Optic nerve sheath meningioma: Neuroradiology

Neuroradiology of Optic Nerve-sheath Meningiomas transcript
Presenter: Indran Davagnanam

What I’d like to do in the next 15 minutes in this whistle-stop tour of meningiomas, is to talk a little bit about the role of imaging, about the modalities that we use, optimised protocols, imaging features in identifying optic nerve sheath meningiomas, some of the pitfalls and differential diagnoses, errors that we encounter as part of the imaging process and what the future will hold for imaging of meningiomas.

This is my standard go to slide; it illustrates what the role of a diagnostic neuro-radiologist, it encapsulates the points that we’ll cover during this talk. We (neuroradiologists) have baseline knowledge that we apply through multiple modalities, which serve to detect, diagnose and where possible, provide reasonable differential diagnoses of pathological processes. Thereafter, it’s about providing value in terms of staging and aiding entry into therapy.

Imaging plays an important part in imaging follow-up and therapeutic response. At our disposal we have a multitude of different imaging modalities. I think earlier on today, Dr. Fergus Robertson -one of my colleagues- would have talked a lot about digital subtraction angiography, so we’re going to dispense with that.

The main workhorses in our institution are CT and MRI, I think most of you will be familiar with utilising those modalities. What you want to actually get, are these really elegant images, both on CT and on MRI, of the well-defined optic nerve sheath complex. The trouble is that you can’t quite define the optic nerves separate from the sheaths on CT, but utilising MRI you have that very elegant discrimination between the optic nerve right in the centre and that sheath that surrounds the CSF containing sub-compartment. This is essentially based on very fine specialised protocols that we do at Moorfields; we’ll talk a little bit more about the protocols that we use later.

Further back, the relevant bit of anatomy is the optic canal. This bit of anatomy poses a diagnostic imaging challenge in terms of trying to get some tissue discrimination. We can see that the canal is defined by its bony outline, but actual tissue discrimination is actually quite poor within the canal, so we really are reliant on MRI to try and delineate, 1.the instrinsic anatomical structures and 2, any pathology that may affect it. You can see very clealy, the intracannalicular segments of the optic nerves on the MRI scan.

Most optic nerve sheath meningiomas are intraorbital and tend to be unilateral, so we rely a lot on the asymmetry on CT and of the asymmetry of the optic nerve sheath complex. There’s some work detailing the morphology of these tumours. I personally don’t think it adds a tremendous amount of value, it can be useful, but it doesn’t add any clinical value.

Tram-track calcification seems to be almost pathognomonic of this particular entity, but that’s variably seen in the case series that have been published, between 20 to 50%. Pneumosinus dilatans, for those of you are unaware of what it is, it’s the disproportionate dilatation of the posterior ethmoidal air cells or sphenoid air cells, which was once thought to be pathognomonic of optic nerve sheath meningiomas or any meningiomata in the skull base region, has since been proven to be not the case.

MRI really is the gold standard because of that tissue discrimination and that ability to assess the  optic canal. Only 8% of all optic nerve sheath meningiomas are isolated canalicular lesions in origin. We do need to be quite mindful of the fact that in patients with a tumour suppressor gene defect, such as NF-2, tumours of this kind, can be bilateral so using that asymmetry on CT is going to fall short in assessment.

Strangely enough in bilateral cases, almost 65% of the cases are exclusively intracanalicular, so you might be missing something within the canals in these cases. Here are some examples of calcification; this is your classic tram-track like calcification, you can see those almost parallel lines of high-density, which are on either side of the optic nerve in the middle. These can be of different morphology, so you can get more punctate areas of calcification and, looking at it from a coronal aspect, you can get variable degrees of “doughnuts”.

CT imaging findings of a typical retrobulbar meningioma, is one of an enhancing tumour, it can be of various morphologies as we talked about and I don’t find it particularly very useful. Here’s one that’s tubular and partly exophytic, hence it doesn’t quite fit into one morphological category.  You get proptosis, the posterior indentation because of the mass effect and therefore the refractive errors that are sometimes associated with it. For the keen-eyed, you can see that there is a soft tissue component there within the intracranial compartment, it’s not that great for the immediate retro orbital regions. Again, we revert to MRI for this, and typically on an MRI what you’ll find secondary signs of a slowly progressive optic neuropathy on imaging; a slender optic nerve on the left here – that’s patient’s left – with some intrinsic hyperintensity, all suggestive of an established optic neuropathy in evolution. You also get the relative disproportionate dilatation of the optic nerve sheath. Contrast really delineates the true extent of the tumour, there is enhancement of the optic nerve sheath, that tram like enhancement and more “doughnuts” in this case, a “doughnut-like” appearance on the coronal imaging.

Remember, meningiomas can be bilateral. Here’s an example of bilateral ectatic and torturous optic nerve sheaths, and again signs of bilateral optic neuropathies with very slender optic nerves. On the post contrast imaging, quite clearly delineated bilateral optic nerve sheath meningiomas.

You may think that in this case, there is a left-sided intraorbital optic nerve sheath meningioma, but again of those of you who are very keen-eyed, there’s a little fleck of calcification there within the optic canal. When you look at the optic canal on MRI, there’s a small bit of soft tissue that’s hyperintense, but on the post contrast MR imaging, the tumour is almost exclusively intracanalicular  with no involvement of the intraorbital segment. The distended and tortuous nerve sheath is due to a compartmentalisation of the CSF.

Nuclear medicine has its role. Dr. Miszkiel talked a little about Gallium Dotatate scans. They’re extremely useful for any tumours containing somatostatin receptors and that goes for these tumours. Other lesions which contain somatostatin which may occur within the orbit, for example metastatic neuro-endocrine tumours, are also positive on Dotate scans, Regardless, it still is quite useful in distinguishing meningiomas from other common intraorbtial tumours such as schwannomas and pituitary adenomas whcih extend into the orbit.

Sonography is not used very much nowadays but you were able to tell, based on the refractivity patterns on the A scan, imaging features which were typical of an optic nerve sheath meningioma. Its utility is now largely historical as it’s not used very much nowadays.

Purely based on imaging, there are a myriad of different imaging mimics, and this what Katherine was talking about, can generate a long list of differential diagnoses. In isolation, this particular case shows signs of an established optic neuropathy, signs of very fine tram-track like enhancement which may be typical of a meningioma. However,  it underlies a more extensive systemic process as you can see. This is sarcoid perineuritis with extensive dural enhancement extending into the optic canal.

Other things may look quite similar. Here we have a case of bilateral papilloedema: you can see the indentation of the posterior globes that go along with that. When you look at the right optic nerve sheath complex, you can see some low signal intensity, which may represent calcification, therefore this may be a meningioma causing papilloedema on one side. But ask yourself, ‘ what’s the cause on the other side?’ So again the keen-eyed in the audience may spot tiny dural disease there, but the reason for the ‘dilatation’ as you can see this on the post contrast imaging, there’s only unilateral involvement of the optic nerve sheath on one side. This is IgG4 disease with dural based disease affecting the transverse sinus on the left resulting in pseudotumour cerebri. Again, an example of sheath dilatation which it isn’t due to a meningioma.

There are other tumours in the retrobulbar region may that may mimic a meningioma. So we’ve got here a nerve sheath tumour schwannoma and a cavernous haemangioma, optic pathway glioma and, in this case not even a tumour. This looks very much like an optic nerve sheath meningioma but pop him onto the scanner, occlude his neck vessels, and lo and behold typical example of an orbital varix. It is completely different from what you would expect clinically from an optic nerve sheath meningioma, but in isolation on the imaging, we can’t tell the difference.

This leads on to the next slide, which summarises a nice paper from JAMA Neurology this year. It showed that a lot of the diagnostic errors in detection of optic nerve sheath meningiomas are due to clinical assessment failure; and that feeds into radiology because what you tell us matters. We scrutinise imaging based on how we have been directed by the clinical differentials. If we are directed differently optic nerve sheath meningiomas may not be high on the differential list.

In this paper, when looking at imaging errors on its own, a lot of the misreads were due to the wrong imaging protocol. Imaging protocols are really important; when you give us the right clinical information, we can protocol it correctly and then we can detect what we need to see. Misinterpretation was also a cause for error, expertise really matters; when we actually put our training, specialist knowledge and experience into it, we can then make the diagnosis. In this paper quite a high proportion, almost 45% of imaging errors, were misreads from non-neuroradiologists.

In terms of staging and prognosis, there have been attempts to use the morphology of these tumours to correlate with outcomes with and without treatment. I don’t really find using morphology very useful. What was useful from this particular series, was that calcification seemed to have conferred a degree of slowing of growth but that could be a chicken and egg situation; whether the tumour in itself expressed calcification and was the nature of a slow-growing tumour or calcification forms a natural boundary for growth and therefore limits its growth. Tass (Tasanee) had spoken a little bit about Schick and his study; well the trouble here is, we’ve got 73 patients and 7 different imaging phenotypes, so we’re sort of getting to the realms of pseudoscience as it’s trying to sub classify a small group of patients with a large number of classification categories. Invariably you’ll get some correlations with a small group one way or the other. It could potentially become very useful but I think we need larger numbers to validate the classification. The nature of the beast is it’s a very rare disease and poorly detected therefore getting large numbers we validate classification systems for prognosis and treatment will be challenging but nevertheless necessary.

Imaging forms the cornerstone of follow-up in conjunction with clinical assessment, visual acuity and visual field testing. It’s quite clear that there isn’t, in some cases, a clear correlation with the size of a tumour or the morphology the tumour. In other cases, it’s a non-linear relationship between therapeutic response and tmour size/morphology. It has been shown in the literature, that after radiotherapy, there is a corresponding improvement in visual acuity but no perceptible change in the tumour size. It’s difficult to square this, so there isn’t a clear one-to-one correlation. In 7% of patients in one series, there had been an improvement in visual acuity with no net change in tumour size, this illustrates the disparity.

Imaging is very useful and it’s very important in terms of follow-up but there are caveats to it. Katherine talked a little bit about Dotatate scans and looking at somatostatin receptor response, so it can be quite useful post-radiotherapy to look at the somatostatin receptor status to adjudge therapeutic response, even when the tumour remains unchanged in size.

‘What does the future hold?’ Well I think this (forum) is a good start, it’s about awareness and education, we need to start telling people ‘Look, you need to actually do a proper clinical assessment, then inform your neuro-radiologists. Then the right imaging protocols need to be done, and then you’ll spot your tumours and we can give it a reasonable differential if it’s not a meningioma’. So this coupling of imaging and clinical management needs to be joined up.

I think what might be useful is in the long run, to have some strategies or guidelines on how we actually manage these tumours, that can take the form of a National/ International registry. We can have a registry of pooled data and imaging. This can be used to create the back-bone of a platform for doing advanced imaging research. For example, looking at the integrity of the optic nerve as a more sensitive marker for progression or treatment response. There can be natural history publication series that we can produce from a curated multinational database. We can then look at imaging and clinical phenotyping in a reasonable manner, proper science really. We explore what can be utilised from this data to determine outcome measures in relation to therapy, surgery, radiotherapy, whatever the case may be. Of course, everyone’s talking about AI, so computational predictive modelling can only work with vast numbers, massive numbers, tens of thousands; you can only start building on that process by curating all of this data in multiple centres in multiple countries.

Three key points then, clinical and imaging needs to work together: you inform us on the clinical aspects, we can then formulate the correct protocols, thereby providing you with the most accurate answer. Specialised protocols will give you a better diagnostic pickup, that goes without saying. Thereafter it’s the right people reading the scans, Neuroradiologists. There is a clear need for education and guidelines, research that will follow on from there.

Q & As section

Question 1
Our neuro-ophthalmologist said the skull base entity advised against a biopsy in a presume optic nerve sheath meningioma, because they suggest that if you opened the membrane then you compromised long term tumour control. My understanding from your talk is that imaging is far from perfect in diagnosis condition, so I was wondering what is the best or the gold standard diagnostic test in order to guide treatment?

Answer
As you said, I think biopsy – tissue in the pot- is the gold standard, but in terms of non-invasive measures, I think imaging is still important. There will be some the exceptions and the caveats as I have mentioned; some of them are incredibly rare and in conjunction with proper clinical assessment and imaging protocols, the pick-up rate and certainty increases substantially. When someone is suspicious of optic nerve meningioma and everything else fits on the imaging, I think there’s a high pick up rate.

Question 2
I was wondering with the imaging protocol, are you using 3T scanners because the resolution is much better? Do you never use 1.5 T?

Answer
There will be some instances where you do have to use a 1.5 Tesla magnet. Normally, we prefer to use 3 Tesla magnets because, just to make a distinction, the higher the field (strength), better the quality of the scan because of the better signal. There are some patients in whom there are implants such as conditional pacemakers, which are then only safe to use at 1.5 Tesla for example. It’s worth remembering that patient factors are really important too. If the patient moves, regardless of whether it’s 3 or 1.5 Tesla you’re not going to get really good diagnostic imaging. It’s about performing the complete protocol, balancing this against the time. The finer the resolution of the scan, the longer it takes, and of course that means it’s more uncomfortable for the patient and they’re less likely to actually comply with the examination.