Orbital AVMs: Surgical management

Orbital AVMs: Surgical management transcript
Presenter: David Verity

This talk is based around the work that we have shared with Fergus Robertson and his team, and I shall take you through some of the patients we have treated recently.

I have no disclosures to make.

The orbital vascular lesions are a complex group of disorders. Broadly speaking, they can be divided into those that are vasoproliferative, such as the haemangiomas, vascular malformations, rare vascular tumours, such as Kaposi’s sarcoma, and finally ectatic ageing vessels. But what we shall cover today is the treatment of arterio-venous vascular malformations.

Vascular malformations can be divided into venous-lymphatic, representing a spectrum, and seen linked on the slide. We shall not deal with naevus flammeus today, but shall examine these difficult arteriovenous malformations which have a high propensity to per-operative blood loss, unless they’re pre-treated, and that’s the reason we collaborate with the interventional radiologists.

Arteriovenous malformations are usually congenital, belonging to a group of disorders called the ‘RASopathies’. These are developmental syndromes caused by germ-line mutations which alter the protein kinases that control signal transduction, leading to the formation of AV malformations. Of interest, neurofibromatosis type 1 is also one of the ‘RASopathies’. They have variable degrees of genetic transmission, and in the case of AVMs are a number of mutations are implicated, including epithelial line and tumour suppressor genes, accounting for the tendency  for these lesions to occur throughout the body.

Orbital Arteriovenous malformations include high flow AV shunts, slow flow, or the so called “dural” (meningeal) intracranial shunt, and high flow intracranial AV shunts: the carotid-cavernous sinus fistula (“CCF”). But today we shall only look at the solely orbital AV shunts.

I shall take you through a few cases, each of which has been treated with a slightly different form of embolising material, guiding us through the history of these materials. This first gentleman had presented to many different centres with his pulsating right upper lid. Perhaps not surprisingly, most of the units that he had visited were conservative in their approach due to the risk of surgical blood loss. Seen on these two slides, it is relatively anterior, with this spongey configuration seen on the CT scan. But as we’ll see from other cases, this can change. When one region is closed, another vascular bed can open up, unfortunately, and one can never say that one operation will be curative. You can see in this video the gentle pulsatility of the upper lid, occurring 60 to 70 times every minute throughout his life, whilst awake and when lying down. He found this desperately troublesome and was very keen on surgical intervention.

This slide shows an angiogram in real time, the contrast injected into the right internal carotid artery. You see the ophthalmic artery here, and the AVM nidus, and then you start to see the contrast flowing out again, shunting into the facial vein. Fergus has shown something similar to this in his talk. That was over the period of about 40 seconds, and on this slide the angiogram is via the right external carotid artery (“ECA”) in the same patient, with a lateral view. You see here the facial artery and the AVM nidus which was at that point there, with the venous drainage through the facial vein, and a faint choroidal blush, with the superior ophthalmic vein shown at the top.

This patient underwent embolization, initially with Glubran, which is 50% histoacryl and lipiodol (making the material visible on fluoroscopy). Here you see the micro-Envoy catheter being passed up, with an intravenous heparin bolus being given to protect the circulation from the embolic effect.  Here you see the Glubran which has been delivered with very minor controlled reflux at that point there. This was all done in anticipation of the surgeon subsequently removing the lesion some four to six weeks later. Here is a very small amount of glue escaping into the cavernous sinus, here is a repeat angiogram after the embolization with almost no filling now. Here is the residual AVM with a very subtle colloidal blush, and the anatomical origin of the retinal artery.

We then took the patient to theatre. The question is often how soon after embolisation should should a patient be operated on. I think it could probably be as soon as one or two weeks later, and I’ll show you a patient later where we have just done that. This particular patient was operated on six weeks later, with, in effect a ‘glue-ectomy’. The pathologist was surprised when they received a mass of glue, having not been informed of the nature of the material, this followed by the advice to ‘pursue clinical pathological correlation’, the concern being that we might have sent the wrong specimen, but that indeed was what it was.

You see here the surgery undertaken through an upper eyelid medial skin crease approach, excising this mass of glue with surprisingly very little bleeding. I would say that the embolisation technique that Fergus performs results in very little per-operative haemorrhage, apart from the  initial incision through the skin. Here you see the embolic material being removed with a  reasonable result; indeed I suspect he was happy with the result because he never returned to clinic, despite my requesting on several occasions. In that particular case, the AV malformation was in abeyance and did not appear to refill from a deeper nidus at all.

Management of the next case involved the embolic material called ‘Onyx’. This gentleman had a very slow-growing, largely anterior, AV malformation, although I shall show you that as we close off one vascular bed, another area opens. Here he is in 2006 with a troublesome, slowly enlarging lesion. Its remarkable, in my experience, how little these patients complain in general, perhaps due to slow progression of the lesion. Here you see him almost ten years later, with significant progression, at which point he proceeded to embolisation of the lesion, and surgery. You see on the lower left image some early ischaemia at the lateral aspect of the lower eyelid.

Of interest, his visual acuity was, and remains, good. The CT scan shows a largely anterior lesion, but be aware that these lesions can have much deeper components, which can be problematic later on. Here is the radiology again demonstrating high perfusion with some saccular spaces. Although it would appear to be relatively anterior, these lesions are not necessarily localised to one area of the orbit, but can be part of larger region of disease. Here is the angiogram here, showing the territories and AVM nidus supplied by the internal and external carotid circulations.

What is Onyx? It is not a surface-active sclerosant agent (these are used to destroy the endothelium in venous lymphatic malformations), but a polymerising glue. Onyx is formed of an ethylene vinyl alcohol co-polymer with micronized tantalum (‘EVOH’). You will know from your periodic table that Tantalum (element number 60) was a god who was destined to stand in a pool of water below the forbidden fruit, forever tantalised by never being able to reach it. Here you see it on the slide – it’s a hard blue-grey, lustrous transition metal, highly corrosion-resistant, which causes less arcing and sparking with diathermy during surgery, although with bipolar diathermy we have not noted this phenomenon.

Here you see the lesion being removed surgically, and you see this mass of black tantalum, prolapsing out of the lid where the surrounding area has become ischaemic. Here you see this being removed before the lid skin crease is opened up. The eyelid is often very floppy and may need shortening. Here you see this ‘bag of worms’ being removed. I remember the first case when there was significant bleeding after the initial incision, thinking “if this is a sign of what is to come, we shall have a very long day of it”. Interestingly once we were through the skin, and in all cases we’ve had, the surgical dissection was surprisingly straightforward given the size of the lesion and the surgeon’s anticipation of haemorrhage from an arterialised lesion.

Now here’s another much smaller case, also being treated with Onyx. This gentleman had a troublesome pulsatile mass at the medial canthus. I approached this with a ‘gull-wing’ incision, a lovely approach to the medial canthus, causing no webbing and healing very well as you can see from the post-op. And here you see these snakey vessels being removed. Likewise this lesion also appeared to be relatively anterior and indeed localised, and did not appear to recur.

But here’s a much more difficult case that we’ve had recently, a lady who is still under treatment. You will remember some of her earlier photographs. This lady always described having a blush in the lower lid. Clearly, she has a ‘RASopathy’, but for whatever reason, perhaps hormonal changes in pregnancy, this lesion began to gallop, exponentially increasing in size soon after pregnancy, and continuing to be problematic. Again this patient was not in the least bit complaining, despite the fact that her eyelids are completely closed. She had a perfectly normal eye with a good vision, but was of course functionally blind due to the eye being closed. Here you see the lesion again being relatively anteriorly, although with experience of the first patient one wonders whether she may have a deeper lesion, as we suspected that at the time of surgery, which I’ll show you in a minute. Here you see the 3D reconstructions of the angiogram, showing this large vascular nidus in the lower lid.

A surgeon would understandably be anxious about approaching such a lesion, Onyx or no Onyx. In fact, she was embolised using the new ‘PHIL’ system (we shall meet PHIL in a minute). And here you see her presenting to the clinic only two or three weeks prior to surgery, with a deterioration over the past few days. It was much better after embolization by Dr. Robertson. But in the last few days it had became more swollen, with some pulsatility. Dr. Robertson felt that all the vessels had been closed off, but that she was hyper-vascular, with some chemosis as you see there. I called Fergus who promptly saw her, and she underwent further embolisation before she came to surgery.

‘PHIL’ stands for Precipitating Hydrophilic Injectable Liquid, this bringing the advantage of being free of tantalum, and therefore not sparking with diathermy. It is also more homogeneous with better radio-opacity and is less dense, allowing the radiologist to see the lesion and cast with better accuracy on fluoroscopy. PHIL was used in this particular instance; this is the lesion immediately pre- embolisation, there you see the nidus and this is the embolic cast. And I was fascinated to see that in fact this could be treated with an anterior injection to permeate all of these snaky vessels with the embolic material.

Now here she is at surgery, and of course there’s only one approach, the lower-lid blepharoplasty-type approach. And you can see here the skin has been opened below the lash line, with the lower lid margin here. And thank you Geoff – we managed to do this one together; we weren’t sure what to expect, but we were surprised to see how well it shelled out. We were able to define the potential planes, with quite a lot of hyaline material around these snaky vessels. Here she is again: most of the lesion has been removed with remarkably little bleeding. Dr Robertson felt that not all of it had necessarily been closed, although from a surgical point of view it seemed to have been. Here you see her preoperatively and postoperatively, with reduction of skin and a pressure bandage placed.

If I was to specify three key points from a relatively small number of patients they are these: (1) orbital AVMs can occur spontaneously, (2) they can follow a mild injury, as in the first patient, and (3) they can develop from a pre-existing vascular anomaly, which can be quiescent for months if not years. That said, they are readily resectable following embolization; polymer DMSO solvent (PHIL) is the current preferred embolic agent having superseded Glubran and Onyx, but do beware that reperfusion can occur if a deeper or residual nidus opens up.

Q & As section

Question 1
Any role for beta-blockers in trying to treat some of these orbital vascular malformations? We see a positive benefits in tracheal lesions and just wondering if you have any experience with that.

Answer
We don’t have that experience. I think it would reduce temporarily the vascularity of it, whether it may make embolization or surgery easier, it’s hard to say. I don’t think that beta-blockers would treat them, as their development and rate of progression may be genetically and hormonally determined. Clearly beta-blockers now have a major role in the capillary haemangiomas, but as far as I’m aware, not in the AV malformations. I do think it would ameliorate the patient’s symptoms perhaps but I suspect it wouldn’t change its rapid progression.

Question 2
What’s the ideal time to operate after embolization, and can embolization be curative without surgery?

Answer
I think that must be the case given that not all embolised intracranial lesions are removed. Naturally if they are closed with Onyx, with tantalum powder, then there is a need to remove it, because of that the tattooing under the skin. We’ve had patients we’ve operated on as late as two or even three months later, but the most recent patient I showed was operated on barely two weeks post embolization. Even at that point, there was good closure and there almost mummified, hyaline tissue throughout the lesion. Thus, answer is that surgery can be undertaken at two weeks post embolization. You could probably leave it much longer, but there is the risk that if one pressure bed is closed off, another one will open up, requiring a further cycle of re-embolisation. Whether one changes this risk by operating earlier waits to be seen.

Question 3
We had a child who had an embolisation and was planned for excision. However that was for an AVM of the ear and unfortunately he became unwell, so couldn’t be operated on. Two years down the line, the lesion didn’t change at all and it’s stayed static and finally completely resolved more or less. Four to two and half years or so, there is no recurrence. I was about to ask the same question to say: Do you ever consider that as an option? To say we can’t do the surgery.

Answer
I think that if you had a PHIL system and it was a small lesion, for example the patient I showed with the gull wing, it might be perfectly reasonable to leave it. If on the other hand you have used Onyx, then you do have to remove it. And if it’s as large as the large patient and again one would have to remove it, but I think there would be a role for selected lesions.