Dr. Marcel Maya at the 2023 Cedars-Sinai Intracranial Hypotension Conference

Dr. Marcel Maya, Professor & Co-Chair, Department of Imaging, Cedars-Sinai in Los Angeles, CA, presented this talk on endovascular approaches to spinal CSF leaks on July 9, 2023 at the 2023 Cedars-Sinai Intracranial Hypotension Conference. The conference was hosted by Cedars-Sinai with generous support from the Spinal CSF Leak Foundation in Kohala Coast, Hawaii.

 

 

 

 

 

Transcript

[00:00:12] Thank you very much, Ian, for the generous introduction, kind words. Good morning, everybody. And I hope you’re enjoying this meeting as much as I am. Yesterday was a terrific day. I learned a lot. And the best thing was when we left the room, we were in heaven on earth, Hawaii. I want to start out with a few things that you might be familiar with or not. This is the official state butterfly, Kamekaha.

[00:00:40] This is black coral, unique to Hawaiian waters. This is a very rare monk seal, Hawaiian monk seal. And the Hawaiian flag has eight stripes. One for each island, eight islands. Anyway this is the team photo. You saw this.

[00:00:59] SIH is really a perfect disease for a neuroradiologist. That’s always what I thought. It involves so many things that are in the realm of neuroradiology, including the cranial spinal anatomy, neuroanatomy, utilizing all these techniques of older techniques and newer techniques of lumbar puncture, straightforward myelogram; then more cross-sectional elements: CT, MRI, advanced MRI; and then interventional aspects of it, the blood patch and the glue injections, and then recently the endovascular approach. And as Ian alluded to, the field moves very incrementally, and it has moved in the last 20, 25 years incrementally, but there are certain things that are just a jolt and a quantum jump, and fistula was one of them. And I feel like this is another one of those quantum jumps in the field.

[00:01:53] You’re all familiar with the leak classification. Just wanted to put it this up there. There are three types of leak. The type one is the ventral, the dural tear, and then the diverticula. And the third one is the subject matter here for endovascular approach, CSF venous fistulas.

[00:02:12] Typically type one may be amenable to blood patches, fibrin glue injections, or mixed injections, but the definitive treatment is usually repair on their surgical open repair. Type two also may respond to percutaneous approaches, but surgery is very effective as well.

[00:02:30] And the type three CSF venous fistula, the literature varies, as was demonstrated earlier, but the surgery has been the definitive treatment for the fistulas. But this new kid in block is the transvenous embolization. And how did that come about? Well, you’ve seen this slide yesterday, the many different types of CSF leak.

[00:02:51] And as we’re looking at these leaks over the years, of course, everybody, including me, we noticed that you know, these drained into some veins, and some of the veins were recognizable. Here’s a DSM showing the intercostal vein. Okay. Just along the rib. And and then there was another patient, here you’re looking at the AP and lateral DSM, and you start seeing what’s called the paraspinal vein, just along the lateral edge of the vertebral body.

[00:03:24] Unfortunately, I was not the one who made the leap to the next step, and it was a credit to Waleed Brinjikji from Mayo, who came up with this novel thought of, why don’t we go up the venous system and try to get it from the vascular side and block the fistula. And that’s what he did.

[00:03:46] A little bit of the anatomy. These develop right along the nerve root cysts and the foraminal epidural veins. Dr. Schievink demonstrated various different sites of the fistulization. It’s hard to preoperatively determine where the fistula is exactly from the DSM. However, they may be anywhere in that, along that spectrum of the nerve, nerve root cysts.

[00:04:12] Then they subsequently drain into the azygos system, most commonly if it’s in the thoracic spine, and the vertebral vein, if it’s in the cervical thoracic junction or the cervical spine. And then the lower leaks, lower fistulas drain into the iliolumbar veins or cava. And although this is a nice drawing from a paper by the Mayo group, things can vary quite a bit.

[00:04:38] We all know from the, even the cerebral intracranial venous anatomy is a lot more variable than the arterial anatomy. But luckily this anatomy was extensively studied in the 1970s by the French investigators. At that time, the cross sectional CTMR was not present yet, those modalities. And to determine spinal pathology exactly where it is, they went by way of the venograms to determine whether it was a disc herniation, where the disc herniation was, whether it was an intradural, extradural tumor, et cetera. And they published this wonderful manuscript demonstrating all this anatomy.

[00:05:18] And we’ll see the anatomy pretty soon with examples here. This was in, back in 1978. Of course, nobody did them, once CT and MR get into the scene, but then it was resurrected when this endovascular approach was discovered.

[00:05:35] These are the basics of our procedure. I like to obtain an MR venogram to get the road map and anatomy. We perform the procedure under general anesthesia. It is comfortable and it allows for precise breath hold imaging when we need to get really detailed imaging of the vascular anatomy. I’ve usually used transfemoral approach, even though we frequently work in the azygos and close to the neck, it’s easier and safer. And we use balloon catheters, micro balloon catheters and usually patients are discharged the same day.

[00:06:10] The reason I like to use the MRV in advance is is because it gives me a good view of the anatomical variations and prepares me for the what’s coming and I don’t have to fish around even an extra few minutes or do a extra few angiograms, or spend more radiation or contrast when I’m fishing, although they are typically located at certain levels. It varies from patient, and I’ll show you some examples of that. And the MRV with Feraheme, which is a unique agent, which stays in the vascular system for a long time, is very well suited for that.

[00:06:53] Here’s an example of anatomy. My catheter is in the azygos vein, and this is the guide catheter, this is the diagnostic catheter inside that triaxial system. So there is a venogram, and that demonstrates the right azygos and hemiazygos. And then further down, you’ll see the branches, the paraspinal vessels, the epidural vessels, and the neuroforaminal vessels that on the AP, you know, surround the neuroforamen.

[00:07:22] This is right where the neuroforamen is. On the lateral, you see the paraspinal vein, and this is right where the neuroforamen is. So the blue depicts the neuroforamen, the red are the paraspinal veins. And the yellow are the epidural veins. So you can see there’s quite a bit of demonstration and networking between the different compartments.

[00:07:42] So it’s important to know the anatomy so that you can deposit the glue where it’s supposed to be, not necessarily where it’s not. Once we have the guide catheter, we go with the super selective balloon micro catheter and take the balloon micro catheter right to the neuroforaminal region, as you can see here, and then do the venogram with the balloon inflated so that there’s no backflow into the system and you can demonstrate all the collaterals and venous network there.

[00:08:13] Here’s an example of a patient with a fistula, the right side at T3 level. Typically the upper thoracic T1, T2, the cervical, those are more challenging because it’s not like the, you go to the azygos and straight highway to the paraspinal. This is a little bit more complicated. And in this case, we had the MRV again to help us with the anatomy and the supply to T3 area. So we saw that intercostal approach. And then we’re able to get in through the micro catheter up to the T3. And then deposit the glue as the glue cast.

[00:08:52] Next one is a complicated case of a patient, a 45 year old woman, but SIH longterm. This was a nine, a 10-year history. And she had several surgeries, and every time the surgery was successful, but a little bit later, the patient became symptomatic. And repeat examinations show new de novo fistulas, and the most recent one was seen in the cervical thoracic spine at a couple of different levels, and then T9-10 level.

[00:09:22] So, there’s an example where we put the catheter into the innominate and select the vertebral vein, and then with the vertebral vein injection, it looks like a mess, but what we’re seeing here is the paravertebral epidural plexus. The foraminal veins here, the epidural ring, as you can see across with the injection that goes to the other side as well. So, once we’re at where we’re supposed to be and inflate the balloon, we deposit the onyx and you see the onyx cast in the foramen surrounding the nerves.

[00:09:58] In patients who have fistulas even higher, it’s a little bit more tricky. This is a 61 year old man who had a headache for seven years, was worse for upright. MR was negative. And there was still some concern that he could have a CSF venous fistula. And after several different myelograms in different parts of the spine, we finally discovered this upper cervical C2-3 fistula. And again, with the same root left vertebral vein, this is about C6-7 level. But next we’re able to extend the microcatheter up into the epidural plexus and through the epidural venous gutter, come up to C2-3, that’s the microcatheter tip, and then deposit the onyx cast. This is, believe it or not, the patient came back with recurrent symptoms, and this is the follow up investigation with the onyx cast in the foramen.

[00:10:52] This here is a patient just to demonstrate how we use the MRV, a 58 year old woman, SIH with headaches. On DSM a T10-11 fistula, and this one shows this vessel here, and there was no left hemiazygos. So, to get to T10-11 you run down to azygos, but then you’re blocked on the left side. So you go to the hemiazygos, but in this one, the hemiazygos was an accessory origin from the innominate, you see the small vessel coming on. So that gave us an indication of how to go about it. And eventually, we’re able to get into that left T8, and then find our way back to the, that accessory vessel, and then deposit the the onyx material in the foramen.

[00:11:43] I believe this is the last case example. This is a 61 year old woman with SIH and headaches. L1 fistula on DSM, Dr. Schievink took her to the OR, but he encountered profuse bleeding. And so he closed up, and then we got in. But again, I got the MR venogram, and I wanted to see how the supply was because L1 sometimes is tricky.

[00:12:08] So we saw from the cava, vena cava, supply around L3, and then we’re able to put the catheter up through the cava and then super selectively go into that L3 and track it up to L1. She did quite well. I have the pre and post images showing the resolution of pachymeningeal enhancement and optic nerve reconstitution of fluid in the nerves.

[00:12:35] So, these are our result. We just looked back earlier. These are not published results, but we have 50 patients, 29 of which are women, females, with a mean age of 55.6. We have in that group five patients with behavioral variant of FTD, and not necessarily because they had fistulas, but they were at the end of multiple attempts of DSMs and different types of treatments which were not responsive, so we thought we could embolize several of them. Three patients with venous fistula with vascular malformations, which were not good candidates or could not have the successful surgical closure repair of the fistula. And these are the distribution of the fistulas. You’ll see I have in this series, there’s higher number of cervical patients, and I will explain why. No major complications were reported on any of these.

[00:13:23] Technically unsuccessful in five patients. Anatomical variations I like to blame too, one of them for sure access was limited from prior embolization, and there is a learning curve early on, maybe some of the cases were due to my learning curve.

[00:13:40] When Brinjikji, Waleed, presented this first time in the paper, he had 40 patients, similar distribution. He has more thoracic levels that he treated. And he had a better follow up with outcome, imaging outcome, uh, HIT follow up, a recorded 36 of 40 successful delivery of fistula glue, and improved in 34 or 40 patients and rebound headaches in seven patients with local pain at embo site, which is a typical effect that I’ve seen in our patients.

[00:14:14] Here the HIT results improvement from baseline to follow up, and 90 percent was recorded as a minimal improvement. 82 percent is very high degree of improvement in the symptomatology. As was alluded previously, the epidural blood patch success is limited in CSF venous fistula. And surgery is the way to go. Even though it says the literature states 75 to 83%, I think it may be higher than that.

[00:14:43] But so why do we do endovascular surgery at this time? Well, patient preferences may dictate that they, many patients think conceptually that this is a less invasive way of treating it, and they may be right, physically and takes a little shorter for recovery.

[00:15:02] Surgical approach is not so desired in cervical and lumbar levels where there is functional nerve roots. And although it can be done, there’s a, this approach may be safer. And then we’ve mentioned the fistula is associated with vascular malformation. Then there are surgical treatment failures, whether it’s a recurrence at the same level of surgery that was treated, or new de novo fistulas at other spinal levels. And there are a certain number of patients who have multiple venous fistula where this approach is pretty straightforward, you go through the same sitting and may target several different levels.

[00:15:42] What are some lingering questions about in the vascular approach? Well, the long term outcomes are still pending. After that paper, I think I had some communication that Waleed is preparing a longer follow up paper with a larger number of patients. And so I’m looking forward to that. I think the recurrence rates at treated levels are higher than surgery. That’s our experience. And so it’s not quite close to the surgical treatment success rate, but it’s high up there.

[00:16:11] The good thing about that is, when there’s failure with endovascular approach, there is a opportunity to treat, and Dr. Schievink and I published that in patients who underwent unsuccessful embolization, there was no problem treating them with surgical repair. And then there’s a question of glue artifact affecting the follow up imaging.

[00:16:30] You saw on that CT there’s some streak artifact, and that may be really bad for CT based workups. A DSM by way of subtraction is a little bit less affected by that. In conclusion, transvenous embolization is safe and effective, less invasive, and well tolerated. The the role will be really based on the referral patterns, patient preferences, importantly. There is a preference for fistulas where there is functional nerve roots and in patients with vascular malformation. Thanks very much for your attention.