Slides
Transcript
Good morning, everyone. Thanks for the introduction. Thank you for the invitation. So, I have no disclosures.
So this is what we often see when we are asked for a second opinion on SIH. We see like comprehensive and extensive MRI studies that may easily take one hour or more, but these images lack some specific MRI sequences to make the diagnosis or to best make the diagnosis of SIH.
In preparing this talk, I was thinking that it might be a good idea to tidy a little bit up with these MRI sequences. It is probably important only to talk about the most important sequences that are most useful to make the diagnosis of SIH and also to bring some streamlined protocols to this audience.
So, what is the relevance of MRI protocols? I guess the relevance is a little bit underestimated. MRI is always the first step in the SIH workup, and it has a great influence on everything that follows. That means especially the decision if and how we perform a following myelography to search for the leak. I haven’t found any data on this, but I’m quite sure that the decision on the MRI protocol and the quality of the MRI protocol has a great impact on the time until the diagnosis of SIH can be made.
But there’s also a problem with these SIH sequences because they are very complex and not easy to understand, and we use a very specific terminology. Moreover, we have different MRI manufacturers that all use their own terminology for these sequences. I provided here a link where you can compare all these terms from each MRI manufacturer in a very clear manner.
Okay, so I want to guide you through the most important sequences of the head and the spine and explain why we use it. I want to make some suggestions for MRI protocols. The T1 gadolinium sequence with contrast is probably the most important sequence of the head because we see a lot of anatomical detail, and we see the enhancement of the dura. We use it every day to calculate the SIH Bern score.
So, we look at the suprasellar distance, we look at the prepontine distance, we look at the mamillopontine distance. We look out for the enhancement of the dura. We look for the engorgement of the sinus, and we look for hygroma or subdural hematomas. There are different features, like major criteria given with two points and minor criteria given with one point, and so we can calculate. Zero to two points is a low probability for the presence of SIH. Three and four is an intermediate, and five or more is a high probability for the presence of SIH.
Then we have another important MRI sequence, which is the T2w FLAIR. This helps us to see hygromas and chronic subdural hematomas, like in this example, in a very sensitive manner. But there’s another feature described in the literature. As we can see also, the dual enhancement without contrast. I think the data are a little bit conflicting, so I would suggest to only use the FLAIR to see the thickening of the dura when we cannot give the contrast agent – so only as an alternative.
Then we have the T2w of the orbit. This is a quite new sequence and the work for SIH, and we are looking here at the perioptic space. This represents the intracranial pressure in a very sensitive way. We need some special sequences for this, and Dr. Schievink recently reported on this MRI sequence and demonstrated that this is very useful in patients which are otherwise MRI imaging negative. When you use these cutoff values, you have a high probability to find a CSF-venous fistula in these patients.
We’ve been using this sequence for a long time, but in the axial view. And we found that it is definitely useful. Like in this patient – she had an SIH score only of three points, but we saw that the perioptic space was completely empty. And this guided us to go on with the imaging. We performed a dynamic myelography, and we saw this clear CSF-venous fistula on the left side. This is also really helpful as a monitoring tool. This patient had a type two leak, and we can see the type two in the CT myelography. We performed a CT-guided fibrin patch, and the other day the patient complained about headache when lying down. What we see now is a massive increased perioptic space, and we also see a kinking of the nerve, representing now a rebound intracranial hypertension.
Then we have the SWI – the susceptibility-weighted imaging, a special sequence that we can see the positions of blood in the cerebellum. We call it infratentorial siderosis. This may occur in patients with a very long course of the disease, usually more than 10 years, and we can see like a signal loss. These patients usually present with ataxia, hearing disturbances, or even dementia. And once we find the leak, we can stop the progression or even improve the symptoms.
The heavily T2-weighted MRI of the spine is probably the most important one. Heavily T2 means we use special sequences with a fat saturation, and we only see the bright signal of the CSF in the spine. And what are we looking for here? We’re looking for the epidural fluid collection, like in this patient in the ventral aspect here, again. And this patient at the sacrum, and this patient had a sacral dural tear. So it is important to always cover the entire spine, including the sacrum. An epidural fluid collection means we have a dural tear somewhere along the spinal axis. If we do not find this, we may consider a CSF-venous fistula.
This is an example of some images, and these are very helpful, very sensitive. We can even see the dura surrounded by the epidural fluid. We can see very small amounts of epidural fluid, and of course, we can exclude epidural fluid with these sequences. This sometimes guided us on where to do the myelography. For example, in this patient with the ventral collection, we put the patient in the prone position and find a ventral dural leak. And here, we had a lateral dural leak. This was, for example, a sacral dural tear.
Then there is the conventional CT myelography. This is a technique we, I think, all have been using for years. And that means we do a lumbar puncture, inject iodine contrast, wait for some time, and do a CT scan in the supine position. Only for the same purpose – to see if there’s epidural fluid collection. There are some studies reported that the sensitivity is comparable. That means we can waive, and I think we have to waive, this technique because it saves radiation exposure, saves lumbar punctures, and saves time.
And there’s something new, recently reported from Dr. Carlton Jones. When we use a regular T2 image of the spine, we may have the chance to find a so-called flow void sign. And this may appear at the site of the leak in patients with a type one leak, when the CSF is running through the dural tear. This is like an artifact with a loss of signal, and we confirmed this with the DSM. Here’s another example with a very subtle finding. And again, we could confirm this.
We have the T2-weighted myelogram. This is only one image. The acquisition is just 20 seconds, and I think this is really helpful. We use it for every protocol, and we can see, for example, like meningeal cysts at a glance. This sometimes helps because, like in this case, sometimes the CSF-venous fistula originates from the biggest meningeal diverticulum. But this is not a rule. It just helps us, for example, on where to start with the myelography, on the right or on the left side.
What we know is patients that are negative on imaging, like this – patients with an SIH score of 1 but presenting with multiple meningeal diverticula – have an increased likelihood for an underlying CSF-venous fistula, like in this patient. And we use this technique also for patients with the sacral dural tear. We regularly find them in our cohort, like in 5 to 6% of patients, and in this selection, we got indications for a sacral dural tear, like at a glance, and all these patients were confirmed to have a sacral dural tear.
Then we have pitfalls, of course. We know that patients with a longer term of the disease will have decreased signs on MRI of the head but will still or may still have a dural leak. So we should always perform an MRI of the head and the spine.
Here are some protocols. When you would ask me, what is the absolute minimum to perform to make the diagnosis of SIH, I would probably say the T1 gadolinium of the head and the T2 space of the entire spine is enough to make the diagnosis of SIH.
When you want to feel more comfortable – this is, at our institution, like 22 or 23 minutes. When you feel more comfortable, you should add the FLAIR, the T2 of the orbit, the T2 HASTE – this is just one image, 20 seconds acquisition time – and the T2 for the flow void sign, which we found to be very helpful. This would be like 37 minutes at our institution.
When you have a patient where you suspect a long course of the disease, you may add the SWI sequence, and this would be like 40 minutes. We created a special protocol in Freiburg. We call it the screening protocol. And these images have a little bit less quality, the slices are a little bit bigger, but it is still very good to make the diagnosis. This entire protocol takes 17 minutes only.
So, in a very clinical, busy day, it is sometimes really helpful to quickly check if the patient may have an underlying SIH or to rule out SIH. And you can do a follow-up. We usually do it like 3 months after targeted treatment, and it is not necessary to perform an entire MRI protocol for the follow-up. For patients that were treated with a dural tear, it is probably enough only to perform a T1 gadolinium of the head, the T2 of the orbit as a monitoring tool, and we only perform one block of the T2 space fat sat centered on the previous leak site. And for patients with the CSF-venous fistula, I think you don’t have to do a spinal MRI for the follow-up at all because these patients never had epidural fluid.
So, my take home message is: I think it’s important to focus on the most important MRI sequences because this will help to more quickly make the diagnosis of SIH. You should always perform an MRI of the head and the spine. And I guess the 3D heavily T2 image of the spine is the most important one because you can waive the conventional CT myelography. Don’t forget to cover the entire spine, including the sacrum.
And for future directions, I think we can make more out of the MRI protocols. I think we should look out for markers that may predict the site of the leak so we can streamline, in the future, the entire diagnostic work. Thanks to the team, and thank you very much for listening.