Transcript
Question: Thank you all. Great talks. I have questions for all of you. I’ll contain myself. I’ll just ask two quick questions. The first one about prevention of rebound intracranial hypertension after treatment. There have been some authors advocating the use of pre-procedure acetazolamide to prevent rebound intracranial hypertension. I was wondering what are your experience with this and does this really work? Is this needed?
Answer:
Dr. Maya: Yeah, for embolizations we don’t use that, but the patients do have it just in case, the prescription. They’re given instructions accordingly.
Dr. Mamlouk: I’ll say when we first started fibrin patching for fistulas, we used to give Diamox for all patients. I would say now we rarely give it. But we provide good patient education and tell them to be aware of what to expect, and surely if they need it then later on that week we will prescribe it for them. So it’s more of an as-needed basis.
Dr. Amrhein: Yeah, our institution it’s similar as well. I think it’s really important to educate patients about it beforehand and to follow up with them because the incidence of RIH is actually quite high in our experience at least— 40%. All of our patients do get prescriptions for both Zofran and Diamox. A lot of times the rebound high pressure happens after they go home the night of the procedure, and so they need to have contact numbers and be aware of that.
Question: My second question is about contrast in the kidneys as a sign for CSF-venous fistulas. In our institution we do DSM, and usually when there’s no clear evidence for leak or fistula on the DSM, we do a standard CT myelography after that. If there’s contrast in the kidneys, this is often taken as a sign that there must be a fistula even if we can’t identify it. But we don’t always have access to CT right away. And sometimes the CT is performed only a half hour, 45 minutes, an hour after the initial intrathecal contrast injection if there’s contrast in the kidneys at that time. Could it be that it has gotten there via the physiological routes? And how soon must the contrast be in the kidneys to be pretty confident that there is a fistula?
Answer:
Dr Kranz: I can weigh in on that question. So that’s a really good question. There are two things that I think the literature has shown us so far. The first is that if you have a CSF-venous fistula, there is a statistically significant difference in the amount of contrast you’ll find in the renal collecting system. The second thing is that contrast in the renal collecting system is a normal physiologic response because you excrete contrast through your kidneys. And so the question is not do you see contrast in the collecting system with fistulas? The answer to that is yes. The question is: Are you able to discriminate a patient with a fistula from a patient without a fistula in a way that has a threshold associated where you can measure and say at this time point if you measure it to be at this value or below, or at this value and above, you’ve ruled in or ruled out the presence of a fistula. And that was the question we tried to address in a paper maybe a year or so ago. And what we found is that when we scan in the early time period, within 30 minutes of contrast injection, that even though there was a statistically significant difference, there was not a clinical threshold that we could develop that would accurately stratify those two groups. So there are bell curves, but the bell curves overlap. You can find a statistical difference, that’s true, but you can’t draw a line that has high sensitivity or specificity in terms of separating those two groups out. So I think other people may have different opinions, but for us, in our hands, I would say we don’t find that to be a particularly helpful sign because there is overlap between normal patients, and there are reasons for false positives. If you take the needle out and contrast leaks into the epidural space, it’s going to get reabsorbed and go into the kidneys early. So there are false positives, false negatives. It’s not quite good enough for us to conclude definitively you do or do not have a fistula.
Question: I’m originally from South Korea, and I have a question to you, Dr. Amrhein. Let’s assume that a patient came to you after three days of onset, and she or he had a very clear sign of brain imaging and the orthostatic headache, and his or her spine MRI shows a severe depletion of the intradural CSF volume. It’s kind of really collapsed, and all the CSF is going out, and they are collected in the extradural space. Let’s assume two scenarios. Would you do dynamic imaging? Would you do the puncture on that collapsed dura? That is the first question. And the second question is that if you have a long waiting list, like the patient should wait, let’s say two months to get an ultrafast image in your institution, would you wait for this, or would you do blind or non-targeted EBP in this case?
Answer:
Dr. Amrhein: Yeah, thanks for that question. If I’m understanding correctly, basically extradural positive extra collection in a patient with known SIH. How do we approach that? That was the first piece, right?
Unknown Speaker: Yeah. First question is any kind of presumed risk to puncture in case of a severely depleted intradural CSF?
Dr. Amrhein: Yeah. No, we will typically go looking, right? So in that scenario where we’ve got a large SLEC, as some of our colleagues call it, we’re looking for either a type 1 or type 2 leak, it’s a dural tear. And some of the things that can be helpful there are you could use Lalani’s sign, where you’re looking for a flow void jet on the MR that might indicate where there could potentially be the site of the dural defect. And then you can target your myelography accordingly. Alternatively, if there are prior imaging exams or if you see that there are osteophyte spurs anteriorly, you might target your myelography in that capacity. And really, what this gets down to is whether you’re going to put the patient decubitus because you’re looking for a lateral leak, or if you’re putting them prone to look for a ventral leak. And a lot of times with dural tears and large volumes of extradural CSF, it’s not an eye test. You know that they have a leak. The challenge is localizing the level of the leak, which requires some dynamic imaging with increased temporal resolution. So that’s usually these DSM techniques that you heard about today, or the ultra-fast CT myelography technique that the Mayo Clinic originally reported, where you’re taking rapid scans back and forth and trying to capture the moment where the contrast leaks out. So at our institution, what we would typically do in that case is go looking for the leak. We wouldn’t do a non-target patch, we would go find it, because in our experience while those ones can often be treated in the early period, many more chronic cases do need surgery. And I can’t send them to my surgeon if I don’t tell them exactly where the leak is. So the diagnostic imaging is really critical in those cases. I’m not sure if that fully answers the question.
Question: I have a question here from our virtual audience. It says Dr. Carlton Jones seems to have a lot of—oh wait, that’s a different one—a lot of slides. Question for Dr. Carlton Jones: how about the learning curve for an interventional radiologist to perform dynamic CT myelogram? Most places do not have trained people to perform and interpret either of these exams, meaning dynamic myelo or DSM.
Answer:
Dr. Carlton Jones: Yeah. I mean when I come to a talk like this, it’s not like I’m going to show some of my early CTMs, which I would be looking at going, where is the leak? There is a learning curve for all types of myelography. But I do think as an initial method it is an easier one to get started on. But it’s not just a sole operator method. We really spend a lot of time educating the techs and the radiographers on positioning, being really anal about that, and also we collaborate with others. And I really can’t express this enough. I have sent many cases to colleagues in the UK, Dan and David, who have been great support, and colleagues abroad, because you shouldn’t have ego about getting these things wrong. You should be sharing these cases with others because there’s learning in that. And when you’re just starting, you really need that scaffolding of support. So I would encourage anybody who is not having success, reach out to me, reach out to others who will be more than willing to help you when you start. And then I think you build up that initial volume and you become better through those marginal gains. I think I was talking with Eike about this. You notice things that perhaps you’re not getting right in the beginning, and then there are small things that you add to your practice from the literature that comes out and from other things that people have told you to improve that. So of course there’s a learning curve to any myelography. I think that CTM is an easy method to start with to perfect it. Like with any intervention, any procedure, you need to add in those marginal gains. Have I answered the question?
Dr. Schievink: Let me ask the virtual audience. Yes.
Dr Kranz: Can I make a comment to that? So I think there are a lot of people who have successfully started practices, and one of the common denominators is that they spent some time at another institution looking to see how they did it. And they don’t always do it the same way that the institution that they visited did it. They may make changes, and those changes may be better than the way, for example, they do it at Cedars. Certainly not at Duke — maybe at Cedar. Just kidding. But I would just emphasize what Dr. Carlton Jones said, which is don’t think that you can or necessarily want to be able to start this practice by yourself just by reading a paper. You need to come and you need to spend time. And a lot of people up on this stage have hosted visitors and are happy to do so. I know Freiburg has a visiting program in November, I think you said Jürgen. We certainly take visitors. I know a lot of places, Cedars has had innumerable visitors and things like that. So you should really reach out to the folks who are already in practice. Find someone who you know, who you like to work with, and visit.
Dr. Carlton Jones: And always remain open to learning because our success rates are not 100%. So there’s always some, you know, going to be new literature.
Question: There’s another question from our virtual audience. It’s for Dr. Beck. Professor Beck, what percentage of patients with a subacute subdural hematoma have SIH as their cause?
Answer:
Dr. Beck: Thank you for that question. We have to discriminate in younger patients below 60 years of age. We have some data that probably a quarter of these patients do have spinal CSF leaks as the cause. We have several risk factors, which are early recurrence and bilateral subdural hematomas have a high chance of having a CSF leak. And since this is so intriguing and changes treatment to a completely different direction, we—thanks to Horst Urbach, Niklas Lützen, Lisa Leffert, and Katharina Wolf, we set up a prospective study. And currently we’re really doing MRI and myelography in all our patients with chronic subdural hematomas for two years. This is a lot of work and this is only being feasible by the enormous willpower of Horst and Niklas, and we are even sharing, I think this is the spirit in Freiburg, we are even sharing a young neurosurgical resident who is now going a couple of days into a neuroradiology and doing myelograms. Only together this is possible. And so we are not as good as we want to be. So we probably have 70% included so far, but it turns out to be that also the 80 years old, 90 years old with chronic subdurals do have spinal leaks, and it’s a prospective study. I’m not fully aware of the data, but it seems to be 8 to 12% positive rate of spinal leaks. This could be a paradigm change, that we need to look for spinal leaks in chronic subdurals. I do not know how to solve this problem. What shall we do? Even do more myelograms? Now, in the most common neurosurgical disease we have, I think a typical neurosurgical center has 200 chronic subdurals a year, or 400. So how can we solve this problem? We cannot, but we cannot ignore it. Only because we do not have a solution we cannot ignore the problem. So continuing work, and I still believe that we do need a biomarker, that we need to find something in the CSF or in the blood to pick out these patients where really invasive imaging is needed. This is an excellent question. We are currently working on that, and we’re trying to go over with our prospective study, and hopefully we will manage together. Thanks.
Question: So if I could just tag on to the part of the other discussion, for the headache neurologists in the group I can say that this morning’s discussion about all of the treatment options and the relative benefits for imaging-confirmed leaks elicits all kinds of emotions ranging from salivation to frustration to irritation, because so many of us work in institutions where none of these things are available, or where there are no specialists in either radiology or surgery who are interested and motivated, etc. So if one of the things that the ISSCL can do for us is help us find ways to get institutional support and demonstrate the value of appointing people who will learn these techniques and the value that clearly exists for patients, that would be an enormous service of that group. The other thing that comes up though is Dr. Carroll’s study showing that even without imaging confirmation, patching has some value. I would like you to say a little bit more about what does three and a half patches constitute exactly. What volume, what level, and how has your thinking about patching in this patient group changed over time? Because for us, again, we have a whole large cohort of patients that seem to meet these criteria but in whom so far imaging has not produced any clear diagnosis.
Answer:
Dr. Carroll: So the first is the patches that we gave to these imaging-negative patients with orthostatic headache were neither nondirected nor targeted in kind of the classic sense because there was no leak seen, but they were image-informed. If somebody had a very large perineural cyst at T4-5 or a large ventral osteophyte at T1-2, their patch targeted that area. So they were image-informed even if you didn’t see a SLEC arising from the osteophyte at T12. So we did try to direct blood and/or fibrin to the spots. Most patients got blood first, and if they were unresponsive, then got fibrin. Although when we went back and looked at it to see if blood or fibrin was predictive of having the response, we didn’t, but they weren’t randomized. So I don’t know what we can conclude from that. My thinking has changed over time in the sense that I feel I am in some ways more cognizant of the risks than when I first started. I think the things that Wouter doctor talks about with the risks are not unimportant. I think they’re really important. I wonder if we’re better now than we were when we only had a couple of years of experience. I think being able to target things that we’re not certain about. Is that a bleb? I don’t know if it’s a bleb, but now I know how to look for a bleb. Any needle in somebody’s prior history, any needle in their back in their prior history, and someone with an orthostatic headache should have a very close look at that dorsal dura. Is there anything funny about it? Over time we’ll learn better ways of distinguishing artifacts from real blebs. What’s a real dinosaur tail sign? What’s an artifact? But there’s always going to be, no matter how good the imaging gets, there are always going to be things that we all see on imaging where we’re looking at it and we’re saying, is that real? Is that not real? And the threshold for calling it real on the study may be higher than the threshold of, you know, in this patient, it’s reasonable to put some fibrin glue on that and see if that’s really the problem. It might not be obvious enough to say I’m going to embolize it or I’m going to operate on it, but it might be good enough to say in this patient with an orthostatic headache this level looks like it might be the problem, and it rises to a threshold of doing something safe like an empiric patch.
Question: Quick comment. I just wanted to acknowledge Farnaz [Amoozegar], who is also a charter member of the ISSCL, who was inadvertently left off the slide. My question has to do with subdural hematomas. Roughly, how often do you see non-traumatic spontaneous subdural hematomas that you think are caused by a leak, but the brain MRI or the brain imaging shows no other signs of a leak? Or you could answer that the other way if you want, that it does show other signs of a leak. Does that make sense?
Answer:
Dr. Beck: Excellent question. And we try to skip that question because what is a non-traumatic subdural hematoma? It depends on how long the resident talks to the patient. Didn’t you hit your head in the last six weeks? Yes, of course. If you insist asking, oh yeah, yeah, sure. I hit my head. So this is absolutely not reliable. So we skip this completely.
Dr. Friedman: Okay. A subdural hematoma bilateral— any whatever?
Dr. Beck: Whatever, whatever. And the other question is, if you have a subdural hematoma you have, I think per se, at least a score of three. So it was the longest discussion, but then we decided okay, we cannot answer this question. It’s a very good question, so scan them all. So every patient with a subdural hematoma in Freiburg gets a spine MRI, and if there is suspicion then we go on with myelography.
Dr. Beck: No. No acute hematoma, no car accident. No, not the acute ones. The acute ones go to the OR.
Dr. Carlton Jones: In some healthcare systems we might not have the capacity to scan every patient with an MR. And obviously your data that comes out will be really important in the future to learn what is the true incidence of this. But we also teach our neurosurgery residents, well actually anyone that’s seeing the patient, that you can reconstruct your CT heads and you will often see the other findings on these scans as well. So it is also about engaging your neurosurgeons or whoever else is picking up, and even sometimes we’ve had some picked up by our accident and emergency physicians to look at these. So I think just raising awareness about that is also important.
Dr. Amrhein: Yeah. But I’ll say that I think the important point maybe is that in the setting of chronic subdural hematomas it’s two things that come to mind. One, I can tell you at least in my practice it’s not zero. Right. There was definitely plenty of— there were several cases that we’ve had float at Duke where all you had was bilateral chronic subdural hematomas without brain sagging or anything else, and then we were surprised to see a fistula. And even some cases where others sort of forced us into going forward with myelography and we were surprised to find— I had a case just like that. The other thing that I wanted to mention is that some of the things that we look for with SIH on a brain MRI with contrast, like dural enhancement, can occur in a reactive fashion to the subdural. So it gets kind of confusing that way because you lose some of your opportunities to see those signs.
Yeah, I think time is against us, so we will draw this session to a close. Just like to thank all of our speakers and Ian and Ajay for moderating.
Question: May I make a last comment only— you guys, lunch is waiting. You make a society and in the bylaws it’s not possible that physiologists might join, and it’s often about diagnostic and therapeutics. But we should also go back to [inaudible]. Wouldn’t it be good to get also basic scientists involved?
Answer:
Dr. Kranz: Yeah, I think that’s a great point. I think that most of the people currently operating in this space are physicians, but we do have provisions in the bylaws for people who are PhD scientists who can join in that context.