Slides
Transcript
Moderator: All right, thank you. And with that, we are going to go on and move along to our next session, which is on CSF-venous fistulas. Next, we have our course director, Dr. Andrew Callen, a neuroradiologist for the University of Colorado Anschutz and director of the CSF leak program, who will present on the nuances of how CSF-venous fistulas present on imaging. And I’m the patient moderator for this section. My name is Aubrey Bolan, and I am a fully recovered CSF-venous fistula patient whose leaks were elusive to initial myelography. My CVFs did not involve paraspinal veins but actually arose from the internal vertebral venous plexus, and they required specialized CT myelography techniques to locate. I currently volunteer on the Board of Directors of the Spinal CSF Leak Foundation, and I’m very excited for this talk because I know it would have been very relevant to my case, and I hope it will be very helpful to many of you listening today. And with that, the floor is yours, Dr. Callen.
Dr. Callen: Thank you, Aubrey. It’s really my honor to share this session with you. I loved your talk last year. I have my screen sharing. I hope that that is good. I’m going to be talking about CSF-venous fistulas today. I’m going to be, again, like Aubrey mentioned, moving beyond the hyperdense paraspinal vein, and I hope that this talk is helpful for everyone listening. It’s sort of a reflection of my own learning throughout this process of performing myelography, of looking for CSF-venous fistulas, and sort of the learning that I underwent during that time.
So my disclosures are that I am on the Medical Advisory Board of the Spinal CSF Leak Foundation as well as Spinal CSF Leak Canada. I will be talking about a myelography positioning device for which I have a patent.
Today, we’re going to be talking about things in three general chapters, although there’s going to be overlap. I think it’s important before we start talking about what CSF-venous fistulas look like, how we find them, to just remember the history of how they were discovered. We’re then going to go into the anatomy of CSF-venous fistulas and their pathophysiology. Then we’re going to move into myelography and talk about the techniques that I use, that others have used, to find CSF-venous fistulas. And throughout all of this, touch on these underrecognized, I think, CSF-venous fistula appearances that are important for us to be aware of.
So let’s start with that history. Let’s talk about the era pre-2013. We would have a patient whose brain MRI showed obvious findings of intracranial hypotension. In this case, severe brain sag. There’s engorgement of the pituitary gland and the venous sinuses, and some dural enhancement, although not in thickening – and I’m not showing that on this image. They would get a spine MRI that showed no fluid collection, and perhaps a regular CT myelogram would be performed where contrast dye was injected into the CSF. The patient was sort of rolled around a couple times and then laid on their back or face down on the scanner, 30 minutes later or so, and a CT of the entire spine was obtained that showed no evidence of a leak.
And so we didn’t know what to do with a patient like this, and we would try blood patching and trying to help in this way, but we were sort of at a loss. But then in 2013, and then published in 2014, something really remarkable happened that changed this field forever, and it was done by Dr. Schievink and his colleagues at Cedars-Sinai, where they used the same techniques that are used in the catheter angiogram lab that neurointerventional surgeons use, for example, to treat or look for a stroke in the brain. And you could see that image of a catheter angiogram, a digital subtraction angiogram. They were performing digital subtraction myelography to look for CSF leaks, and they noticed that there were these small veins that were taking up contrast and there was CSF leakage this way. And this was a paradigm shift in our understanding of how these leaks could occur, that they didn’t just have to be through a hole in the dura with a resulting fluid collection around the spine, but instead could go into the venous system itself.
And it wasn’t too long after that that they made the, that Dr. Schievink and Dr. Farb and their colleagues described, that actually if we lay a patient on their side, rather than laying them face down, that we will see these fistulas more frequently. Just that simple change, having that contrast layering on the side of the thecal sac, will increase our hit rate for finding these fistulas substantially. And this was important because it helped us understand that most of these fistulas are occurring at or around the nerve root sleeve complex or with their associated meningeal diverticula. In 2016, Dr. Kranz and the group from Duke described that we could see these on CT as well, and they described this sign, the hyperdense paraspinal vein sign.
And so by putting this patient in the CT scanner, who had a CSF-venous fistula here shown in panel B on the digital subtraction myelogram, that we were actually able to see this hyperdense paraspinal vein, the segmental vein here along the spine. And this perhaps to somebody who does a lot of DSM, you could say, oh yeah, that’s definitely a CSF-venous fistula, but I think to most radiologists and maybe to most people, this is something that is very compelling, to see this image on a CT scanner. You know, you’re like, oh, that’s absolutely abnormal. There’s no contrast in the veins, or there shouldn’t be, and so that’s bright and therefore it’s a leak. And this, I think, these things together are really the story of the CSF-venous fistula and how we came to understand what it is and what it should look like.
So that led to this incredible change where we took patients like this one, this 60-year-old man who had years of an orthostatic headache and had six different diagnoses trying to explain what was going on with him, that when he had a regular myelogram at one point that said no evidence of a leak, and by performing a myelogram with the techniques that we’re going to talk about today. Instead he just had one diagnosis, and that was intracranial hypotension due to a CSF-venous fistula. I’m showing both images here at the same level, and that fistula was just not evident when the myelogram is performed in a conventional fashion.
So it seems simple, right? We just basically do a myelogram and turn the patient on their side and then look for a hyperdense paraspinal vein. It really should be easy. I mean, here are four examples of hyperdense paraspinal veins. Here are some ones that I think are actually quite pretty, and it is interesting for as horrible a disease as it causes, they can be sort of beautiful. Here’s a very long one coming down to the level below. Here’s that classic segmental vein appearance. And one here exiting the foramen as well.
Except we still can miss them. And these can happen in different sorts of ways, and I’m going to try to break down the ways in which we can miss CSF-venous fistulas into two general categories. In this patient here, when I took her for her first myelogram, I found nothing despite her having an overtly positive brain MRI, and on the second image we were able to find the CSF-venous fistula. We repeated it, and there were some differences in the technique that we performed on that myelogram, which we’re going to talk about today.
So I think that some of why we might miss CSF-venous fistulas is related to those errors of technique. But we could also find the fistula. We could successfully delineate the fistula on the myelogram, but if we’re not aware of the less commonly recognized appearances of fistulas, we may not actually see them. So here’s a patient with severe brain sag who had a myelogram, and I’m telling you right now that there is a CSF-venous fistula on this image. But, if you’re looking at this, looking for the hyperdense paraspinal vein, you may be like George here and squinting your eyes, saying, “I have no idea what you’re talking about. This seems like voodoo.” But indeed, there’s this very subtle opacification of the internal epidural venous plexus and these small veins around the nerve root sleeve. And when this was ligated, the patient’s brain had a profound response. It always really strikes me when these tiny little findings can cause such severe, profound changes both on the brain MRI and clinically. Patients can present in coma, for example, from a tiny little CSF-venous fistula. Now, we’re probably not capturing the entire extent of that fistulous drainage, but nonetheless, it is always remarkable to me.
And so, I think that CSF-venous fistula detection really can be broken down into two main pillars: that of technique, and are we doing our myelograms in the best way to optimize the detection of these fistulas; and that of perception. Are we looking in the right places? Are we making sure to scrutinize all possible egresses into that epidural venous plexus, both internal and external?
And so let’s talk about perception first. I think that when we talk about perception, we have to break this down further and first think about the pathophysiology. What are CSF-venous fistulas, and why do they form?
This is a very interesting question that doesn’t have perfectly clear answers, but I’ll try to get at some of them today. And then, where can they occur? Really understanding where they can occur is mostly a question about anatomy, understanding the venous anatomy in the spine, and so we’ll talk about both of these things now.
It’s important to remember that there is some degree of normal CSF resorption that occurs in the spine. It’s estimated that around a fifth of the total CSF resorption, just physiologic normal resorption, occurs in the spine, and it seems to occur mostly in these arachnoid granulations that are these structures that are intimately associated with the veins, where there could be physiologic CSF resorption. These are in the spine along these nerve root sleeves, and there’s a higher predilection of these in the lower thoracic spine, which in the thoracic spine, in the lower thoracic spine in particular, which may account for the distribution that we see where most of these tend to occur in the thoracic spine, and a lot of them in the lower thoracic spine.
And so the sort of overwhelming, prevailing theory is that there is some damage to the arachnoid villus arachnoid granulation complex that will then lead to unregulated flow of CSF into the venous system. And so this kind of makes sense, right? That we have this normal process that has gone awry, and there’s been some damage there.
But then why does that damage occur? And I think that there are two general categories for this that I like to think of them in. One is intrinsic factors, and one of the prevailing theories in regard to intrinsic factors is a pre-existing high-pressure phenomena. And then there’s extrinsic factors: trauma, or needles, or surgery perhaps, or inflammation, or even degeneration in the spine, and we’ll get at all of those now.
So one thing that has been described over and over again by us and others is that the majority of patients with a CSF-venous fistula will have a normal or even elevated opening pressure. And so this begs the question, what was their pressure pre-leak? Was it perhaps even higher than what we’re measuring once we’re doing that myelogram to look for the CSF-venous fistula, and maybe the CSF-venous fistula acted as sort of a release valve, and that pressure led to damage of that arachnoid granulation, and this is what occurred. And this is further corroborated by the fact that we’ve observed that there’s a higher average BMI seen in patients with CSF-venous fistulas. This is something that has been known to be associated with IIH – idiopathic intracranial hypertension. Not all patients, but this has been shown in several studies, and so it further lends to this idea that perhaps a pre-existing higher pressure state could be contributory.
There’s also been some very interesting studies recently showing that there can be extrinsic damage or injury to the dura that can lead to formation of a CSF-venous fistula. Dr. Madhavan, who’s going to talk a little bit later today, published this very nice paper of a series of patients who had CSF-venous fistulas arising from trauma, traumatic spinal pseudomeningoceles. This is a tear or a partial tear of the dura where a CSF-venous fistula is occurring in close proximity once they perform their myelogram. This is very interesting because it lends the question, when we think about what Dr. Beck was talking to us about, when we have an area of dural injury, is there perhaps either a secondary vascular uptake that becomes the primary pathophysiologic phenomenon, or is there something more that we need to get to?
Dr. Caton, who was actually my former co-fellow, he published this very beautiful image in JAMA recently, of an iatrogenic cerebrospinal fluid venous fistula. You can see that this needle tip during the lumbar puncture was very close to the ventral dural margin, and then subsequently the patient had a CSF-venous fistula into the basivertebral vein here at that same location. So this is very interesting, and it’s something that we’ve observed well.
And Dr. Beck described this very nicely in the preceding talk, but I just want to show one example of a patient in whom we saw this quite profoundly. So this is a patient who was 40 years old, had chronic post-dural puncture headache. We looked at this myelogram, and in addition to that very subtle sort of dinosaur tail sign that we see here, there was also this more focal area of T2 hyperintensity at the L4–5 disc space in the dorsal epidural space. And we wondered, could this be a potential bleb here? This is exactly where this patient had their lumbar puncture.
This patient did go to surgery, and I want to thank Dr. Lennarson for providing these and the other operative images that I’ll show in this talk. This was a very profound set of images to me, where we have this bleb that you can see here just sort of hiding under this blanket of this large vascularized membranes, very engorged veins, and as this was further dissected down, there was this one vein that was sort of attached to the apex of this bleb here. When that was removed, there was actually CSF coming out of that little place where the vein was plugged in. We see this quite frequently, and really I do think that what Dr. Beck said is so true, that there is something happening on the venous side with this sort of vascularization that occurs around the area of a puncture where a CSF-venous-fistula-like pathophysiology can arise. And so even though we’re talking about spontaneous leaks here, I think there is important overlap.
It’s important to talk about meningeal diverticulum when we talk about CSF-venous fistulas, and these are a very interesting phenomenon that almost every patient asks me about when I see them in clinic. We’re of course talking about these focal outpouchings along the dura, along the nerve root sleeves, and they’re outpouchings of the dura and arachnoid, and these are the membranes that contain our CSF. They typically do occur around nerve root sleeves, and they can be completely incidental, but in some patients they can mark a point of dural fragility. In this patient, they have multiple large irregular meningeal diverticula, and a couple of important observations have been made, one of which actually just came out recently. It’s a very simple concept, but it’s an important one that Dr. Mamlouk published, that the CSF-venous fistulas are usually occurring near the biggest diverticulum. We observe this too. Here are two examples of CSF-venous fistulas that are occurring in close proximity to very large meningeal diverticula. I wrote “cyst” there because meningeal diverticula wouldn’t really fit in that little space.
But we have some preliminary data which is currently submitted, not yet accepted for publication, so take it with a grain of salt. We looked at the relationship of finding a CSF-venous fistula to the size and number of diverticula, and we found that both larger and more diverticula were associated with a higher chance of finding a CSF-venous fistula. This is very important because pre-CSF-venous fistula, the meningeal diverticula were really thought to be mostly incidental, and there was not thought to be a strong relationship between a dural tear, for example, and the presence of meningeal diverticula. But that relationship may be different in the context of CSF-venous fistula.
Another thing that we observed in these data was that older patients had larger and more diverticula. So perhaps the diverticula themselves are a marker of degeneration of dural fragility over time, and this could be an important insight into why these fistulas occur. We also previously did a study looking at the relationship of spinal osteoarthritis and CSF-venous fistulas. We frequently observed that we would see CSF-venous fistulas at an area of disproportionate spinal osteoarthritis. This is wear-and-tear arthritis, disc disease, facet degeneration.
And here’s a couple examples of that, of patients in whom we see the fistula at an area where this patient has a large disc bulge here off to the side with some osteophyte formation. This patient has some facet arthrosis as well. It’s been well described that osteoarthritis is linked to inflammation, and a lot of those inflammatory factors are actually pro-vascular. So is it possible that this is playing a role in some of these patients? It could be completely coincidence, right? I mean the lower part of our thoracic spine is not connected to the sternum. The ribs are free floating. It allows for that twisting motion that allows us to do things like play golf, swing a golf club to any of my fellow golfers out there. And so there’s more degeneration there combined with a higher prevalence of arachnoid granulation, for example. Perhaps it’s just a coincidence, but nonetheless this observation was made, and we found that it was a statistically significant association.
We also looked at five different patients in whom the nerve root sleeve was resected and sent for histopathology and found that many of these patients had chronic vascular and dural changes that looked more like a chronic inflammatory picture at that dural lining at the nerve root complex. So it may be that there is a sudden insult that occurs, or it may be that this is sort of a chronic inflammatory process that occurs over time.
Okay, so let’s move into the anatomy of CSF-venous fistulas and again really the anatomy of spinal veins. When we think about the spinal venous system, I think we can break it down into three general compartments. The first is the internal epidural venous plexus. This is the venous system that occurs in the intracanalicular space within the spinal canal itself, the veins that surround the dura itself. There’s very complex anatomy here, but it’s very closely opposed to the dura and to the vertebral body margin and can be a blind spot. So we’ll go over some cases of that.
Then we have the external epidural venous plexus, going from those foraminal veins into the segmental veins, posterior intermuscular branches, all of the classic paraspinal veins that we talked about at the beginning of this lecture that we’re perhaps more familiar with. And then, very closely related to that but I’m going to put it in its own distinct category here, is the basivertebral or intraosseous veins. These are also very important veins that I think can very easily be missed if we’re not carefully looking for them, because they’re within the bone themselves. They’re not away from the spine, easily detected by our eye. These are the three categories, the three ways that my search pattern is designed when I am looking for a CSF-venous fistula.
So here’s some examples of each of those. Here’s some internal epidural, an internal epidural CSF-venous fistula. You can see on the axial image here, it just looks like a little dot. And if you’re looking for a hyperdense spinal vein out here, you’re not going to see anything. But instead, for these, for all of my myelograms, I also like to look at the coronal view. This is the view where it’s like we’re face-on to the patient, and we can see much more clearly that there is this linear vascular structure here arising from the inferior margin of this nerve root complex, going into that internal epidural venous plexus. So this has to be a part of your search pattern, just that little dot next to the dura.
We have our external epidural venous plexus. Again, here is that segmental vein opacifying that classic hyperdense paraspinal vein look. And then we have our basivertebral intraosseous veins, and these can be very, very tricky.
So I want to go over that patient that I just showed an example of very closely. So let’s look at this first image here from the first set of images we took of this patient’s myelogram, and there’s really nothing here. I’m not seeing any CSF-venous fistula, but let’s now compare that to 45 seconds later. I think that comparison will help us point out exactly the pathology that we’re seeing in this patient.
Number one, look at that basivertebral vein that is picking up contrast. It isn’t very bright, but compare it to what it looked like before the dye started going into that vein. This could very easily be scrolled past very quickly and just thought, oh, this is just the cortex of the vertebral body, it’s not abnormal uptake. But we also see that there is some intraosseous drainage as well of this CSF-venous fistula, which was not present just 45 seconds earlier. And in this patient also, there is contralateral paraspinal segmental drainage.
So in this patient, they actually had two of the three categories that I described that you really have to be careful to look at here. Now maybe you would have picked up on this, but would you have picked up on the basivertebral and intraosseous components if that was the only finding?
It’s also important to remember that even if we’re dealing with just an external epidural venous plexus drainage, that it’s not necessarily that just because I see this hyperdense paraspinal vein at T6–7 in this patient that the CSF-venous fistula is necessarily arising from that level. In this patient, if we look carefully, we could see that the fistula was actually arising at the level below at T7–8. We see some early opacification just anterior to the nerve root sleeve complex. And then as the myelogram progresses, we see that opacification into an ascending paraspinal segmental vein to the level above. And then we see this very robust opacification on the third phase of that T6–7 vein. But that is not where that fistula is arising from. This is very important when you think about treatment planning and what you’re going to target as the origin of that fistula.
Dr. Carlton Jones provided me this example, so I wanted to give her credit, and she also was the senior author on this very nice multi-institution series of CSF-venous fistulas where there’s more than one.
Now we’re talking about pitfalls in radiology, reasons why radiologists might miss something. To any radiologist in the audience, you’re very familiar with the term satisfaction of search. That refers to the fact that we make a finding and then we say, great, we pat ourselves on the back and we’re done. We did it. We found the pathology. In every patient with a CSF leak, you cannot be done after you make that one finding because they can have more than one. Particularly in the context of CSF-venous fistulas, there could be more than one leak.
So in this patient here, they had multiple fistulas. Here’s the right T7–8 and right T10–11. And this is thought to be sort of a rare phenomenon, but the more you do this work, the more you will find. Here’s an example from one of our patients. This patient had a fistula at T7–8 with that external epidural venous plexus draining into the paraspinal vein, and also at T11–12, and also at L1–2 into the internal epidural venous plexus.
And so you could really imagine that after you found this one at T7–8, you just think you’re done. And maybe even after you found the T11–12, wow, that really must be it. Especially if you’re not trying to be looking very carefully for these internal epidural venous plexus fistulas. But we really, really have to make sure that we’re looking at the entire spine, the entire study, even after we’ve made one finding.
There’s this phenomenon which I think is incredibly interesting, which I think deserves a lot of attention and more discussion, which is the fact that sometimes we will put a patient, for example, left side down, but we will see drainage onto the other side. Here’s an example of a patient who has drainage into that internal epidural venous plexus. And then we have drainage into the basivertebral vein. And remember, they’re on the left side down, but that drainage is now coming across. And as I’m scrolling up to the level above, we’re actually very soon going to see drainage out of the other side, the side that’s anti-dependent, coming out into the paraspinal segmental vein as well. And here it is right there. And coming up even further, okay?
So even though your eyes may be fixed on the part that is dependent, after all, we’re putting that side down or having the contrast layer there, it’s very possible that there is that paradoxical drainage of the fistula, which is very important to remember. Here’s another example of that. So this is a patient. Let’s look at their right-side-down imaging. We see some opacification here of these foraminal veins.
And then on the left-side-down image, we appreciate at the same level that there is this intermuscular posterior branch here which is opacifying very robustly too. And so depending on whether or not you caught the fistula at any given time, it may be that it is going to drain to the other side, the side up. And this is an important place to put your eyes as well when you’re looking for these fistulas.
Now, we’ve talked about the fact that CSF-venous fistulas don’t have epidural fluid collections, and so is there any utility at all in using MRI once we’ve said there’s no collection? Do we just forget about the MRI? Well, I think that they can help, and they can help when we’re looking at our myelograms, and they can help in two sort of ways.
First, I like to have the myelogram up next to, and side by side with, the MRI of the spine with that T2 fat-suppressed imaging, usually 3D, but as long as there’s some sort of axial T2 fat-suppressed imaging available, then I can troubleshoot these foraminal densities. In this patient here, I see a little bit of high density layering out here, and I’m wondering, oh, could that potentially be a CSF-venous fistula? But then I look and I see that actually there’s this sort of boot-shaped meningeal diverticulum nerve root sleeve here, that is just laying in the depending portion, of when the patient is on their side.
Whereas this patient here has this bi-lobed meningeal diverticulum, but then there’s this very dense linear structure emanating out of its apex which is not present on the MRI, and that increases my confidence that this is indeed a CSF-venous fistula and not just partial filling of that meningeal diverticulum.
I had this very interesting case recently where actually the MRI helped me rule in a CSF-venous fistula. Every time I find one of these, I go back to the MRI and say, is there any sort of clue that I could have had that this was where the fistula was? And usually the answer is no. But in this patient, who had a very, very profound, highly positive brain MRI, was actually very critically ill in our ICU, I knew there had to be a CSF-venous fistula somewhere.
On first pass, I didn’t see anything, but I was looking at this one meningeal diverticulum that had this tiny little nubbin just coming out of its posterior margin here, and I said, well, could that be a fistula? I’m not sure. But I compared the immediate phase, the patient’s in the same position, and the contrast in this structure here seemed to wash out 45 seconds later. I said, oh, that’s unusual. If this was just a little lobe of that diverticulum, like we see here, then it should just retain that contrast. It should still be sitting in there. It shouldn’t go anywhere. Why is it going somewhere?
And then I looked back at the MRI, and there’s actually a large vein that’s running back here, kind of coming right from that meningeal diverticulum. It actually even looked more bright to me than the other veins here. Perhaps there was more fluid or CSF in that vein. But that gave me more confidence that this was actually going into a very large vein that I was not appreciating. And so, in this way, the MRI actually helped me rule in a CSF-venous fistula.
I think it’s important when we talk about these sort of subtle or indeterminate findings to touch on this fact that sometimes we’re just not 100% sure. This is a patient I did a dynamic myelogram on, and to me this looked too dense, this internal epidural venous plexus here. But let’s be fair, I’m windowing this very harshly to make this very dramatic. Right? This looks more dense to me than the other levels. But am I 100% sure this is the CSF-venous fistula? No. But I think it’s suspicious. I can’t find anything else that’s going on. I want to communicate this to the patient, but I also know that they’re suffering immensely, that they are very hopeful and almost desperate that I’m going to find something, and I don’t want to give them false hope. I don’t want to pretend like this is definitely the answer and all we need to do is fix this and everything’s going to be better, because I’m just not sure.
And this is why we came up with this certainty score. We administered a survey to patients and referring providers, and the majority of patients felt like they didn’t totally understand their myelogram report, and almost half of patients felt like their reports contained some sort of uncertain or possible but not definite findings. And the vast majority of patients and the overwhelming majority of referring providers said they would prefer some sort of structured report where we’re saying, listen, this is this category of certainty, we’re not 100% sure, but we’re somewhere in between, and this is what we think it is.
And you know, this is a very important topic to me, and I felt very grateful to collaborate with patients and members of the CSF Leak Foundation on this study to get their input, because this is something that we need to work towards, I think, as a community is better communication so that we all understand how to navigate this world of uncertainty.
We’re going to end this talk with technique. We’re going to talk about density, positioning, timing, pressurization, respiration, and scanner technology. We know now that the number one priority in dynamic CT myelography is to get the densest possible contrast along the dependent thecal sac. And we’re going to do that in a couple ways. We’re going to use very high-density contrast. We’re going to use more contrast than we might use for a normal CT myelogram. And we’re going to position the patient just so. We’re not going to just have their head all the way down. We’re going to prop their head up so that dense contrast stays in the spine as much as possible so that we get a very sensitive exam.
I like to use this tilt table device that we developed here, where I’m able to control the contrast bolus. I’ll start very conservatively with just a little angulation to get that dye to not rush into the head, and if I need to, then I can angle the patient a bit more and get the dye to move up a little bit more cephalad towards the head as needed.
One of the benefits of CT myelography for fistulas is that we can study both sides in the same day. And we can do this by doing two different lumbar punctures. But of course, the catch-22 of CSF leak work is that to look for a fistula, we have to do a lumbar puncture and potentially give the patient a post-dural puncture headache. So there have been a couple other techniques described. We can roll the patient very carefully with a needle in place. Dr. Carlton Jones described this, and therefore only needing one puncture. I like to do it slightly differently. Again, with our tilt table here, I puncture the patient while they’re horizontal. I then elevate the hips with the needle in place. We’ll do my run on one side. I will then lower them with the needle in place again. Then I can inject additional contrast with them horizontal, take out the needle, roll them to the other side, and then I have fresh contrast in the lumbar cistern that I can use for the contralateral exam, which ends up with a very beautiful contralateral study, which is very very nice as well.
There are a couple other tips and techniques that I think can help in some cases. Pressurization is one that I do routinely and was very nicely described by Dr. Madhavan and colleagues in this case series that was recently published. That patient I showed you at the beginning of the talk where we couldn’t find a fistula at first, the difference between the two exams is that we pressurized on the second exam. We raised their pressure with sterile saline aliquots up to about 23 centimeters of water, and we saw that fistula.
What I want you to pay attention to is look at the distension of the nerve root sleeve in between these two exams. With pressurization, that nerve sleeve complex was more distended and allowed for more penetration of that dense contrast into the nerve root sleeve complex.
And indeed, if we just compare the appearance of a conventional myelogram to that of a lateral decubitus one, just gravity alone will let that meningeal diverticulum and nerve sleeve complex be more distended. And I think this is part of what these techniques, these adjunct maneuvers, will help us do is allow for that deeper penetration of contrast into the nerve sleeve to the potential fistulous source.
Resisted inspiration is a technique that’s been described in some small series with success. Here’s an example where we did this. We have a patient taking a deep breath in through a 1 cc syringe with the plunger pulled out. We didn’t see anything without the resisted inspiration, but with the resisted inspiration, we start to see filling of this fistula. So it was very useful. But again, pay attention to the nerve root sleeve. Look how much more distended it is with resisted inspiration. I believe these are two sides of the same coin, and we’re getting at the same phenomenon with this, is that we’re allowing that distention of the nerve root sleeve by increasing the CSF pressure.
I believe that timing can be important in certain cases. Here’s an example of a patient where, on the immediate run with no resisted inspiration, the first scan after we injected, we see everyone here is an expert at that little dot, that internal epidural venous plexus fistula, all these sequential images moving down. But then four minutes later, we did a scan with resisted inspiration, and I don’t see that anymore. It’s gone. And so I think that there are certain cases where timing can make a big difference. These fistulas can just be fleeting, and we may miss them if we don’t get more than one scan.
I had this very, very unusual case. I actually showed many of my colleagues. I didn’t know what to think of this. But this patient, who had many years of orthostatic headache, on their immediate run actually had this very small, subtle serpiginous opacification arising from the dorsal dura at T12–L1 that on their immediate phase was visible, and then 60 seconds later with resisted inspiration this was gone. And so this was what I think is a very small CSF-venous fistula arising from the dorsal dural margin, which is a very unusual place for these to arise. But again, the timing here played a key role in my ability to see this and find this in this patient.
We did a study looking at how frequent, how often, timing makes a difference. And most of the time we do see these on the early phase, but occasionally we only see them on the later phase. We found that about one in eight of our patients, if we had only done one phase, we would have missed potentially their fistula. So I do think that obtaining at least one more phase is a good idea.
Finally, I want to end on something that I think everyone has a lot of interest in, which is the concept of photon-counting CT. Touching on a little bit of the technology and the details of why it’s useful. Here’s an example of a patient who I did on an energy integrating detector CT, the normal regular CT, where I couldn’t really see much with confidence that anything was there. But much more clearly on the photon-counting CT, there was a CSF-venous fistula here. This technology is very promising and exciting.
I like to think of the technology for this in very simple terms. I’m not a physicist. I like things that are very simple. I like to think of an energy integrating detector CT as walking outside, looking at the ground and seeing that it’s wet and seeing a big puddle and saying, “Oh, it must have rained.” And the bigger puddle, the more it rained. It’s more wet on the ground.
Whereas otherwise, I could go outside and have somebody hand me a sheet that told me exactly how many raindrops came down on the street just before I walked outside. And that’s sort of the difference, that the technology is so different in that these semiconductor pixels in the photon-counting CT will log each photon or energy individually. And so it gives us much more detailed information. It gives us less noise, that artifact that we see that we could think might be a fistula or might not be. It gives us multi-energy data that increases our contrast resolution, gives us these intrinsic monoenergetic views, and decreases our artifacts, and does all of this at an equal or even lower radiation dose.
This is a very exciting technology, and this is not an example of a CSF leak but just an example of how much sharper we can see the anatomy here on these two images at the same slice in this patient.
It can be very helpful for internal plexus fistulas in particular. Here’s a photon-counting CT example of a very tiny opacification of this internal epidural venous plexus which was just not at all visible on the energy integrating detector CT.
But it also can be helpful for bigger fistulas. This is a patient who I did a myelogram on the energy integrating detector CT, and I wasn’t sure what this was. I didn’t really think it was a fistula per se on this study. But on the photon-counting CT, you really do see opacification of this quite nicely. I could say with confidence that this was their CSF-venous fistula here draining backwards along the back of the rib into those posterior intermuscular branches.
So in conclusion, I want to just remind you of a couple things. First of all, to anybody out there who is suffering who doesn’t have an answer yet, I think there is so much to be optimistic about in this field. We are learning so much every day, every week, every month about what’s going on, and stay tuned because I think that there’s a lot to look forward to in the science of CSF leak. And CSF-venous fistulas are not just hyperdense paraspinal veins. You need to know your spinal venous anatomy in order to dictate a detailed search pattern when you look at myelography, and myelography for CSF-venous fistulas is an exam that requires real-time troubleshooting, maybe using additional views, using respiratory techniques, or pressure augmentation to really get the most out of your exam and provide the highest quality study possible.
And with that, I will say thank you, and thank you to my incredible team at the University of Colorado. I couldn’t do any of this work without them.