Sphenoid wing meningioma: Neuroradiology

Neuro-radiology of Sphenoid Wing Meningioma
Presenter: Katherine A. Miszkiel

I have no disclosures.

We shall be looking today at a range of of MR and CT imaging, much of which you will be very familiar with, but hopefully I’ll be able to tell you a few things you may not know.

Meningiomas are the most common primary orbital brain tumour, with about 20% centred on the sphenoid wing (SWM). They have two main growth patterns: the nodular or ‘en-masse’, or the more diffuse ‘en-plaque’ pattern, which can be distinguished radiologically, although there is significant overlap. They are classified according to their site of origin along the sphenoid ridge: lateral, middle or clinoidal. The clinoidal ones are further subdivided into three different types, according to their relationship to the optic canal.

CT scan has been around for a long time and we all use it. It’s great for bony anatomy and architecture, standard volumetric acquisitions, with thin sections. You can give contrast if you need to look at the soft tissues reformatted on bone window settings, high-resolution bone algorithms if you need more detail, and other reformats and 3D reconstructions as required.

We can also look at arteries and veins, and although venous structures may not be so important in terms of the sphenoid wing meningioma, for other meningiomas they are, for example in relation to the superior sagittal sinus. We are all familiar with hyperostosis and sclerosis that occur with sphenoid wing meningiomas, and it’s thought that 90% of cases represent tumour infiltration rather than a reactive phenomenon. We must also remember that occasionally we see osteolytic meningiomas, and one can be caught out. Tumoural calcification occurs in up to 20% and that can mean a better prognosis because it suggests the tumour has been present a bit longer.

We have had MRI since the 1980s, with 1.5 and 3T MRI scanners readily available, giving good contrast and spatial resolution, particularly with intravenous gadolinium, but you’ve got to do the study well. Remember with orbital imaging to do fat-saturated sequences, and additional MRA and MRV sequences can be done to look at the vascularity. More recently, diffusion-weighted sequencing and susceptibility-weighted imaging are included in the portfolio of sequences that are part of the brain study. We all hear about these advanced MR imaging techniques -spectroscopy, perfusion, etc, but are they useful in investigating meningioma? Again we’ll touch on that a bit later.

As radiologists, we’re quite confident at looking at meningiomas; it’s a well circumscribed dural based mass, with variable T1 and T2 signal characteristics. T2 is usually iso- to hyperintense, and a dural tail is not uncommon in almost three quarters of cases. An indication of vascularity may be present on routine imaging , and there may or may not be vasogenic oedema. Up to 50% of tumours have some vasogenic oedema, but it doesn’t correlate with tumour size.

Although this case (MRI images) isn’t a sphenoid wing meningioma, I thought it was a really nice example of the vessels that could be seen in a meningioma. You may see these serpiginous flow-void structures, which are enhancing structures in a tumour, which suggest high vascularity, and you may need to have a discussion with your interventional colleague if it’s large, and consider preoperative embolization.

Vascular tumours: if you were to do an angiogram in this case, you would see this very prolific tumour blush supply from distal anterior cerebral arteries and middle cerebral artery branches, but we don’t do angiography that commonly in meningiomas. Actually when I went round the other day to the angiographic suite, I had trouble finding a case where we had done an angiogram in the past year – so not commonly done, but it still can be useful.

Often we can see tumour encasement (and reduction in calibre) of normal vessels just by looking at structural cross-sectional imaging. Here this cavernous meningioma is clearly encasing the carotid artery and narrowing it. You can see the enhancing tumour on either side extending into the chiasmatic cistern and adjacent to the supraclinoid artery. If you do a CTA examination with MIT reformats, you can get some additional vascular information without needing to do angiography.

So let’s look at some typical and atypical cases. This first case was a patient who presented in 2011 with a bit of proptosis and had CT imaging, which showed mild bony sclerosis of the greater wing and anterior clinoid process, and a little soft tissue along the lateral wall and along the roof. You don’t see any obvious intracranial abnormality in the CT scan, but of course the MRI will show you that there is intracranial disease with more extensive dural tail which may be a reactive phenomenon but may also indicate direct tumour invasion of the dura. The other aspect to consider when you’re looking at the MRI is whether there is any enhancement in the intradiploic space of the sclerotic bone, because that suggests there is bony infiltration, that is, intraosseous disease.

We know from textbooks and our own clinical experience, that meningiomas are often relatively slowly progressive and sometimes, when you look at annual scans you don’t pick up that subtle incremental rise in tumour size and you need to look at it over a longer time interval. But in this poor patient – the same patient that we saw in 2011 – by 2018 this bony hyperostosis had really progressed, and you can see that on the 3D reformats as well. The dural tail had become a more nodular mass rather than simply ‘en-plaque’. In the MRI it’s hyperintense to brain, slightly heterogeneously, enhancing a little bit lobulated on the deep margin, with a bit of a dural tail.

Diffusion-weighted sequencing can be quite helpful in looking at solid masses in the brain but we often use it in the orbit as well, if we can get good enough views.
Diffusion-weighted sequence imaging relies on the free random motion of water molecules, Brownian motion, between the intracellular and extracellular space. Then if you’ve got a lesion with increased cellularity, Brownian motion cannot occur because that motion is restricted. We say there’s some restriction of that Brownian motion, and that restriction will show up as an area of T2 hyperintensity on the diffusion-weighted trace image. You can’t quantify this, and if you want some quantification, you convert this image into an apparent diffusion coefficient (ADC) map. This diffusion trace image almost looks a bit like a FLAIR sequence; whereas on the diffusion ADC map the fluid in the ventricles and subarachnoid spaces shows up as bright, an area of restricted diffusion will show up as an area of darkness, and in this case similar darkness hypointensity as the brain parenchyma. It gives us an idea of cellularity and you can measure that apparent diffusion coefficient and give it an actual number.

The second case demonstrates vasogenic oedema which can occur in up to 50% of patients, although it doesn’t necessarily correlate with tumour size. This patient does not have a huge mass, but a rather extensive area of vasogenic oedema. Looking at the bone we don’t see massive hyperostosis, indeed you actually wonder whether there’s a little bit of lysis there. There wasn’t a corresponding CT scan though. But just to show you the difference between the restricted tumour showing restricted diffusion, and the vasogenic oedema (where you’ve got loss of the blood-brain barrier in that area of brain involvement): in the latter there is increased water molecule movement between the intracellular and extracellular space, indicating free diffusion, or free diffusivity, of water. Thus we can use diffusion-weighted sequencing to help us.

This third case, patient who had radiotherapy many years previously for a posterior fossa tumour surgery and then radiotherapy (and note the secondary basal ganglion thalamic calcification). The patient presented with a little proptosis, and we saw this very subtle area of bony sclerosis, perhaps a little rarefaction, and this slightly atypical, spiculated appearance to the medial margin of the bone, protruding into the orbital soft tissue mass.

We were worried here, ‘Is this a radiation related sarcoma?’ This would be an odd appearance for the bony changes in sphenoid wing meningioma. We did our usual imaging, again you see some restricted diffusion in the orbital component of the mass, you can see its relationship to the optic nerve which is medially displaced, and there’s enhancement in the tumour as expected, avidly vascular, so we expect them to enhance. There’s also this additional ‘en-plaque’ meningioma on the contralateral side.

We mentioned susceptibility-weighted imaging before as a more recent introduction to the imaging protocol. This sequence is a very heavily T2-weighted sequence looking for the presence of blood products and paramagnetic substances such as melanin, manganese and even calcification. You can sometimes use this to look at the calcification in the tumour, you can see it’s actually more obvious on the SWI sequence. Here we’ve got a radiation related cavernoma; this is the conventional T2-weighted sequence showing a hemosiderin deposition and a central area of hyperintensity and a larger area of low signal susceptibility on the SWI sequence, because you get exaggeration of the blood products there and can see this blooming effect.

We don’t do perfusion in our routine clinical practice for meningiomas because we’re usually fairly confident in the diagnosis but, as expected, because of their vascularity, they would show elevated relative cerebral blood volume. Similarly, meningiomas are known to have a typical spectrum on MRS with an alanine peak and sometimes a glutamine peak. But again, in clinical practice we don’t tend to do it unless you are looking for atypical meningiomas.

Now meningiomas express somatostatin receptors, and 68-Gallium Dotatate CT imaging uses that principle; the radio-labelled gallium attaches to a somatostatin receptor and it’s taken up avidly by meningiomas. In this image, you can see the sphenoid wing meningioma showing up, it shows you very clearly the intraosseous involvement, some contralateral disease, and actually shows you a tiny meningioma along the tent here, which would normally be very difficult to pick up, you might just think it’s a confluence of vessels. So Dotatate imaging is another modality we can use to help us if we’re not sure, or if we’re looking for subtle recurrence.

Radiologists love differential diagnoses, the longer the better. I’m going to quickly tell you the three common ones. First differential diagnosis: don’t forget metastases. It’s easy sometimes forget them when you’ve got a slightly atypical history. This poor gentleman, no longer with us, had prostatic CA with a short history of swelling and proptosis, which would be a bit atypical for a sphenoid wing meningioma. Looking at the bone which is a bit sclerotic, it could easily be a meningioma – again, with a nodular mass and intracranial vasogenic oedema- but it turned out to be prostate. So don’t forget metastasis even if it looks very similar to a meningioma. I think in this other case, we would say that this is not a meningioma, with too much bony lysis’ – this was in fact breast CA with a large ‘en-plaque’ mass intracranially as well as intraorbitally, and with vasogenic oedema. So always consider metastases.

Another differential diagnosis is fibrous dysplasia. We not uncommonly get referred cases where there is uncertainty as to whether it represents one or the other. But there are very different bony characteristics on CT: in fibrous dysplasia there is bony sclerosis and expansion, the cortex is intact and smooth, and there areas of ground-glass opacification – which is the classical thing that’s always mentioned in the textbooks – and you may have vascular lytic lakes but you don’t get a dural tail on MRI, so no dural tail in fibrous dysplasia.

Finally third differential diagnosis. Is this case an odd meningioma? We’ve got a slightly expansile lesion, centred on the greater wing of sphenoid, with a mixture of lysis and sclerosis. But it looks odd, with some central spiculated calcification. Could it be a sarcoma? It is something very odd, with enhancement around the periphery. This is where your diffusion-weighted trace image can help. There’s your globe, and the soft tissue mass, but we know meningioma have restricted diffusion so they should be dark on ADC. However, in this case it is bright on the ADC map, with free diffusion. Therefore, this is not a meningioma; this tumour was followed up for many years, and we think it’s an atypical sphenoid wing haemangioma.

In summary MRI is still the imaging modality of choice, with limited clinical use of the advanced imaging techniques but, who knows what the future is going to bring with AI, 7T imaging, MR elastography, all those things in the future when I’m retired. Don’t forget the wide differential diagnosis, which I’m sure your radiologists will love to discuss in detail.

Thank you.

Q & As section

Question 1
About embolisation, I’ve given up doing preoperative embolisation in meningiomas and we’ve actually written up cases scalp necrosis following embolisation. It’s not without risk and I think it’s by large unnecessary.

Answer
That’s a fair comment and I when I spoke to Fergus, I wanted a picture for the presentation, and I remember we did them much more frequently, but Fergus said ‘we haven’t done one for a long time’ and it’s out of vogue at the moment, so practice does change.

Question 2
In one of your cases, you touched on by laterality. In your experience, what proportion of sphenoid wing meningioma patients develop either simultaneous or sequential bilateral disease?

Answer
If you see simultaneous disease, you’re thinking about how they had radiotherapy previously, thinking of multifocal meningioma post-radiotherapy, or obviously want to make sure they haven’t got neurofibromatosis. Even in patients that don’t have any genetic risk factors that we know of we see a fair proportion of patients with multifocal intracranial meningioma, and obviously that lesion can spread across the midline and involve the contralateral side. If you were not quite sure whether the contralateral cavernous sinus was thickened or not, 68-Gallium Dotatate PET-CT maybe very useful. We’re also using Gallium Dotatate imaging with Joan Grieve, Neurosurgeon, when we’re looking for a difficult pituitary lesions, if we’re looking for some post-op focus of, for example Cushing’s adenoma, that we can’t see. We’re looking for that because the gallium dose is also taken up by neuroendocrine tumours, so it’s great for meningiomas and neuroendocrine tumours. If you do normal FDG PET imaging, you’ve got the activity of the cortex and the rest of the brain, which you don’t get on the Gallium Dotatate, so you don’t see that interference – its easier to pick up.