Optic nerve sheath meningioma: Diagnostic challenges

Optic Nerve Sheath Meningiomas: Diagnostic Challenges
Presenter: Tasanee Braithwaite

Over the next 15 minutes, I’m going to talk about primary optic nerve sheath meningiomas, some of the diagnostic challenges that they can present and some mimicking conditions. Further, I will talk to you about optical coherence tomography (OCT) as a relatively novel imaging biomarker for the diagnosis and management of this disease. I have no disclosures.

Primary optic nerve sheath meningiomas (ONSM) arise predominantly from the meningoepithelial cells lining the intraorbital and intracanalicular optic nerve. The optic nerve sheath can also be involved secondarily by meningiomas creeping along from the olfactory groove, the sphenoid wing, and the tuberculum sella. They are true neoplasms and they like to follow a path of least resistance, so they typically grow to encapsulate the optic nerve, and as they do so, the optic disc can be seen to swell. This probably occurs as a result of axoplasmic transport hold up. With ongoing compression, there can be demyelination of the axons and eventual axonal loss; and at that point we start to see optic atrophy and the consequent impacts on visual function.

A morphological classification system has been developed by Schick and colleagues, and this system is based on description of the different shapes that they can adopt. There are fusiform, tubular and globular shapes within the orbit. The meningiomas can creep through the optic canal and towards the cavernous sinus. They can extend for more than a centimetre towards the chiasm and across to the contralateral optic nerve.

In terms of the epidemiology, if we consider meningiomas generally, the prevalence is at least 100 per 100,000 population, based on pathologically confirmed cases. But autopsy and imaging studies reveal a very much higher prevalence in the population, particularly in the female population, which may be as high as 2% to 3%. Atypical and malignant forms form less than 5% of this total. The chief risk factor -which has been established- is previous exposure to ionizing radiation. This emerged from studies in the 1950s, where children with tinea capitis were treated with radiation caps, and they had a tenfold higher incidence of meningioma compared to the general population. Only 1-2% of all meningiomas arise primarily from the optic nerve sheath. And as you can see from this graph, the incidence goes up with age and is much higher in women. The little blue diamonds in this graph illustrate the female to male ratio, which peaks at around the 35 to 44 year age bracket, when the ratio is about 3.5 women to 1 man who are affected.

Our understanding of the natural history of ONSM has been informed by numerous case series in the past century, and by a particularly large review of the literature in 1992 by Dutton. At that time, there were 488 cases of optic nerve sheath meningioma reported worldwide, and a subsequent case series from Amsterdam and Vancouver added a further 88 cases (Saeed et al, 2003). We see from these series that the median age of onset is around 40, that women are affected more than men, about 4 to 7% of cases are children and in those children, 2/3 have associated neurofibromatosis type 2. Sometimes, the meningiomas particularly present in women during pregnancy, indicating a hormonal influence.

In these case series, the chief presenting symptoms are decreased acuity and other impacts on visual function, such as loss of colour vision and field defects; patients can also present with transient visual obscurations which may be gaze evoked; and sometimes with pain which can confuse the initial picture in terms of the differential diagnosis. Diplopia can also result, and in about 5% of cases the presentation is bilateral and sequential because of spread to the contralateral optic nerve. On the other hand, regarding presenting signs, approximately half of patients present with optic disc swelling and another half present with optic atrophy. Optociliary shunt vessels, which connect between the retinal and choroidal circulation, are tremendously helpful to suspecting this diagnosis, and these occur in about 25 to 30%. Patients can present with proptosis, although this is generally mild when the tumour is originating primarily from the optic nerve sheath rather than from the sphenoid wing. This probably results from a mechanical restriction of the extraocular muscles, secondary to the stiff optic nerve sheath with a smaller proportion coming from true cranial neuropathy.

The prognosis in the Saeed case series of 88 cases which had a seven year follow-up, was that approximately 25% developed vision loss to ‘no perception of light’ during that interval. Dutton proposed that, given long enough follow-up, essentially everybody would progress to no perception of lights. The pathognomonic triad of signs- reduced visual acuity, optociliary shunt vessels and optic atrophy – was seen in only 25% of patients. Whilst optociliary shunt vessels can be seen in other conditions, such as central retinal vein occlusion or chronic glaucomatous optic neuropathy, other clinical features in those conditions readily differentiate them.

Thinking about some diagnostic challenges, we have this case. This is a patient I saw in my clinic recently, a 51-year-old lady who had presented elsewhere six years ealier, with gradual onset painless mild vision impairment in one eye to 6/9 visual acuity initially, with some reduction of colour vision. She had a degree of optic atrophy and this unilateral optic atrophy was investigated, looking for nutritional and toxic causes, even though it was asymmetric; the vitamin B12 was checked, folate, and the methylmalonic acid and homocysteine (which can indicate functional B12 deficiency) were normal. She didn’t have a significant alcohol intake, she had no history of sexually transmitted infections, and her Treponema serology was negative. A chest x-ray was normal, as was her bone profile, and serum ACE. After two years of follow-up at that time, her vision function remained stable at 6/9 and so she was discharged with a diagnosis of idiopathic mild optic atrophy. But she represented this year because there’d been a much more rapid progression of vision loss over four months, down to 6/36 with 4/17 Ishihara plates. The imaging done originally was an MRI head without contrast, it had shown no gross compressive lesion, but repeat imaging recently with gadolinium contrast and with fat suppression and dedicated orbital views, revealed the classic “tram track sign” of the optic nerve sheath meningioma.

In this context, I would like to discuss the role of Optical Coherence Tomography (OCT). Essentially on the left side, we have the normal configuration of the retinal nerve fibre layer. This scan takes a 360 degree circle around the optic nerve head, looking at the thickness of the layer between the red line and the blue line, which is the retinal nerve fibre layer. And underneath the scan, is this schematic trace of the 360-degree scan. It providing essentially a traffic light system; green is normal, the line in green provides the mean RNFL thickness of a normal reference population. The retinal nerve fibre layer thickness contour follows the ‘ISNT rule’, with the neuroretinal rim thickness generally being greatest inferiorly, then superiorly, nasally, and the temporarily rim being thinnest. The patient’s retinal nerve fibre layer thickness is depicted by the black line, and you can see on the other side that in this patient, the thickness of the retinal nerve fibre layer is dipping into the red line and that’s indicated here.

Spectral domain OCT is the tool that we have been using for some time to look at thinning of the retinal nerve fibre layer, particularly in glaucoma patients. With enhanced depth imaging, we’re now able to look at much greater detail at the cross-section of the nerve head, and you can see really fantastic detail down to the choriocapillaris and to Bruch’s membrane, and some detail on the position of the cribriform plate here. This was not really possible five years ago with the standard time domain OCT imaging that we were doing. OCT is tremendously valuable because it’s non-invasive and it’s quick to do. For example, at Moorfields, we are doing hundreds and hundreds of these OCT scans daily on patients across our services.

A further particularly valuable application in neuro-ophthalmology, is segmentation of the macula volume OCT scan. Here you can see that the layers have all been depicted in these different colourful lines, and particularly useful to look at in neuro-ophthalmology, is the retinal ganglion cell layer. It’s useful because in that layer are the cell bodies which project their axons through the optic nerve to synapse in the lateral geniculate nucleus and the thalamus and at other locations. If you have axonal injury in the optic nerve, you get retrograde degeneration usually over about a month or two. So if you have a severe optic neuritis or compressive optic neuropathy, about a month or two later, you see thinning of the retinal ganglion cell layer and this tells you that there’s been irreversible damage to axons; and the visual acuity and the vision function that you measure at that point is what you might expect to the patient to persist with at best. Whereas when we find that this layer is still healthy and thick, and a month or two has gone by, we can be hopeful that there hasn’t yet been permanent injury to the optic nerve. So it’s a very useful tool for preoperative assessment, for post-operative management and for decisions on how quickly something is progressing. In this patient, on the right-hand side, you see the normal retinal ganglion cell thickness. It’s like a yellow healthy doughnut, but on the left side you see a sort of red wasteland of thinning, because this patient has had marked loss of that ganglion cell layer.

At other times we see that an indolent perioptic lesion which has resulted in mild optic atrophy, which we presume to be an optic nerve sheath meningioma, in fact, isn’t. In this second case, the patient presented with 3/60 vision, a relative afferent pupillary defect, and complete loss of colour vision. Their radiological assessment of the orbits, by MRI with gadolinium and fat suppression, was completely typical for an exophytic right optic nerve sheath meningioma, extending through the optic canal and via the superior orbital fissure, to involve the anterior cavernous sinus. This patient was monitored for two years, her vision function remained very poor and then all of a sudden, she had spontaneous recovery: first to 6/9 and then to 6/6 with complete recovery of her colour vision! This is the persisting field defect which is much smaller than it had been years before. The patient, with a ‘disappearing meningioma’ (which of course can not have been a meningioma at all) continues to do well. She had no systemic symptoms, and extensive investigations were all negative for other causes of an inflammatory or infectious compressive optic nerve mass.

This just goes to show that sometimes the diagnosis can be challenging. OCT technology wasn’t available when this patient presented 10 or so years ago, but it would have helpfully revealed that despite the bad visual function, the retinal ganglion cell would have looked pretty good. Maybe there was a block to axonal transport, maybe there was a degree of demyelination, but there can’t have been irreversible loss of axons. With reduction in the size of this presumed inflammatory mass, the axons were able to recover their function, and so the vision function recovered.

Sometimes optic nerve sheath meningiomas can be mistaken for optic neuritis and vice versa. This is a picture of an optic neuritis case where you see marked swelling and elevation of the optic nerve head. The retinal nerve fibre layer here is very oedematous and thickened. The machine hasn’t actually done a particularly good job here of segmenting the layers accurately, so you can see that that blue line is not actually tracking where the retinal nerve fibre layer is, and inaccurate segmentation is one of the issues that we face in imaging swollen optic nerves. It is important to manually check the segmentation when the OCT is ‘abnormal’. But this patient with a painless optic nerve head swelling and marked vision loss was thought of as having an atypical optic neuritis, so not a typical demyelinating picture. She was treated with urgent high dose steroids. There were clues in the systemic history- she had profuse night sweats and a dry tickly cough. FDG-PET imaging subsequently revealed extensive lymphadenopathy and she  underwent biopsy via EBUS (an endobronchial biopsy) which confirmed the diagnosis of sarcoidosis.

Some other examples. When colleagues Dr. Petzold and Dr. Plant at the National Hospital for Neurology and Neurosurgery reviewed their case series of over a hundred patients with optic neuritis, they found one case who in the end, 10 years later, was re-diagnosed as having an optic nerve sheath meningioma. This reminds us that 5% of the time, the presentation of optic nerve sheath meningioma can be much more swift. In those cases, you need to particularly worry about a higher grade malignancy.

These are two further cases from the literature, to exemplify the differential diagnosis. In these cases, the eventual diagnosis was of a malignant melanoma that had metastasised, and one that had originated de novo in the orbit; and those patients progressed rapidly and died soon after presentation. Other mimics of optic nerve sheath meningioma can include IgG4-related disease. This is a patient who presented with proptosis, vision impairment and disc pallor, and she was treated with radiotherapy with some regression. But the lesion regrew 12 months later and multiple lesions elsewhere on the CT with an elevated IgG4, gave a clue to the diagnosis. Her subsequent histopathology confirmed IgG4-related disease, which is not typically very steroid responsive.

In addition, in the differential can be Rosai-Dorfman disease, in which – in simple terms- macrophages are seen to be eating lymphocytes. Typical histopathology was seen in the cervical lymph node biopsy. The last case reveals that sometimes a trial of steroids can be very helpful, if an inflammatory mass is in the differential diagnosis, to avoid the need for an unnecessary biopsy. Here you can see an optic nerve sheath meningioma, not an inflammatory problem, but the optic nerve sheath meningioma had resulted in a sort of focal compartment syndrome and with the trial of steroids, the patient had some improvement in their visual function. Other conditions in the differential, like Rosai-Dorfman disease, might also be expected to improve with a steroid therapy trial.

In summary, OCT is a very valuable emerging imaging biomarker. Systemic enquiry can help to target other investigations, and an empiric trial of steroid can be helpful where the diagnosis is uncertain.

Q & As section

Question 1
A quick question about OCT. I think you showed that there was some thinning of the retinal nerve fibre layer in these conditions. Does it follow any particular distribution? At Moorfields, research has demonstrated a role for automating OCT segmentation and artificial intelligence analysis, and I am wondering if optic nerve sheath meningioma is something that could masquerade as a different type of optic neuropathy.

Answer
With respect to optic nerve sheath meningioma, I think the pattern of retinal ganglion cell thinning doesn’t help us very much with the differential diagnosis. At Moorfields, there was a landmark paper published last year in Nature Medicine. In this research, colleagues in the Medical Retina Service at Moorfields, in collaboration with Google Health, formerly DeepMind, used OCT images to train an artificial intelligence algorithm to diagnose 50 different retinal diseases, in order to determine patients who might need a referral for treatment of a sight-threatening problem affecting the macula. The algorithm had comparable diagnostic accuracy to the Professor of Medical Retina at Moorfields. It was an amazing breakthrough but required huge numbers of OCT images to be annotated by experts, to label what the different scans were showing in order to train the algorithm effectively.

I think there could be a role for this sort of technology for the monitoring of optic nerve-involving pathologies, but for that to be possible a few hurdles need to be overcome. Firstly, we don’t yet have enough of these OCT images to train an algorithm, as we’ve just been really collecting these sorts of images in neuro-ophthalmic practice for the last few years and you need a really large volume. Secondly, the data needs to be annotated by experts and we’re still gaining understanding of what the clinical correlation is with the features that we’re finding. There are also some issues with segmenting the layers around the disc accurately, because it’s a thicker structure and even with enhanced depth imaging, the technology of OCT doesn’t quite get deep enough.

Question 2
Can you tell us something about the role of HRT in these tumours, as we heard that it was a risk factor in sphenoid wing meningiomas?

Answer
There are some cohort studies which have looked at whether HRT drives the incidence of meningioma, and Geoffrey said earlier that it was felt to be a risk factor. Some of those large studies have been done in America. There’s the 1.3 million women study and another large cohort study, and they found an increase in the relative risk of about two for women who had pre-menopausal HRT therapy, compared to women who were post-menopausal and had never had HRT therapy. But other studies have been certainly less conclusive on the role of oral contraceptives, for example. What’s emerged from this epidemiological data is that a lot more research will be needed in the future, thinking carefully about study designs. To best answer the question, these will need to stratify recruited patients by whether the meningioma is expressing estrogen receptors, progesterone receptors, and/or androgen receptors, and gather robust data on the duration and dose of hormonal therapies, and whether they’re estrogen or progesterone based hormonal therapies. I think that epidemiological data hasn’t yet been collected and it may lead to a transformation in practice, potentially in the way that the expression of these androgen and hormonal receptors in breast cancer has transformed therapy, so to conclude – a pitch for ‘more evidence needed’ here.