Endoscopic approaches to the orbit

Endoscopic Approaches to the Orbit transcript

Presenters: Benjamin S. Bleier, Suzanne K. Freitag.

We have some disclosures: I have textbook royalties and Horizon Pharma Ad. Board. Dr. Bleier has the following disclosures.

At Massachusetts Eye and Ear Infirmary, we have a centre for thyroid eye disease and orbital surgery. Ben and I are the co-founders of this, but our membership and people who work with us, are very similar to the people sitting in this audience. Our Center is filled with radiologists, interventional radiologists, pathologists, neurosurgeons, ENTs, and other people that we work very closely with, so we’re thrilled to be part of today. I can say 10 to 15 to 20 years ago in my practice, when I would see tumours like this, I would cringe and hope that patients didn’t need surgery. But now we are very happy to perform surgery on these patients with our collaboration.

As many of you know, the management of these tumours generally is observation, unless they are symptomatic or having optic neuropathy, as you can see here demonstrating choroidal folds, double vision, exposure keratopathy, and visual field changes. Traditional orbital surgical approaches for these patients were transcutaneous or transconjunctival, which you can see diagramed on the left.  The transcaruncular approach has really revolutionised the approach to the medial orbit.   We were often offering transcranial approaches to these orbit apex tumours.

The decision of which approach is based on the size and location of the lesion, with special attention to the location of the lesion with regard to the optic nerve. Sometimes it’s hard to tell and this is where we depend on our radiologist but generally, if you’re doing a pure orbital surgery, if the lesion is medial to the optic nerve, you might consider a medial approach, if lesion is lateral to the optic nerve, you might consider a lateral approach. However that’s a little bit thrown out the window, as Ben will explain in a few moments. Further, we always want to be careful of the optic nerve, as that is the biggest danger problem in the orbit. We traditionally operate on the side where the lesion is, but now we’re doing some crossover and want to be cognisant of course at the optic nerve. We’re doing these cases endoscopically through the nose for a large part.

‘And why do we do endoscopic surgery in general?’ Well, it gets us access to difficult to reach areas, it magnifies and illuminates our surgical field in a way that we’re not used to as orbit surgeons, it minimises cutaneous incisions, it minimises damage to surrounding structures, and allows patients a rapid recovery from their surgery. We don’t have effective trans-orbital endoscopes, but transnasal endoscopic surgery is a very expanding field.

The indications for these transnasal endoscopic approaches that we’re doing include removal of orbital apex masses and other masses in the orbit, removal of foreign bodies, diagnostic biopsies and decompression. We set up our room as such.  It’s usually more crowded with humans, but this is the basic endoscopic setup with the ability of two surgeons to work through the nose. We use a stereotactic image guidance system, which most people in this room are likely familiar with.

The surgery starts with a standard endonasal approach to surgery with a maxillary antrostomy, ethmoidectomy, and sphenoidotomy. The contralateral nasoseptal mucosal is harvested and kept as a flap; the blood supply is preserved, the flap is stored in the nasal pharynx during the surgery, and it’s very critical to our reconstruction of the medial orbital wall later in the surgery as you will see. A large posterior septectomy is performed, so instruments can be passed through both nostrils; there’s not enough space through one and we like the angle of going contralaterally. The diagram here shows on the patient’s left side there’s a little red mass, that’s the lesion. The endoscope is passed through the contralateral nose, which is driven by the assistant surgeon. A retractor is placed through the contralateral side, driven by the surgeon. A dissecting instrument is placed through the ipsilateral side and used by the surgeon, and this leaves the assistant a spare hand for suction and other needed assisting.

One of the questions that comes up here  – and also as you’ll see we’re presenting like we operate here as a team – is who is an endoscopic candidate and as Dr. Freitag mentioned, traditionally the thought was if the tumour is medial to the optic nerve then they can go the medial approach may be appropriate, but anything that crosses the nerve may be dangerous. The issue is that because we’re doing what we would call a binarial approach or by through both nostrils. In addition to allowing multiple surgeons to operate, this also gives us a unique angle from the contralateral side and so what this allows us to do, is to create this sort of plane where we’re coming and sneaking up under the optic nerve.

As I mentioned, the traditional sort of teaching of dividing the orbit into medial and lateral sort of goes out the window, and now essentially it’s only tumours that are in the lateral and superior quadrant of the orbit are absolute contraindications for a primary endoscopic approach, although an assisted approach could still be considered. But because we can sneak up under that, we can actually get to almost three-quarters of the orbit from an endoscopic perspective. This isn’t just theoretical – these are three examples of cavernous haemangiomas, all of which extend lateral to the optic nerve, at least have components that extend lateral which we were able to completely resect.

When we think about the endoscopic approach, one of the issues that we had to develop was an understanding of the endoscopic anatomy of the medial intraconal space, and this was not something that was well established in the literature. But if we think about this space for the ophthalmologists and oculoplastic surgeons this may be more familiar, but we think about drawing a vertical line through the inframedial muscular trunk of the ophthalmic artery. To us this divides the orbit into an anterior compartment and a posterior compartment, where if lesions arise anterior to this, then we don’t have to really involve the orbital apex, the ocular motor nerve and so forth.

Further, we have to think about retracting the medial rectus muscle, thus it becomes important to create this hypothetical line that goes through the axis of the medial rectus; and so lesions that are inferior to that, we will retract the medial rectus superiorly, and lesions that arise superior to this line the medial rectus will be retracted inferiorly. Now this becomes important because with inferior retraction, we get closer to the anterior and posterior ethmoidal arteries and their neurovascular bundles, and so technically those lesions in the anterior-superior quadrant can be more complex. What we did, was we took those concepts of compartmentalising the medial intraconal space and created a staging system.

This staging system (CHEER) was developed specifically for cavernous haemangiomas that were being endoscopically managed, but we do feel that this can also tend to start to be expanded for other lesions. The stage I are lesions that are purely extraconal. What you’re seeing here, is a figure of the left orbit and as we go into the upper stages, we’re going to divide the orbit into those compartments based on that inferomedial muscular trunk and based on the whether it’s superior or inferior to the medial rectus muscle. A stage II lesion, is going to be a lesion that’s anterior to inferomedial muscular trunk take off of the ophthalmic artery, and inferior to that line through the axis of the medial rectus muscle, and here we’re going to show examples of all of these lesions. Stage III is anterior and superior. Stage IVA is a lesion in the orbital apex but posterior to the inferomedial muscular trunk, and stage IVB are lesions that can be in the apex or extend or are exclusively within the optic canal, like you can see here. Stage VA lesions, are lesions which extend into the pterygopalatine and infratemporal fossa, and so this increases complexity with more vascular dissection. Then finally stage VB would be lesions like this extending through this superior orbital fissure intracranially.

In general, when we think about CHEER stage, we increase complexity as we go from low to high stage, and this has been validated with an international group who have helped to develop the staging system. Now let’s just look at a couple examples. This is a patient with a stage IVA cavernous haemangioma which is going to be, in this case, exclusively resected endoscopically.

What you’ll see here is that figure where we’re using again this binarial binostril approach. You can see here, by definition we have three instruments: we have the scope that’s looking at what’s going on and then the bimanual dissection. We make a very controlled incision in the periorbita, because we want to use the periorbita to retract the extraconal fat particularly in the anterior half of the orbit. Then we have to identify the medial rectus muscle; and then in these apical lesions, we’re going to go inferior to the medial rectus.

We’re going to go through this transseptal approach to retract the medial rectus into the nose. As we do that, what we’re going to see are these bridging vessels. These are the vessels of the ophthalmic artery that serve the medial rectus muscle and our lesion is going to be behind those. As we do a more posterior dissection, you can actually see here the oculomotor nerve trunk supplying the medial rectus, in this case the inferior division. As we continue our dissection, we start to see the cavernous haemangioma and the reason these lesions, as we’ll talk about a little bit more in a moment, are a nice lesion to start doing these types of dissections with, is because they have a nice capsule, they don’t tend to invade the surrounding tissue. What we want to do is essentially a 270 degree dissection avoiding dissecting and retracting directly on the optic nerve.

As you can see, there are a variety of ways to manage intraconal fat and bleeding during these dissections, but we really want to avoid removal of fat and we want to avoid any type of cautery. What we do here is we place a cottonoid that’s saline soaked, and this allows us to not only retract tissue but also to suction on the cottonoid without suctioning directly on the fat, so we can wick away blood, and we can really get nice exposure of that lesion. Once we have that 270 degree dissection, we’re going to gently retract in the direction of the optic nerve, just to let those final fascial bands lyse and then we’re going to pull it out. If you’re doing this and you’re feeling a significant amount of resistance, then you have to go back and continue your dissection, but here you can see that just pops out nicely. This was one of the first cases we did and that’s the point where Dr. Freitag tells me “just pull”.

Now, I never would have done that on my own and this is the type of multidisciplinary type of work that we need, because there’s a lot of expertise from both fields that come into play. Here you’re going to see where we reconstruct the orbit with this nasal septal flap, and then this is the appearance of the orbit after three months, so you can see the flap is nicely healed in with a little bit of just granulation tissue up there, and what’s nice about this as you can see that the extraocular function is completely intact, and also there’s nice retention of orbital volume, and that’s important with respect to the orbital reconstruction.

What we’ve shown you so far and the CHEER staging system, was all developed around the diagnosis of cavernous haemangioma; and these are of course the most common orbital tumour in adults, they have minimal adhesion to the orbital structures and this is a really great place to start learning these techniques, you can lyse them easily from surrounding adhesions with a cotton tip applicator and then you can grasp or cryo the lesion and simply remove it.

But unfortunately not all orbital tumours are cavernous haemangioma; so we see solitary fibrous tumours and schwannomas as the next most common primary solid tumours in the orbit, and they are much more adherent to surrounding structures. In these cases we consider an approach from the orbit side concurrent to the endoscopic approach. A trans-orbital approach combined with the transnasal endoscopic approach has many advantages: the lateral portion of the tumour is dissected from the surrounding structures via the orbital side, the trans-orbital tumour exposure is much better with the medial orbital wall opened, and the tumour is then pushed from the orbit into the nose and we’ll demonstrate this momentarily.

This was a recent case we had where you can see, at least from the ENT perspective, a pretty dramatic presentation with this left-sided proptosis. When we scan this patient, she had quite a large solitary fibrous tumour filling the majority of the orbit, but if you look at the characteristics of the lesion, as I look at this from an ENT perspective, the lesion extends above the skull base and is really adherent to the superior aspects of the orbit, and potentially even superior rectus and superior oblique. When I look at that, I said “these are not areas that I can comfortably dissect from an endoscopic perspective”, and again that’s why we start to think about the combined approaches.

This is an example of that patient where we do both: an endoscopic and trans-orbital approach, so we can optimise aspects of both. Again we’re going to make a controlled incision in the periorbita. Now unlike that first case we showed you, this lesion because it’s in the superior orbit, it’s sort of that equivalent of a CHEER stage III. We’re attracting the medial rectus inferiorly, and we’re getting good exposure and we’re dissecting the medial superior aspects of the lesion. There you see the transcaruncular approach. Then here, Dr. Freitag and I are both working to dissect both the medial and superolateral aspects of the lesion, and you can see the light coming through on the endoscopic side. Now where this is a little bit sped up, but what you can see here is, we start to deliver the lesion and as we do that, we continue to do dissection, we don’t just pull blindly but we slowly deliver that through the nose. Now here you can see coming from the right nostril through the septum looking into the left nasal cavity, so that’s called a transseptal approach, but we can repair that without creating a permanent perforation in the septum, and then this is the nasal septal flap that we put in place to reconstruct the orbital wall. This is the preoperative scan, this is the post-operative MRI.

Now interestingly, we just saw this patient.  She wrote both of us about a four-page letter, and one of the things I want to note in the letter she said “There was, I saw many doctors about this, there was no consensus but the only overall consensus was: don’t go through the nose” so we thought that was kind of funny. But here we see the pre-op and the post-op imaging, now you can see there is a little even with the orbital reconstruction there is a little bit of enophthalmos here because of just the mass of that tumour, so reconstruction of the orbit really is an important feature of this.

Talking about some technical considerations of this type of surgery, there are issues that are common to all orbital surgery, which is the orbital fat prolapse and the need for absolute haemostasis. But there are some aspects that are unique to this transnasal or combined approach, and that is the difficulty in retracting the medial rectus muscle and a large medial orbital wall defect that’s needed for access. As Ben mentioned, we have to be careful of the neurovascular bundle as it inserts in the posterior aspect of the medial rectus. We have been using just manual instrument elevation and retraction, although there has been some discussion in literature and at meetings of using a suture or another device, but this way every time we take the scope out or change instruments, we are releasing the medial rectus and it perfuses;  and we’re doing well with regard to double vision issues.

There’s also the medial wall defect management, and Ben’s already discussed this a bit, but these large defects predispose people to enophthalmos and diplopia, we do feel that immediate reconstruction is beneficial, and we feel that this flap is our best option. We rarely get enophthalmos in our cases.  I think with this solitary fibrous tumour we just showed you, she had a lot of orbital fat issues just because the tumour was filling so much of the orbit.

The bimanual technique requires this large posterior septectomy and this septal tissue is really useful for the reconstruction: it provides support for the orbital contents and a barrier between the orbit and the sinus. And this was a really cool scan we wanted to show you; so this was a small lesion in the optic canal which was symptomatic: the patient had visual field loss and optic neuropathy from it, and so we removed it and we just happened to get a slice on the post-op imaging showing this nasal septal flap in position.

Our key points are: the transnasal endoscopic surgery provides a safe, effective and efficient approach to many orbital lesions, careful patient selection is critical for safety and success, and a multidisciplinary collaboration between rhinologic and orbital surgeons maximises experience.

Q & As section

Question 1
Do you find that cavernous haemangioma striation-regarded as freely able to be removed? I find that there tend to be two sorts: there are those that come out really easily and then there’s a more fibrous version, which tends to incorporate things like passing nerves vessels, which are harder to remove, Have you noticed that with your endoscopic approach?

Answer
We haven’t had any particularly sticky haemangiomas in our combined efforts. I’ve certainly met them personally, but we approach all of these surgeries with the patient that there is an option to go in trans-orbital if needed, so all patients are expecting that possibility.

Question 2
I understood, from our ENT colleagues, that restoration of nasal function (as opposed to just simple anatomical closure of surfaces) takes about nine to twelve months to re-establish; for example the mucous carpet flow, properly across a reconstructed area, and the airflow patterns, clearly are never going to be the same following an ethmoidectomy. Do you find that can actually cause long-term problems? Does it leave any deficit?

Answer
We have extensive experience with manipulating the nose for endoscopic neurosurgical approaches and we know that those patients do actually very well. There’s excellent quality of life outcome studies that show that nasal function is not adversely affected. In fact, we published a study on after orbital decompression, where we do bilateral complete sinus surgery, expose a whole bunch of extraconal fat and those patients’ quality of life outcomes scores that are specific to their nasal outcome, are actually better than pre-op, because they’re so wide open. When you have normal nasal mucosal function, it heals beautifully, and that takes about three months not nine months. Now if you have patients that are radiated, I’ve had other types of tumours, that’s a whole different ballgame but in these cases, these patients have completely normal nasal function.

Question 3
How do you secure the nasal septal flap at the end? How do you keep that in place?

Answer
Because this is a vascularised flap, it’s exuding blood, exuding fibrin and so forth, so it’s quite sticky. What we do, is we essentially just put absorbable packing over it and then it sticks to the underlying tissue quite nicely. Again, from a historical perspective, these flaps were originally developed in the mid 2000s for reconstruction of large skull based defects after neurosurgical approaches, so we have a lot of experience on how they work, how long they take to heal and so forth. We use sort of soft packing because we don’t want to actually put rigid or firm packing in the nose, and get a compartment syndrome in the immediate post-operative period.

Question 4
The soft packing, how long do you leave on this?

Answer
Well it’s absorbable, so it just washes out on its own.

Question 5
You mentioned reconstruction because of your concern about enophthalmos, but do you have your own control group where you didn’t reconstruct and found that enophthalmos was a problem and then started reconstructing?

Answer
Just our experience with orbital decompressions and fractures knowing the risks.

Question 6
A. As I understand it, doing the nasal septal flap leaves a fistula. I wondered what problems you see with that.

1. For the posterior defects, because often we need a convexity to reconstruct the normal shape of the orbital bone. I wonder whether the nasal septal flap is actually achieving that reconstruction which you would need, and whether it really is necessary and maybe reserved the nasal septal flap for another occasion for something else, but perhaps reconstruct the convexity if you need to, if indeed it is a problem long term.

Answer
A. I think the term fistula may not be anatomically appropriate, maybe you mean a perforation or a septectomy.

1. So a posterior septectomy, meaning an opening between the right and the left nasal cavity in the back half of the septum, is totally asymptomatic because you’re so deep in the nose that airflow humidification is not really affected. When we have lesions that are in the mid to anterior orbit, we actually do what’s called a transseptal approach, meaning we temporarily elevate both sides so we have access, but then as you saw in that second video we showed that, so there is no permanent hole in the septum or perforation at the end. The side where the flap comes from, re-epithelialises. Once everything is healed, which again can take about up to three months, both sides are fully re-mucosalised, there is no long-term perforation for the mid to anterior lesions.