Hello, I'm Rob Naftal, and welcome to the AANS-CNS section on Pediatric Neurological Surgery. Thank you for attending the 2021 AANS Virtual Annual Meeting. We look forward to an exciting event and connecting with you. To submit a question, please enter it into the Q&A box on the right side of your screen. This will go directly to the faculty. To engage with your colleagues, please use the chat box on the right side of your screen. Self-report CME for the live virtual event in MyAANS. Thank you to the AANS leadership for their thoughtful and exhaustive planning of the 2021 AANS meeting. As the meeting transitioned from Vancouver to Orlando, organizers remained nimble and fortunately planned for a virtual contingency plan. In keeping with our mission, they made the difficult and appropriate decision to transition to virtual. We are thankful to still be able to come together today. On behalf of the Pediatric Section National Meetings Committee, I want to personally thank Dr. Greg Oliveria. When the meeting transitioned to Orlando, he was instrumental in assisting with planning speakers and content. Dr. Oliveria is here to introduce our HAARP lecturer. Thank you, Greg. Thank you, Rob. It's my pleasure to introduce Dr. Amy Smith. Amy is the Director of the Pediatric Neuro-Oncology Program and the Division Chief of Pediatric Hematology Oncology at Orlando Health Arnold Palmer Hospital for Children. She is the Study Chair of the ACNS 0831 Phase 3 Randomized Trial of Post-Radiation Chemotherapy in Newly Diagnosed Ependymoma in Children and Young Adults. And Amy is the Chair of the Ependymoma Subcommittee of the COG. It's my pleasure again to introduce Amy Smith for our HAARP talk, Ependymoma, a Paradigm Shift. Hello, thank you for that nice introduction, Dr. Oliveria, and thank you very much to the committee for inviting me to give this talk today. So I'm going to, I try not to make too many slides. And as we start off, I would like for each of you to try to imagine you and me as individuals in a few different scenarios. Imagine me as a mother. You imagine that you meet me in the emergency room and you're seeing my one year old with fussiness and crossed eyes and vomiting that's been going on now intermittently for a few weeks. My child has that top MRI. Another day, you imagine that you meet me and you're seeing my three year old who has poor balance, a head tilt, and vomiting. And the MRI shows that middle picture. And in the last scenario, we meet and you're seeing my 12 year old who's been complaining for a few months on and off about headache and neck pain and now more recently, some double vision. What do you tell me? In all scenarios, you're going to tell me that my child has a brain tumor and that he or she needs to have surgery to remove it. What's the diagnosis? In all three scenarios, it's an ependymoma. You all well know that 30% of ependymomas occur in the supratentorial compartment of children and 60% occur in the infratentorial compartment. And then a lesser proportion occur in the spinal cord, but we're not going to talk about spinal cord ependymomas today. And so the questions I ask and that we all should ask ourselves is, are these tumors the same disease? No, I think we know that now. Do they get the same treatment? Essentially yes. And do they have the same outcome? And the answer is no. I'm going to spend most of my time reviewing this timeline just to kind of take us back, because I think it's always interesting to take a little look at history and that this tumor, which was described in the 20s, and then really there are small case reports and brief reports up until the 70s and in the early 80s, where surgery was really the approach for just as it is for all brain tumors. And sometimes radiation was given, but really until the late 70s, early 80s, those patients were never referred to oncology. And the reports that are published are really have very short follow-up and really don't give us a good idea of what the outcome was. And so in the 90s, there were several small trials published, and I say small, up to between 50 and maybe up to 100 patients. And these trials looked at a variety of things, trying to delay radiation for very young children by using chemotherapy approaches after surgery, using radiation at relapse, and or using surgery-only approaches in the older kids using radiation after surgery. And essentially in 1994, the baby POG trial was published showing some exciting results that the two-year overall survival was 74%. But when you follow those patients out to four years, which was published in 1999, the overall survival was only 36%. And additionally in the 90s, there was a real effort to look at what was the outcome related to surgery. So there was a retrospective review of 31 patients that showed about a 50% survival at five years. However, almost 75% of those patients had relapsed. When you went to look at the extent of resection in those patients, there was a definite survival advantage for patients who had a complete resection. You can see 69 versus 46%, but there were low resection rates. I put that there's only two out of 20 posterior fosses who had complete resections. There were obviously more to get those statistics, but the others were supertentorial in nature. That was published in 1996. So there's a really long gap there. So since the mid 90s, the dogma has really been that surgery and radiation are the approaches and chemotherapy doesn't work in apendymoma. And I would say over the last couple of decades, the focus has been on how do we increase the rate of complete resection, safe, complete resections? How do we safely give radiation? How do we give more conformational radiation to these patients? So that's really taken up the lion's share of the last couple of decades. In 2003, the COG opened a trial, ACNS0121, that was the most rapidly accruing trial. And at that time, we had the most patients enrolled. And then subsequent to that, we've opened and now closed ACNS0831 that you can see opened in 2010 and closed in 2019. And I'll go through the details, a little bit of both of those trials as we go through the talk. The other thing that's been really interesting is that prior to 2005, most of the prognostic factors that were really assessed in all the different papers, and they continue to be assessed, is looking at extent of resection, location of tumor, the grade of the tumor, whether age has an impact and whether radiation has an impact. But in 2005, Michael Taylor and his group published a really good paper to show that there were candidate stem cells of appendymoma that correlated to the location of the tumor, to whether they were in the supratentorial, the infratentorial, or the spinal cord compartments. And then there was really an explosion over the next, really up until now, of biology that's come out. And so in 2010, Richard Gilbertson's group published a paper showing that cross-species genomics matches driver mutations and correlates with the cell compartments, and they were able to model appendymomas, so sort of following on the 2005 paper. And then in 2014, the Heidelberg group published a paper showing that there were two distinct molecular subgroups of posterior fossa appendymoma, and then a separate paper showing that the C11-ORF-RELA fusion drives NF-kappaB signaling in supratentorial appendymoma and then Steve Mack and his group published a paper showing that these tumors are epigenetically modified and that you could identify different subtypes of appendymoma in infancy. So just to go over the largest COG trial up until the most recent one, this is ACNSO-121, and again it was open from 2003 until 2007 and enrolled 378 patients. You can see that it was all the patients had central pathology review performed on all the tumors. It was done rapidly in real time for supratentorial patients, and the trial was divided and stratified into four groups. The first one was patients with completely resected supratentorial differentiated or classic histology appendymoma because there had been a couple of small series only up about 25 patients showing that it was possible to observe these patients with surgery alone, and so this was done in a formal fashion on this trial. Stratum-2 was for patients who had a subtotal resection, any histologic grade, any location, and the question was really being asked, could you use short course chemotherapy to improve your resection rates for these patients? Based on an earlier paper that I mentioned at the last slide that had been published in the mid to late 90s showing that there was benefit to giving short course chemotherapy to improve resection rates. The third stratum was for patients with near total or GTR-2, any grade, any location. Those patients went on to get radiation therapy with a CTV of one centimeter, and then stratum-4 were patients with a gross total resection one if you were anaplastic and supratentorial, infratentorial, any grade. That paper was published in 2019 in JCO, and you can see the results here that for the event-free survival at 10 years is 55 percent, and the overall survival is 71 percent. I didn't go into a lot more slides for this trial because it has been published for a couple of years now, and you can go through the details, but the conclusions I will mention is that really that trial was really groundbreaking in that it showed that you could safely deliver conformal radiation therapy to very young children down to the age of 12 months. The trial also showed that observation of supratentorial ependymomas that were resected with grade 2 histology really could not be considered a standard. It was, there were only 11 patients on that arm of the trial, and there have not been any deaths, but there were some relapses, and those patients were salvaged with repeat operation and radiation therapy. Chemotherapy at the end of that trial could be considered as a bridge to second surgery as there was some improvement in outcome for those patients if they could get to a complete resection before radiation therapy, and it addressed some issues that had been published several times before that there was a lot of variability in the histopathologic assessment, but in this trial, it showed that there was not as much variability as we had predicted, and in this trial, survival outcomes did correlate with grade. It also showed, and they began to apply some of the biology that had come out during the lifespan of this trial, that 1q gain is a marker of poor prognosis showing increased local and distant progression, and that there was prognostic significance between having a posterior fossa group A and a group B tumor, or sorry, was not shown after having radiation, and that relay fusion tumors do not have a uniformly poor survival, and this went against a biology and retrospective paper that had been published in 2014. The next generation trial that I've led over the last several years is 0831, and this trial was really built on the backbone of the prior study to do a similar stratification, again, just based off of extent of location of the tumor, extent of resection, and grade. The primary question of the study was, do patients benefit from adding maintenance chemotherapy after radiation therapy for patients who have had at least a near total resection? That was the primary question. We did decrease the radiation field from 1 centimeter to a half a centimeter. There were 479 patients enrolled. 451 patients were eligible. 330 patients were randomized, and 325 of those patients were eligible. You can see there were 164 patients randomized to radiation plus chemotherapy, and 161 randomized to radiation only. There was an arm for patients who had residual tumor. We non-randomly assigned those patients to get maintenance after radiation therapy, and then we did a similar thing to a one-to-one where we had patients who were supratentorial, grade two, completely resected, who were observed, and we had 37 of those patients. The one very important and interesting thing historically and is really impacting us is that there was a high rate of non-compliance for patients who were randomized to the arm of chemotherapy after radiation. I really believe that this is in large part because there was bias in our community about whether or not chemotherapy worked because as you saw a few slides ago, that really had become the dogma because all the studies that had been done, though they were small, had really not shown benefit of chemotherapy except for one trial that had been run in the mid-'90s, but only the abstract had been published until about 2012. This presents a significant challenge for us with this trial. It also impacted, I think, the rate of accrual. It took us nine years to accrue the necessary number of patients. Because we anticipated there would be trouble with compliance, we planned an as-treated analysis from the very beginning in the design of the study. Here's the primary analysis of the trial. This is for all randomized patients with intent to treat or as randomized. The blue line represents patients who got radiation followed by maintenance chemotherapy, and the red line represents those who had radiation only. You can see that it looks slightly better, but it does not meet statistical significance with a p-value of 0.07. Because we had planned ahead of time to do a secondary analysis to look at patients as treated, not just as randomized, you can now see a separation of the curves. The patients in blue are those who got chemotherapy after radiation. The patients in red, they only include patients who got radiation only. But again, thinking about that, that's including patients who had been randomized to get chemo, but they never got any. Anybody who got a single dose of chemotherapy stayed up in its randomized in the blue arm. This does reach statistical significance of 0.01. Then we begin to break this down a little further. This is looking at the event-free survival for randomized patients with complete resection after initial surgery, again, based on an intent to treat, so as randomized. You can see here, the data do suggest that the addition of maintenance chemotherapy is beneficial for newly diagnosed patients who have a complete resection up front. What's interesting here is that for this group, 27 percent of those assigned to chemotherapy did not receive any chemotherapy. To me, this also implies, because again, this is a curve that shows you as randomized, not as treated, it shows the separation of the curves. Then in a sub-analysis here, when we're looking at patients who are as treated, not as randomized, the p-value becomes quite strong at 0.004 between the two groups. Then this slide is also interesting because if we look at by location, I want to start with the black line at the bottom. These are posterior fossa patients who got radiation only, just about 50 percent or a little less. Then the red line in this curve are for posterior fossa patients who got radiation plus maintenance chemotherapy. Those are the two comparisons. The supratentorial patients who got radiation only is the gray line, and those who were supratentorial who got radiation and maintenance chemotherapy is the brown or auburn line. This is a curve that actually looks at the patients with residual disease and compares the patients on 0831 with the patients treated on ACNS 0121. Basically, the comparison is those on 0121 got radiation only, even though they had residual disease, and patients with residual on 0831 got radiation and chemotherapy, but there was no difference. For patients with residual disease, there's no benefit to adding chemotherapy for patients who have residual. The recommendations that we walked away from after this trial is that patients with localized M0 ependymoma who undergo a complete resection, either as defined by a gross total or a near total resection, which is less than five millimeters of residual at the time of initial surgery, they appear to benefit from the addition of maintenance chemotherapy if it's given after if it's given after radiation. We are currently recommending chemotherapy for this group. The challenge right now is that there are molecular analyses going on to look at the subgroups because we do know that it's very possible that benefit of chemotherapy will be maybe limited to certain subgroups. So even in the slide before where we see that whether you are posterior fossil or supertentorial, if you get a complete resection, you appear to benefit from chemotherapy, but we do not yet have the breakdown of the more newly described molecular subgroups, which will be out in the next month, I hope, maybe before the COG meeting. And it's also important that we need to, as clinicians, be considering the toxicities of chemotherapy, such as myelosuppression and ototoxicity, which we're looking at very carefully in this group. And patients with M0 ependymoma who are unable to achieve a complete resection, those patients do not benefit from chemotherapy. And I think that's just as important to remember, to not expose these children to treatment that's not going to benefit them. So now just to take a summary slide and talk a little bit, go back and talk about the biology that we have been learning, is really looking at the molecular subgrouping. That's been the main discovery in the last several years is that there are really nine molecular subgroups of ependymoma. This is determined by their DNA methylation and gene expression profiles and by their specific genomic alterations. So for patients who have infertentorial tumors, there are two main groups. There's the posterior fossa group A, which reveals a loss of H3K27 trimethylation mark, and posterior fossa B. Those patients have retained H3K27 trimethylation mark. There are some clinical features about these tumors, the PFA and the PFB, and that the PFAs occur mostly in young children and PFBs occur mostly in adults and down into the teenage years. So back to the first slide where we were talking about if I have a 12-year-old with that posterior fossa tumor, it would be important whether that's a PFA or a PFB tumor. The supertentorial tumors are divided into two categorized biologic groups, the C11-ORF95-RLA group, which makes up about 70% of all supertentorial ependymomas and the YAP1 fusion ependymomas. One important comment that we learned from 0831, they also saw this on a one-to-one with central review, is that supertentorial ependymomas can be tricky and that most of the tumors that turned out to not be a ependymoma on central review were in the supertentorial compartment. And some of them have unique molecular findings that identify them as other entities that fall into hygrogliomas or neuroglial tumors in other families. And then the spinal cord tumors are divided into myxopapillary and classic histology and the subependymomas biologically fall out into each compartment as well. So I think as we look forward, it's gonna be really, it is critically important that we accept stratification for all of our future trials. It's necessary that there is molecular testing available to really all people who are taking care of these children. And our group looked at this recently in the state of Florida and found that there were actually a large number of centers that do not necessarily have molecular testing performed on their tumors. And so as there are additional labs coming online and this is becoming more available, I think this is an important thing for you all as surgeons, as you work with your teams and for all of us to make sure that our patients have molecular testing available. I didn't go show the papers, but there are a couple of really, really interesting papers now that are not just looking at bulk sequencing of these tumors, but actually isolating single cell and doing single cell RNA sequencing to understand the spectrum of cells that occur within these ependymomas and showing that there's really a lot of dynamic going on. The balance of that is very important. And this is probably going to become very important for us as we identify different treatment strategies that it's not going to be just like we were not too long ago, we're only looking at location of the tumor. Well, maybe we're not just looking at whether it's a RelA fusion, we may also need to be looking at what is the true cellular makeup, what's the balance of the cells that are going on in that particular tumor. Additionally, we really have got as a community to focus on making more models, better models that really reveal what's going on in the child and sharing those with one another, increasing our data sharing. And so that we're really learning from every patient all the time. And so encouraging your hospitals, wherever you work to make sure that tumor tissue is being obtained and saved properly and shared with the scientists working on this all over the world. So that's all I have right now. If anyone has questions, please feel free to ask and I'd love to have some discussion. Thank you. Dr. Smith, thanks so much for that wonderful presentation. Thanks for being our hard lecturer today. And it's nice to have someone who has such an expertise in the ependymoma to join us. One question I have is someone who has as intimately involved with the previous trials, what do you see off of that last slide that you showed us when looking forward, what's gonna be the next trial and how are we setting up for that? So we're really working on that quite a bit right now. I think one potential thing that we need to answer is since chemotherapy does look like it works, at least for some, there is toxicity that comes along with that. That doesn't come without a price. So one question will be to define very clearly as best we can, who are those subgroups of patients? Is it really everybody across the board who's had a complete resection or are there some biological subgroups? And then I think once we know that, which we will pretty soon, then I think we need to be asking the questions, are there things, what sort of toxicities are we seeing? How severe is the hearing toxicity that we've encountered? Are there things that we can do to mitigate that while we get the benefit from the addition of the chemotherapy? Because that's something that our ependymoma patients really haven't had to face a lot to date since we've not been using those agents. And it's still very difficult in regards to targets, which is why I made the comment about it's important for models because even though we have all of this wonderful work to show us how to subgroup the patients, the drivers and looking at targets to intervene is really still lacking. There's an exciting trial going on. I'm sure that most of you know about the CAR T cells that's happening, I believe, in Seattle. So there are some, but those are, again, rare subgroups, multiply relapsed patients. There are just still, there's a paucity, I think, of what to do in the biologically driven groups next. So a lot of work to be done. Yeah, thank you so much, Amy. This is great. One question I had, and you sort of started to answer it, I think, which was, yeah, you mentioned a lot of the single cell resolution data that are coming out now and sort of how do you envision not only in a pandemoma, but in other brain tumors, translating those kinds of data into clinical trials? Oh, Todd, I wish I knew. I wish I knew. I really, I would love if anybody else has a thought about that. I honestly am not quite sure how we're gonna make that leap, but we are going to need to make that leap because it is fascinating that you can identify a molecular, even a molecular driver, but if the volume of cells you have in your tumor is really not very high and it's the other cells that are more influential, we are gonna have to find a way to, so I don't know if that's gonna be through in some integrating AI a little bit to help us understand that dynamic. I'm not quite sure. Do you have a thought? I know you work- One of the things that, I mean, yeah, with craniofragilma, we've been playing with like sub-network models where you're looking at the communication pathways between cell populations, but I feel like that is very nascent in terms of getting to the level of sophistication needed to translate that, and that's kind of why I was curious about what your thoughts are, because I feel like neuro-oncology is usually a number of steps ahead of at least where I tend to be thinking in terms of biological things, and so I was really curious where you were on that. For me personally, I think about those things, but I'm a bit stuck there, honestly. I don't do any single cell work myself, so those people who are working in that space would probably be much more helpful to give comment there. Yeah. Dr. Smith, we have a question from our online audience. Dr. Boop is asking, did chemotherapy have benefit in kids with a 1Q gain? So we don't know yet because the biology is under analysis actually right now as we speak, so in the next month, four to six weeks, I hope we'll know the answer to that. Okay, great. One more question from online. It's in reference to the noncompliance. So we noted the 30 plus percent of noncompliance. How do you adjust the results when such occur? Would you consider that more like a crossover? Any ideas? So what we did in the way we had planned it and when I was presenting the as-treated arm, essentially what we did is we said, okay, if you were randomized to the chemotherapy arm but you didn't get even a single dose, then we did that secondary analysis and just included those patients with those who got randomized to radiation only. So even there, even taking those patients out of the chemotherapy arm, you could still see the benefit of chemotherapy. So we do think that the effect is real, but of course it does make things a little challenging since that wasn't the way the original trial was set up, the way the trial was powered initially. Thank you. Amy, I'd like to ask you just one other kind of more general question, which you commented on the fact that there were challenges around adherence to the assigned regimen and related to bias. And I think that's something that we really run into with neurosurgical RCTs as well. And so I just kind of wanted to ask what your perspective is on ways to address challenges like that in any kind of randomized trial. You know, one thing I wish I would have done from the very beginning, which I have to confess, senior person suggested to me and I didn't really know how to do it, so I didn't do it. But I do think it's probably important as we're designing these things, especially now with all the social media stuff going on is finding ways to understand the biases that are going on in the community and to ask people what they want and really starting from that a little bit. I don't know if we've really done that in medicine is to find out what people, what questions they want, what things they are willing to do before we launch into these things. So not to say that we wouldn't have done the trial because I think obviously we should have. And I think that it is showing, it's a surprising result. So I'm glad we did it, but I do wonder if we would have done better had we engaged the community upfront in maybe some of these parent forums to try to understand really where they were coming from and maybe deploy a group of us to have done some community engagement along the way. Because I think it wasn't until children started on this, like relapsing a lot. And then I think people started to saying, oh, uh-oh, maybe I should have put them on the trial. So you're really thinking along the lines of community engagement as opposed to engaging our colleagues upfront during trial design. So you think that the adherence issues had to do with the patients a little bit rather than necessarily the neuro-oncologists themselves? I think it had to do with both. And I think in our discipline, as surgeons and neuro-oncologists and radiation oncologists, we care for our patients in such an integrated way that it's not just, I think the neuro-oncologists could feel one way, but if the surgeons and the radiation oncologists feel differently, the parents are interacting with all of us. So I do think that you're right, we should engage our colleagues in this. Like I was a little intimidated to come talk to you guys today because you're all neurosurgeons, but I think it's important that we talk with one another and with our radiation oncology colleagues and not just stay within our own small communities. But I also really think that we've got, I mean, we all know it, our families all talk to each other. These are rare diseases and they're out there saying how they feel and what their personal experience has been. Yeah, no, I think you're right. I mean, I think that really engaging with them is a great point and something that I know that then in what we do, we can think much, much more about that. Thank you again so much. This was really great. Both you and Greg, thank you both. And yeah, don't be intimidated by talking to a bunch of neurosurgeons. Big nerds are not intimidating people. Thank you all so much for having me. I really do appreciate it very, very much. So. Awesome. Thank you to our audience. Have a good one. Have a wonderful afternoon. Thank you. Thanks, Amy. Bye. So I think next, we'll just go ahead and move straight into our abstract presentation sessions. And I'll just introduce all four speakers right now and then we'll move through those talks sequentially. So we'll start with Dr. Malala and then move on to Dr. Gianpiccolo, Scherer, and then the final abstract presentation will be by Dr. Phan. And at the end of all of those presentations, we'll have a 10 minute discussion Q&A session. All right. So my presentation, my name is Arkham Malala. I'm a current PGY-4 at the UPMC Department of Neurosurgery. And I'll be discussing my presentation. Graph length does not impact the degree of revascularization with PLC and angiosis for Moyamoya disease in children. So no relevant disclosures. So just an overview of the talk today. Essentially, as I'm sure you all know, Moyamoya disease is a chronic progressive cerebrovascular disease that's characterized by progressive occlusion of the distal ICA and formation of collaterals. It's fairly rare, more common in Japan. And the most common scale that it's used is Suzuki grading. And there's multiple forms of surgical intervention. And the purpose really of this talk is to focus on PLC and angiosis. It's one of many different kinds of interventions possible for Moyamoya disease. But it's one that is practiced here, particularly at the Children's Hospital of Pittsburgh. So we've been able to study this. And so the key question is, does the length of the donor superficial temporal artery, does that affect the degree of revascularization as measured on postoperative angiogram? So we performed a retrospective cohort study of 19 patients who had a total of 27 PLC and angioses. Several of these patients had bilateral angioses who ranged from age two to 21. We collected, retrospectively collected this data and our primary outcome was Matsushima revascularization grade on one year post-op angiogram. So the cohort was a mean age of 10, but there was quite a bit of range between two and 17 years old. It was 70% female. The most common diagnosis was Moyamoya disease in about half the population was, as you can see, there were multiple other diagnoses represented. Then the primary presentation was ischemic event neurological deficit. But as you can see, there were multiple other forms of presentation as well. So the cohort range between Suzuki grades two and five with Suzuki grade four by far being the most common presenting grade or grade prior to surgery. So this is just a sort of overview of our operative and postoperative courses. Our blood loss was about 50 CCs. We ensure the patients remain well hydrated with reasoning about an average of, or median of two liters of fluid and 95% of patients were discharged to home or the prior setting that they were at. So the Matsushima grade outcomes were 90% A and 7% B in this cohort. So you can see here the intraoperative length, this was measured prior to the placement of synengiosis did not vary significantly between groups. There was a difference, but it was not significant on the Mann-Whitney testing suggesting that there may not be a difference between the lengths between these two groups. We observed something similar with angiographic length on postoperative imaging as well. So while this is a preliminary study and certainly larger studies are going to be required in the future to confirm this, this does suggest that the length of the donor artery may not affect the degree of revascularization. Now, and our primary interpretation of this is not that intraoperatively we should try for the largest possible STA, but the presence of a short STA should not preclude a patient from being a candidate for PL synengiosis. And so this raises the question, if not length, then what is causing the synengiosis? And we know from the literature that there's a huge milieu of local cytokines and other factors that can induce a vascular perforation. These can be coming directly from cerebral tissue. They can come from the dural tissue. There may be even factors in CSF. These include VEGF, platelet-derived growth factor and other key factors. And it's important to remember that even the technique like multiple boreholes, which there's no vascular connection at all, can be a successful treatment for Moyamoya disease, suggesting that really it's these vascular factors that are driving the revascularization that we see. This is definitely an area of further growth, and we're gonna continue investigating this area and collecting more patients in this series. And the key question that I think we want to keep in mind is does demand drive supply and not the other way around? And that's it. Oh, okay, I understand we're taking questions at the end, so. Okay, hello everyone. I'm Davide Giampiccolo speaking today of Professor Francesco Sala, and I'm the first author of this work, which has just been published in Child Nervous System. So I have no confidence of interest to disclose. So cerebellar mutism or cerebellar mutism syndrome is the most common complication of posterior fossal surgery in children. So this comprises a complex set of language, movement, attentional, emotional, and executive disorders. The central component of which is an initially profound, but usually, even if not always reversible speech disorder. So non-motor disorders of cerebellar mutism suggest a dependency to the cerebral cortex. It has therefore been linked to disruption of the cerebellodental thalamocortical pathway, which is the main efferent pathway connecting the cerebellum to the cerebral cortex. Under this perspective, a damage to the cerebellum reflects also damage to the cortex, and cerebellar mutism can therefore be considered disconnection syndrome. So in neuroimaging studies, it is well established that CMS may be caused by cerebellocortical disconnection. However, while imaging can document the occurrence of CMS, it can show where this track lie. Interoperative neurophysiology may provide online testing of this pathway through simulation. So we have, as you see here, recently reviewed studies on this topic, and we have to say that unfortunately, we cannot find any new monitoring technique to preserve this pathway interoperatively. I think what's surprising for us is that, interestingly, neurophysiological methods for identifying this pathway exist, but in the extraoperative setting. So this has mostly been done with TMS, transcranial magnetic stimulation, and it works with two coils, one simulating the cerebrum, and the other one simulating the cerebellum. So it works like this. Briefly, you have a cortical stimulation, so you stimulate directly the M1, and you record this potential at the level of end muscles as multiple potentials. And this is considered as a baseline piece, the one that you see here in the black square. Then you use a conditioning stimulation, which means that you stimulate the cerebellum a few milliseconds before, normally five to eight milliseconds before cortical stimulation. As you can see here in the box in blue, the MEPs evoked is much, much smaller than the one evoked during the test stimulus, so by stimulation only the cerebral cortex. And this short lattice has been linked specifically to the cerebellum cortical pathway, and is very, very adapt to the connection between the cerebellum and the cortex. But what is most important is that by doing this type of paradigm in patient that have been brain damaged in the cerebellum, in the dentate, in the muscle thalamus, and the superior cerebellar peduncle, so along the cerebellodental thalamocortical pathway, this effect is absent. So our plan was to adapt this type of stimulation to intraoperative monitoring in the posterior fossil children. So, and in this aim, we aim to translate this dual cerebellar cortical stimulation in intraoperative setting with the final aim at the end to prevent cerebellar amputation. So we use a conditioning test stimulus paradigm to see if it was possible to monitor this pathway inducing a variation of MEP amplitude. So, so far we have done it in 16 patients, six of which were pediatric. And it works like this. First, 20 stimuli as test stimulus are obtained from the end muscle, so contralateral to the corkscrew that you see here. And then the stimulus is combined with other 20 stimulus. So you can see the striplextra displaced on the cerebellum, and you do a combined stimulation with a short tissue. To conclude, you also do a testing piece. You repeat them to be sure that the inhibition or the facilitation you will find is related to the conditioning and not to some other aspects of the surgery. So then we acquired all the stimulation points in the cerebellum. We normalized to the MNI space, and we obtained group data. So this is our look in a patient with a pediatric patient with a douloblastoma. You can see how the points are acquired in your navigation. You can see now the changes. This is the test stimulus, so only cortical stimulation. And this is how it works when you give a cerebellar conditioning. So moving to the results, you can see our results on the left. And you can see that the sites for conditioning were really interesting for us because first I had the two different areas, one in anterior cerebellum, basically covering the lobule four to six, and another one in the posterior cerebellum covering the paramedial lobule. But what was interesting is that these were matching the expected somatotopies in animal models, but also in task-based fMRI in human here in the right area. The other thing is that the short latency that we found, as you see here in yellow, are consistent with a direct cerebral-cortical connection. So to sum up, cerebellar-cortical stimulation is feasible in both adult and pediatric patients intraoperatively. The short latency of modulation, so eight milliseconds, are coherent with a direct cerebellar-cortical connection, most likely representing cerebellar than totalomocortical pathways. Therefore, it is possible that cerebellar-cortical stimulation might be valuable for monitoring this pathway and possibly preventing the syndrome, but further research is needed to evaluate this in clinical practice. Thank you to our group. Hello, I'm Andrea Scherr. I'm a pediatric neurosurgeon at Nemours Children's Hospital in Orlando, Florida. I will be talking about most of the time, a big head is just a big head, the problem of macrocephaly. This study was born out of a collaboration with neurosurgery, pediatrics, and radiology in the Nemours system in Thomas Jefferson University. I have no financial disclosures. For our background, macrocephaly is a common condition seen by primary care doctors, affecting 2.3 percent of infants. We use the definition of a head circumference over two standard deviations above the mean or above the 97.7 percentile on World Health Organization growth charts. The differential diagnosis includes benign conditions, such as benign familial macrocrania and enlargement of the subarachnoid spaces of infancy as well as potentially dangerous conditions, such as hydrocephalus, subdurals, arachnoid cyst, brain tumor, Chiari, and metabolic disorders. Currently, there are no good evidence-based guidelines for primary care physicians to lean to. Imaging has been historically low yield for identifying neurosurgical conditions with the primary indication of macrocephaly. A prior study by Tucker showed only eight out of 538 imaging studies were positive for neurosurgical conditions. Our aim was therefore to add to the current evidence related to macrocephaly, and to identify risk factors and indications for imaging in order to develop a practical scoring system with high sensitivity and specificity to improve the yield of imaging and primary care for macrocephaly. Our method was a case control study. All imaging studies for the indication of macrocephaly in the Nemours system in children 24 months or younger from the time period of October 2011 to February 2020 were reviewed. 1,381 unique patients were identified. Out of those patients, 86 were deemed concerns after review by a senior radiologist and senior neurosurgeon who were blinded to the clinical history. These images were further reviewed by four fellowship-trained pediatric neurosurgeons, and exclusion criteria were then applied. Out of that 86 concern cases, 46 were then deemed concern cases meriting a neurosurgical referral, follow-up, or intervention. Out of these 46, only 15 were surgical. For our control cases, these were selected randomly from the remaining studies in a ratio of 4-1. The exclusions were applied and matched 3-1 by imaging date. This was performed by two independent blinded investigators. The number of controls was 138. The exclusion criteria applied to concern and control cases included those charts that had no head circumference measurements, if there was a history of perinatal intraventricular hemorrhage, and any indications for imaging apart from macrocephaly, including any neurologic abnormalities, seizure, nystagmus, history of TBI, brain surgery, or CNS infection. Prematurity was not an exclusion criterion. As for our results, out of the 46 concern cases, factors that were found to be associated with concern included the head circumference z-score, the head weight z-score difference, the head length z-score difference, history of prematurity, developmental delay, rapid acceleration of head growth, irritability, a bulging fontanelle. Not associated with concern was actual macrocephaly, age, delivery method, sex, feeding disturbance, parental macrocephaly, skull deformity, sleep disturbance, and year of imaging study. The diagnoses found included chronic subdural hematoma, possible hydrocephalus, Chiari malformation, and a possible benign brain tumor. We then attributed the following points to these signs and symptoms. Six was given to prematurity, four points to a bulging fontanelle, and two points for any delay. The head weight z-score difference was a raw number. Out of our 15 out of 46 surgical cases, 53% came through the emergency department rather than through the primary care referral system. These cases were associated with the following signs and symptoms. A bulging fontanelle, irritability, sleep disturbance, age, the head circumference z-score, the head weight z-score difference, and the head length z-score difference. Surgical cases were not associated with actual macrocephaly, a history of rapid acceleration of head circumference, skull deformities, mode of delivery, feeding disturbance, parental macrocephaly, prematurity, developmental delay, and sex. The diagnoses included chronic subdural hematoma, hydrocephalus, and possible benign brain tumor. The point scoring for the surgical cases were three for a bulging fontanelle, two for disturbed sleep, and then a raw numbers given to the head weight z-score difference. Here is a table of our sensitivity and specificities and thresholds for the concern cases on the left. And box and whisker plots showing the distribution of predictive scores for control cases on the left and concern cases on the right. Similarly, for our surgical cases, here are the thresholds for sensitivity and specificity for our surgical cases, as well as box and whisker plot showing the distribution of predictive scores for the control cases on the left and surgical cases on the right. So in our study, very few patients had concerns as defined by pediatric neurosurgeons, only 3.3%. Even fewer of these, only 1.1% required surgical intervention. Findings that were associated with being flagged as concern was developmental delay and prematurity. Findings that were associated with needing an operation, including a bulging fontanelle, sleep disturbance, and a larger head circumference z-score compared to a smaller weight z-score. Our aim was to identify a scoring system that would eliminate 27 to 28% of negative scans with 93 to 96% sensitivity, which could then lead to a electronic health record prompt that would be automated, providing thresholds with sensitivities and specificities in a tabular form so that clinicians could weigh the decision rule against their pretest probabilities. Our limitations included a lack of validation by a prospective study in other settings. Missing data was also problematic in that a number of growth charts from practices not associated with our EHR were missing. There was also a lack of parental head circumferences. There was also considerable variation among our pediatric neurosurgeons on what imaging was flagged as a concern warranting a neurosurgery referral, follow-up, or intervention. In conclusion, macrocephaly is a very common problem encountered in the primary care setting with few guidelines for management. Only a very small percentage of children under the age of two with macrocephaly have neurosurgical issues. A future direction for our study would be a prospective observation of a large primary care-based cohort. Thank you very much for your attention. Hello, everyone. I'm Zui. I am an MD-PhD student at Yale. I finished my PhD with Dr. Christopher Cawley. I'm now an MS3 going on clinical clerkship. I'll be applying to neurosurgery residency next year, so I look forward to meeting some of you and potentially even working in the OR with some of you. I'm very honored to be able to give this talk, Trim 71 Mutations Cause Congenital Hydrocephalus by Impairing Prenatal Neurostem Cell Regulation. I have no conflicts to disclose. So I'd like to start with this case. Basically, this is a case that was a Yale New Haven patient where she's basically a three-year-old girl coming in with macrocephaly developmental delay on brain MRI. We saw this massive hydrocephalus with startling CSF turbulence in the ventricles. So when we see this sort of imaging, the first thing we think about is basically treat the, we see a lot of fluid inside the ventricles, so we try to fix that. So obviously we put in a shunt. So here you can see the shunt catheter and the CSF turbulence is resolved, but the ventricles are still very big. There are no changes in the cortex and the patient remains developmentally delayed. So this is a really prime example of the problem in the hydrocephalus field where it's really believed to be a fluid disorders, but even when we treat the underlying CSF circulation defect the patients don't get better. And what I'm going to try to convince you by the end of this talk is that we need to also think about this disease as a developmental brain disorder and think about the vessel that's holding the fluid and not just the fluid itself. So again, the key point is that neurosurgical shunting can address some consequences of hydrocephalus, namely survival, maintaining survival, but it doesn't target the underlying developmental defect. So what I'm interested in understanding and also what the Kali lab is studying is what is the mechanism of human congenital hydrocephalus? So as you all know, there's primary hydrocephalus and there's acquired hydrocephalus from injury, infections, tumors. For my project, I am interested in the primary hydrocephalus asking, basically using human genetics to understand this. And so the questions that I ask are what are the genes that cause disease in humans? And then using mouse and stem cell models to understand how do these genes cause human disease? So the overall approach is we actually start with human patients. We sequence now 289 hydrocephalus patients and their parent trios. We come up with a list of potential candidate genes and then we combine this with bioinformatics to first learn where are these genes operative? When and where are these genes operative in the developing human brain? And so first, surprisingly, what we found was that these genes really converge in biological processes related to nervous system development, but not really enriched for processes related to CSF physiology. By looking at the expression of these genes over time, we see that they are expressed very early in human development, corresponding to periods of neocortical neurogenesis and preceding the functional maturation of CSF circulation components, which actually happened in late prenatal or after birth. So here we looked at what regions of the developing cortex are these genes converging in? And they are actually converging in these, in the ventricular wall of the neocortex. And it's important to remember that before birth, the ventricles are lined by germinal neuroprogenitors that then generate neurons that migrate to the prospective neocortex. Ependymal cells actually come after birth. So then we looked at what are the cell types, the relevant cell types using single cell. And we found that these genes are really converging in neuroepithelial cells, which is the earliest neurostem cell population of the human brain. So of all the genes that we found, TRMM71 was the most commonly mutated gene in the cohort. And what's also interesting about TRMM71 is that it is expressed the highest in the earliest neurostem cell populations of the human brain. And here's some neuroimaging of some of those TRMM71 patients. So basically these were missense mutations, de novo heterozygotes. And you can see already at birth, there's massive hydrocephalus. And here's some prenatal imaging demonstrating hydrocephalus at around 18 to 20 weeks of pregnancy. Here you can see already ventriculomegaly. And again, at this time point, there are actually no ependymal cells. It's all neuroprogenitors along the ventricular wall. This is also a time point of peak neurogenesis. So based on the fact that TRMM is the most commonly mutated gene and expressed the highest in neurostem cells, this was a gene that I decided to focus on. So here's some background on this gene. Basically, it was originally discovered in worms as a gene that regulates developmental timing. And in the mouse, it's expressed. Early in the nervous system, there is some conflicting and arguments about whether it may also be expressed in the postnatal ependymal cells. When the gene is knocked out, these mice have neurotube defects and some defects in neurogenesis, although this is really understudied and the mechanisms are still poorly understood. And particularly how it leads to hydrocephalus is not known. So for my project, we actually took the human mutation, put it in a mouse and try to see whether to look at the developmental defects. So first, we found that a subset of the heterozygous mice actually have communicating ventriculopathy at birth detected by MRI here. It gets worse three weeks later. And what you can appreciate is that at birth, it's essentially the intracranial volume is comparable to wild type, but with big ventricles. And then at three weeks later, the ventricles become huge and the intracranial volume also expands. So it basically looks like a situation where there's primary cortical hypoplasia that precedes the macrocephaly appearance later. When we cross the heads together to generate homozygotes, these mice all have neural tube defects, very similar to the knockouts. And the importance of this is that it tells us that the point mutation is really doing something very early in brain development at the level of the neural tube formation. Then I went and looked at what is the expression of TRMM71 in the developing mammalian brain. Using immunohistochemistry, I found that TRMM71 is expressed in E9.5 neural tube epithelial cells. So these are all, so at this stage in development, the neural tube only has neural stem cells, and this is where TRMM71 is expressed the highest. With the onset of neurogenesis, TRMM71 expression is gone. And notably, TRMM71 is not expressed in the postnatal ependymoma or choroid plexus. Based on those expression patterns, I use CRELOX technology to knock out TRMM71 in neural stem cells, and particularly in this line, TRMM71 is knocked out only in neural stem cells that are specified to become the dorsal neocortex, but not ependymal cells, choroid, or meninges. And in both of these, using both of these CRE lines, I observed hydrocephalus in both models. So this tells me that knocking out TRMM71 in the dorsal neocortex stem cells is sufficient to cause hydrocephalus. So this hints at a cortical developmental problem that leads to the ventriculomegaly. I then examined what is the functional impact of the TRMM mutation on neural stem cells development, and I found that in the mutant neural tubes, there is less neuroproliferation, but more neurodifferentiation. So it suggests that the stem cells are not able to divide as normally. And this results in basically a smaller neocortex at birth. So here I stain for various markers of the cortical layers, and you can see that the cortex is thinner. Now, the layers are formed in the right way, but they're basically just way fewer cells in each of these layers. So this supports that the mutations decrease cortical neurogenesis. And the physiologic impact of this decrease in cortical neurogenesis is that the brain becomes floppy. So I'm using here atomic force microscopy in collaboration with biophysics, where I take brain tissue and it basically indent into it and measure how much is the tissue responding. And I can calculate basically the stiffness and viscoelastic properties of the brain. And what these calculations are actually telling me is that these brains have pathologically high compliance and low stiffness. So that means that this decrease in neurogenesis makes the brain floppy, and it's more susceptible to deformation even when there's no primary initiating defect in CSF circulation. So this is a summary slide overall. There's a lot more data that I have not told you about, but basically the mutations are clustered in an RNA binding domain. This disrupts the ability of TRIM to regulate RNA targets important for neurodevelopment. This results in premature neurodifferentiation at the expense of proliferation. So basically less neurogenesis overall. In homozygotes, this causes a failed neurotube closure, but in the head mice and also in humans, there's reduced cortical thickness and stiffness leading to this floppy brain. And this we hypothesize is the reason for prenatal onset hydrocephalus in TRIM mutant patients. From a broad level, from a broader view, why is this work is important? So the overall approach is basically bedside to bench, where we're starting with human patients and doing basically gene discovery in these patients, understanding what are the genes causing disease. We can integrate this with a large scale data sets of the human brain transcriptome. That allows us to really understand where are these genes converging in brain development and using in vitro and in vivo models of disease, we can get a molecular pathophysiologic understanding of human hydrocephalus. This hopefully can lead to better molecular classifications of disease by genes and not just communicating versus obstructive. This could also be helpful for treatment stratification and hopefully molecularly targeted therapies without the need for neurosurgical shunting. And finally, this also informs mechanisms of human brain development. And so this is a question I always get asked is that then what is the link between hydrocephalus and microcephaly given that defective neurogenesis is seen in my model? So what I would like to show is that actually in the literature, there are genes that can cause both the classic hydrocephalus appearance. So small cortex, big ventricles and microcephaly, but also the classic microcephaly appearance where there are big ventricles and microcortex, but no expanding microcephaly that we'd expect to see and classic hydrocephalus. And here's a case where this is a gene that causes microcephaly and humans but when knocked out in the mouse, because is predominantly hydrocephalus. Basically what I'm proposing is that these two conditions may be linked, where the common conversion pathway is impaired neurogenesis micro and safely so small cortex in large ventricles but in. In some cases, this can lead to the progressive microcephaly the others it leads to the classic microcephaly appearance. This is sort of a summary overall basically I'm proposing that, from an even broader level, ventricle omega Lee could actually be a structural biomarker of altered brain development. I just my acknowledgement slides I think everyone in my lab, and the double a and s for this opportunity, my med school and the MD PhD program. I'd like to end on this slide with my contact information and this is just a sneak peek of all the techniques that I'm working on to characterize CSF circulation in my mouse model and to really understand how impaired neurogenesis affects these classic CSF circulation pathways. Thank you very much. Thank you to each of our four abstract presenters, thank you for all that you do the double a and s. And thank you for that amazing science you just presented to us. Please continue to enter your questions into our q amp a, that's what we're going to start with our questions. First, for Dr. Malala from Dr. Scott, how long a length of STA actually lay in contact with the brain in your patients. Over the years I had seen many failed operations with very small craniotomy and short STA links. Hence, our recommendation to get as long a length of STA as, as large a craniotomy as possible. Yeah, that's a really awesome point that's something that that I, for the sake of time didn't include but basically always make a very large, as large a craniotomy as possible to afford the maximum amount of contact. It's, if I had to estimate probably five or five to seven centimeters at minimum is lying on the cortex, and it raises the question, is there a minimum, and I don't know the answer to that, that's something that maybe there is a floor above what you need to get sufficient to make this work. So that's that's an area that we're that we need to look into further. Great. Next question is for Dr share. Can you use your parameters to evaluate macrocephaly perspectively. What are the next steps. That is a potential, you know goal in the future. So one of our collaborators on this study was is a pediatrician now resident work very hard on this project and definitely there's interest there and in a primary care based cohort study going forward. Right. Thank you. Niccolo. Let me ask you a question about at the end of your presentation you had mentioned the next steps for linking your findings to cellular mutism. What are those steps and how are you going to take those. I think that now we know that this technique is feasible also in the interoperative setting, the next step is to use this to monitor. So, I think, start to select just short latencies because in our study we selected them all. Just like those places trip on and theorists are abandoned at the start of the surgery, and see how these impact the post surgical outcome. I think that's also a very extensive pediatric neuropsychological assessment, both on speech and multiple emotional and executive aspects. Excellent. Thank you. Another question from our audience. Great discussion. Nice. And this one sorry is for Dr sweet. Great discussion nicer view of the molecular genetic issues related to brain and spinal cord development. How do you address where environmental factors make the situation more susceptible. Thanks. This is a great question. So obviously my work is really focused on the primary hydrocephalus. There's another arm of the lab that's focused on post hemorrhagic hydrocephalus and post infectious. However, in my mouse model there is actually incomplete penetrance, where it's not all the mice getting hydrocephalus. So, in terms of what are in those cases, why some lead to hydrocephalus and not others is really, it's not known right now. My hypothesis is that it's likely related to inflammation related things in the environment or there are also could be epigenetic factors, but I at this moment I do not have a satisfactory answer to your question, because I that will require quite a lot of work. And also, we remember that with post hemorrhagic hydrocephalus even that doesn't always so interventional hemorrhage doesn't always cause hydrocephalus in all patients, so it's likely that you have these basically rare point mutations that increase hydrocephalus plus hemorrhage or plus a little bit of something else, and that that lowers the threshold that basically this is the threshold for hydrocephalus, if those two things get combined together. You get bigger ventricles. Thank you. Dr. Cher, can you talk more about any type of validation studies that you may need to do. And then also perhaps about implementation, whether this would be a better tool for neurosurgeons or pediatricians. Yeah, so basically, looking at it from a primary care point of view with, you know, neurosurgeons is kind of like, you know, providing their recommendations. The goal is to use it more in a primary care setting. Of course, real world, what I find a lot of pediatrician referral any macrocephaly gets an automatic referral to neurosurgery and then any imaging is up to the neurosurgeon. So, you know, I have been using what are findings in my own clinic when I see macrocephaly to aid me in deciding whether or not to image. Great, thank you. Dr. Melilla, one more question for you, you mentioned at the end, perhaps demand is driving supply. How can we measure that. That's an interesting question. And I think that's the crux of why this is hard to study. So I think there's two sort of converging pathways you can look at this. So one is in our in our patients that undergo revascularization, the tumor world has been way ahead of us in terms of understanding the transcriptome and the proteome so we can do CSF sampling, blood sampling, and try to just characterize what is going on with the tumor versus not. So that's one arm, and I think it's inevitable that this is that it's there's limitations in terms of what we can understand and sample in our patients so then we have to have animal models of revascularization. I think that really gives you the potential to take tissue from, you know, from from animal models and do single cell RNA sequencing. And so I think there's a lot of potential of importing tools that have been used in other disciplines to improve this but it's a it's a really interesting field overall. Right, thank you. I just had one question for for Julie with regard to you know you mentioned that trim seven is down regulated by the time you get into the postnatal ependymoma and that one way to potentially offer therapy would be molecularly targeted. In terms of thinking about that is that something that you think of in sort of the fetal setting or would there be some kind of downstream intervention that you could consider. Yeah, so I think the first downstream consideration is that I didn't include this in my talk but we're basically finding that these defects in brain development, basically lead to obstructive hydrocephalus, eventually. So from a neurosurgical standpoint, you there is still a role for neurosurgical something and CSF diversion, because if you leave it alone, the mice actually die. So, there needs to be a consideration of some way to get the CSF out. And what we could propose is actually maybe thinking about ETV first before ventricular VP something. So, that's point number one. And from the fetal intervention standpoint. We're now really interested in drug screening platforms. So basically trim 71, it actually interacts with several other hydrocephalus risk genes. So this is the power of molecular genetics where you're not really just looking for specific genes, but also pathways, and then using bioinformatics and drug screening, predict and study how different even FDA approved drugs affect these large pathways and start the interventions early. And then third, I'm actually on OBGYN clerkship right now so this is really interesting to think about. So there's now a lot of prenatal diagnostic for whole exomes, even for whole exome sequencing. There's prenatal interventions, and now with gene therapy. We really think that for all these patients that are prenatally diagnosed this gene therapy can be delivered in the fetal setting, and to restore the genetic defect. Of course this will require much additional work to understand what are the, all the other genes doing what's the safety of these fetal interventions but I, at least for the patients I think this hopefully can lead to more molecular understanding of this complex disorder and not just communicating versus non communicating. Thank you. That's great. Thanks. Yeah, I think that was the first reference to the OBGYN clerkship in med school that we've had in the double a and s this year. Thank every one of you so much for your time for your hard work for contributing to the science of neurosurgery and for joining us in the pediatric session today, we're going to transition now into our next segment of our session. So, one of the fun formats that has been developed by the double a and s is something called clinical quandaries. And this is intended for us to be able to look at a particular case where it seems like everything is being done correctly throughout the whole case but yet the underlying problem is still very frustrating and difficult to solve. Many of us have all seen this type of case before. And the one that was chosen for today's case is the recalcitrant spinal cord syrinx. We have asked Dr. Lance govern alley who is the chief of pediatric neurosurgery at the University of Florida to join us and take us through this case. And then it should stimulate a lot of interesting discussion afterwards so thank you Lance for joining us. Thank you Rob very much for the invite and for inviting me to speak today. I thought long and hard about which case to present today I think I picked a doozy for conversation it's definitely one of the most challenging cases I've had over the past 10 years and your description of the theme for clinical quandaries I think fits this very well. So I have no disclosures. I met this child. She was already five years old. She's female she was born with a myelomeningocele. She had shunted hydrocephalus had already been status post multiple revisions her ventricles do enlarge with a malfunction in the past when she had malfunctioned during her sequence she had one interventricular hemorrhage requiring temporary external ventricular implantation. At three months old her prior neurosurgeon had performed a key or decompression that was extra durable. The occiput see one and see two bone was removed but the door was not opened. At that time, the indications for surgery were swallow issue strider vocal cord Paris's, and it was reported that she had a slow improvement over time after that surgery. Time zero so I'm going to this is a bit of a complex timeline so I'm going to list this as time zero and then do t plus one year t plus two years etc that's I'm going to try to bring you through this chronology. So when I met her at time zero she was not at her baseline she was having issues unfortunately which is of course not necessarily the time where you want to meet a complex patient. She was having a chronic cough. She had a swallow study that showed deep penetration within liquids, it was worse than her prior baseline ENT endoscopy showed left palate weakness right vocal cord Paris's cooling of secretions, and a sleep study showed central sleep at the end again these, this was not her baseline cranial imaging showed that her ventricles are small and at baseline she did not have any or prior shot malfunction signs or symptoms which were lethargy emesis and disconjugate phase again she did not have those. And again her ventricles enlarge with a prior malfunction and here they were small and at their baseline size. Not surprisingly, given her diagnosis of myelomeningocele she did have a significant key or two malformation. It was possible that she had some tonsil necrosis, shown here with my mouse right now. And then I'll show you in a moment she was found to have a whole record syrinx with a potential connection to the key or region. So here's her spine imaging at the time. You can clearly see that there's a large syrinx going through most of the spinal cord it's a little hard to tell here where the wall the spinal cord is it the wall, it's actually here so that is all syrinx and not extra spinal space. Not surprisingly, given her congenital diagnosis her spinal cord ends at the prior closure site. And then you can see on the T2 that she's got these, the T2 changes up top that connect the key or region to the syrinx so not necessarily a fluid cavity, but that kind of T2 edema pre syrinx state, potentially there. So we decided that we were going to go to the operating room for a redo key or decompression, we were going to do a duraplasty this time and try to open things up as best as we could and multiple levels. Not 100% surprisingly we found that tonsils were just inseparable from the brainstem, they were drawn down and fuse to the brainstem and also fuse to each other so there was no frame and agenda here, which obviously raises the suspicion for CSF dynamics, contributing to her problem. Given that she had this syrinx and I thought it was an imperative to connect her fourth ventricle to the subarachnoid space. I did use ultrasound guidance to dissect into her fourth ventricle through the tonsil or parenchyma. Her fourth ventricle was very small but I was able to do this again ultrasound guidance was very helpful. I placed a stent from a small pudens ventricular catheter silastic tube from the fourth ventricle to the cervical subarachnoid space. There wasn't too much difficulty in placing this again it wasn't the easiest catheter placement but certainly not the most difficult either and again it was ultrasound guidance. Very difficult postdoc course, her swallow was worse. She required a tracheostomy. She then developed ICU myopathy. On top of it all, imaging showed that the stent was in the intended position perhaps there was a little bit of T2 signal in one portion of the ventricle so we're pouring over every cut here but it really didn't look horrible at all it was in the fourth ventricle. We wound up taking her for a shot and exploration to completely rule out any potential contribution, given all the issues she was having. We found sluggish proximal flow but likely from her small ventricles and it was likely a working shot with the assessment there and again, preoperatively that's what we thought because their ventricles were small baseline and she didn't have a malfunction symptoms. She did eventually improve. It was not clear that she was better than pre op, obviously not very satisfying for for anybody involved at the, at the time. Okay, so now we're at t plus one year, so she's still not great clinically, some days are better than others. Her flow now worsens at the one year time point. Her syrinx, I'm about to show you now is larger. Given the prior experience nobody is eager to reoperate in the key area region, anytime soon. So here's her imaging at the time so what we've got here on the left is pre op from the key ID compression at time zero the image you've already seen. We've got t plus one year which is where we're currently at so post up from the key ID compression. Things are, despite all the duraplasty and moves things are still very tight in the QR region. You can't see the stand on this particular cup and on other cuts you can. She still has some of this pre Syrian state so certainly not the wide open beautiful region we'd like to have but again, given the prior issues. Nobody is excited about jumping back in there at the time. Her syrinx however is bigger. So this is the upper portion of her staring she does have this little septum between her two portions but they seem to increase and decrease as a unit so they do seem to communicate with each other as is typical with a serious with a septum. So, on the right we've got our newer image and the, even though I didn't measure the prior one you can see that the syrinx is larger. So, first, even though again ventricles are small and she didn't have a shot malfunction symptoms given all she had before we wanted to rule out Sean contribution as best as we possibly could so we took her for a formal exploration and found a working shot. You can you could argue that we could have just done a shot tap at the time but remember again last time we saw sluggish proximal flow, but man and metrics were good at that time so I didn't think that was going to be sufficient enough in this particular case. So we decided. This time, perhaps the syrinx was enlarging. Maybe there was a contribution from tethering to her syrinx enlarging. So, we, and even if that wasn't necessarily the cause. So we found a way to make this syrinx become smaller without again, jumping into the key area region that this can obviously be debated and potential point for discussion. So we, we performed a spinal cord de tethering, and at the time we placed a syringo subarachnoid stent into the lower part of the lumbar lumbar syrinx which was very easy because her parenchyma was extremely thin. So we decided to do a post op from this, she was quote excellent for six weeks for her parents. Then she started to wax and wane again, but she was still better than pre op. We decided, let's just make sure she's not having seizure activity given the waxing and waning that workup was negative. I think it was worth doing our post op imaging now which is that one year and three months so on the left, we have our pre op de tethering and our post op de tethering. And the main difference here is you can see that the syrinx that was present before is very much collapsed compared to prior you can't see the actual stent tube on this image but you can see some of the artifact from it. So that was encouraging. And at the same time point higher up the syrinx is clearly smaller, including the portion that's above the septum, and here's the higher version syrinx definitely smaller. And parents are content with how she's doing. So now again, I've got our post op to tether picture on the left, shifting it over I'm showing you now where we are t plus two years. The lower syrinx is even smaller than before here you can see the stent tube in the syrinx still in position. Then we're going to ascend now. The lower syrinx below the septum is is still smaller and where I want it to be. You can see however that the higher syrinx above that septum is seems a little larger. And here's the medullary syrinx which really wasn't present before we had our T2 signal change up there again which physiologically probably connected the two spaces but we did not have a syrinx cavity there before. Again, for the reasons I described before, shunt exploration shunt is working. Still nobody's eager to reoperate on the key or a decompression key or a region. So we decided, let's make that upper syrinx smaller since it's not necessarily fully communicating with the lower part anymore. So we decided to do an upper thoracic laminectomy for placement of a syringo-subarachnoid stent into the lower part of the cervical syrinx. That went very well interoperatively no issues. Now she has three stents, FYI. Her swallow post-op perhaps 10% improved compared with pre-op. So not an exciting change clinically but her parents are content with where she is right now. Post-op imaging shows the cervical syrinx is smaller. However, the medullary component is larger, paradoxically. So here's our pre-op stent on the left, post-op stent on the right. You can see our stent position here where I'm marking with the mouse. This syrinx is clearly smaller. I was hoping it was going to have a positive effect on the medullary syrinx, at least make it stable, but unfortunately, as you can see, the medullary syrinx is larger. We see her back again. She's stable clinically. The parents are content, given her history. Nobody's excited to continue to operate for radiographic changes when she's stable. So we decide to observe. We re-image at this point. I don't have this image for you, but the cervical syrinx is even smaller. The medullary component is stable. And long discussion, consensus for observation, she's been through a lot. She's in a decent place right now. Fast forward to T plus five years. She presents with two months of left facial nerve palsy. Her medullary syrinx is larger. So left image is where we were before, T plus two years, three months, now T plus five years. And you can see this medullary syrinx is larger. She's symptomatic again. Now she needs to go for an operation. It's decided that she undergoes a repeat Chiari at this point. We've tried the other things. She has a medullary syrinx fenestration with no stent or repeat duroplasty. Post-op her left facial nerve resolves. Her medullary syrinx is smaller. And then since then, which is two years, she's had no additional issues. Here's our pre-op fenestration on the left. And this is the five-year, one-month post-op image showing the medullary syrinx smaller on the right. So that's the case. Some potential discussion possibilities for the group. There have been suggestions for surgery over time. The surgical procedures performed. The order of the surgical procedures performed. And then the question, why did the medullary syrinx enlarge after the cervical syrinx stent where the cervical syrinx got smaller? Just some potentials for discussion. And thank you very much. Here's my contact information if you'd like to get a hold of me. And I hope this complicated case will stir some discussion. Great. Lance. I kept expecting that case to end and it kept going. Gave me the chills by the end. Yeah. I'm sure that those are always challenging. There were a lot of themes throughout your presentation and I appreciate that last slide you put together about discussion points. There's a lot of them. Can you just, I think one of the most interesting fundamentals just to go back to is the role of your shunt revisions in these cases. Can you talk about how and when you decided to revise the shunt and what role that had in this? Yeah. So given, I mean, in general, as we all know, she's a myelomeningocele patient. It's the shunt first. We need to rule it out. At the first surgery that I performed when I met her, I went by her history. She was having none of her usual shunt malfunction symptoms. Her ventricles were at baseline. So I decided not to do a formal shunt assessment at that time. As you remember, I did do a 12 days post-op when she was doing great and it confirmed what I thought, that she wasn't having issues, ventricles were small. But then given how difficult her post-op course was the first time, I said, I want to be, and I had mentioned this, I want to be 100% sure that this shunt is working. Now, again, like I was saying, you can argue that a shunt tap could have been sufficient, but she's got baseline small ventricles. We all see at times when we do a shunt revision, there might be slow proximal flow, but then we put a manometer up to it and it has great manometrics and it's just because she's got baseline small ventricles. I really didn't think I would have the level of clarity I wanted with her shunt function before diving in again for these repeat procedures over time. So that's why I decided to do that. I think that's a really interesting aspect of this too. I mean, it's a little bit of an aside, but I've only had a handful of kids where you have a shunt and a syrinx and you're trying to figure things out, but I've really, I can't recall one where I did the shunt exploration, which I invariably do as well. And I actually found a shunt malfunction. And I'd be really curious to know from the group how often people actually see that. I know it's really something that's been taught to everybody, but I think that's an interesting question and maybe worth looking at. I've definitely had hydrocephalus patients present with a syrinx and then I do a hydrocephalus surgery and the syrinx goes away, which obviously begs the corollary that a malfunction can do it. I think that's more common in what I've seen. What do you think, Rob? Yes, I think that is definitely more commonly what I've seen, but I can definitely remember some patients, especially spina bifida patients, where revising the shunt when the syrinx went up did lead to a decreasing in size or an improvement in symptoms. Lance, we've got a few questions over here in our Q&A. I'm going to read one of those off to you right now. This is from Dr. Duhigg. Did you look for an indolent infection? Occasionally you can find one of these recurrent cyst problems like this. Yeah, we did. That's an excellent point. She never had any clinical signs of infection otherwise and CSF sampling had been done at multiple time points here from the shunt because we obviously couldn't get it from the lumbar system. I don't know the details of where we got it from are skipping me a little bit, but we definitely ruled out CSF infection over time. She didn't have it. Yeah, good. And then one more question, we've got a couple more coming in from Dr. Bruce. Is it clear from the imaging that the first decompression did not result in an adequate CSF space posterior to the herniated vermis? Was the operative plan to place a bigger CSF graft to establish a CSF space? If not, why not? Yeah, so the operative plan was to make it look beautiful post-op, you know, to make it so that you could drive a truck through there. And we performed a, you know, we had a redo more bony opening. We performed a wide, the Y-shaped wide opening. Again, it was below the level of C1, the bone was already exposed there. And we performed a large duraplasty and it was not very exciting to see the post-op result where it looked like we hadn't done anything. She still looked tight back there. There are these subset of kids more in the QR1 population where you do the same operation every time and some of them just grow into the space and look like you didn't do anything, which is not satisfying at all. And she was one of these. So yeah, I wish that face would have looked great post-op. And then, you know, the T plus five years, again, another duraplasty. And it's not like you can drive a truck there either. And for full disclosure, that was another neurosurgeon because I have moved institutions in the prior, the interim there. So T plus five was another neurosurgeon who also did try to do that. And it looked like that beautiful post-op scan we all want to see. Yeah. I often wonder in these, when you have a shunt in that's draining so much, you have really small ventricles, whether that's contributing to preventing a big CSF space from developing back there if there's any over drainage, but who knows? Another question from Dr. Scott, Lance, a patent cervical subarachnoid space is key when the stents in the fourth ventricle are placed. Was she scarred from prior infection there? Note how badly deformed her brainstem is, stretched midbrain, really short tentorium, all bad prognostic signs. Yeah. So due to the expert guidance of my fellowship mentor, Dr. Scott, I made sure that the, every time I put one of these stents in, I mean, this is Dr. Scott's pudens ventricular stent, that the outlet of the stent is the placement of that is just as important as where you put it proximally. And it was in free subarachnoid space, I'm a hundred percent sure of that. And I made sure it was long enough if there were, you know, she got multiple of these, if there were adhesions, I made sure to get into a good cervical subarachnoid space and make sure visually that that stent went into it. You know, was there another septum even further down that I couldn't see? It's I can't rule it out. It's always possible, but definitely a consideration when placing the stents, plural. Yeah. I was thinking of names for your case. And at one point I thought three stents and a shunt would work pretty well after you mentioned the three stents. One question I had for you is, do you think there's a difference between putting in a syringo subarachnoid shunt and a syringo peritoneal or syringo plural shunt? Does that matter? Well, you can see that the cervical and the lumbar ones worked really well and they stayed working well. So I think, I think those are good enough. I try not to involve another body cavity if we don't need to, and the corollary of that being ETV versus BP shunt. And obviously they're easier to revise if you need to. And I think they are sufficient. Yeah. Yeah. Here's another question here for you from the Q&A. When you eventually re-explored the Chiari, did you find dorsal tethering to the dura or did the patient just need a larger duroplasty and fenestration of the syringobulbium? So again, as I, as I said, full disclosure that T plus five was after I had moved institutions. So that was another neurosurgeon. My eyes weren't on it. I am told that it was very scarred. The parenchyma was very scarred to the prior area because now this is the third time, well, the first surgery by the prior neurosurgeon wasn't intradural. This is the third time somebody was there and the parenchyma was scarred to the patch. And then there's a couple of comments that may lead to some commentary here in the Q&A. One is you were behind the eight ball from beginning as a patient had initial suboccipital decompression in addition to a cervical laminectomy. And I've tried to avoid suboccipital decompression in this population as can result in cerebellar slump, which can make the CSF blockage at the frame and magnet worse, but congratulations on eventual excellent anatomic results. Thank you. And then one more comment. I appreciate your decision to surgically explore the shunt. I think the new medullary cyst, you would have gone to, without the new medullary cyst, you would have gone to the redo Chiari. I've definitely had shunt malfunction that caused cervical syrinx enlargement. I'm curious what you think, what do you think led to the enlargement of the syringa vulva? You know, I think it's one of the more perplexing questions I've had over the past 10 years, that particular question for this patient. And I haven't forgotten about it. I really can't tell you. I mean, if we had Joe Madsen here, he could put up some graphs in pulsatility, but I am not sure why that medullary syrinx got larger. You could argue that it just got larger because she really needed to know the Chiari decompression that wasn't done. However, it was stable post-op for three months. You could potentially argue that I altered the cranial versus spinal gradient of things by putting another stent in, even though I didn't go to another cavity and that worsened the physiology of the foramen magnum, but that's as best as I've got. That is a tough case, and I know it's one that many of us have encountered in one variety, one flavor or another. Thank you for going through this with us, Lance. This has been a great case and gives us a lot to think and talk about. It was a good clinical quandary. Sounds good. Thank you for the invite. We're going to move on. Yeah. Thank you so much. We're going to move on to our next. Thanks, Lance. All right. Yeah, so this is a session that I've been looking forward to that I think is really interesting, and I apologize for my connectivity problems that relate to my video, but this is two presentations that are going to involve and more influential in the pediatric neurosurgery literature over the years. The first by Dr. Bookland, and then the second one by Dr. Marr, and then we'll follow that up with panel discussion of both of those presentations. Good afternoon. I'm Marcus Bookland, and I want to thank everyone for the opportunity to present some of the work that has come out of our lab at the University of Connecticut and Connecticut Children's. I am humbled to consider that it was influential. I will tell you it was an enlightening study for us. The paper that I was asked to talk about was this one from 2020 that came out in the JNS-PEDS on the association between peripherally circulating microRNA expression levels in pediatric brain tumor patients and the microRNA expression levels within matched tumor tissues. I have no disclosures to report for this particular case. Just to give everyone a little bit of background on this particular topic, I'll just remind everyone that microRNA are very short segment RNA molecules that are used by all cells to help regulate intracellular processes. They can alter gene expression at the translational and the post-translational level to affect the different protein expression levels. Tumor cells happen to be particularly adept at utilizing microRNA to alter intracellular pathways, particularly to promote cell growth, as well as inhibit pro-apoptotic pathways and enhance migratory and invasive pathways. They do this not only within their individual cells, they will also package up these microRNA exosomes and other microvesicles and secrete them into the interstitium surrounding them. Several studies have shown that these microvesicles bearing microRNA may actually alter the microenvironment of tumors. Many other researchers over the past five to 10 years have looked to try to identify these microRNA in peripheral biofluids such as saliva, urine, and blood as a means of potentially developing diagnostic tests for tumors, since the microRNA profiles that these cells produce are often quite unique. Our own lab back in 2018 identified a set of upregulated microRNA in pediatric brain tumor patients who had juvenile palliative gastrocytomas that were highly predictive of the presence of a juvenile palliative gastrocytoma when compared to controls. In that same group, we also showed that resection of those tumors led to a normalization of those microRNA profiles within the patient's bloodstream to normal levels, and at least in one case within that cohort that we evaluated, we had a patient who did not show a complete resolution in their microRNA profile to normal levels, but was subsequently found to have residual tumor on follow-up imaging, indicating that peripherally circulating microRNA might be used both as a diagnostic tool to screen for tumors in pediatric brain tumor patients, but also to look for recurrence or treatment response down the line. However, there's a question that vexes many of us working with microRNA as potential biomarkers, and that is, what is the source of that microRNA that we're studying? This is important information to have if you're considering using this tool as a full liquid biopsy diagnostic, something that imparts both taxonomic information as well as molecular information. That was the question that we looked to answer in this paper being presented here today. To try to glean some insight into what was going on in our pediatric brain tumor patients, we took a cohort of 10 patients, three JPAs, four group 3 medulloblastomas, and three group A ependymomas, all posterior fossa tumors. And we enrolled patients in the study such that each one had a blood sample taken within six hours of their index surgery. Then we also sequestered a sample of their tumor tissue to the lab as well for analysis. Paired samples underwent extraction of their microRNA using an 84-microRNA panel consisting of microRNA that were known to be often aberrantly expressed in brain tumors. We normalized all samples to normal tissue controls. In the case of the tumors, we use cerebellar tissue. In the case of the serum samples, we used HMAT serum controls. We also used a synthetic microRNA spiking control to help deal with any differences in microRNA detection that might be engendered from handling of the serum samples. Our tissue microRNA data showed a very strong taxonomy or categorization of microRNA profiles around the histopathologic diagnosis of the tissue. This is despite the fact that the overall cohort is relatively small n, but this is reported extensively in the literature that microRNA tend to be highly cell-specific and are expressed in a very cell-type specific manner. Looking at our serum microRNA samples, we did see some of the same family of microRNA elevated in our juvenile phallocytic gastrocytoma patients as we had in our previous paper. We did not find any significant alterations in microRNA expression profiles in the serum of metroblastoma patients. However, interestingly, we did see an elevation in miR-93, miR-137, and miR-138 in our endomomas. miR-93 in particular has now been reported in the past and may, if larger studies support this finding, be a potential biomarker for group A ependomomas in the future. Certainly, miR-93 is a microRNA that is known to associate with chromosome 7, which is often duplicated and contains a variety of aberrantly expressed genes known to associate with ependomomas. The crux of the paper, though, was whether or not we could find an association between the microRNA expression in the serum of these pediatric lymphoma patients and the tumors that were matched to them. In the end, we found no such co-clustering or correlation between serum microRNA and tissue microRNA. That is to say, even though there were some significant and, at least for the JPAs, reproducible patterns of microRNA expression within the bloodstream of those patients, the microRNA that were elevated in the serum did not correlate with the degree of elevation or suppression in the tumor tissues themselves. The results of the study were ultimately negative, but informative. We had to say that, in the end, we cannot correlate these aberrations in microRNA expression in the serum with the tumors that they're associated with. This may be because the study itself was too small, but it may also just be that the microRNA released by these tumors into their interstitium did not cross the blood-brain barrier in significant volume. It may be that normal patient-derived microRNA are drowning out the microRNA that may or may not be reaching the peripheral circulation. Alternate compartments, such as the CSF, may need better biofluids for detecting tumor-derived microRNA in these patients. When you look at other tumors that have been studied in this manner, the best reservoirs are often biofluids that are in direct contact with the tumors of interest, such as urine for bladder cancers or saliva for salivary cancers. We do not believe this research completely debunks the notion of using microRNA as a diagnostic, but it does temper the interpretation of any such data derived from serum-based microRNAs. We don't feel that our current work supports the notion that this data is reflective of molecular changes directly from the tumor, but maybe more post-derived changes. I want to thank my colleagues who have supported this work. Dr. Komikova at the University of Connecticut, Dr. Song at Hartford Hospital, and all my wonderful collaborators in neurosurgery and oncology in the research department at Connecticut Children's. Thank you again for the opportunity to share this work with the AANS and the neurosurgical community. Hi. My name is Cormac Marr. It's good to be with you all this afternoon. I want to thank the AANS, and in particular, I want to thank Rob and Todd for allowing me to speak today and for the kind invitation. I don't have any disclosures for this work. I'll talk today briefly over the next 10 minutes about a paper we wrote on ventricular volumetrics. We've been interested in volumetrics in other areas, particularly Chiari and the spine now for many years, but this was really our first foray into ventricular volumetrics. I think for this audience, it's not controversial to say that ventricular volume impacts our daily practice every day. Just to provide one example, it's a patient of mine seen here on the left side of the screen with a essentially normal brain MRI scan, enchanted hydrocephalus, presented to an outside emergency department, and we received a phone call that there was mild ventricular dilation. I think you can appreciate that the ventricles, in fact, are quite significantly dilated compared to the baseline. Of course, we didn't have the new scan and the outside hospital didn't have the old scan. This, I think, was quite limiting. I'm sure we've all seen examples like this on our own practices where we're using very imprecise terms such as mild ventricular dilation or moderate ventricular dilation instead of defining it more precisely as we could. Here's the other side of that coin. That's two completely normal children without hydrocephalus. But you can see that the ventricular volumes are very different in these children, both normal certainly, but with very different ventricular volumes. Again, can we be more precise? Instead of saying mild ventricular megaly, can we measure? The answer is, yes, we can. Here's one example of some 3D volumetrics that we've done in the ventricles. Again, spend the next few minutes discussing with you how we do that. The computer program that we use essentially uses K-means clustering. One of our residents, Siri Kalsa, wrote a really nice program for us that uses these K-means clustering. Essentially, it's a way to automatically cluster data points, making signal observations on every slice of the scan and then partitioning these clusters into the nearest mean. Then this is done automatically by the computer without any user input at all. I'll show you what that looks like in reality. Here's, again, just a typical MRI scan. We run it through the computer program that identifies different clusters of signal. In the MRI scan, this is done automatically. Then the semi-automatic part comes next, where we simply have to click on what we're identifying as the ventricle. This is an important step. It does make it semi-automatic rather than automatic, but it is an important step, we think, for quality control in the sense that the computer, even with a very good program, can occasionally identify similar signal outside of the ventricle, such as areas of strokes, arachnoid cysts, and so on. Sometimes we'll exclude small CSF spaces that are in the ventricular system, such as fourth ventricles and so on, that don't appear to be connected onto the images. That's because the segmentation with the computer program does automatically attempt to join on these two-dimensional slices, the adjacent CSF spaces, and turn them into a three-dimensional volume. But again, when the ventricles get very, very small, it does require, at least in our program, some human quality control to make sure that that's been done properly. We use this program essentially in two ways. Number 1, to try to define what normal is, because we can't go any further if we don't understand and more properly define what normal is. Then number 2, or practically, for us as pediatric neurosurgeons, is trying to define abnormal, especially with respect to Duchenne failure. The first step, again, defining normal. Our methods were very simple. We just simply used our MRI database. We limited our analysis to children, ages newborn to 18 years. We tried to deliberately exclude conditions that we thought would lead to abnormalities of the brain volume or ventricular volume. For instance, we excluded people with brain neoplasms. We excluded known diagnoses of hydrocephalus for this stage of the analysis. We excluded those with congenital brain malformations, cerebral cysts, and so on. We ended up with 26,000 patients. Three and a half thousand of these had a brain MRI that was usable. We decided to pick the most recent 50 brain MRI scans for every age cohort that we were interested in studying. That was four different age cohorts for the 0-1 age range because a lot changes in the first year of life. Two different age cohorts for ages 1-3, and then simple annual cohorts for ages 3-18. Here's what we found. Here's essentially the data for the normal ventricular volume part of this analysis. It's a growth curve, which you can see, not surprisingly, the ventricular volumes increased dramatically in the first year of life, and then generally increase from ages 0-18, as we would expect based on head circumference and so on. Interesting thing, which I think was unexpected, at least for me, is that we found this little dip between ages 1-3, which I'm highlighting here for you. This was unexpected. In general, growth curves increase throughout life, but the ventricular volume did seem to decrease. Again, with the large number of patients that we studied, we felt pretty secure about this finding. We struggled to explain that. I'll confess that our data can't really explain it, but I think this paper really does. This is a really excellent paper published a couple of years ago by Drs. Peterson, Worf, and Steve Schiff at Penn State. They studied normal human brain volume growth over the first 18 years of life, and you can see their growth curve here. When you combine what these doctors found with respect to brain volume growth, and you look at what we know about, say, head size growth, you're looking at head circumference, and you put the two together, what you see is that over the age 1-3 interval, when we notice the decrease in ventricular volume, you see a less than seven percent increase in head circumference, but a really dramatic increase in brain volume. Again, we can't say for sure this is the reason for this small dip in ventricular volume between ages 1-3, but I think it is potentially explainable here because for those two years, brain growth does seem to outpace the relatively slower growth of head circumference. When you look at the next couple of years of growth, you see that in general, ventricular volume is holding pretty steady between ages 3-5 before increasing again. Again, I think this is explainable looking at the paper from Drs. Peterson, Worf and Schiff, they saw a five percent increase in brain volume over that time versus a two percent increase in head circumference over that time according to WHO data. Again, not much difference between those two, and I think that goes along with the fact that the ventricular volume tends to stabilize over those years and then gradually increases again. When you look at those curves that I showed you, where we tried to define percentiles, essentially making normal curves, normative curves for ventricular size across all pediatric age groups, you can see that we really did have quite a tight fit of our growth curves to the data. You see just QQ plots here with a very strong correlation coefficient. There were about 1,400 patients in that study, so we felt good about those correlation coefficients that the data were able to generate. That's the normal part of the study. Objective number 2, let's calculate volume change thresholds that detects shunt failure. Back to my case example, that would be really helpful if the outside hospital, instead of saying mild ventricular megaly, could have given me a number. What we chose to do is take 100 of our own patients that were known to have a well scan and a failure scan confirmed in the operating room, compared that to a similar sized group of patients that had well scan to well scan comparisons with no history of shunt failure or shunt surgery. Not surprising when we found a tremendous difference in ventricular volume change between these two groups, essentially no change in the well to well comparisons and a really significant increase of ventricular volume in well to failure. So again, not terribly surprising and perhaps not even useful in and of itself. I think potentially the more useful finding though was trying to define a threshold for volume change where shunt failure became very predictable. And so we looked at area under the curve trends for different thresholds. And what we arrived at is between about 10 and 14 and a half CCs of increased ventricular volume. We had a very high sensitivity for shunt failure, 98%, and a very high specificity, 95 and a half percent. So that was a relatively small group of about 200 patients who are in the process of validating that now on a new group, a much larger group of new patients. So again, back to this case example, the recent visit with the so-called mild ventricular dilation. Here's how that looks when it's plotted out on a semi-automatic volume graph. And you can see before the shunt placement, there were about 100 CCs of CSF in the head, went down to less than 10 CCs on numerous well child visits. And when we got the phone call from the outside hospital, you can see the ventricular volume was now over 70 CCs. So I would suggest that rather than mild ventricular megaly and almost seven times increase in ventricular volume would be a good way of properly diagnosing shunt failure in a more quantitative way. So just to quickly summarize, I do believe that a rigorous quantitation technique is lacking in our assessment of ventricular volume. And the ventricular volume is really important in our day-to-day practices. So we tried to develop and validate a 3D computer-aided ventricular volume calculator, first to establish normal growth curves for ventricular volume, and next to aid with the diagnosis of shunt failure. I've shown you the curves, as well as the suggested threshold based on our first analysis of between 10 and 14 CCs being very sensitive and specific. There are a number of limitations to this work. Anytime you're making normal curves, you need a very large sample size. We had about 1,300 patients, which I think was sufficient for a start, but I would like to see this work validated in much larger studies in the future. This is a pediatric group. It doesn't speak to adults. The semi-automatic measurement is a limitation in the sense that I think it would be much easier for referral hospitals, outside hospitals, to have a computer program that would generate this number without user input. As of right now, I don't believe such a program exists in a clinically satisfactory way. So we've used the semi-automated technique because I think it's simply more reliable. And then most important limitation, these data have to be interpreted in a clinical context. And I think in general, the usefulness of this, if it is useful, will not be for experienced pediatric neurosurgeons. It will be for non-neurosurgeons, perhaps in more remote locations, trying to communicate the results of scans to experts. Defining a cut point for abnormal, I think is always risky. And again, we need more and more data to do this properly. And we need to take into account the multiple clinical issues that can come up in the assessment of these patients, cortical atrophy, syndromes, big heads, little heads, developmental assessment, and most importantly, the physical examination of the child. Finally, I'd like to thank my collaborators on this study. Siri Kalsa is one of our chief residents at Michigan, and he's the one who wrote this computer program and did so much of the work for this study, as well as many of the other morphometric studies that we've been doing. And also a medical student, Noah Cutler, who was very, very instrumental with gathering the data. So again, thank you to the meeting organizers for allowing me the opportunity to share the data with you today. Dr. Bookland and Dr. Maher, thank you guys so much. Those were both fundamental papers that we all experienced in the last year, and we appreciate you coming to us and letting us hear from the horse's mouth exactly what you did. So thank you so much. My first question is for Dr. Bookland. At the end, I heard you go through some different explanations of why your study may have been negative. I am curious about whether or not you think that maybe certain tumor types that serum microRNA might be detectable, or there's certain tumor types you think that might release higher concentrations. I certainly think that there is probably an association between the amount of surface area in contact with the parenchyma and the amount of microRNA that would then be accessible to the circulation. We have seen in our work higher concentrations of microRNA overall in brain tumors that had a larger amount of surface area in contact with cerebral parenchyma versus ones that were mainly just hanging in the CSF had little to no elevation in overall microRNA concentrations in their serum. I'm not sure if we know enough about these tumors to say that one particular tumor may release more microRNA into its microenvironment than others. Although certainly when it comes down to ependymomas, looking at cell-free DNA, they don't seem to release a whole lot of that. Assume that perhaps microRNA would be the same since they're all bundled together often in microvesicles. Thank you. We do have one question in our Q&A. It's from Dr. Bruce. This is for you, Dr. Meier. Did the outside hospital comment on the periventricular edema? Was there a sign that it's a sign of shunt malfunction? And will this program be easily exportable to the community? Answer to the first question is no. There definitely was some periventricular edema, which I think, again, for all of us is easy to recognize, but that wasn't commented on. And I think we've all seen cases in our practice where that's been the case. Also edema along the shunt catheter and so on can be easily missed. So again, I think for an experienced pediatric neurosurgeon, these are frequently easy calls to make, but that's not necessarily the target group for this. The second question is, I would love to think that it's easily exportable. We're working with another hospital right now in the West Coast as part of our attempt to validate this, and we'll see how successful they are with this. I hope it is. And that remains to be seen. Dr. Meier, how do you think this becomes implemented? Is it partnering with industry, partnering with radiology? How does this become something that can be universally used? Yeah, well, I think in terms of that computer program, I think inevitably it will be partnering with industry and radiology. To me, it's not so much about this specific program as much as the data itself. So what is the volume threshold that's relevant? I think that's something that will be hopefully a true thing regardless of what program people are using and whether people are using this program or some vastly different, perhaps superior program. But I think defining normal volumes will be a true thing no matter what you're using to measure it, as long as the program is accurate. And the same thing with defining thresholds for failure, that should be a true thing no matter how you're measuring it. Thank you. Dr. Bookland, I just had one quick question. It's a little bit of a sort of side question, but do you know of any work where people have tried as a proof of principle to look at like pituitary tumors or using blood-brain barrier disruption to see if they can get more signal to sort of give a sign that that's actually where the problem is in terms of identifying microRNA in the serum? You know, that's a great idea. I am not aware of anyone actually trying to do that, but it would be a fascinating way to try to increase the sensitivity of peripherally circulating biomarkers. One of the challenges always with these biomarkers is that you're trying to find something that's easy to implement. You want to use it as a screening tool in lieu of the things that we already have. So the more complexity you add to it, the less likely it is going to be actionable in the real world. But those sorts of early studies, even if they are more difficult, may be the gateway to developing those diagnostics down the line. Okay. Yeah, and Dr. Mara, one question I had in terms of another use potentially for the software is in the shunted patient with headaches who you're evaluating for potentially a slit-like ventricle syndrome or something like that, can you envision sort of applications in that population as well? Yeah, absolutely. I think that would be really relevant. I think, yeah, I think that would be a really good use for this. The other thing that we're exploring right now is normal pressure hypercephalus in a different age population, obviously, so not maybe as much of a concern to this group, but a very important issue for our adult colleagues. And I think that if we can, again, define what normal is in that older population and then try to set perhaps a diagnostic threshold, I think that would be really interesting. Well, thank you to both of you guys for joining us and sharing your science with us. We will move on to the next part of our presentation today. Thank you. Thank you. Well, the next part of our session is another one of those interesting formats that the AANS has developed for us. This one is called Following the Patient Diagnostic Challenge. The idea behind this is that many of our patients don't live just in a single episode with a neurosurgeon. They pass through a complex process in the hospital, encountering multiple different services. And we wanted to go through that experience with a common diagnosis that we might see in pediatric neurosurgery. So we're lucky today to have Dr. Tina Duhaime join us. She served as the Director of Pediatric Neurosurgery at Mass General Hospital until this year. She is the Nicholas Zervas Distinguished Professor of Neurosurgery at Harvard Medical School. She is an expert in abusive head trauma. And we have asked her to tell us about the journey of the acute presentation of suspected abusive head trauma. Importantly, during her presentation, she will be taking advantage of some of the technology we have in this platform. So there will be poll questions that she asks and over on the right side of your panel, you're about to answer those questions and interact with the presentation throughout it. So Dr. Duhaime, thank you so much for joining us today. I'll turn it over to you. All right. Okay, thanks very much. We've had a little bit of technical difficulties here and I'm not sure I'm gonna be able to see the results because currently I don't see the chat, which is how I'm gonna get the results. So we'll do our best. But we're gonna go through quickly three sort of classic cases from beginning to end. And we're gonna rush through this so that we have time for your votes on these. So let's start with case one. This is a 23-year-old that's at an outside hospital. You get a call from their ED. The story in quotations that they tell you is the baby was dropped in an infant carrier down some outside steps. By history, the baby was told to the ED doc. She cried immediately. They got a CAT scan there and they tell you that she has a small subdural. They see no skull fracture and the baby looks a little irritable. They wanna send it to you. So you say, yes, the baby arrives about four hours after the injury. The CAT scan that was done at the outside hospital was done about one hour after injury. You see the baby. The baby has a boggy bump on the head. The fontanelle is flat. Baby looks pretty alert. It's a 23-day-old, so it doesn't have much repertoire. But the baby cries when you handle her and she moves all around. So this is your scan that you have from the outside hospital. It's an outside scan done an hour after the fall. And then the question is, what are your next steps? Here are your options that you may have others, but these are the options you're gonna be given for this particular presentation. So are you gonna observe in the ED and call the child protection team? Are you gonna repeat the CAT scan or maybe a tailored MRI, rapid MRI of some kind? Are you gonna admit to the hospital, order a skeletal survey and a retinal exam? Or are you gonna consider surgery with maybe a burr hole over that clot? 100%? Okay. I'm hearing my results from our technical support here, so thank you folks for your help. Oh, actually, I can see it up in there. It's a little small. Okay, what's the biggest one? Okay. All right. Have we given it enough time? Are we stable now? Okay. The way we'll do this is that I'll tell you what actually happened, and then we'll give you some rationale, but obviously there are many ways that we can approach these. So this baby got a repeat image because the first one was done quite quickly. By the time we got the image, it was now about six hours post-op, and the reason we got this image is that I suspected this was not a subdural. It was a pretty good story for a skull fracture with a venous epidural, and it's very common in my experience that outside hospitals and sometimes even my hospital misreads venous epidurals as subdurals and starts the whole chain of events that leads to a child protection consult. This is a rapid MRI, and you can see, I don't know if my cursor shows, but you can see that there is preserved subarachnoid space that confirms that this is a venous epidural. The reason it fakes people out is that it can be crescent-shaped, and in particular, if your CAT scan is done in the axial plane and they don't necessarily do reconstructions or there's motion artifact, they will miss this kind of a skull fracture, which is what this child had. It's absolutely in the plane of the way they took the CT images, which is why they couldn't see it on the CAT scan. This baby had a good story for a fall out the stroller onto concrete with a linear skull fracture with a venous epidural that was misidentified as a subdural. This baby got a skull film to confirm the fracture. You could have maybe done a CT as a repeat instead of your MRI, but we try to reduce our radiation. The baby got a social work consult like all children do in this age group. There were no red flags, so this baby was observed overnight and then discharged, and this was not considered necessary to escalate to child protection. In our hospital, when the social workers do consults, they talk to the child protection team who looks at the images, talks to us, and decides if they want to escalate from there. Now, I will say that this child had already been reported to child protection at the outside hospital. We have to always work hard to intervene on those cases. Next question for the audience is what follow-up do you order? Number one, follow-up with you and or the pediatrician. Don't do any imaging. Number two, office visit with a rapid MRI. Number three, office visit with a CAT scan or a skull film to look for a growing skull fracture. Number four, follow-up with child protection with a skeletal survey at two weeks. Those are your choices. There may be others. All right, well that looks like stability. So most people, well there's a few. Most people are saying, follow the baby up, no imaging is needed, but a few people want to do follow up imaging, oh it's still changing on us, okay. And more people want a rapid MRI than a CAT scan, it's dynamic. Okay, so I'll tell you what was done, I'm not necessarily saying it's the right thing. For a child like this, we have kids just follow with us or the pediatrician, usually us because the pediatricians are not typically very comfortable seeing kids with skull fractures and bleeds, they get nervous about it, but the point to be made is you could certainly get another study. We actually did get another study in-house because the reading for the second study that was done in the ED was that the clot was slightly bigger, so before the baby went home we got a third one which is probably overkill, so we didn't do another one, and I'm not sure I would have with a well-appearing baby. But the reason that we didn't get any kind of imaging with radiology to look for growing skull fracture is that growing skull fractures virtually never happen from low height falls with an MRI that shows no contusion, no injury underneath. Those typically require a higher set of forces that tear the dura or at least tear the arachnoid, and this is not the context where you have to worry about a growing fracture, and so I mean maybe somebody out there has seen an exception to that rule, but we tell pediatricians and parents that they don't have to worry about that particular complication in this context. Okay, we're going to go to this slide which just is to say that there is data that for the exact same injury, exact same history, exact same risk factors, kids from minority backgrounds have three times the likelihood of being reported to child protection than non-minority kids, and this kid was a minority kid and had a child protection referral made at the outside hospital because as soon as they hear the word subdural, you know, those balls get rolling, and it's understandable that pediatricians don't want to miss a child abuse, but for accidental injuries, the rate is three times higher in minority kids for the same kind of injury, so just keep that in mind, and this paper luckily is over a decade old, and I'm hoping that things change over time. Okay, next case. This is a two-month-old with several outpatient episodes of twitching. It was thought maybe to be reflux, now went to the outside hospital for the third time with the same symptoms, and they decided maybe the baby had a seizure and got a CAT scan that showed a subdural, so I'm just going to show a quick video of this baby. This video was actually taken by the parents, and in a minute you'll see you'll get a neuro exam. So that's the baby. That's what the baby looks like, and what I'm going to tell you is that that set of symptoms was missed in several visits to the pediatrician and to the local hospital. So this baby was having seizures and got treated for the seizures and came to our PICU as a direct transfer. The baby was intubated because the seizure medication was sedating. The eyes, and by the time we saw the baby, that should have worn off so we could get a pretty good exam, and the point I want to make is that the eyes opened, but the baby did not cry to trapezius pinch. The fontanelle was full. It wasn't absolutely tense, but it was clearly above the plane of the skull, and this was the scan with this blood here and blood interhemispheric and a tiny bit of blood on some of the upper cuts over the hemisphere, but most of it was interhemispheric. So questions. What are your next steps? Do you do an MRI? You've got a baby whose airway is controlled. They're treating the seizures. They're talking about long-term monitoring and so forth. Do you do an MRI? Do you put in an ICP monitor and manage the baby medically? Do you do a fontanelle tab? Do you go to the OR for subdural drains? You've got a baby who's in status, really. Do you do a craniotomy for subdural? So those are your choices. Okay, so I'm not seeing the slide choices. Slide didn't advance. Well, those are my, yeah, I didn't advance the slide. Did we not get the Slido on this one? Oh, got it. And what should I do? Just keep going? Oh, there we go. Okay. Got it. Got it. Yeah, I got a pretty good view of that. Sorry to the audience. If you hear me talking and it isn't to you, I'm talking to the wonderful tech support who are helping us out here. Okay, I think we'll hold that one so maybe a few things will change but most people want an MRI as their next step. I will tell you that what we did in this baby, it was a baby in status with a full font now, and blood, and I just wanted to reduce the pressure so I did a font male tap at the bedside. And you could have done other things too we were going to get an MRI but not as the next step. The first thing I wanted to do was reduce any kind of intracranial pressure, because the baby was in status. So if I know tap is, you know, generally easy to do this baby had a nice big font now it was a simple thing to do and you could withdraw some fluid and, you know, see if that made a difference in the status because I was trying, we're trying to approach multiple directions. So I'm only putting this up here to emphasize nobody uses the scale I use it but nobody else uses it. Maybe Susan Durham uses it she, she helped write it, but the one point to be made by this scale which is another infant coma scale, is that when you have a baby, even if the eyes are open, particularly these little babies in the first few months of life. And you pinch the baby hard, and you get no change in facial expression that's that's a sick baby. And I have found this particular thing which I learned from from epilepsy neonatal epilepsy specialists I didn't, I didn't come up with this myself. But it's extremely useful to assess babies where there may be inflicted injury and they, they, they can look like that baby did they can be moving they can be making funny noises they look like they're looking around that baby was in status and was sick. And that's why we tapped the font now there was no facial expression to trapezius pinch, and that baby was not on payments at that time. So I wanted to get the pressure off as fast as we could and that's why that baby got a font now tap now, I don't have class one data to tell you that was the right thing to do. But my typical routine for these sub girls is to tap the font now, a few times and and only go to surgery if the final taps, you know, if the funnel keeps filling up and clearly, it's not working to do serial taps after, you know, about three will tend to go to surgery. But again, there are no class one data for which step to take. Okay, so next steps I know. So they've got a funnel tap set for culture because there was this question of maybe you know there was infection we suspected child abuse in this case, because there was no history as a baby with an inter hemispheric subdural and seizures so that points you towards child abuse but it isn't necessarily child abuse it could have been something else. So, that's what we did. So work up baby did get an MRI which many of you wanted you can see it there and what the MRI shows you is that there are extensive abnormalities. These are what we consider extensive patchy the patchy pattern. The baby had long term monitoring and the seizures in this setting are often difficult to control requiring multiple agents. And then when the baby was well enough to get careful x rays had multiple skull fractures, some bucket handle fractures and the father confessed to throwing the baby the father had PTSD from being a veteran we think and he threw the baby across the room on several occasions and he admitted to that. So next questions. The family wants to know data about prognosis, let's just go back to that MRI for one second. So you see all those diffusion abnormalities they want to know, you know, mothers savvy woman she wants to know what's going to happen to my baby so you're thinking you're going to answer in your own style. But when you think about the prognosis of this baby to yourself, not what you tell the parents necessarily. Here's what you're thinking babies are resilient babies probably going to do better than you'd expect because it's little there's a lot of plasticity, probably going to have mild deficits and do pretty well. Or you're thinking, pretty moderate deficits special schooling, you know, maybe dependent, or this is going to be a baby with bad outcomes severe disabilities dependent So, those are your questions which are you going to pick. Okay, so most people think that that's going to be pretty good, baby's going to do pretty well. So here's how the baby did. At 10 months, there was a fair amount of atrophy, particularly occipitally. At six years of age, you can see those areas of atrophy. So this is the so-called patchy pattern of damage seen, I think, most easily on DWI early on, but you can see these as areas of encephalomalacia. You see them sometimes on flare, depending on what point in time you get your imaging. This baby grew up to have cortical visual impairment and field cuts, significant visual processing deficits, needs an IEP. She has behavioral problems with aggressive outbursts, low frustration tolerance. She's on medication for that. She's clumsy. She does not have currently a seizure disorder. But this is not a great outcome. This is not mild deficit. It's not at the most severe end of the spectrum, but kids with the patchy patterns typically have something in the middle. But as they get older, their deficits can become more apparent, their social deficits, their processing deficits. So they can look good when you see them a year out, two years out. When they are 5, 6, 7, 8, 9, 10 years of age, their deficits often become more apparent. So my advice is don't oversell the resiliency of the very immature brain to this kind of injury. These kids almost always are a year or two behind in school at best. Most of them have behavioral and emotional regulation problems, even apart from their socioeconomic or social issues because they were victims of child abuse, which of course disrupts the whole family. So a diffusion abnormality like that early on is a predictor of significant deficit that you have to be prepared for. Okay, last case. This is a two-year-old found vomiting and unresponsive at home. When you see the baby comes in by EMS, the baby is unresponsive with a dilated pupil and was noted during transport to have active seizure. So baby is intubated because of the seizure and this is what you see. So question is next steps. Seizure meds, ICP monitor, medical management, put a ventriculostomy plus that medical management, do a craniotomy, do all that other stuff plus a decompressive hemicranioctomy. And I put other there because there are many different ways to manage this. I'm going to go back just to show you that scan again. And what you can see is a thin subdural, pretty extensive in terms of up and down, but pretty thin. There is some shift. The brain parenchyma looks pretty good without a lot of gray-white abnormality at this point in time. So those are your choices. Okay, so most people want to treat medically, a few people have ventriculostomy, very few people want to go to surgery. There's a few now. Okay, I will tell you that this particular case changed my mind about this, because actually I will be honest, one of my residents felt very strongly that we should do a heavy craniectomy and I wasn't sure what to do. I really wasn't sure, but the baby had focal deficits and a dilated pupil. And the hard thing was, was the dilated pupil due to the seizures, but it was a unilateral dilated pupil on the side of the clot. So I actually did do a hemicraniectomy and immediately after the hemicraniectomy, this was not the biggest hemicraniectomy, so don't judge me on that, this was early in the hemicraniectomy days, but it was low in the temporal lobe, we at least got that part right. But we went from the OR to the MRI scanner, and this is the diffusion study. And really that whole hemisphere had been affected. It's just that we got all this stuff so early that we couldn't really see it. And this entire hemisphere went on to swell dramatically and then atrophy over time. So when a kid comes in with shift, a subdural, a dilated pupil, could you have managed this medically? Maybe, but I will tell you that I was awfully glad we had the room for this hemisphere to go out because I think when you have that room, it spares this contralateral frontal lobe from getting a subfall scene herniation and wiping out, now you've got a bifrontal deficit. So I have taken over the years to doing early hemicraniectomies in any kid that is a sick child with focal deficits. Sometimes I'll do the MRI first to see if I see that diffusion abnormality, if I can get one like right away, but I don't hesitate. Now, I am well aware that, and we're going to get to a question on this, that when you have a big hemicrania in a little kid, it can be a real management problem and there are many complications that others than me have documented very well. So we don't take it lightly, we don't do it in everybody, but I think that early decompression will spare that contralateral frontal lobe, which makes a big difference, I believe, in outcome. A bifrontal injury is much worse than a unilateral injury. So I'm not going to do slider questions for this, but it always is an issue when you replace the skull and how you replace the skull. At 17 days, the swelling was going down, but I waited a little bit longer till this atrophy kicked in. There's always this problem with fluid and some kids have problems where they need to shunt, there's a high incidence of that. At five years of age, what you can see is the hemisphere is gone, but the contralateral hemisphere is pretty well protected. No subfall scene herniation, nothing that really went across the midline. Now, she was little, she was two at the time, just about two, and she required a couple of revisions of her cranioplasty. And we all know, you folks out there know that these cranioplasties in these little kids can be a bear. Nonetheless, I don't regret anything that might have, you never know, it's the existential crisis, what happened if you didn't do that, sparing that contralateral frontal lobe, because this kid actually did quite well. She's like a kid who had a hemispherectomy for seizures early in life. She's hemiparetic, but cognitively, she's quite good. And I think that's it. Thanks, Dr. Duhaime, that was a great presentation. And I was trying to write down all the different pearls that I learned from you during that presentation. I really appreciate you for the education about how those epidurals can mimic subdurals and how we need to be careful about identifying those as a risk factor for traumatic use of head injury or not. How to do a GCS examination on an infant, something that most of us even as trained pediatric neurosurgeons don't take advantage of. The patchy findings on MRI, again, same thing with all the challenges we have in reconstructing these hemicraniectomy defects. So thank you very much. And also thank you for taking advantage of the technology and using the PULSE. I think that was a lot of fun for everybody, especially at the end of the day to keep everybody engaged. One question I have for you is when you're tapping the fontanelle in one of these children, what are your goals? What are you tapping? How much are you taking off and how are you doing that? Yeah, so, I mean, it's nothing dramatic. It depends on the fontanelle. Come in laterally, see it through the skin so it doesn't leak or you reduce the risk of leaking. And I just let it drip out. Oftentimes, if it's a nice big fontanelle, like it is in a little baby, I'll literally gently put a little pressure on the opposite side of the fontanelle and let the fluid come out until I've basically got a slightly sunken fontanelle. Occasionally I'll do two sides if I think the two sides may not communicate. Is that what you guys do? We tend to do it with an angiocath and we'll leave the angiocath sitting there and it will often work for a day or two. I mean, once in a while we'll get lucky and it works longer. We haven't had somebody like inject IV fluid in through a presumed scalp catheter yet, but that may happen one day. But we found that, yeah, I think similarly, it's been nice to sort of serve as a pop-off valve for some of these kids who really just have a transient pressure problem that doesn't require a permanent diversion and needs a little intervention. One question I did want to ask which you started to allude to. Yeah, no, go ahead. I was just going to say, I have on occasion put an angiocath in. Yeah, go ahead, go ahead. Your turn. When, you alluded to it a little bit, but when have you, or when did you do the cranioplasty on that patient? And what have your sort of steps been in terms of like, do you leave drains in when you do do those cranioplasties? Do you try and do them earlier? I know there's starting to be a decent literature around timing and what the risks and benefits of that are, but what have you done in this population? Yeah, I'm an early cranioplasty person. I don't know that the data necessarily firmly supports that, but I just think there's less distortion of the tissue. It's just easier. So I do my cranioplasties in kids in this setting as soon as I'm able to, as soon as there's enough room. I have occasionally put subdural drains in or a ventriculostomy and then wean the ventric, but most often I try to wait until I think I won't need that because I don't want to increase the risk of infection. I will tell you that in the early days of my career for little babies, we used suture, heavy suture to hold it in place. I think if it's mobile, it doesn't heal as well. So now, even though it's more expensive, I'll use absorbable plates and screws to try to get a better fixation. And maybe that helps them resorb less, but it's still a crapshoot. This kid had a partial resorption that required a cranioplasty and they're bare. What do you guys do? I think we're a little bit over the place. You know, yeah. Go ahead, Rob. Also, we always go back and put the bone back, but we're almost already planning our cranioplasty for six months later. So it seems like the right thing to do, even though it doesn't seem to work out in this patient population very well. Do you use any kind of material when you're doing your duraplasty to try to make the cranioplasty easier down the road? Do you lay anything specific on the surface of the dura or the brain? I have used gel film in the past to keep the dura from sticking if I think I'm going to have to go back and do cranial surgery. But for these kids, I just put, you know, whatever collagen du jour is that we have available, the collagen, you know, synthetic collagen stuff. So usually when you go back, if you're lucky, you don't even get in there if you've waited enough time that it closes. I've got two questions here from our audience, both are from Dr. Bruce. The first one is, why do you think anterior subfalcing herniation affects the contralateral anterior cerebral? Yeah, I think it's just that when they have mass effect and they have what's been called frontal compartment syndrome, and you don't allow that frontal lobe to move laterally, it will move medially when that whole hemisphere swells. And it will pinch off, you know, the frontal polar branches of the anterior cerebral artery under the falx, and you'll get a wedge shape. Literally, it's an infarct on the opposite side. So I do feel that early decompression, maybe there are other ways to do it, but it's the most effective way I've seen. And those kids, there's a night and day difference, at least in my practice, from what the frontal lobe looks on the contralateral side when you do that compared to when you don't. That's a good segue into his next question, which is, is there any good evidence that hemicraniotomy has any value in trauma in children? Well, that's like asking, you know, what's the meaning of life? I mean, in order to really know, you would have to, you know, have enormous numbers of kids and have some way to stratify and compare them. And I believe hemicranioctomy helps you sleep better. Maybe that's not enough justification, but you pretty much take care of ICP. In the setting, I'm going to put a caveat, where you know from your MRI and from the clinical scenario that this is going to be a unilateral, what we used to call big black brain, now we call panhemispheric severe damage, almost always in child abuse, but occasionally in accidental subdurals. This pattern is so predictable when you have that early MRI finding that that whole hemisphere is going to swell, and then it's going to atrophy. And it's very profound, and it's stem to stern, top to bottom, front to back. And when you have something that points to that, to me, hemicranioctomy is the safest way to protect the rest of the brain. So to answer the question of, are there data? There are several papers on this. There are people that showcase reports. Do we have prospective, randomized, or even good retrospective stratified comparisons in the infant subdural, toddler subdural, big black brain scenario? No, we don't. But I can tell you, I used to not do it, and then I switched to doing it. And so far, other than recognizing that there are going to be complications, or that there's a risk of complications, I would rather there were complications to the skull than that there were complications to the contralateral hemisphere. Let me ask you a question about the hemicranioctomies. So the case you showed was a two-year-old. And to me, that's a very different operation than a hemicranioctomy, maybe in a three-month-old. Do you have a lower age threshold where you consider it differently, or does the age not matter to you? Yeah, good question. So it turns out that the data are pretty good that the unilateral form of this kind of injury is much more likely in the older kids, and the bilateral form is much more likely in the younger kids. Now, you notice there was no hemicranioctomy in that baby with Apache 1. Those kids you're going to be able to control medically in almost all cases. And as you know, the little kids will expand their sutures. So it's very rare. It has been very rare. In fact, I can't think of a bifrontal crani or a hemicrani in any kid under the toddler age that I've done for this scenario. So, you know, there may be instances where people, there certainly are instances where people do hemicraniectomies or bifrontal craniectomies for babies, like little babies, infants. I just have not had to for this scenario because it's rarely that it's unilateral. And I don't believe that bilateral hemicraniectomies, when you have bilateral, terrible damage of the, you know, infant, typical infant stage where both hemispheres are wiped out, I don't think that helps anybody. And frankly, most kids don't need it because they just expand their sutures. Thank you. Kind of referencing the first case you showed, a case that could be mistaken as abusive head trauma when it wasn't, and to some of the biases that go into how those cases get brought up as abusive head trauma. What do you think the role is for us as pediatric neurosurgeons in advocacy or having more involvement in helping guide the team on whether we think that this could be the result of abuse or not? Personally, I think the pediatric neurosurgeon who is informed about child abuse, mechanisms of injury, you know, head injury in young children can play an enormous role. Oftentimes people are so afraid of missing something that they go overboard on the other direction. That doesn't mean you don't report if you think, you know, we should never intimidate people to going against their conscience or what they feel is their professional duty. That said, it's a constant struggle to educate people about cases like that first one. It's not a subdural, it's an epidural. There is a skull fracture. Yes, this kid is a minority, that doesn't mean it's child abuse. I mean, I'm not trying to be glib here, but it's so, these patterns are so common where people are afraid of being judged for missing a child abuse case. So I think I personally am on our child protection team and I, you know, they often, pretty much always when it's a head injury in a baby, they will call me and we'll discuss it. So I think the pediatric neurosurgeon can play a huge role. And if you're not sure, there are people out there besides me, but I'm always willing to get a call to say, what do you think of this case? And it doesn't mean I'm always right, but some of us have a particular interest in this and we're happy to, just like, you know, I would call somebody else about a brain tumor or whatever. Talk to your colleagues who have an interest in this and maybe know the literature a little more than you do or whatever, because, you know, none of us sees a ton of these or if we do, my condolences, but it's an area of, you know, there's a literature, there are some facts, there are some unknowns and sorting through what you know and what you don't know and what is known and what isn't known can be daunting for the neurosurgeon. But people look to us for the answers. So I think we do have a role to play and I think you can reach out to your colleagues around the country. Sometimes it is really helpful to have somebody else to talk to. Excellent. Well, thank you so much, Dr. DeHaan, for coming to us, sharing your expertise with us and walking us through this case. All right, that brings us to the end of the session. I want to thank every one of you for joining us. I've already received feedback from the staff that's helped made this possible to let us know that this has been the most interactive session of all of the subspecialty sessions, the AANS. I think that speaks to the cohesive nature of our subspecialty and to just the type of people that the way we work together and the way we interact together at meetings. I want to apologize to some of the participants who were scheduled for the second session that was supposed to occur in Orlando. It's unfortunate that we had to cancel one of the sessions and I apologize to those who may have already prepared and didn't get to share with us. I'm sure we would have learned a lot from you. Thank you to the AANS and for the staff kind of behind the scenes who made this happen. This is a huge lift, a pivot in the last two weeks to make this happen. I wish that we were all there in person together, but it's nice to be together right here virtually. This has been a great experience and I thank each one of you. I hope next year that we can all come together in person in Philadelphia. This brings the AANS to an end. Thank you guys once again for joining us today. ♪♪

pediatric neurological surgery

ependymoma

brain tumor

treatment

early detection

targeted therapy

clinical trials

research

data sharing

microRNA expression

correlation

ventricular volume

hydrocephalus

head trauma

diagnosis