Slides
Transcript
Thank you very much Eike, and thanks for the kind words. Good morning everybody. Very nice to be here. I want to start by thanking the organizers for inviting me. Let’s move on.
First I want to acknowledge Wouter, we’ve been partners for over 25 years now, and he’s really pushed me and others in the department in this direction, and I’m so happy that we’ve been able to do this work together. He’s a good friend as well. So thank you, Wouter.
I also want to acknowledge our team of other neuroradiologists, neurologists, anesthesiologists, and more importantly our team that really makes all of this possible. All of this imaging doesn’t happen by chance. I know we, directed by our IR techs, are really amazing in terms of going with the flow.
We heard about different types of leaks—type 1, type 2, type 3—and today I’m going to talk about embolization, which is related to CSF-venous fistula. Just a simple summary slide showing different types, and obviously type 3 is on the right side. It’s been more than 10 years now that we stumbled on this fistula finding. And we’ve heard about blood patches, fibrin glue injection, spinal surgery—we’ll hear about it next. But the new kid, so to speak, in town is transvenous embolization, and that’s my topic.
And you can see by this collage that there are many different types of CSF-venous fistula, but most of them are recognizable by opacifying a vein of some sort. And when we look at it on DSM with AP and lateral projections, we can see these are either intercostal veins, as shown on this case on the AP and lateral, clearly recognizable intercostal vein, and a paraspinal vein on this case on the AP and lateral.
What is obvious to us now but wasn’t obvious at the time, that we’ve seen for so many years, became obvious to this person. His name is Waleed Brinjikji, he’s at Mayo, he’s an interventional neuroradiologist. He started thinking, “Why can’t I go and try to fix this—go to the vein through my angiographic skills?” And then he started looking at what venous anatomy looks like. This is a publication from 1978, pre-MRI, pre-CT, where French investigators are trying to diagnose disc herniations by doing venography and by way of displacement of the veins, and that’s where the inspiration came and the further work. And this is now what we do every day.
Basic anatomy: there is a fistula around the spine in the thecal sac, and that goes into the paraspinal or intercostal veins. It can also be internal or external epidural plexus to the basivertebral vein, various different combinations. But basically we go to the azygos system for thoracic levels, the vertebral vein for the cervical levels, and the iliolumbar and other veins I’ll describe for the lumbar segments.
How do we do it? I do it by first getting an MR venogram with Feraheme. We’ll talk more about this. We do all the cases under general anesthesia for patient comfort, and my comfort, and my precision in terms of seeing the images and the fine network of vessels.
Through the groin approach, heparinization during the procedure, and we use a tri-axial system which involves an intermediate catheter, a guide catheter, and then a small balloon microcatheter for the pressure cooker technique, where we inflate the balloon and then inject the Onyx, or uh, 18, or Onyx-34. We found this to be very useful – the MR venogram. That is a robust technique which can be done without special attention to timing of injection. It stays there for a while and then it continues to opacify the arteries and the veins, and it helps us with the procedure planning, shows us the anatomical variance. We feel and we know that it shortens the procedure time, reduces the contrast load, and can lead to some unexpected findings which we’ll talk about.
Once we have this road map, and you can see this network of vessels, we can go to the respective levels by catheterizing the azygos or the vertebral vein.
Now when surgeons like Marcus, Wouter, or Jürgen talk and show their intraoperative cases to the non-surgeons, it all looks red and beige and variations of tan, and so we don’t know what we’re looking at really. I don’t want this to be like this, and I hope I can clarify this. These are angiographic pictures—hopefully it’ll be clearer. This is the catheter – it’s in the azygos vein. And you see the azygos and hemiazygos obviously green, and then we have another injection. Again the catheter is in the azygos vein, AP and lateral projections, and there we start seeing the anatomy that’s relevant. We see that the pink-red is the paraspinal vein, that blue triangular depicts the neural foramen, and the light teal is the intercostal vein. The yellow is the epidural plexus. Just going back, you see the actual images like that.
So our target is that neuroforaminal epidural plexus with the Onyx. And here you have an actual case where we drop the microcatheter and balloon to the level of the epidural neuroforaminal. And then you see the glue cast sort of exactly like the angiogram, the venogram that was done previously. And the more penetration you have into that neuroforaminal or epidural place, the more successful the procedure is. Although it’s not 100%, we’ll see examples of that no matter how much you pack it in. Some patients do come back.
Here’s an example in the cervical area. Cervical is interesting because you don’t have to find individual paraspinal vein. Once you get up to the vertebral vein and advance your microcatheter into the epidural plexus, then you can run up and down the cervical levels, whether it’s C1, C2, or up to T1 or T2. And then we can successfully place the glue in this case.
I want to show a few cases where MRV helped us and how it guided us, and also show examples of how I do a case. This is a 61-year-old woman with an L1 fistula, and Wouter went into the OR to fix it, and he backed out because he encountered profuse bleeding. This doesn’t happen very often, maybe once or twice over the last 15–20 years. So he said, “Why don’t I go ahead and try to fix this?” We did the MRV and it showed that this paraspinal vein actually came out of the cava. Typically, the literature says anatomically it should be iliolumbar. So that kind of informed our catheterization. We put the catheter into the cava, we selected the lumbar interspinal, advanced the microcatheter, and put the glue into the respective levels at L1 and L2, a couple of levels, and the patient did well afterwards.
With this study that we did with the MRV—that was the initial presentation series—which were consecutive cases done with MRV and angiogram correlation, the most common so-called variant was a lumbar segmental vein that drained into the cava. In the literature it’s usually iliolumbar vein. The combination of iliolumbar vein both on the right and left was very uncommon in our series, only about 7%. The next common finding was something draining into the accessory hemiazygos and azygos, and I’ll show an example of that.
This is a 57-year-old woman with a left-sided T10 fistula. You can see that there is really variant azygos anatomy. On the left side where we want to go there is no connection with the azygos. So it connected, drained into the brachiocephalic vein. That gave us the roadmap for directly engaging that vessel and dropping our catheter down that route.
This is another case, and this is a challenging and frustrating case earlier in the series—a 75-year-old woman who had multiple CSF-venous fistulas. The first one was treated successfully and then she recurred. We found another one at T7. We did the MRV. You can see that at T8 and T9 there are nice paraspinal veins, but there is no paraspinal vein at T7 and T6, and the angiogram/venogram confirmed that. We spent a lot of hours, we tried to get through the lower levels and go up the epidural, but in the thoracic spine it is not as easy as in the cervical spine to take that ladder up. So in the end we couldn’t do it, and we had to punt it to surgery.
This is a relatively uncommon situation, but it did happen and it does happen from time to time. This is a patient with bilateral T11-12 and T12-L1 fistula, and we did the MRV. We showed that there was duplicated IVC, and both of these paraspinal veins came off of the different moieties of the duplicated IVC. The big one on the right side, but the smaller one on the left depicted by the asterisk, and G stands for the gonadal vein. So having this foreknowledge in advance made the procedure really straightforward from our perspective, where we knew where we were going to go directly, and we were able to cast with Onyx.
Finally, I can’t pass up the opportunity to show this situation where early on when we started doing these procedures we thought, why don’t we try to treat these patients with behavior variant FTD even, and despite the fact that we didn’t find any fistulas. We studied some patients and we found that in four of these patients on MRV we demonstrated azygos stenosis. We went ahead and treated those with azygos stenting, and two patients had remarkable improvement, and one patient had mild improvement. This MRV shows that stenosis in the azygos.
The results of this particular series are that a lot of fistulas that I treated, we treated, were in the cervical spine, just by referral pattern, because many of them in the thoracic spine were treated preferentially by surgery. Although we did treat quite a few in the thoracic spine as well, some of these cases are the FTD variant without fistulas as indicated.
Let’s look at the overall results. This is a recent meta-analysis and includes 15 studies that involve 321 patients with 354 fistulas. They conclude that both embolization and surgery are quite successful in improving headaches and overall symptoms in about 90% of the patients. The remainder of the symptomatology is about 60% resolution. The rates of complications or recurrence are not statistically different in each group.
The major factor that they found in this analysis was the duration of symptoms prior to treatment. In other words, if the patient was suffering for a longer period of time, then the treatment was not effective in terms of relieving the symptoms, and this was discussed earlier this morning. An important subject.
The major complication of rebound hypertension, rebound headaches, was found to be higher in the embolization group, and there is no clear explanation for this because they both work with the same mechanism. It remains to be seen whether there’s something about embolization that predisposes to higher rates of rebound headaches.
In the embolization group – minor complications, which are actually what we observed in our series, not so many Onyx emboli in the lungs, but intramuscular extravasation, back pain, and extensive embolization, which is usually clinically not relevant and not symptomatic.
Why choose endovascular over surgery? Well, it’s less invasive. It’s physically, mentally, psychologically much more acceptable to most patients. Although there are some patients who like the concept of having surgery and being done. Serious complications are rare—a really safe procedure. You’re on the venous system, not on the arterial system. And unless you go through a major vein and cause major hemorrhage, it’s a safe procedure. It’s amenable to treat patients with multiple fistulas, as was discussed. This is something that’s being recognized more and more—at a minimum 10% in many publications. And it’s also probably better for levels of eloquent nerve roots, like cervical or lumbar, where surgery may or may not have some complications afterwards.
A small subset of patients with vascular malformations—obviously the surgeons want to stay away from that. And if the surgery has been done and there is recurrence at the same level, there may not be a good option for resurgery or appetite from the patient or from the surgeon. And new fistulas at other levels—we discussed that.
Why would we do surgery over embolization? Prefer surgery. Well, as I mentioned, patients like the concept of getting there, treating it, and it is in practical sense a little bit more resistant to recurrence, and the success rates are higher in surgery. That’s our experience at least. Also, when you do an embolization and there’s a recurrence, it’s usually not easy to go back and retreat it because you already blocked those veins. That’s another reason to choose surgery. And then some patients you’re not able to access, you don’t have access like I demonstrated earlier in one particular case.
There are still questions out there. What are the long-term outcomes? Why do we have more rebound headaches? I don’t know if that’s a real finding or a statistical fluke. And the effect of glue on recurrence on imaging—obviously with glue streak artifact, imaging for recurrence becomes problematic, maybe less so in DSM compared with CT-based studies. There are potential new embolic agents which have less streak artifact, so we’ll see.
So my conclusion: it’s a safe and effective way to treat CSF-venous fistula, well tolerated. The practice will depend on where you are, what your expertise level is, the patient preference, and there are certain areas where one is preferred over the other, definitely in certain subset of patients. Thank you for your attention.