unisolablility was the presence of drainage vessels too numerous or too deep to be compressed.

Clinical benefits of the isolation procedure

As hypothesized, the clinical effects were significantly better for isolable lesions than for unisolable lesions.

Local adverse events (LAE)

Ethanol has a highly permeable nature, passing easily through fine channels to potentially cause injury to normal tissue. Muto et al.²⁴ reported a case of lip necrosis that occurred 10 days after injection of 7 mL of ethanol into a lip mass with ethanol dose (area size was estimated using published diameter data [r] as π × r² × k, for k = 1 [round] when r < 6; and k = 2/3 [oval] when r ≥ 6, because the larger lesions are most likely oval than round shapes) of 1 mL/cm². Plots of the data of leraldi et al.²⁵ show that lesions with LAE had significantly larger mean area and total dose, and significantly smaller dose per unit area, than those without LAE (85.6 ± 75 versus 39.0 ± 55 cm², 24.4 ± 11.3 versus 16.53 ± 0.9 mL and 0.41 ± 0.2 versus 0.89 ± 0.71 mL/cm², respectively). We speculate that the occurrence of LAE could be related to the presence of communicating veins. Our case of an ulcer that developed peripherally to the lesion may indicate the effect of superfluous ethanol that migrated via a channel. We consider that the present technique can reduce the risk of unnecessary local injury.

Angiography method

The ability to visualize communicating vessels can be strongly influenced by the method of angiography. We identified communicating veins in all cases,¹⁷ though in numerous other studies lesions could be defined without visualization of the communicating channels. Puig et al. reported that 30% of VMs are isolated malformations without peripheral drainage (Dubois type I).^22,23 Fine communication vessels can be difficult to identify on conventional fluoroscopy.¹⁵ To improve detect-ability we used continuous injection of contrast media and DSA observation (with recording) until the surrounding communication vessels were seen.¹⁷ This observation must be completed in the first injection to avoid the masking effect of residual contrast medium.

Feasibility of the isolation procedure

There are limitations to the applicability of this technique. The isolation procedure must be as practical as readily achievable with simple manual compression or ligation. Ligation is a useful alternative to digital compression when it is carefully performed without intraoral bleeding. Anatomically it is not easy to apply effective compression to draining veins that have a deep course. Sometimes the physician lacks an assistant to help apply manual compression during the procedure. In such cases it may be better to use a foam or viscous type of sclerosant to reduce the ethanol dose and control the draining flow.^5,9 Cine mode recording during direct injection cisternography enables optimal visualization of the draining vessels.¹⁷

Assessment of optimal dose

A small amount of ethanol should be sufficient only to degenerate the vascular intimal layer of a well-evacuated cistern. In the present study the mean ethanol dose required for complete shrinkage was 0.7 mL/cm². Mean dose in the isolable/Ec subset (a) was larger (0.76) than that in the unisolable/Ec subset (c) (0.58) (P = 0.053). As mentioned earlier, taking the operators risk-benefit decisions into account, the doses administered to lesions that achieved Ec effects may have included supra-optimal doses. In an Ethibloc study, the ethanol dose is calculated as 0.32 ± 0.1 mL/cm².²⁶ The optimal dose for ethanol sclerotherapy using this procedure may be less than 0.7 mL/cm², though further studies are required to identify the optimal dose.

Limitations of this study

The present study was performed on a small number of head and neck VMs at a single institution. It was initiated based on the concept that all VMs have draining vessels that can be visualized by an appropriate fluoroscopic procedure;¹⁷ however, this assumption has not been proven in other areas of the body. Treatment of lesions in other sites needs to be evaluated. Thus, further studies are required before this technique can be recommended for general application.

Conclusions

A manual flow-control technique with a compression evacuation and isolation procedure for VM ethanol sclerotherapy was feasible in 57% of our patients. The clinical results were better in the isolable group. We consider that this technique can improve the safety and effectiveness of ethanol sclerotherapy for VMs.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Footnotes

Conflict of interest statement

The authors have no conflicts of interest to declare.

References

Young

. Pathogenesis of vascular malformation. In: Mulliken

, Young

, eds. Vascular birthmarkings, hemangiomas and malformations. Philadelphia, PA: W.B. Saunders Co, 1988: 107–13.

Kohout

, Hansen

, Pribaz

, Mulliken

. Arteriovenous malformations of the head and neck: natural history and management. Plast Reconstr Surg 1998; 102: 643–54.

Hein

, Mulliken

, Kozakewich

, Upton

, Burrows

. Venous malformations of skeletal muscle. Plast Reconstr Surg 2002; 110: 1625–35.

Shireman

, McCarthy

, Yao

, Vogelzang

. Treatment of venous malformations by direct injection with ethanol. J Vasc Surg 1997; 26: 838–44.

Gorriz

, Carreira

, Reyes

. Intramuscular low flow vascular malformations: treatment by means of direct percutaneous embolization. Eur J Radiol 1998; 27: 161–5.

Rautio

, Saarinen

, Laranne

, Salenius

, Keski-Nisula

. Endovascular treatment of venous malformations in extremities: results of sclerotherapy and the quality of life after treatment. Acta Radiol 2004; 45: 397–403.

Lee

, Do

, Byun

, Choo

, Kim

, Huh

. Advanced management of venous malformation with ethanol sclerotherapy: mid-term results. J Vasc Surg 2003; 37: 533–8.

Berenguer

, Burrows

, Zurakowski

, Mulliken

. Sclerotherapy of craniofacial venous malformations: complications and results. Plast Reconstr Surg 1999; 104: 1–11; discussion 2-5

Siniluoto

, Svendsen

, Wikholm

, Fogdestam

, Edstrom

. Percutaneous sclerotherapy of venous malformations of the head and neck using sodium tetrade-cyl sulphate (sotradecol). Sound J Plast Reconstr Surg Hand Surg 1997; 31: 145–50.

10.

Choi

, Han

, Ok

, Cha

, Chang

. Craniofacial cavernous venous malformations: percutaneous sclerotherapy with use of ethanolamine oleate. J Vasc Interv Radiol 2002; 13: 475–82.

11.

Berthelsen

, Fogdestam

, Svendsen

. Venous malformations in the face and neck. Radiologic diagnosis and treatment with absolute ethanol. Acta Radiol Diagn (Stockh) 1986; 27: 149–55.

12.

Yakes

, Luethke

, Parker

. Ethanol embolization of vascular malformations. Radiographics 1990; 10: 787–96.

13.

Hammer

, Boon

, Mathurin

, Vanwijck

. Ethanol sclerotherapy of venous malformations: evaluation of systemic ethanol contamination. J Vasc Interv Radiol 2001; 12: 595–600.

14.

Mason

, Michna

, Zurakowski

, Koka

, Burrows

. Serum ethanol levels in children and adults after ethanol embolization or sclerotherapy for vascular anomaliesl. Radiology 2000; 217: 127–32.

15.

Liu

, Liu

, Wang

, Zhang

, Zhao

. Clinical study of sclerotherapy of maxillofacial venous malformation using absolute ethanol and Pingyangmycin. J Oral Maxillofac Surg 2009; 67: 98–104.

16.

Yun

W-S

, Kim

Y-W

, Lee

K-B

. Predictors of response to percutaneous ethanol sclerotherapy (PES) in patients with venous malformations: analysis of patient self-assessment and imaging. J Vasc Surg 2009; 50: 581–9.e1

17.

Kishi

, Morita

, Terada

, Sato

. Radiographic findings on venous malformations: communications to systemic veins, satellite lesions and serial appearance of cisterns. Phlebology 2012: doi: 10.1258/PHLEB.2012. 011137

18.

Boxt

, Levin

, Fellows

. Direct puncture angiography in congenital venous malformations. AJR Am J Roentgenol 1983; 140: 135–6.

19.

Enjolras

, Wassef

, Chapot

. ISSVA classification. In: Enjolras

, Wassef

, Chapot

, eds. Color Atlas of Vascular Tumors and Vascular Malformations. New York: Cambridge University Press, 2007: 1–11.

20.

Rigamonti

, Drayer

, Johnson

, Hadley

, Zabramski

, Spetzler

. The MRI appearance of cavernous malformations (angiomas). J Neurosurg 1987; 67: 518–24.

21.

Hovius

, Borg

, Paans

, Pieterman

. The diagnostic value of magnetic resonance imaging in combination with angiography in patients with vascular malformations: a prospective study. Ann Plast Surg 1996; 37: 278–85.

22.

Dubois

, Patriquin

, Garel

. Soft-tissue hemangiomas in infants and children: diagnosis using Doppler sonography. AJR Am J Roentgenol 1998; 171: 247–52.

23.

Puig

, Aref

, Chigot

, Bonin

, Brunelle

. Classification of venous malformations in children and implications for sclerotherapy. Pediatr Radiol 2003; 33: 99–103.

24.

Muto

, Kinehara

, Takahara

, Sato

. Therapeutic embolization of oral hemangiomas with absolute ethanol. J Oral Maxillofac Surg 1990; 48: 85–8.

25.

Ierardi

, Mangini

, Vaghi

, Cazzulani

, Carrafiello

, Mattassi

. Sclerotherapy of peripheral venous malformations: a new technique to prevent serious complications. Vasc Endovasc Surg 2010; 44: 282–8.

26.

Fradis

, Podoshin

, Simon

, Lazarov

, Shagrawi

, Boss

. Combined treatment of large head and neck capillaro-venous malformation by a fibrosing agent. J Laryngol Otol 1989; 103: 390–8.